{ "count": 108, "guidelines": [ { "slug": "acs-sdoh-linkage-checklist-rwe", "name": "ACS and Area-Level SDoH Linkage Checklist", "short_definition": "A checklist for linking ACS-derived contextual variables and neighborhood indices to patient-level RWE datasets while preserving geography, time, linkage, and ecological-validity caveats.", "long_description": "**What it is** - This guideline is the checklist layer for linking American Community Survey\n(ACS), Area Deprivation Index, Social Deprivation Index, CDC/ATSDR SVI, or similar area-level\nsocial-context variables to patient-level RWE datasets. The companion concept explains ACS/SDoH\nlinkage; this guideline states what must be specified before those variables are used as\nconfounders, effect modifiers, equity strata, descriptive context, or model inputs. The key\ndistinction is that area-level measures are contextual proxies, not individual-level facts.\n\n**When to use** - Use it when claims, EHR, registry, survey, or linked studies attach\nneighborhood or geography-derived measures through address, ZIP Code, ZIP+4, census tract, block\ngroup, county, or site geography. It is most important when the SDoH variable affects confounding\ncontrol, equity stratification, transportability, missingness interpretation, or risk adjustment.\nUse it before analysis because geocoding precision, linkage date, ACS vintage, and geography\nchoice determine the exposure assigned to each person.\n\n**What it requires / checklist domains** - Specify address source, geocoding method, geography,\nACS vintage, linkage date, and crosswalk. Report linkage success and compare linked versus\nunlinked patients. Preserve linkage-quality flags in the analytic dataset. Align ACS 5-year\nwindows or index releases to the clinical observation period. Avoid ZIP-only linkage when tract\nor block-group heterogeneity matters; if ZIP is all that exists, label the limitation. Pre-specify\nwhether the measure is a confounder, effect modifier, equity stratum, descriptive variable, or\ncontextual covariate. Evaluate sensitivity to geography, vintage, and linked/unlinked inclusion.\n\n**When NOT to use - limitations and common misapplications** - Do not interpret an area-level\npoverty, deprivation, or vulnerability score as the patient's income, education, race, or housing\nstatus. Do not link current address to historical outcomes without considering residential\nmobility and timing. Do not mix vintages or geographies without a crosswalk plan. Do not ignore\ndifferential geocoding failure, because unlinked patients can be systematically different. Do not\nadjust for SDoH variables mechanically if they are mediators or colliders for the estimand.\nArea-level linkage can improve contextual validity, but it can also introduce ecological fallacy,\ntemporal mismatch, and selection from linkage failure.\n\n**How it maps to this catalog** - This guideline cross-references\n`acs-sdoh-area-level-linkage-rwe` for the linkage concept,\n`sdoh-social-determinants-of-health` for the broader construct,\n`linked-data` and `tokenization-privacy-preserving-record-linkage-rwe` for linkage mechanics,\n`generalizability-transportability-external-validity-rwe` for target-population interpretation,\n`selection-bias-sensitivity-analysis-rwe` for linked/unlinked selection, and\n`missing-data-pattern-table-rwe` for linkage failure and geography missingness. Use this checklist\nfor the linkage appendix and source table; use the concept for operational examples.", "primary_category": "Guideline", "tags": [ "guideline", "checklist", "acs", "sdoh", "area-level-linkage", "deprivation-index" ], "aliases": [ "ACS checklist", "SDoH linkage checklist", "area deprivation linkage checklist" ], "applies_to_study_types": [ "health_equity", "claims_analysis", "ehr_study", "registry_linkage", "comparative_effectiveness" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "url": "https://www.census.gov/programs-surveys/acs/about.html", "citation_text": "U.S. Census Bureau. American Community Survey (ACS).", "year": 2026, "authors_short": "U.S. Census Bureau", "notes": "Official ACS data source." }, { "role": "explain", "doi": "10.1056/NEJMp1802313", "url": "https://doi.org/10.1056/NEJMp1802313", "citation_text": "Kind AJH, Buckingham W. Making Neighborhood-Disadvantage Metrics Accessible - The Neighborhood Atlas. New England Journal of Medicine. 2018;378(26):2456-2458.", "year": 2018, "authors_short": "Kind and Buckingham", "notes": "Area Deprivation Index / Neighborhood Atlas source describing ACS-derived neighborhood-disadvantage measures and public access." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "acs-sdoh-area-level-linkage-rwe", "notes": "Checklist for ACS and area-level SDoH linkage." }, { "relation_type": "applies_to", "target_slug": "sdoh-social-determinants-of-health", "notes": "Contextual linkage checklist for SDoH analyses." } ], "index_definitions": [], "checklist_items": [ "Specify address source, geocoding precision, geography, ACS vintage, and linkage date.", "Report linkage success and compare linked versus unlinked patients.", "Treat ACS-derived measures as contextual variables, not individual-level facts.", "Align ACS 5-year windows to the clinical or claims observation period.", "Avoid ZIP-only linkage when tract-level heterogeneity matters; state the limitation if ZIP is all that exists.", "Pre-specify whether SDoH variables are confounders, effect modifiers, equity strata, or descriptive context.", "Retain linkage-quality flags in the analytic dataset and sensitivity analyses." ], "regulatory_relevance": [ "hta", "fda" ] }, { "slug": "ahrq-methods", "name": "AHRQ EPC Methods Guide for Effectiveness and Comparative Effectiveness Reviews", "short_definition": "The methodological reference that governs how AHRQ Evidence-based Practice Centers scope, conduct, appraise, synthesize, and grade the strength of evidence in comparative effectiveness reviews — including reviews that incorporate observational and real-world evidence.", "long_description": "**What it is.** The **AHRQ Methods Guide for Effectiveness and Comparative Effectiveness Reviews** (commonly the\n\"AHRQ EPC Methods Guide\" or \"CER Methods Guide\") is the master methodological reference maintained by the **Agency for\nHealthcare Research and Quality (AHRQ)** through its **Effective Health Care (EHC) Program** and the network of\n**Evidence-based Practice Centers (EPCs)**. It is a living, chapter-based guide — launched in the AHRQ Series of\n*Journal of Clinical Epidemiology* papers (2010) and continuously updated as standalone EHC chapters — that prescribes\nhow an EPC should run a comparative effectiveness review (CER): formulating the question with the PICOTS framework,\nsearching and selecting evidence, assessing the risk of bias of individual studies, synthesizing results\n(qualitatively and via quantitative/meta-analytic methods), grading the **strength of a body of evidence** across the\ndomains of study limitations, directness, consistency, precision, and reporting bias, and rating the applicability of\nthe findings. It is a **process-and-appraisal guide for evidence synthesis**, not a primary-study reporting checklist\nand not a risk-of-bias instrument in itself.\n\n**When to use.** Use the AHRQ EPC Methods Guide when you are **producing or reviewing a systematic review / comparative\neffectiveness review of an intervention or exposure**, especially one commissioned under or modeled on the EHC Program\n(EPC technical briefs, full CERs, USPSTF evidence reviews, CMS coverage evidence reviews). It is the governing standard\nwhen a review must (1) compare ≥2 interventions head-to-head, (2) integrate evidence of mixed design — RCTs plus\nobservational/RWE studies drawn from claims, EHR, or registries — and (3) deliver a defensible strength-of-evidence\ngrade for decision makers. Decision rules for which guide applies: use **AHRQ EPC Methods Guide** for the *conduct and\nappraisal logic* of an EHC-style CER; use **PRISMA 2020** to *report* the systematic review and **PRISMA-P** to register\nits protocol; use **ROBINS-I** (which has superseded AHRQ's earlier RTI Item Bank approach) for per-study risk-of-bias\nof non-randomized studies cited within the CER; use a **GRADE**-based scheme or AHRQ's own strength-of-evidence grading\nfor the certainty rating. For the primary observational studies you are appraising, the relevant *reporting* standards\nare STROBE / RECORD-PE / HARPER — not this guide.\n\n**What it requires.** The guide enforces a structured chain of methodological decisions, each of which becomes the lens\nthrough which included RWE studies are appraised:\n- **Question framing & scope** — an explicit **PICOTS** statement (Population, Intervention, Comparator, Outcomes,\n Timing, Setting) and analytic framework linking interventions to intermediate and health outcomes.\n- **Evidence search & selection** — comprehensive, reproducible, dual-reviewer searching with pre-specified eligibility,\n including gray literature and regulatory/registry sources.\n- **Risk-of-bias appraisal of included studies** — design-appropriate assessment; for non-randomized/RWE evidence this\n means scrutinizing confounding control, exposure and outcome misclassification, selection/immortal-time bias, and\n time-zero alignment *as reported by the primary study* (the CER judges whether the study handled these, it does not\n itself analyze patient-level data).\n- **Data-fitness and applicability** — whether the underlying data source and population support the question, and how\n well the body of evidence applies to the target decision (applicability/PICOTS match).\n- **Quantitative synthesis** — when pooling is appropriate, pre-specified meta-analytic methods, heterogeneity\n assessment, and handling of sparse or observational data; otherwise a structured qualitative synthesis.\n- **Strength-of-evidence grading** — a transparent rating across **study limitations, directness, consistency,\n precision, and reporting/publication bias**, yielding High / Moderate / Low / Insufficient.\n- **Sensitivity & bias analysis** — assessment of how robust conclusions are to study-level bias, including reporting\n bias and, where included RWE warrants it, quantitative bias considerations.\n\n**When NOT to use — limitations and common misapplications.**\n- **It is not a primary-study reporting checklist.** Authors of an individual claims/EHR cohort study should report with\n STROBE/RECORD-PE/HARPER and register a protocol; the AHRQ guide governs the *review of* such studies, not their\n conduct. Treating it as a single-study checklist is a category error a senior reviewer will flag immediately.\n- **It is not, by itself, a risk-of-bias instrument or a quality score.** It directs you to design-appropriate tools\n (now ROBINS-I for non-randomized studies); it does not produce a numeric quality score, and a high strength-of-evidence\n grade is a statement about the *body of evidence*, not a guarantee that any included observational study is causal.\n- **It is not an HTA reference case or an economic-evaluation standard.** For value/cost-effectiveness or payer\n decision frameworks, use NICE / CADTH / ICER reference cases and CHEERS for economic reporting — AHRQ EPC reviews\n inform but do not replace these.\n- **Completing the steps does not make the underlying observational evidence causal.** Synthesis discipline cannot\n repair confounding or selection bias baked into the primary RWE studies.\n- **Checklist-as-theater** — going through the chapters without the dual-review rigor, pre-specification, and honest\n strength-of-evidence grading they demand produces a CER in name only.\n\n**How it maps to this catalog.** The AHRQ guide is a wrapper that, when a CER ingests real-world evidence, cross-walks\ndirectly onto this catalog's implementing concepts:\n- Question framing → **picots-framework-rwe** and **comparative-effectiveness-research-cer-methods** /\n **cer-observational**; the review object itself is a **systematic-review** (with **meta-analysis-obs** when pooling\n observational studies).\n- Data fitness-for-use → **fit-for-purpose-data-assessment-rwe** (does the source support the question?).\n- Appraising whether an included RWE study controlled confounding and aligned time zero → **target-trial-emulation**,\n **active-comparator-new-user**, **high-dimensional-propensity-score-hdps-rwe**, and **time-zero-index-date-alignment-rwe**.\n- Appraising estimand clarity and outcome/exposure misclassification in cited studies →\n **estimands-ate-att-intercurrent-events-rwe**, **diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe**, and\n **claims-outcome-algorithm-ppv-sensitivity-rwe**.\n- Appraising attrition/missing data in cited studies → **attrition-and-loss-to-follow-up-rwe**.\n- Strength-of-evidence robustness and sensitivity/quantitative bias analysis →\n **quantitative-bias-analysis-toolkit-rwe** and **e-value-sensitivity-analysis**.\n\n**Applied note (claims/EHR/registry RWE).** When an EHC-style CER pulls a claims- or EHR-based cohort study into its\nevidence base, the guide does not ask the reviewer to re-run the analysis — it asks the reviewer to judge, from what the\nprimary study reports, whether time zero was aligned at initiation, whether an active comparator and new-user design\ncontrolled confounding by indication, whether the outcome phenotype was validated (PPV/sensitivity), and whether\nattrition was differential. Those judgments then feed the study-limitations domain of the strength-of-evidence grade.\nThis is the practical seam between the AHRQ synthesis layer and the RWE-method concepts catalogued here.", "primary_category": "Guideline", "tags": [ "guideline", "comparative-effectiveness-review", "systematic-review", "evidence-synthesis", "strength-of-evidence", "risk-of-bias", "ahrq-ehc", "epc" ], "aliases": [ "AHRQ CER Methods Guide", "AHRQ EHC Methods Guide", "AHRQ EPC Methods Guide", "Methods Guide for Effectiveness and Comparative Effectiveness Reviews", "CER Methods Guide", "AHRQ Effective Health Care Program Methods Guide" ], "applies_to_study_types": [ "systematic_review", "cer_observational", "meta_analysis_obs" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "doi": "10.1016/j.jclinepi.2008.06.009", "url": "https://doi.org/10.1016/j.jclinepi.2008.06.009", "citation_text": "Slutsky J, Atkins D, Chang S, Collins Sharp BA. AHRQ Series Paper 1: Comparing medical interventions: AHRQ and the Effective Health-Care Program. Journal of Clinical Epidemiology. 2010;63(5):481-483.", "year": 2010, "authors_short": "Slutsky et al.", "notes": "Opening paper of the AHRQ Series that launched the EPC Methods Guide, defining the Effective Health Care Program's comparative effectiveness review mission and the methodological challenges it must address." }, { "role": "explain", "doi": "10.1016/j.jclinepi.2009.05.005", "url": "https://doi.org/10.1016/j.jclinepi.2009.05.005", "citation_text": "Helfand M, Balshem H. AHRQ Series Paper 2: Principles for developing guidance: AHRQ and the Effective Health-Care Program. Journal of Clinical Epidemiology. 2010;63(5):484-490.", "year": 2010, "authors_short": "Helfand & Balshem", "notes": "States the foundational principles (analytic frameworks, stakeholder input, transparency) that govern how every chapter of the Methods Guide is developed and applied." }, { "role": "explain", "doi": "10.1016/j.jclinepi.2009.03.009", "url": "https://doi.org/10.1016/j.jclinepi.2009.03.009", "citation_text": "Owens DK, Lohr KN, Atkins D, et al. AHRQ Series Paper 5: Grading the strength of a body of evidence when comparing medical interventions—Agency for Healthcare Research and Quality and the Effective Health-Care Program. Journal of Clinical Epidemiology. 2010;63(5):513-523.", "year": 2010, "authors_short": "Owens et al.", "notes": "Defines the EPC strength-of-evidence grading scheme (study limitations, directness, consistency, precision, reporting bias) that is the analytic endpoint of an AHRQ comparative effectiveness review." }, { "role": "explain", "doi": "10.1016/j.jclinepi.2011.08.004", "url": "https://doi.org/10.1016/j.jclinepi.2011.08.004", "citation_text": "Chang SM, Bass EB, Berkman N, et al. The Agency for Healthcare Research and Quality (AHRQ) Effective Health Care (EHC) Program Methods Guide for Comparative Effectiveness Reviews: keeping up-to-date in a rapidly evolving field. Journal of Clinical Epidemiology. 2011;64(11):1166-1167.", "year": 2011, "authors_short": "Chang et al.", "notes": "Documents the living-document nature of the guide and the process by which chapters are revised, explaining why the maintained EHC chapters supersede the original Series papers." }, { "role": "use", "doi": "10.23970/AHRQEPCMETHGUIDE3", "url": "https://effectivehealthcare.ahrq.gov/products/collections/cer-methods-guide", "citation_text": "Morton SC, Murad MH, O'Connor E, et al. Quantitative Synthesis—An Update. Methods Guide for Effectiveness and Comparative Effectiveness Reviews. Rockville, MD: AHRQ; 2018. AHRQ Publication No. 18-EHC007-EF.", "year": 2018, "authors_short": "Morton et al.", "notes": "Current maintained chapter on quantitative/meta-analytic synthesis; the EHC collection URL is the authoritative, continuously updated home of all Methods Guide chapters." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "systematic-review", "notes": "The guide governs the conduct and appraisal of systematic reviews / comparative effectiveness reviews." }, { "relation_type": "applies_to", "target_slug": "cer-observational", "notes": "Directly applicable when a CER integrates observational/RWE evidence into its synthesis." }, { "relation_type": "applies_to", "target_slug": "meta-analysis-obs", "notes": "The quantitative-synthesis chapter governs meta-analysis of observational studies within a CER." }, { "relation_type": "see_also", "target_slug": "comparative-effectiveness-research-cer-methods", "notes": "The catalog concept that implements the comparative-effectiveness-review methodology this guide standardizes." }, { "relation_type": "see_also", "target_slug": "picots-framework-rwe", "notes": "Implements the PICOTS question-framing step the guide requires at the outset of every review." }, { "relation_type": "complements", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "Operationalizes the data fitness-for-use judgment the guide demands when appraising RWE evidence." }, { "relation_type": "complements", "target_slug": "target-trial-emulation", "notes": "The design lens for judging whether an included RWE study aligned time zero and controlled confounding." }, { "relation_type": "complements", "target_slug": "quantitative-bias-analysis-toolkit-rwe", "notes": "Supports the sensitivity / strength-of-evidence robustness assessment for observational evidence in the review." }, { "relation_type": "see_also", "target_slug": "claims-analysis", "notes": "Apply when the body of evidence under review is built from claims/EHR/registry studies." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "hta", "journal" ] }, { "slug": "ahrq-registries", "name": "AHRQ Registries for Evaluating Patient Outcomes: A User's Guide", "short_definition": "AHRQ's authoritative how-to manual for planning, building, operating, analyzing, and assessing the quality of patient registries for evaluating real-world outcomes, safety, and effectiveness.", "long_description": "**What it is.** *Registries for Evaluating Patient Outcomes: A User's Guide* (the \"AHRQ Registries Guide,\" currently the 4th edition, Gliklich, Leavy & Dreyer, eds., AHRQ, 2020) is the field's most comprehensive **operating manual** for patient registries. It is not a one-page reporting checklist; it is a multi-chapter best-practice handbook that walks an investigator through the full life cycle of a registry — defining purpose and stakeholders, governance and oversight, sample-size and design, selection of data elements, patient identification and recruitment, data sources and linkage, data collection and quality assurance, analysis, interpretation, and long-term operation, transition, or closure. It is maintained and published by the U.S. Agency for Healthcare Research and Quality (AHRQ) through its Effective Health Care Program, with successive editions and topic-specific addenda (e.g., registries assessing the safety and effectiveness of medical products, linking registries to other data, 21st-century interoperability). It is constructive (\"how to build and run one well\") rather than purely evaluative, which is what distinguishes it from registry critical-appraisal checklists.\n\n**When to use.** Reach for the AHRQ Registries Guide whenever the deliverable is a **registry as an evidence-generation instrument** — a disease/condition registry, a product (drug or device) registry, a pregnancy/exposure registry, a post-authorization safety study (PASS) or post-marketing requirement run as a registry, or a registry-based study feeding an FDA/EMA submission, an HTA/payer dossier, or a peer-reviewed manuscript. Use it at the *design and build* stage, before data collection, when you must justify governance, data-element definitions, outcome ascertainment, and quality processes prospectively. Decision rules versus siblings: choose the AHRQ Guide when the task is to **plan, build, or operate** a registry end-to-end; choose **ENCePP Guide/Checklist** when you are conducting or reporting a (often EU) pharmacoepidemiologic study more broadly; choose the **FDA RWE framework / FDA non-interventional guidances or CIOMS RWD/RWE** when the task is to align registry output with a specific regulatory submission (those frameworks *consume* what the AHRQ Guide helps you *produce*); choose **ISPE GPP / SCOPE or GRACE** when you need to *appraise the quality* of a finished study rather than construct a registry; and choose **STROBE / RECORD / RECORD-PE** for the *reporting* manuscript that comes out of the registry. In practice the AHRQ Guide is used alongside, not instead of, those documents.\n\n**What it requires.** The Guide enforces substantive, registry-specific domains, each of which has a concrete real-world-data implementation in this catalog: (1) **Purpose, stakeholders, and governance** — an explicit primary question, oversight structure, and a transparent plan (links to study protocol/SAP elements). (2) **Design and fitness-for-purpose of data** — registry design choices and an honest assessment of whether the captured data can answer the question (data fitness-for-use). (3) **Patient identification, enrollment, and selection** — eligibility, recruitment, representativeness, and time-zero alignment so follow-up begins at a defined, unbiased index. (4) **Data elements, phenotypes, and outcome ascertainment** — standardized, validated condition/exposure/outcome definitions and adjudicated or validated endpoints rather than ad hoc code lists. (5) **Data quality, completeness, and follow-up** — source-data verification, monitoring, and pre-specified handling of attrition, loss to follow-up, and missing data. (6) **Analysis and confounding control** — appropriate comparison strategies (often active-comparator, new-user, or target-trial-emulation logic), confounding adjustment, and clearly specified estimands and intercurrent events. (7) **Interpretation, sensitivity, and quantitative bias analysis** — assessment of residual confounding, selection, and misclassification, plus transportability/generalizability of the registry population. (8) **Quality assessment of the registry itself** — the Guide's evaluation chapter provides a structured framework to grade a registry's rigor.\n\n**When NOT to use — limitations and common misapplications.** (a) The AHRQ Registries Guide is a **best-practice manual, not a validated risk-of-bias instrument and not a numeric quality score**; do not report \"AHRQ score = X\" — its evaluation chapter is a structured appraisal aid, not a weighted scale like a STROBE-derived score or a ROBINS-I rating. For formal risk-of-bias use ROBINS-I/ROBINS-E; for reporting completeness use STROBE/RECORD. (b) **Following the Guide does not make a registry causal.** A well-governed, high-quality registry that lacks a sound comparator, time-zero alignment, and confounding control still cannot support a causal claim; the operational excellence the Guide demands is necessary, not sufficient. (c) **Wrong document for the stage:** using the AHRQ Guide as your manuscript reporting checklist (STROBE/RECORD is correct), or using a reporting checklist to *design* a registry (the AHRQ Guide is correct), is a category error. (d) **Checklist/manual-as-theater:** citing the Guide in a protocol while none of its data-quality, phenotype-validation, or governance practices were actually implemented. (e) **Edition/scope drift:** applying generic registry advice to a regulatory product/safety registry without consulting the relevant addendum (medical-product safety and effectiveness, or registry-to-data linkage) that adds the regulatory-grade requirements.\n\n**How it maps to this catalog.** Each Guide requirement is implemented by a concrete concept here: *fitness-for-purpose of data* -> `fit-for-purpose-data-assessment-rwe`; *phenotype/algorithm definition and validation* -> `diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe` and `algorithm-validation`; *outcome ascertainment/adjudication* -> `endpoint-adjudication-chart-review-rwe`; *time-zero/index alignment* -> `time-zero-index-date-alignment-rwe`; *estimands and intercurrent events* -> `estimands-ate-att-intercurrent-events-rwe`; *attrition, follow-up, and missing data* -> `attrition-and-loss-to-follow-up-rwe` and `missing-data-pattern-table-rwe`; *comparison and confounding control* -> `active-comparator-new-user`, `propensity-score-methods-psm-iptw`, and `target-trial-emulation`; *sensitivity and residual bias* -> `quantitative-bias-analysis-toolkit-rwe`; *external validity of the registry population* -> `generalizability-transportability-external-validity-rwe`. The registry study-type definitions are `disease-registry`, `product-registry`, and `pregnancy-registry` (the latter typically paired with `mother-infant-linkage-rwe`), and registry analyses frequently lean on `claims-analysis` for linkage to complete utilization and mortality data.\n\n**Applied note (claims/EHR/registry RWE).** Registries are rarely self-contained. A disease or product registry usually captures clinical detail and adjudicated outcomes well but is incomplete for full medication exposure, healthcare utilization, and death; linking to administrative claims (with continuous-enrollment requirements) and a death index is the standard remedy and is exactly where the AHRQ Guide's data-source, linkage, and quality chapters earn their keep. For a regulatory product/safety registry, treat the Guide's medical-product addendum as the floor: pre-specify validated phenotypes and adjudicated endpoints, document time-zero and the comparator strategy, quantify attrition and missingness, and carry residual-confounding sensitivity analyses — the same machinery a reviewer expects from any defensible non-interventional RWE study.", "primary_category": "Guideline", "tags": [ "guideline", "registry", "methodological", "pharmacoepidemiology", "real-world-evidence", "data-quality" ], "aliases": [ "AHRQ Registries Guide", "Gliklich Registries Guide", "Registries for Evaluating Patient Outcomes", "AHRQ Patient Registries User's Guide", "AHRQ Registries for Evaluating Patient Outcomes" ], "applies_to_study_types": [ "disease_registry", "product_registry", "pregnancy_registry" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "doi": "10.23970/AHRQEPCREGISTRIES4", "url": "https://doi.org/10.23970/AHRQEPCREGISTRIES4", "citation_text": "Gliklich RE, Leavy MB, Dreyer NA, eds. Registries for Evaluating Patient Outcomes: A User's Guide. 4th ed. AHRQ Publication No. 19(20)-EHC020. Rockville, MD: Agency for Healthcare Research and Quality; 2020.", "year": 2020, "authors_short": "Gliklich, Leavy & Dreyer (eds.)", "notes": "Canonical AHRQ statement; multi-chapter best-practice manual covering registry design, operation, data quality, analysis, and a structured evaluation framework." }, { "role": "explain", "url": "https://www.ncbi.nlm.nih.gov/books/NBK562575/", "citation_text": "Registries for Evaluating Patient Outcomes: A User's Guide (4th ed.) — full text. Rockville, MD: Agency for Healthcare Research and Quality; 2020. NCBI Bookshelf.", "year": 2020, "authors_short": "AHRQ / NCBI Bookshelf", "notes": "Open full text of all chapters and addenda, including data-element selection, linkage, quality assurance, and registry evaluation." }, { "role": "use", "url": "https://effectivehealthcare.ahrq.gov/products/registries-guide-4th-edition/users-guide", "citation_text": "AHRQ Effective Health Care Program. Registries for Evaluating Patient Outcomes: A User's Guide, 4th Edition — product page and prior editions/addenda.", "year": 2020, "authors_short": "AHRQ Effective Health Care Program", "notes": "Maintained AHRQ landing page with the current edition, prior editions, and topic-specific addenda (medical-product safety/effectiveness, registry linkage)." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "disease-registry", "notes": "Primary operating manual for designing, building, and assessing a disease/condition registry." }, { "relation_type": "applies_to", "target_slug": "product-registry", "notes": "Governs drug/device product registries; pair with the Guide's medical-product safety/effectiveness addendum for regulatory use." }, { "relation_type": "applies_to", "target_slug": "pregnancy-registry", "notes": "Applies to pregnancy/exposure registries; outcome capture typically requires mother-infant linkage." }, { "relation_type": "see_also", "target_slug": "disease-registry", "notes": "Study-type definition the AHRQ Guide operationalizes." }, { "relation_type": "see_also", "target_slug": "product-registry", "notes": "Study-type definition for drug/device registries governed by the Guide." }, { "relation_type": "see_also", "target_slug": "pregnancy-registry", "notes": "Study-type definition; combine with mother-infant-linkage-rwe for infant outcomes." }, { "relation_type": "used_with", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "Implements the Guide's requirement to judge whether registry data can answer the question." }, { "relation_type": "used_with", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Implements standardized, validated condition/exposure phenotype definitions." }, { "relation_type": "used_with", "target_slug": "algorithm-validation", "notes": "Implements validation of registry phenotypes and outcome algorithms." }, { "relation_type": "used_with", "target_slug": "endpoint-adjudication-chart-review-rwe", "notes": "Implements adjudicated/validated outcome ascertainment the Guide expects." }, { "relation_type": "used_with", "target_slug": "time-zero-index-date-alignment-rwe", "notes": "Implements unbiased enrollment/index-date (time-zero) alignment." }, { "relation_type": "used_with", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Implements explicit estimand and intercurrent-event specification for registry analyses." }, { "relation_type": "used_with", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Implements the Guide's attrition/loss-to-follow-up and follow-up-quality requirements." }, { "relation_type": "used_with", "target_slug": "missing-data-pattern-table-rwe", "notes": "Implements pre-specified missing-data documentation and handling." }, { "relation_type": "used_with", "target_slug": "quantitative-bias-analysis-toolkit-rwe", "notes": "Implements sensitivity and quantitative bias analysis for residual confounding/selection/misclassification." }, { "relation_type": "used_with", "target_slug": "generalizability-transportability-external-validity-rwe", "notes": "Implements assessment of how well the registry population transports to the target population." }, { "relation_type": "used_with", "target_slug": "target-trial-emulation", "notes": "Causal-design discipline for registry-based comparative effectiveness/safety questions." }, { "relation_type": "used_with", "target_slug": "active-comparator-new-user", "notes": "Comparator/new-user strategy for confounding control in registry comparisons." }, { "relation_type": "used_with", "target_slug": "propensity-score-methods-psm-iptw", "notes": "Confounding-adjustment machinery for registry-based comparative analyses." }, { "relation_type": "used_with", "target_slug": "claims-analysis", "notes": "Linkage to administrative claims for complete exposure, utilization, and mortality capture." }, { "relation_type": "used_with", "target_slug": "mother-infant-linkage-rwe", "notes": "Required to capture infant outcomes in pregnancy/exposure registries." }, { "relation_type": "see_also", "target_slug": "study-protocol-or-sap-elements", "notes": "The Guide's governance/design requirements feed directly into the registry protocol and SAP." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "hta", "journal" ] }, { "slug": "amcp-format", "name": "AMCP Format for Formulary Submissions", "short_definition": "The US managed-care industry standard for the structure and content of evidence dossiers submitted by pharmaceutical and biologic manufacturers to support formulary consideration by health plans and pharmacy benefit managers. Version 5.0 (2024) is the current edition. It sets expectations for clinical, economic, and patient-reported outcomes evidence — the primary document payers request when evaluating a new product for formulary placement in the United States.", "long_description": "**What it is** — The **AMCP Format for Formulary Submissions** (AMCP Format) is the de facto US payer\ndossier standard published and maintained by the **Academy of Managed Care Pharmacy (AMCP)**. First\nissued in 2000 and now in its **fifth edition (Version 5.0, 2024)**, it specifies the content and\norganisation of a comprehensive evidence package — a \"formulary dossier\" — that manufacturers\nprepare when seeking formulary consideration from US managed care organisations, pharmacy benefit\nmanagers (PBMs), and integrated delivery networks (IDNs). Unlike HTA submissions in Europe\n(NICE, G-BA, HAS, JCA), the AMCP Format is not a regulatory requirement but an industry\nconvention: adoption is voluntary, yet the document has become the default request from US payer\npharmacy and therapeutics (P&T) committees. Version 5.0 updates the structure to reflect the\nexpanded role of **real-world evidence**, patient-centric outcomes, health equity considerations,\nand digital/cell-and-gene-therapy product types. The dossier is organised around a product **Section 1**\n(product information: indication, mechanism, dosing), **Section 2** (clinical evidence: pivotal trials,\nsubgroup data, indirect comparisons, RWE), **Section 3** (economic and outcomes evidence: budget impact,\ncost-effectiveness models, HEOR dossier), and **Section 4** (product value and patient-reported outcomes),\nplus a supporting appendix. The format does not adjudicate evidence quality — it defines what evidence\nto *organise and present*, leaving P&T committees to judge adequacy.\n\n**When to use** — The AMCP Format governs US payer dossier preparation. Apply it when: a manufacturer\nis seeking formulary placement, preferred tier status, or step-edit removal from a US health plan,\nPBM, or integrated delivery network; when a payer's medical or pharmacy director has requested an\nevidence dossier for a P&T committee review; or when a HEOR, market access, or medical affairs team\nneeds a standardised framework for assembling and organising the full clinical-economic evidence\npackage for a US submission. Decision rule: for **US managed-care formulary submissions**, use the\nAMCP Format (version 5.0 for new submissions); for **European HTA submissions** (Germany, France, UK,\nEU JCA), the relevant national or regional guidance applies (G-BA AMNOG, HAS CEESP, NICE reference\ncase, **EU HTA JCA** regulation); for **value-based contracting or payer engagement outside a formal\ndossier**, the AMCP Format sections can still structure the evidence narrative but are not used\nverbatim. The AMCP Format complements rather than replaces clinical study reports — it is an evidence\n*organisation* standard, not a study-conduct or reporting guideline.\n\n**What it requires (checklist domains)** — The AMCP Format enforces a structured dossier with these\nkey components that P&T reviewers expect: *Section 1 — Product information*: FDA-approved labelling\n(prescribing information), mechanism of action, proposed formulary tier, and any REMS requirements.\n*Section 2 — Clinical evidence*: (a) **Pivotal controlled trial data** (efficacy, safety, subgroup\nanalyses, and head-to-head comparisons where available); (b) **Indirect treatment comparisons** and\n**network meta-analyses** for placing the product relative to formulary alternatives when no head-to-\nhead trials exist (see **ISPOR Indirect Comparisons** guideline); (c) **Real-world evidence** —\nclaims, EHR, registry, or patient-reported outcomes data demonstrating effectiveness and safety in\nbroader or plan-specific populations, a rapidly growing dossier section given FDA's increased emphasis\non RWE; (d) disease burden and unmet need in the target managed-care population. *Section 3 — Economic\nand outcomes evidence*: a **budget impact model** (BIM) customisable to a plan's patient population\nand cost structure; a **cost-effectiveness model** (cost per QALY or cost per outcome); supporting\nHEOR analyses (adherence, resource utilisation, productivity). *Section 4 — Value summary and PROs*:\npatient-reported outcomes evidence (instruments, minimally important differences) and a value summary\nnarrative. *Appendix*: full clinical study reports, model technical reports, and supporting literature.\nVersion 5.0 adds explicit sections on **health equity**, **digital therapeutics**, and **cell and\ngene therapies** to reflect product-type expansion.\n\n**When NOT to use — limitations and common misapplications** — The AMCP Format is a US managed-care\nconvention; it does not substitute for European HTA dossier requirements, FDA regulatory submissions,\nor NICE technology appraisals. Common misapplications: (1) **Confusing the AMCP dossier with a value\nfile** — a value file is an abbreviated sales-aid document for account managers; the AMCP Format is\na comprehensive P&T evidence package requested by medical and pharmacy directors, not a marketing\nleave-behind. (2) **Treating compliance with the format as evidence of quality** — the AMCP Format\norganises content; it does not set the evidentiary bar. A dossier can follow the template exactly\nwhile relying on a poorly designed indirect comparison or an unvalidated claims outcome algorithm;\nP&T reviewers use tools like **ISPOR Indirect Comparisons**, **GRADE**, or **GRACE** to appraise\nindividual sections. (3) **Ignoring the RWE sections** — many manufacturers still treat Sections 2c\nand the economic sections as optional; US payers are increasingly requesting RWE to understand\nreal-world effectiveness and adherence in their own populations. (4) **Using an outdated version** —\nVersions 3.0 (2010) and 4.0 (2016) are no longer the current standard; Version 5.0 (2024) updates\nthe structure for modern product types and evidence expectations. (5) **Applying the AMCP Format to\nEU markets** — European JCA, NICE, and national HTA bodies have their own submission requirements\nthat differ substantially in PICO scope, comparator selection, and economic reference-case assumptions.\n\n**How it maps to this catalog** — Each AMCP Format section draws on a cluster of methods-catalog\nconcepts for its evidence. Section 2 clinical evidence on RWE effectiveness and safety relies on\nthe core pharmacoepidemiology concepts: **active-comparator-new-user** and\n**target-trial-emulation** (rigorous observational design), **propensity-score-methods-psm-iptw**\n(confounding control in claims/EHR studies), and **fit-for-purpose-data-assessment-rwe** (confirming\nthe data source is adequate for the managed-care population). Outcome validity — a common P&T concern\nin claims-based RWE — is addressed by **claims-outcome-algorithm-ppv-sensitivity-rwe** and\n**diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe**. Section 2b indirect comparisons are\ngoverned by **ispor-indirect** (the ISPOR Indirect Treatment Comparisons guideline in this catalog)\nand **prisma-nma** (PRISMA-NMA for the systematic review underpinning the network). Section 3 economic\nmodels follow **ispor-modeling** (good practices for economic modeling) and **ispor-bia** (budget\nimpact analysis). Patient-reported outcomes in Section 4 are appraised via **cosmin-criteria**,\n**cosmin-reporting**, and **isoqol-standards**. The full evidence hierarchy that P&T committees\napply to appraise AMCP dossier submissions can be structured using **grade** and **grace** for\nobservational evidence quality. The US payer-specific design choices underlying a strong AMCP\ndossier RWE section are also informed by **ispor-suitability** (RWD suitability framework) and\nthe **fda-rwe-framework** and **fda-rwd-ehr-claims** guidances (increasingly cited by payers as\nbenchmarks for RWE credibility).", "primary_category": "Guideline", "tags": [ "guideline", "dossier", "formulary-submission", "payer", "managed-care", "budget-impact", "cost-effectiveness", "real-world-evidence", "us-market-access" ], "aliases": [ "AMCP Format", "AMCP dossier", "AMCP Format for Formulary Submissions", "AMCP Format 5.0", "AMCP Format 4.0", "formulary dossier", "payer dossier" ], "applies_to_study_types": [ "cer_observational", "cohort_retrospective", "economic_model", "budget_impact" ], "data_sources": [ "claims", "ehr", "registry", "linked", "primary" ], "citations": [ { "role": "introduce", "doi": "10.18553/jmcp.2024.30.4-b.s1", "url": "https://doi.org/10.18553/jmcp.2024.30.4-b.s1", "citation_text": "AMCP Format for Formulary Submissions 5.0. Journal of Managed Care & Specialty Pharmacy. 2024;30(4-b Suppl):S1.", "year": 2024, "authors_short": "AMCP", "notes": "Current (Version 5.0) AMCP Format statement published as a supplement to JMCP; updates the dossier structure for real-world evidence, health equity, digital therapeutics, and cell/gene therapies." }, { "role": "explain", "doi": "10.18553/jmcp.2016.16092", "url": "https://doi.org/10.18553/jmcp.2016.16092", "citation_text": "Pannier A, Dunn JD. AMCP Format for Formulary Submissions, Version 4.0. Journal of Managed Care & Specialty Pharmacy. 2016;22(4):429-432.", "year": 2016, "authors_short": "Pannier et al.", "notes": "Version 4.0 overview describing the dossier structure; contextualises the evolution from earlier versions to the current 5.0 standard." }, { "role": "use", "url": "https://www.amcp.org", "citation_text": "Academy of Managed Care Pharmacy (AMCP). AMCP Format for Formulary Submissions — maintained dossier standard, templates, and guidance resources.", "year": 2024, "authors_short": "AMCP", "notes": "Official AMCP organisation page; source for the current dossier format, training materials, and updates to submission standards." } ], "relations": [ { "relation_type": "used_with", "target_slug": "ispor-indirect", "notes": "Section 2b indirect treatment comparisons and network meta-analyses in AMCP dossiers should follow ISPOR Indirect Treatment Comparisons good-practice guidance." }, { "relation_type": "used_with", "target_slug": "ispor-bia", "notes": "Section 3 budget impact models follow ISPOR Budget Impact Analysis guidelines." }, { "relation_type": "used_with", "target_slug": "ispor-modeling", "notes": "Section 3 cost-effectiveness models follow ISPOR Economic Model Good Practices." }, { "relation_type": "used_with", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "The Section 2c RWE component requires demonstrating that the data source is adequate for the managed-care population being evaluated." }, { "relation_type": "used_with", "target_slug": "claims-outcome-algorithm-ppv-sensitivity-rwe", "notes": "Claims-based RWE in the dossier requires validated outcome algorithms; P&T reviewers increasingly flag unvalidated diagnosis-code definitions." }, { "relation_type": "see_also", "target_slug": "target-trial-emulation", "notes": "The target-trial framework is the design discipline that produces credible RWE for Section 2c of the dossier." }, { "relation_type": "see_also", "target_slug": "grade", "notes": "GRADE provides the evidence-certainty framework P&T committees apply to appraise clinical evidence across the dossier." }, { "relation_type": "see_also", "target_slug": "grace", "notes": "GRACE is the appraisal tool used to judge the quality of observational CER evidence in Section 2c of the dossier." }, { "relation_type": "see_also", "target_slug": "ispor-suitability", "notes": "ISPOR RWD suitability questionnaire helps assess whether the data source for Section 2c RWE meets evidentiary expectations." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "hta", "journal" ] }, { "slug": "amstar-2", "name": "AMSTAR 2", "short_definition": "A 16-item critical-appraisal instrument that rates the methodological quality of a systematic review of healthcare interventions (with randomized and/or non-randomized included studies) and assigns an overall confidence rating of high, moderate, low, or critically low based on seven critical domains.", "long_description": "**What it is** — AMSTAR 2 (\"A MeaSurement Tool to Assess systematic Reviews,\" version 2) is a critical-appraisal\ninstrument that judges the *methodological quality* of a completed systematic review. It was published by Shea and\ncolleagues in the BMJ in 2017 as a major revision of the original 2007 AMSTAR, broadening the tool so it can appraise\nreviews that include randomized trials, non-randomized studies of interventions (NRSI), or both. AMSTAR 2 has 16 items;\nseven are designated **critical domains** because flaws in them undermine the validity of a review's conclusions\n(Item 2: a pre-registered protocol; Item 4: a comprehensive, reproducible literature search; Item 7: a justified list of\nexcluded studies; Item 9: an appropriate risk-of-bias assessment of the *included* studies; Item 11: appropriate\nmeta-analytic methods; Item 13: accounting for risk of bias when interpreting results; Item 15: an adequate investigation\nof publication/small-study bias). The remaining nine items are non-critical. AMSTAR 2 is maintained by its developer group\n(Bruyère / Ottawa) at amstar.ca and is widely adopted by HTA agencies and methods journals. Critically, it does **not**\nproduce a numeric summary score; the developers deliberately replaced the original AMSTAR sum-score with a structured\nrating of *overall confidence in the results of the review*.\n\n**When to use** — Apply AMSTAR 2 when you must appraise an *existing* systematic review (your own, or one you are\nciting/synthesizing) of healthcare interventions, and you need a defensible, transparent statement of how much confidence\nits findings warrant. This is the appraisal layer for evidence syntheses in HTA/payer dossiers (NICE, CADTH, and similar\nbodies routinely expect AMSTAR 2 for submitted SRs), in peer-reviewed overviews and umbrella reviews, and in regulatory\nevidence packages that lean on published syntheses. Decision rule for choosing AMSTAR 2 vs siblings: use AMSTAR 2 to\nappraise the *quality of the review itself*; use **ROBIS** if you specifically want a risk-of-bias (not quality) judgment\nof the review's process; use **PRISMA 2020 / PRISMA-P** to *report or register* a review (those are reporting checklists,\nnot appraisal tools); use **GRADE** to grade the *certainty of the body of evidence* for each outcome (a different\nquestion from the review's methodological quality). AMSTAR 2 fits SRs of interventions — including reviews that synthesize\nreal-world evidence from claims, EHR, or registry sources — but is not intended for reviews of diagnostic accuracy,\nprognosis, or purely qualitative/scoping reviews.\n\n**What it requires** — AMSTAR 2 enforces, item by item: an *a priori* protocol with pre-specified PICO and analysis plan\n(Item 2); explicit study-design eligibility justification (Item 3); a comprehensive search of at least two databases plus\nsupplementary sources, with strategy and date reported (Item 4); duplicate study selection and data extraction (Items 5–6);\na complete list of excluded full-text studies with reasons (Item 7); adequate description of included studies in PICO\ndetail (Item 8); a satisfactory **risk-of-bias assessment of the included primary studies** using an appropriate tool\n(Item 9) and reporting of funding sources of those studies (Item 10); appropriate methods for statistical combination,\nincluding heterogeneity assessment (Item 11) and assessment of the impact of risk of bias on the pooled estimate\n(Item 12); accounting for individual-study risk of bias when discussing results (Item 13); explanation and investigation\nof heterogeneity (Item 14); an adequate investigation of publication/small-study bias when meta-analysis was performed\n(Item 15); and disclosure of the reviewers' conflicts of interest (Item 16). The overall rating follows an explicit rule based on\nweaknesses in the **seven critical domains** (Items 2, 4, 7, 9, 11, 13, 15): no critical weakness and no more than one\nnon-critical weakness = **high** confidence; more than one non-critical weakness (but no critical flaw) = **moderate**;\nexactly **one** critical flaw (with or without non-critical weaknesses) = **low**; **more than one** critical flaw (with\nor without non-critical weaknesses) = **critically low**. The count of *non-critical* weaknesses never pushes a review\nbelow moderate — only critical-domain flaws drive a Low or Critically-low rating. For an SR that pools\nreal-world data, Item 9 is the hinge — it requires that the included observational studies were appraised with a fit-for-\npurpose risk-of-bias tool (e.g., ROBINS-I), which is where design transparency, time-zero alignment, confounding control,\nand attrition would actually be scrutinized at the primary-study level.\n\n**When NOT to use — limitations and common misapplications** — AMSTAR 2 is an *appraisal* tool for *systematic reviews*,\nand most failures come from using it as something it is not. (1) It is **not a reporting checklist**: completing AMSTAR 2\ndoes not satisfy PRISMA 2020, and a review can be well *reported* yet score critically low, or vice versa. (2) It is\n**not a risk-of-bias instrument for primary studies** — do not apply AMSTAR 2 items to an individual RCT or cohort study;\nthat is the job of RoB 2 or ROBINS-I, and AMSTAR 2 *invokes* those tools at Item 9 rather than replacing them. (3) It is\n**not GRADE**: AMSTAR 2 judges how the review was conducted, not the certainty of the evidence for a given outcome; the two\nare complementary, not interchangeable. (4) **Do not compute a numeric AMSTAR 2 score** — summing \"Yes\" answers\nresurrects the discredited original-AMSTAR sum-score behavior and obscures that a single critical-domain failure should\ncollapse confidence; the developers and subsequent commentary (e.g., Lorenz et al. 2020) emphasize the structured overall\nrating, not an arithmetic total. (5) It is the **wrong tool for the design** if the object of appraisal is a scoping\nreview, narrative review, or a review of diagnostic-test accuracy. (6) **Checklist-as-theater**: marking items \"Yes\"\nwithout verifying the protocol, the excluded-studies list, or the risk-of-bias assessment of included RWE studies produces\na hollow rating that senior reviewers will see through. (7) Appraising a *review of observational evidence* does not make\nthat evidence causal — a high AMSTAR 2 rating certifies the review's methods, not the internal validity of the underlying\nreal-world studies.\n\n**How it maps to this catalog** — AMSTAR 2 lives at the *synthesis* layer, one level above the primary-study methods this\ncatalog documents, so the mapping is by delegation rather than item-for-item. Item 9 requires that the systematic review\nappraised its included studies with an appropriate risk-of-bias tool; when those included studies are real-world\nobservational analyses, that tool (typically ROBINS-I) is what then interrogates exactly the methods catalogued here. An\nAMSTAR 2 appraiser of an RWE-inclusive review should therefore expect the underlying primary studies to have implemented:\na defensible design and time-zero alignment via `active-comparator-new-user` and `target-trial-emulation`; confounding\ncontrol via `high-dimensional-propensity-score-hdps-rwe`; a clearly specified target estimand and handling of intercurrent\nevents via `estimands-ate-att-intercurrent-events-rwe`; validated outcome/exposure definitions via\n`diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe`; and transparent handling of follow-up loss via\n`attrition-and-loss-to-follow-up-rwe`. Item 4's data-source and search transparency, and the data-fitness considerations\nthat determine whether claims/EHR/registry sources can answer the review's question, connect to `claims-analysis`. Applied\nnote for claims/EHR/registry RWE: when you appraise a meta-analysis that pools, say, claims-based cohort studies, a \"Yes\"\non Item 9 is only credible if the review actually examined whether the included studies used incident-user, active-\ncomparator designs with valid phenotype algorithms and pre-specified estimands — i.e., the review's risk-of-bias step\nreached down into the concepts above. If it did not, Item 9 is a critical weakness and overall confidence drops to low or\ncritically low regardless of how polished the review's prose is.", "primary_category": "Guideline", "tags": [ "guideline", "critical-appraisal", "quality-assessment", "systematic-review", "meta-analysis", "evidence-synthesis", "risk-of-bias" ], "aliases": [ "AMSTAR 2", "AMSTAR-2", "A MeaSurement Tool to Assess systematic Reviews 2" ], "applies_to_study_types": [ "systematic_review", "meta_analysis_rct", "meta_analysis_obs" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "doi": "10.1136/bmj.j4008", "url": "https://doi.org/10.1136/bmj.j4008", "citation_text": "Shea BJ, Reeves BC, Wells G, et al. AMSTAR 2: a critical appraisal tool for systematic reviews that include randomised or non-randomised studies of healthcare interventions, or both. BMJ. 2017;358:j4008.", "year": 2017, "authors_short": "Shea et al.", "notes": "Canonical statement paper defining the 16 items, the seven critical domains, and the high/moderate/low/critically-low overall confidence rating. Supersedes the original AMSTAR (Shea et al., BMC Med Res Methodol 2007; 10.1186/1471-2288-7-10), whose numeric sum-score the developers deliberately abandoned." }, { "role": "explain", "doi": "10.1016/j.jclinepi.2019.10.006", "url": "https://doi.org/10.1016/j.jclinepi.2019.10.006", "citation_text": "Lorenz RC, Matthias K, Pieper D, et al. AMSTAR 2 overall confidence rating: lacking discriminating capacity or requirement of high methodological quality? Journal of Clinical Epidemiology. 2020;119:142-144.", "year": 2020, "authors_short": "Lorenz et al.", "notes": "Methods commentary on how the overall confidence rating behaves in practice and why a single critical-domain failure (not an additive score) governs the rating." }, { "role": "use", "url": "https://amstar.ca/", "citation_text": "AMSTAR — A MeaSurement Tool to Assess systematic Reviews. Maintained instrument, item guidance, and online appraisal form (amstar.ca).", "year": 2017, "authors_short": "AMSTAR developer group", "notes": "Maintained tool, scoring guidance, and the online rating instrument used to apply AMSTAR 2." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "systematic-review", "notes": "Primary use case — appraising the methodological quality of a completed systematic review." }, { "relation_type": "applies_to", "target_slug": "meta-analysis-rct", "notes": "Applies to reviews that include and pool randomized controlled trials." }, { "relation_type": "applies_to", "target_slug": "meta-analysis-obs", "notes": "Applies to reviews that include non-randomized/observational studies, where Item 9 should invoke ROBINS-I for the included RWE studies." }, { "relation_type": "see_also", "target_slug": "claims-analysis", "notes": "Item 4 (search/data-source transparency) and data-fitness judgments connect here when the review synthesizes claims/EHR/registry evidence." }, { "relation_type": "see_also", "target_slug": "target-trial-emulation", "notes": "Methods that AMSTAR 2 Item 9 expects the included real-world studies to have used; appraised at the primary-study level via ROBINS-I, not by AMSTAR 2 directly." }, { "relation_type": "see_also", "target_slug": "active-comparator-new-user", "notes": "Design quality of included observational studies that a credible Item 9 risk-of-bias assessment should have scrutinized." }, { "relation_type": "see_also", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "Confounding-control method expected in well-conducted included RWE studies; relevant to Item 9 and Item 13 judgments." }, { "relation_type": "see_also", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Estimand clarity in included studies informs whether the review pooled comparable contrasts (Items 9, 11)." }, { "relation_type": "see_also", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Validity of outcome/exposure definitions in included claims/EHR studies that a credible risk-of-bias assessment should have considered." }, { "relation_type": "see_also", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Follow-up completeness in included studies, relevant to risk-of-bias appraisal at Item 9." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "hta", "journal" ] }, { "slug": "bradford-hill", "name": "Bradford Hill Considerations for Causation", "short_definition": "A set of nine viewpoints (strength, consistency, specificity, temporality, biological gradient, plausibility, coherence, experiment, analogy) proposed by Austin Bradford Hill in 1965 for judging whether an observed exposure-outcome association is best read as causal rather than as the product of chance, bias, or confounding. It is a structured heuristic for causal argument, not a reporting checklist or a quality score.", "long_description": "**What it is.** The Bradford Hill considerations (often loosely called the \"Bradford Hill criteria\")\nare nine *aspects of an association* that Sir Austin Bradford Hill set out in his 1965 presidential\naddress to the Section of Occupational Medicine of the Royal Society of Medicine, \"The Environment and\nDisease: Association or Causation?\". They are: (1) **strength** of the association; (2) **consistency**\nacross persons, places, circumstances, and investigators; (3) **specificity** of the\nexposure-outcome link; (4) **temporality** (the exposure must precede the outcome); (5) **biological\ngradient** (dose-response); (6) **plausibility** given existing biological knowledge; (7)\n**coherence** with the natural history and biology of the disease; (8) **experiment** (does removing\nor reducing the exposure reduce the outcome?); and (9) **analogy** with established cause-effect\nrelationships. Hill was explicit that these are *viewpoints to aid judgment*, not a checklist and not\na set of hard-and-fast rules — \"None of my nine viewpoints can bring indisputable evidence for or\nagainst the cause-and-effect hypothesis and none can be required as a sine qua non.\" The framework has\nno maintaining standards body (it is not an EQUATOR reporting guideline, a Cochrane risk-of-bias tool,\nan ISPOR good-practice report, or an agency guidance); it is a foundational piece of epidemiologic\nreasoning that has been re-read and formalized many times, most usefully through the modern\ncounterfactual and directed-acyclic-graph (DAG) lens.\n\n**When to use.** Reach for the Bradford Hill considerations at the *interpretation and discussion*\nstage of a non-interventional study, when you have a well-estimated association and must argue —\ntransparently and against alternative explanations — whether it supports a causal claim. They are\nappropriate as a structuring device for the causal-argument section of a peer-reviewed manuscript, an\nHTA/payer dossier that leans on observational comparative effectiveness, an FDA/EMA submission where\nRWE is offered as causal evidence, a safety-signal causality assessment, and pharmacoepidemiology\nprograms that triangulate across designs and data sources. Decision rule: use Bradford Hill to\n*organize and stress-test a causal narrative across a body of evidence*; do **not** use it as the\nprimary tool for any of the upstream jobs that have purpose-built instruments. For *reporting*\ntransparency use STROBE/RECORD(-PE) or HARPER; for *risk-of-bias of a single non-randomized study*\nuse ROBINS-I; for *certainty of a body of evidence* use GRADE; for *the design that licenses a causal\nestimate in the first place* use a target-trial emulation with a pre-specified protocol. Bradford\nHill complements these — it does not replace any of them.\n\n**What it requires (read for real-world data).** Treated rigorously, the nine considerations demand\nevidence that maps directly onto modern RWE practice. **Temporality** is the one non-negotiable\nconsideration and forces explicit *time-zero alignment*, a clean lookback/washout, an incident (new-user)\nexposure definition, and attention to immortal-time and induction/latency windows so the exposure\ndemonstrably precedes the outcome. **Strength** must be a *well-confounded-adjusted* estimate, not a\ncrude one — so confounding control (active-comparator new-user design, propensity or high-dimensional\npropensity scores, DAG-guided covariate selection), a clearly stated estimand and handling of\nintercurrent events, and quantification of residual confounding (E-value, negative controls) all sit\nunderneath \"strength.\" **Biological gradient** requires defensible dose/duration measurement from\nfills or administrations. **Consistency** is the engine of triangulation: reproducibility across\ndatabases, designs (cohort, case-control, self-controlled), and populations, which in turn depends on\nvalidated phenotypes/algorithms (with PPV/sensitivity), fit-for-purpose data assessment, and honest\naccounting of attrition and missing data so that \"consistency\" is not just shared bias. **Experiment**\nin RWE maps to quasi-experiments and natural experiments (policy changes, formulary shifts) and to\nthe as-if-randomized logic of target-trial emulation. **Plausibility, coherence, specificity, and\nanalogy** are judgment-laden and weak as discriminators, but they are where mechanism, prior\nevidence, and sensitivity/quantitative-bias analyses are marshaled.\n\n**When NOT to use — limitations and common misapplications.** The dominant failure mode is treating\nthe nine considerations as a *checklist or additive score*: counting how many are \"met\" and declaring\ncausation above some threshold. Hill warned against exactly this; the considerations are not\nindependent, not equally weighted, and most are neither necessary nor sufficient. Specific traps:\n(1) **Specificity** is largely obsolete — most exposures have many effects and most diseases many\ncauses, so its absence says little. (2) **Consistency** can reflect a bias shared across studies\n(e.g., the same misclassified claims phenotype reused everywhere), so replication is reassuring only\nif the studies do not share the same flaw. (3) **Plausibility/coherence/analogy** are bounded by the\nknowledge of the day and invite confirmation bias; absence of a known mechanism is not evidence of no\neffect. (4) Using Bradford Hill as a substitute for design — applying it to a crude, confounded\nassociation to launder it into a causal claim — is the cardinal misuse; it cannot repair confounding\nby indication, immortal time, or selection bias baked into the study. (5) It is **not** a risk-of-bias\ninstrument (use ROBINS-I), **not** a certainty-of-evidence grading system (use GRADE), and **not** a\nreporting checklist (use STROBE/RECORD/HARPER); presenting a \"Bradford Hill table\" in lieu of those\nis checklist theater. (6) Reverse causation and collider/selection structures can mimic several\nconsiderations at once and are best surfaced with explicit DAGs, not narrative coherence.\n\n**How it maps to this catalog.** Each consideration is implemented by specific concepts here.\n*Temporality* -> `time-zero-index-date-alignment-rwe`, `washout-clean-lookback-period-rwe`,\n`new-user-design`, `immortal-time-bias-handling`, `exposure-lag-induction-latency-window-rwe`.\n*Strength (properly adjusted)* -> `active-comparator-new-user`, `propensity-score-methods-psm-iptw`,\n`high-dimensional-propensity-score-hdps-rwe`, `dags-backdoor-criterion-drug-studies`,\n`estimands-ate-att-intercurrent-events-rwe`, with the strongest causal scaffold being\n`target-trial-emulation`. *Biological gradient* -> `time-updated-exposures-cumulative-dose-rwe`,\n`exposure-episode-construction-rwe`. *Consistency / triangulation* -> `meta-analysis-obs`,\n`self-controlled-case-series`, `case-control`, plus the validity substrate of\n`diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe`,\n`claims-outcome-algorithm-ppv-sensitivity-rwe`, `algorithm-validation`,\n`fit-for-purpose-data-assessment-rwe`, and `attrition-and-loss-to-follow-up-rwe`. *Robustness behind\nthe judgment* -> `e-value-sensitivity-analysis`, `negative-control-outcomes-rwe`,\n`empirical-calibration-negative-controls-rwe`, `quantitative-bias-analysis-toolkit-rwe`,\n`unmeasured-confounding-probabilistic-bias-analysis-rwe`. Applied note for claims/EHR/registry RWE:\nbefore a single Bradford Hill consideration is invoked, the underlying estimate must rest on a\nvalidated outcome phenotype (report PPV and, where feasible, sensitivity), a real (not missing)\ndrug-free washout supported by continuous enrollment, time-zero set at initiation, and a pre-specified\nestimand. \"Strength\" then means the confounding-adjusted comparative estimate with an E-value and\nnegative-control calibration attached; \"consistency\" means the result holds across at least one\nindependent database or design that does not share the same phenotype/data weakness. Used this way,\nBradford Hill is a disciplined argument layered on top of a defensible design — never a shortcut\naround one.", "primary_category": "Guideline", "tags": [ "guideline", "causal-inference", "causation", "pharmacoepidemiology", "triangulation", "methodological" ], "aliases": [ "Bradford Hill criteria", "Hill criteria", "Hill's criteria for causation", "Bradford Hill viewpoints", "criteria for causation" ], "applies_to_study_types": [ "cohort_prospective", "cohort_retrospective", "case_control", "meta_analysis_obs" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "doi": "10.1177/003591576505800503", "url": "https://doi.org/10.1177/003591576505800503", "citation_text": "Hill AB. The environment and disease: association or causation? Proceedings of the Royal Society of Medicine. 1965;58(5):295-300.", "year": 1965, "authors_short": "Hill", "notes": "The original 1965 address introducing the nine considerations; Hill explicitly frames them as aids to judgment, not a checklist or set of necessary rules." }, { "role": "explain", "doi": "10.1186/1742-7622-2-11", "url": "https://doi.org/10.1186/1742-7622-2-11", "citation_text": "Höfler M. The Bradford Hill considerations on causality: a counterfactual perspective. Emerging Themes in Epidemiology. 2005;2:11.", "year": 2005, "authors_short": "Höfler", "notes": "Reinterprets each consideration through the modern counterfactual framework, clarifying which are weak (specificity, analogy) and why temporality is the only essential one." }, { "role": "use", "doi": "10.1186/s12982-015-0037-4", "url": "https://doi.org/10.1186/s12982-015-0037-4", "citation_text": "Fedak KM, Bernal A, Capshaw ZA, Gross S. Applying the Bradford Hill criteria in the 21st century: how data integration has changed causal inference in molecular epidemiology. Emerging Themes in Epidemiology. 2015;12:14.", "year": 2015, "authors_short": "Fedak et al.", "notes": "Contemporary application showing how triangulation across data sources and mechanistic evidence operationalizes the considerations, and warning against checklist-style use." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "cohort-prospective", "notes": "Use to structure the causal-argument/discussion of a prospective cohort once the association is estimated and confounding addressed." }, { "relation_type": "applies_to", "target_slug": "cohort-retrospective", "notes": "Common setting for claims/EHR cohorts; temporality and strength carry the weight, specificity and analogy little." }, { "relation_type": "applies_to", "target_slug": "case-control", "notes": "Useful for triangulation, but watch reverse causation and recall/selection structures that can mimic several considerations." }, { "relation_type": "applies_to", "target_slug": "meta-analysis-obs", "notes": "Consistency across studies is the central consideration here; reassuring only when studies do not share the same bias." }, { "relation_type": "see_also", "target_slug": "target-trial-emulation", "notes": "The design that actually licenses a causal estimate; Bradford Hill interprets results, it does not substitute for emulating the trial." }, { "relation_type": "see_also", "target_slug": "dags-backdoor-criterion-drug-studies", "notes": "DAGs make the temporality, confounding, and collider structure explicit, replacing narrative \"coherence/plausibility\" with a formal causal model." }, { "relation_type": "used_with", "target_slug": "e-value-sensitivity-analysis", "notes": "Quantifies how robust the \"strength\" consideration is to unmeasured confounding." }, { "relation_type": "used_with", "target_slug": "negative-control-outcomes-rwe", "notes": "Calibrates the estimate underpinning \"strength\" and detects residual systematic error before causal interpretation." }, { "relation_type": "used_with", "target_slug": "time-zero-index-date-alignment-rwe", "notes": "Operationalizes the non-negotiable temporality consideration in real-world data." }, { "relation_type": "see_also", "target_slug": "estimand-analysis-traceability-rwe", "notes": "For evidence appraisal, modern risk-of-bias (ROBINS-I) and certainty (GRADE) tools, plus a traceable estimand framework, supersede checklist-style use of the nine considerations." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "hta", "journal" ] }, { "slug": "cadth-methods", "name": "CADTH (CDA-AMC) Guidelines for the Economic Evaluation of Health Technologies: Canada", "short_definition": "The Canadian HTA reference case (4th Edition, 2017) — a methods manual that specifies the recommended, standardized approach for economic evaluations submitted to Canada's Drug Agency (CDA-AMC, formerly CADTH), covering perspective, comparators, time horizon, discounting, modelling, effectiveness inputs, costing, and uncertainty.", "long_description": "**What it is** — The *Guidelines for the Economic Evaluation of Health Technologies:\nCanada* (4th Edition, 2017) is the Canadian health technology assessment (HTA)\n**reference case**: a normative methods manual that prescribes the recommended,\nstandardized methods a submitter should follow when conducting an economic evaluation\nfor review by Canada's Drug Agency (CDA-AMC, the agency formerly known as CADTH). It is\nthe Canadian counterpart to the NICE reference case in England and ICER's value\nframework in the US. The reference case exists to make submissions *comparable* across\ntechnologies and decision problems — it fixes the analytic conventions (perspective,\ncomparators, time horizon, discount rate, outcome metric, handling of uncertainty) so\nthat reviewers can interpret an incremental cost-effectiveness ratio without first\nhaving to relitigate every methodological choice. It is maintained by CDA-AMC and was\ndeveloped through a public, multi-stakeholder consultation process. It is a *methods*\nguideline (how to build and parameterize the model), not a reporting checklist and not\na critical-appraisal/risk-of-bias instrument.\n\n**When to use** — Use this reference case whenever you are preparing or appraising a\ncost-effectiveness, cost-utility, cost-minimization, or budget-impact analysis intended\nfor a **Canadian HTA or pan-Canadian reimbursement decision** (CDA-AMC Reimbursement\nReview / pCPA-bound submissions). It governs the base-case structure: a public\nhealth-care-payer perspective, a time horizon long enough to capture all relevant\ndifferences in costs and outcomes, QALYs as the primary outcome for cost-utility,\nprescribed discounting of costs and effects, and explicit characterization of\nuncertainty. Decision rule for *which* guideline applies: use **this** document for the\nCanadian economic base case; use **nice-reference-case** for England/NICE submissions\n(different perspective and discount conventions); use **cheers-2022** to *report* the\ncompleted economic evaluation (CHEERS is the reporting checklist, not the methods\nreference case); and use the CDA-AMC RWE reporting guidance (Tadrous 2024, cited below)\nwhen the *inputs* to your model are drawn from real-world data and must be reported to a\nCanadian regulator/HTA standard.\n\n**What it requires** — The reference case enforces a defined set of methodological\ndomains: (1) a clearly framed decision problem and target population; (2) the choice of\nevaluation type (cost-utility/CEA preferred, with cost-minimization or budget impact\nwhere appropriate); (3) all *relevant* comparators reflecting current Canadian practice,\nnot just the convenient one; (4) the public health-care-payer perspective for the base\ncase (with broader societal perspectives as scenario analyses); (5) an appropriate time\nhorizon and prescribed discounting of both costs and effects; (6) transparent,\nvalidated decision-analytic modelling; (7) effectiveness and safety inputs derived from\nthe best available evidence with explicit handling of indirect/network comparisons and\nextrapolation; (8) Canadian-context measurement and valuation of health (QALYs via a\nrecognized utility instrument) and of resource use and unit costs; and (9) rigorous\ncharacterization of uncertainty — deterministic *and* probabilistic sensitivity analysis\n(PSA), structural sensitivity, and scenario analyses — plus consideration of equity. For\nreal-world-data inputs feeding the model (single-arm-trial external comparators from\nclaims/EHR/registry, real-world costs, real-world utilization), the same fitness-for-use\nand bias discipline expected of any RWE applies: data fitness assessment, validated\nphenotypes, transparent time-zero and estimand definitions, confounding control, and\nquantitative bias/sensitivity analysis around the effectiveness and cost parameters.\n\n**When NOT to use — limitations and common misapplications** — (1) **It is not a\nreporting checklist.** Completing a base case that conforms to the reference case does\nnot satisfy a journal or HTA reporting requirement — report the study with CHEERS 2022\n(`cheers-2022`). (2) **It is not a critical-appraisal or risk-of-bias tool.** It tells\nyou how to *build* a defensible economic model, not how to *grade* someone else's\nobservational evidence — use ROBINS-I / GRACE / AMSTAR-style instruments for that. (3)\n**It is jurisdiction-specific.** Applying the Canadian perspective, discount rate, and\ncosting conventions to a NICE, IQWiG, or US/ICER submission is a category error; each\njurisdiction has its own reference case. (4) **Conforming to the reference case does not\nmake biased inputs valid.** A pristine model architecture fed by an unadjusted, immortal-\ntime-biased, or poorly phenotyped real-world effectiveness estimate produces a precisely\ncomputed wrong answer — the reference case constrains the *model*, not the *credibility\nof the RWE feeding it*. (5) **Reference-case-as-theater:** presenting only the base case\nwhile burying unfavorable structural and parameter uncertainty in an appendix defeats\nthe purpose; PSA and scenario analyses are mandatory, not decorative. (6) Do not confuse\nthis economic-methods reference case with CDA-AMC's separate *Guidance for Reporting\nReal-World Evidence* (a reporting standard for RWE submissions) — they are complementary,\nnot interchangeable.\n\n**How it maps to this catalog** — The economic-modelling machinery is implemented by\n`health-economic-modeling-methods-rwe` (model structure and parameterization),\n`discounting-costs-effects-rwe` (the prescribed discounting of costs and effects), and\n`probabilistic-sensitivity-analysis-hea-rwe` (the mandatory PSA characterizing parameter\nuncertainty). The Canadian costing requirements are operationalized by\n`healthcare-costs-pppm-pppy-pmpm` (real-world cost aggregation),\n`all-cause-vs-attributable-costs-rwe` (whether to count all-cause or disease-attributable\ncosts), and `cost-outlier-handling-rwe` (treatment of catastrophic/outlier costs in\nskewed real-world cost distributions). When real-world data supply the effectiveness or\nexternal-comparator inputs, fitness and bias are governed by\n`fit-for-purpose-data-assessment-rwe` (data fitness-for-use),\n`quantitative-bias-analysis-toolkit-rwe` and `negative-control-outcomes-rwe`\n(quantitative bias / residual-confounding probes around the effectiveness parameter),\nwith comparative-effectiveness inputs typically built via `active-comparator-new-user`\nand `target-trial-emulation`. The Canadian-data note: claims/administrative inputs for a\nCanadian model differ structurally from US Medicare/commercial claims (single-payer\nprovincial coverage, different fee schedules, different drug-plan capture); contrast with\n`medicare-ffs-ma-commercial-claims-differences-rwe` before importing US-derived utilization\nor cost estimates into a Canadian reference-case base case. The closest sibling guideline\nin this catalog is `nice-reference-case` (the England/NICE analogue), with `cheers-2022`\nas the reporting layer.", "primary_category": "Guideline", "tags": [ "guideline", "methodological", "hta", "reference-case", "economic-evaluation", "cost-effectiveness", "canada", "cda-amc" ], "aliases": [ "CADTH Methods Guide", "CADTH economic evaluation guidelines", "Guidelines for the Economic Evaluation of Health Technologies: Canada", "CDA-AMC reference case", "Canadian HTA reference case" ], "applies_to_study_types": [ "cost_effectiveness", "cost_utility", "cost_minimization", "budget_impact" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "url": "https://www.cda-amc.ca/guidelines-economic-evaluation-health-technologies-canada-4th-edition", "citation_text": "CADTH. Guidelines for the Economic Evaluation of Health Technologies: Canada. 4th ed. Ottawa: CADTH (now Canada's Drug Agency, CDA-AMC); 2017.", "year": 2017, "authors_short": "CADTH (CDA-AMC)", "notes": "Canonical Canadian HTA economic reference case (4th Edition, 2017). Agency methods manual; no single journal statement paper, so the stable CDA-AMC landing page is the authoritative source." }, { "role": "use", "url": "https://www.cda-amc.ca/sites/default/files/pdf/guidelines_for_the_economic_evaluation_of_health_technologies_canada_4th_ed.pdf", "citation_text": "CADTH. Guidelines for the Economic Evaluation of Health Technologies: Canada — 4th Edition (full PDF). Ottawa: CADTH; 2017.", "year": 2017, "authors_short": "CADTH (CDA-AMC)", "notes": "Full text of the 4th-edition guidelines, including the reference-case table of recommended methods." }, { "role": "use", "doi": "10.1016/j.jclinepi.2024.111545", "url": "https://doi.org/10.1016/j.jclinepi.2024.111545", "citation_text": "Tadrous M, Aves T, Fahim C, et al. Development of a Canadian Guidance for reporting real-world evidence for regulatory and health-technology assessment (HTA) decision-making. Journal of Clinical Epidemiology. 2024;176:111545.", "year": 2024, "authors_short": "Tadrous et al.", "notes": "Companion CDA-AMC/Health Canada guidance for reporting real-world evidence inputs to Canadian HTA and regulatory decisions; use alongside the economic reference case when RWE supplies effectiveness, utilization, or cost parameters." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "cost-effectiveness", "notes": "Reference case for the Canadian cost-effectiveness base case." }, { "relation_type": "applies_to", "target_slug": "cost-utility", "notes": "QALY-based cost-utility analysis is the preferred form under the reference case." }, { "relation_type": "applies_to", "target_slug": "cost-minimization", "notes": "Permitted where comparators are established to be of equivalent effectiveness." }, { "relation_type": "applies_to", "target_slug": "budget-impact", "notes": "Budget-impact analysis accompanies the economic evaluation for Canadian reimbursement review." }, { "relation_type": "alternative_to", "target_slug": "nice-reference-case", "notes": "The England/NICE reference case is the analogous jurisdiction-specific HTA methods standard; perspective, discounting, and costing conventions differ — do not interchange them across jurisdictions." }, { "relation_type": "see_also", "target_slug": "cheers-2022", "notes": "CHEERS 2022 is the reporting checklist for the completed economic evaluation; this reference case governs the methods, CHEERS governs the report." }, { "relation_type": "used_with", "target_slug": "health-economic-modeling-methods-rwe", "notes": "Implements the decision-analytic modelling the reference case requires (structure, parameterization, validation)." }, { "relation_type": "used_with", "target_slug": "discounting-costs-effects-rwe", "notes": "Implements the prescribed discounting of costs and health effects." }, { "relation_type": "used_with", "target_slug": "probabilistic-sensitivity-analysis-hea-rwe", "notes": "PSA characterizing parameter uncertainty is a mandatory component of the base case." }, { "relation_type": "used_with", "target_slug": "healthcare-costs-pppm-pppy-pmpm", "notes": "Operationalizes real-world cost measurement feeding the economic model." }, { "relation_type": "used_with", "target_slug": "all-cause-vs-attributable-costs-rwe", "notes": "Informs whether all-cause or disease-attributable costs enter the model." }, { "relation_type": "used_with", "target_slug": "cost-outlier-handling-rwe", "notes": "Handles skewed real-world cost distributions and catastrophic outliers in costing inputs." }, { "relation_type": "used_with", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "Governs fitness-for-use of real-world data supplying effectiveness, utilization, or cost inputs to the model." }, { "relation_type": "used_with", "target_slug": "quantitative-bias-analysis-toolkit-rwe", "notes": "Quantifies residual confounding and bias around RWE-derived effectiveness parameters." }, { "relation_type": "see_also", "target_slug": "medicare-ffs-ma-commercial-claims-differences-rwe", "notes": "Caution before importing US claims-derived utilization/cost estimates into a Canadian single-payer reference-case base case." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "hta", "journal" ] }, { "slug": "care", "name": "CARE (CAse REport guidelines)", "short_definition": "A 13-item EQUATOR reporting checklist for transparent, complete write-up of a single clinical case report (or small case series), centered on a structured patient timeline. It is a reporting tool, not a design, risk-of-bias, or causal-inference instrument.", "long_description": "**What it is.** The **CARE (CAse REport) guidelines** are a 13-item reporting checklist for clinical case\nreports, developed by consensus in 2013 (Gagnier et al.) and elaborated in 2017 (Riley et al.). CARE is hosted by the\n**EQUATOR Network** alongside its sibling reporting guidelines (STROBE, CONSORT, PRISMA, STARD). Its purpose is narrow and\nspecific: to make the narrative of a *single patient* (or a very small series) transparent and complete enough that a\nreader can judge what happened, in what order, and why the author believes the case is informative. The signature CARE item\nis the **timeline** — a chronological figure or table that aligns presentation, diagnostics, interventions, and outcomes —\nwhich is what distinguishes a disciplined case report from an anecdote. The checklist also covers title, key words, abstract,\nintroduction, de-identified patient information, clinical findings, diagnostic assessment (and diagnostic challenges),\ntherapeutic intervention, follow-up and outcomes, discussion (strengths/limitations against the literature), patient\nperspective, and informed consent. CARE governs *how a case is written up*; it makes no claim about study validity, effect\nestimation, or generalizability.\n\n**When to use.** Use CARE when the unit of evidence is a clinical narrative about one patient (occasionally a handful):\nan unexpected adverse drug event, a novel presentation, an unusual response to therapy, a rare disease, or a diagnostic\npuzzle. The primary decision context is **peer-reviewed journal submission** — most journals that publish case reports now\nrequest the CARE checklist. Decision rule for choosing CARE versus a sibling guideline: if the report describes **one\npatient's clinical course as a story**, CARE applies. The moment the design involves a *defined cohort, comparison group,\ndenominator, or analytic estimate* — even a descriptive multi-patient series analyzed as data — you have left CARE's scope\nand entered STROBE/RECORD/MOOSE territory. If the case is an individual adverse-drug-reaction report destined for a\npharmacovigilance system or a regulatory safety database, the governing framework is **ICH E2D / CIOMS / GVP**, not CARE,\nalthough CARE's timeline discipline still improves the narrative.\n\n**What it requires.** CARE enforces narrative completeness and chronological transparency rather than analytic rigor.\nThe substantive item groups that matter most:\n- **Structured timeline (the defining item):** a chronological alignment of symptom onset, encounters, diagnostics,\n interventions (with dates, doses, durations), and outcomes — the single best defense against post-hoc storytelling.\n- **Diagnostic assessment and challenges:** the tests performed, differential diagnosis, reasoning, and what made the\n diagnosis difficult — so a reader can appraise diagnostic certainty.\n- **Therapeutic intervention and outcomes with follow-up:** exactly what was given and what followed, including adverse\n events and the duration and completeness of follow-up.\n- **Discussion against the literature:** why this case adds knowledge, and an explicit statement of limitations —\n crucially, the inferential limits of an n-of-1 observation.\n- **De-identification, informed consent, and patient perspective:** ethical and patient-centered reporting requirements.\nFor a case identified or assembled from an EHR, CARE-quality reporting demands that the chart-derived timeline be\nreconstructed faithfully (encounter dates, medication start/stop, lab values) rather than summarized loosely.\n\n**When NOT to use — limitations and common misapplications.**\n- **CARE is a reporting checklist, NOT a risk-of-bias tool and NOT a quality score.** A fully completed CARE checklist says\n the report is *transparent*, not that the case is *valid* or that any inference from it is sound. There is no total score.\n- **A single case cannot support a causal or comparative claim.** With n=1 there is no comparator, no denominator, and no\n control of confounding; CARE compliance does nothing to license \"Drug X caused Y\" or \"X is more effective than Z.\"\n Treating a polished case report as causal evidence is the most common and most damaging misuse.\n- **Wrong unit / wrong guideline.** A *case series* analyzed as multi-patient data is observational research and should be\n reported under **STROBE** (or **RECORD/RECORD-PE** for routinely collected health data, **MOOSE** for synthesis), not\n CARE. Using CARE to dress up a small cohort hides the absence of a denominator and design.\n- **Pharmacovigilance confusion.** An individual case safety report (ICSR) for a suspected adverse drug reaction is\n governed by **ICH E2D / CIOMS / EMA GVP** causality and minimum-criteria conventions; CARE is not a pharmacovigilance\n standard and should not be substituted for one in a safety submission.\n- **Checklist-as-theater.** Submitting the CARE checklist while omitting the timeline figure, dates, doses, or consent\n defeats the purpose; the timeline in particular is frequently the missing item that reviewers should demand.\n\n**How it maps to this catalog.** CARE sits at the descriptive, hypothesis-generating edge of the evidence hierarchy, so\nmost analytic RWE concepts in this catalog do **not** implement it — and saying so plainly is part of using CARE correctly.\nThe genuinely relevant cross-references are narrow:\n- **case-report** — the study-type concept CARE directly governs; start here.\n- **case-series** and **descriptive-epidemiology-rwe** — the adjacent designs you move to when the unit becomes multiple\n patients, at which point STROBE (not CARE) is the reporting guideline.\n- **endpoint-adjudication-chart-review-rwe** — the chart-review discipline that underpins a faithful CARE diagnostic\n assessment and timeline when the case is reconstructed from records.\n- **safety-signal-case-definition-rwe** — relevant when a case report seeds a pharmacovigilance signal; note the handoff to\n ICH E2D/CIOMS for the formal ICSR.\n- **diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe** — applicable *only* in the narrow situation where an EHR\n phenotype was used to surface the index case; this is a genuine stretch and should be cited as such, not as a core\n requirement. CARE does **not** map to target-trial emulation, propensity scores, hdPS, estimands, or active-comparator\n designs — those require comparison groups CARE explicitly lacks.\n\n**Applied note (EHR/chart-based case reports).** When a case is drawn from an electronic record, CARE's value is forcing a\ndefensible timeline: pin each event to a source date (encounter, order, fill, lab result), distinguish ordered from\nadministered therapy, and reconcile discrepant dates before narrating causation. The same chart-review rigor used for\noutcome adjudication in cohort RWE is what makes an EHR-sourced case report trustworthy — but it remains a single\nobservation, and the discussion must say so.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "case-report", "equator", "descriptive" ], "aliases": [ "CARE", "CAse REport guidelines", "CARE Statement", "CARE checklist" ], "applies_to_study_types": [ "case_report" ], "data_sources": [ "ehr", "primary" ], "citations": [ { "role": "introduce", "doi": "10.1016/j.jclinepi.2013.08.003", "url": "https://doi.org/10.1016/j.jclinepi.2013.08.003", "citation_text": "Gagnier JJ, Kienle G, Altman DG, Moher D, Sox H, Riley D; CARE Group. The CARE guidelines: consensus-based clinical case report guideline development. Journal of Clinical Epidemiology. 2014;67(1):46-51.", "year": 2014, "authors_short": "Gagnier et al.", "notes": "Canonical consensus statement defining the 13-item CARE checklist (co-published across multiple journals in 2013)." }, { "role": "introduce", "doi": "10.7453/gahmj.2013.008", "url": "https://doi.org/10.7453/gahmj.2013.008", "citation_text": "Gagnier JJ, Kienle G, Altman DG, Moher D, Sox H, Riley D; CARE Group. The CARE guidelines: consensus-based clinical case reporting guideline development. Global Advances in Health and Medicine. 2013;2(5):38-43.", "year": 2013, "authors_short": "Gagnier et al.", "notes": "Co-publication of the original CARE statement in its 2013 simultaneous-release venue." }, { "role": "explain", "doi": "10.1016/j.jclinepi.2017.04.026", "url": "https://doi.org/10.1016/j.jclinepi.2017.04.026", "citation_text": "Riley DS, Barber MS, Kienle GS, et al. CARE guidelines for case reports: explanation and elaboration document. Journal of Clinical Epidemiology. 2017;89:218-235.", "year": 2017, "authors_short": "Riley et al.", "notes": "Item-by-item elaboration with rationale and worked examples, including the structured timeline." }, { "role": "use", "url": "https://www.equator-network.org/reporting-guidelines/care/", "citation_text": "CARE — CAse REport guidelines, EQUATOR Network (maintained checklist, timeline template, and translations).", "year": 2013, "authors_short": "EQUATOR Network", "notes": "Authoritative landing page for the current checklist and writing templates." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "case-report", "notes": "CARE is the reporting checklist for single-patient clinical case reports." }, { "relation_type": "see_also", "target_slug": "case-report", "notes": "The study-type concept CARE directly governs; the natural entry point for implementation." }, { "relation_type": "alternative_to", "target_slug": "case-series", "notes": "Once the unit is multiple patients analyzed together, the report moves from CARE to STROBE-style observational reporting; CARE does not cover denominators or comparison." }, { "relation_type": "see_also", "target_slug": "descriptive-epidemiology-rwe", "notes": "Descriptive multi-patient summaries leave CARE's scope and require design/denominator reporting CARE does not address." }, { "relation_type": "used_with", "target_slug": "endpoint-adjudication-chart-review-rwe", "notes": "Chart-review discipline underpins a faithful CARE diagnostic assessment and timeline for EHR-sourced cases." }, { "relation_type": "see_also", "target_slug": "safety-signal-case-definition-rwe", "notes": "A case report may seed a pharmacovigilance signal, but the formal individual case safety report is governed by ICH E2D / CIOMS / GVP, not CARE." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "journal" ] }, { "slug": "cheers-2022", "name": "CHEERS 2022", "short_definition": "The Consolidated Health Economic Evaluation Reporting Standards 2022 — a 28-item reporting checklist for the transparent reporting of any health economic evaluation (cost-effectiveness, cost-utility, cost-benefit, cost-minimization), maintained by ISPOR and listed on the EQUATOR Network. It governs how an economic evaluation is reported, not how the underlying clinical or real-world evidence was generated.", "long_description": "**What it is** — CHEERS 2022 is the **Consolidated Health Economic Evaluation Reporting\nStandards**, the consensus reporting guideline for health economic evaluations. It is a\n**28-item checklist** developed and maintained by an ISPOR (Professional Society for Health\nEconomics and Outcomes Research) Good Practices Task Force (the \"CHEERS II\" task force) and\nis registered on the **EQUATOR Network** library of reporting guidelines. The 2022 update\nreplaces the 2013 statement, expanding and clarifying items on perspective, model structure,\nheterogeneity and distributional effects, engagement of patients and stakeholders, and\napproach to uncertainty. Crucially, CHEERS is a **reporting** instrument: it specifies *what\nmust be disclosed* so a reader can understand, appraise, and reproduce the economic\nevaluation. It is not a how-to for conducting an analysis, not a methods or reference-case\nstandard (those are NICE/CADTH/ICER/ISPOR methods guidance), and not a critical-appraisal or\nrisk-of-bias tool. The canonical statement (Husereau et al., 2022) was co-published in\n*Value in Health* and the *BMJ*, with a companion Explanation and Elaboration paper that\nexpands each item.\n\n**When to use** — Apply CHEERS 2022 whenever you report a **health economic evaluation that\ncompares costs and consequences of two or more courses of action**: cost-effectiveness\nanalysis (CEA), cost-utility analysis (CUA), cost-benefit analysis (CBA), or\ncost-minimization analysis (CMA), whether trial-based, model-based, or built on real-world\ndata. It applies across decision contexts — a peer-reviewed journal submission, an HTA/payer\nreimbursement dossier (NICE, CADTH, ICER, G-BA, PBAC), a value-message file, or the economic\nsection of a regulatory or market-access package. Decision rules for *which* guideline:\n(1) if the deliverable is a full economic evaluation, CHEERS is the reporting standard;\n(2) if the deliverable is the **underlying comparative clinical or RWE study** feeding the\nmodel (the effectiveness, safety, or HCRU inputs), report *that study* with STROBE or\nRECORD/RECORD-PE for observational data, CONSORT for trials, and pre-specify it with\nHARPER/STaRT-RWE — CHEERS does not cover those inputs; (3) for a **budget-impact analysis**,\nCHEERS is not the design standard — follow the ISPOR BIA Good Practice — though CHEERS-style\ntransparency on assumptions, perspective, and time horizon is still good practice; (4) for a\nsystematic review of economic evaluations, use PRISMA, not CHEERS, for the review itself.\n\n**What it requires** — The 28 items map to the structure of an economic evaluation and force\ndisclosure of the choices that determine its credibility. Substantively, CHEERS 2022\nrequires: a clear **title/abstract** identifying it as an economic evaluation; a stated\n**research question, target population, subgroups, setting, and comparators**; the\n**perspective** (e.g., healthcare-sector vs societal) and its justification; the **time\nhorizon** and the **discount rate** for costs and effects; the choice and structure of any\n**model** with rationale and assumptions; the sources and methods for **clinical\neffectiveness, health-state utilities/preferences, resource use, and unit costs**, with the\nmeasurement and valuation approach for each; handling of **currency, price date, and\nconversion**; the analytic methods, including how **heterogeneity, uncertainty, and\ndistributional effects** were characterized; full **results** (incremental costs and effects,\nICERs, net benefit) with deterministic and probabilistic **sensitivity analysis**; a\ndiscussion of findings, limitations, generalizability, and **equity**; and disclosure of\n**funding, conflicts of interest, and stakeholder/patient engagement**. For evaluations\nbuilt on real-world data, this means the report must make transparent *how RWE inputs were\nused in the model* — which estimands, data sources, and time horizons the effectiveness and\ncost inputs came from, and how their uncertainty was propagated — even though CHEERS does not\nitself adjudicate the validity of those inputs.\n\n**When NOT to use — limitations and common misapplications** — CHEERS is a **reporting\nchecklist, not a risk-of-bias instrument and not a quality score**: a fully \"CHEERS-compliant\"\npaper can still report a biased, poorly-specified, or non-credible analysis. The single most\ndangerous misapplication for RWE-informed economics is **treating CHEERS as a substitute for\nreporting the underlying real-world study**. Completing CHEERS on a CEA whose effectiveness or\ncost inputs come from a claims/EHR analysis does **not** discharge the obligation to report\nthat observational study to STROBE/RECORD-PE standards (design transparency, time-zero\nalignment, phenotype/algorithm validation, confounding control, attrition, sensitivity\nanalysis) — CHEERS does not ask for, and does not certify, any of that. Other failure modes:\nusing CHEERS where a **methods/reference-case** standard is required (CHEERS reports a choice;\nit does not tell you the *correct* perspective or discount rate for a given jurisdiction —\nthat comes from NICE/CADTH/ICER); using CHEERS as the design guideline for a **budget-impact\nanalysis** (separate ISPOR BIA good practice applies); **checklist-as-theater**, where item\nnumbers are ticked in an appendix while the corresponding content is vague or absent; and\napplying CHEERS to a study that is **not an economic evaluation at all** (a cost-of-illness or\nHCRU descriptive study is not a comparative economic evaluation and is better served by\nSTROBE plus costing good practice). Completing the checklist also does not make an\nobservational comparative-effectiveness input causal.\n\n**How it maps to this catalog** — CHEERS sits at the *reporting* layer over the economic\nevaluation; the implementing concepts in this repo supply the methods whose reporting it\ngoverns. The four evaluation types are the design concepts CHEERS reports:\n**cost-effectiveness** (cost-effectiveness), **cost-utility** (cost-utility),\n**cost-benefit** (cost-benefit), and **cost-minimization** (cost-minimization); the adjacent\n**budget-impact** (budget-impact) is reported under ISPOR BIA guidance, not CHEERS proper.\nCHEERS items on model structure and analysis map to **health-economic-modeling-methods-rwe**,\n**markov-transition-probabilities-rwe**, **partitioned-survival-models-rwe**, and\n**discrete-event-simulation-rwe**; the survival-input item maps to\n**survival-extrapolation-hta-rwe**. The utilities/preferences item is implemented by\n**qaly-utility-mapping-rwe**; the discounting item by **discounting-costs-effects-rwe**; the\nresults/outcome metrics by **icer-net-monetary-benefit-rwe**; and the uncertainty item by\n**probabilistic-sensitivity-analysis-hea-rwe**. The cost and resource-use inputs map to\n**healthcare-costs-pppm-pppy-pmpm**, **all-cause-vs-attributable-costs-rwe**, and\n**hcru-healthcare-resource-utilization**. When the effectiveness or cost inputs are drawn\nfrom real-world data, the *inputs* (not CHEERS) are governed by **target-trial-emulation**,\n**fit-for-purpose-data-assessment-rwe**, and **claims-analysis** — report those to STROBE/\nRECORD-PE, then report the economic evaluation that consumes them to CHEERS.\n\n**Applied note (claims/EHR/registry RWE).** When a CEA or CUA is parameterized from\nadministrative claims, EHR, or registry data — e.g., real-world progression and discontinuation\nfeeding survival extrapolation, or PPPM costs feeding the cost arm — a credible submission\nreports *two* things to *two* standards. The economic evaluation (model structure,\nperspective, time horizon, discounting, ICER/NMB, PSA) is reported to CHEERS 2022; the\nreal-world inputs (data source fitness, phenotype/algorithm validation, time-zero, attrition,\nconfounding control, the estimand and how its uncertainty was carried into the model) are\nreported to STROBE/RECORD-PE and pre-specified with HARPER/STaRT-RWE. Conflating the two —\nticking CHEERS and omitting the provenance and validity of the RWE inputs — is the most common\nreason an HTA reviewer rejects a real-world-evidence-based economic model as non-transparent.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "health-economic-evaluation", "cost-effectiveness", "hta", "ispor", "equator" ], "aliases": [ "CHEERS 2022", "CHEERS-2022", "CHEERS II", "Consolidated Health Economic Evaluation Reporting Standards 2022" ], "applies_to_study_types": [ "cost_effectiveness", "cost_utility", "cost_benefit", "cost_minimization" ], "data_sources": [ "claims", "ehr", "registry", "linked", "primary" ], "citations": [ { "role": "introduce", "doi": "10.1016/j.jval.2021.11.1351", "url": "https://doi.org/10.1016/j.jval.2021.11.1351", "citation_text": "Husereau D, Drummond M, Augustovski F, et al. Consolidated Health Economic Evaluation Reporting Standards 2022 (CHEERS 2022) Statement: Updated Reporting Guidance for Health Economic Evaluations. Value in Health. 2022;25(1):3-9.", "year": 2022, "authors_short": "Husereau et al.", "notes": "Canonical CHEERS 2022 statement and 28-item checklist (ISPOR CHEERS II Good Practices Task Force). Verified via Crossref (Husereau; Value in Health; 2022)." }, { "role": "introduce", "doi": "10.1136/bmj-2021-067975", "url": "https://doi.org/10.1136/bmj-2021-067975", "citation_text": "Husereau D, Drummond M, Augustovski F, et al. Consolidated Health Economic Evaluation Reporting Standards 2022 (CHEERS 2022) statement: updated reporting guidance for health economic evaluations. BMJ. 2022;376:e067975.", "year": 2022, "authors_short": "Husereau et al.", "notes": "Simultaneous co-publication of the CHEERS 2022 statement in the BMJ; identical checklist. Verified via Crossref (Husereau; BMJ; 2022)." }, { "role": "explain", "doi": "10.1016/j.jval.2021.10.008", "url": "https://doi.org/10.1016/j.jval.2021.10.008", "citation_text": "Husereau D, Drummond M, Augustovski F, et al. Consolidated Health Economic Evaluation Reporting Standards (CHEERS) 2022 Explanation and Elaboration: A Report of the ISPOR CHEERS II Good Practices Task Force. Value in Health. 2022;25(1):10-31.", "year": 2022, "authors_short": "Husereau et al.", "notes": "Item-by-item explanation and elaboration with rationale and worked examples for each of the 28 checklist items. Verified via Crossref (Husereau; Value in Health; 2022)." }, { "role": "use", "url": "https://www.ispor.org/heor-resources/good-practices/cheers", "citation_text": "CHEERS 2022 Statement and checklist resources, ISPOR Good Practices / EQUATOR Network — maintained checklist, fillable template, and translations.", "year": 2022, "authors_short": "ISPOR / EQUATOR Network", "notes": "Maintained landing page for the current checklist and fillable reporting template; the practical artifact authors complete for a submission appendix." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "cost-effectiveness", "notes": "CHEERS 2022 is the reporting standard for cost-effectiveness analyses." }, { "relation_type": "applies_to", "target_slug": "cost-utility", "notes": "CHEERS 2022 is the reporting standard for cost-utility analyses (QALY-based)." }, { "relation_type": "applies_to", "target_slug": "cost-benefit", "notes": "CHEERS 2022 is the reporting standard for cost-benefit analyses." }, { "relation_type": "applies_to", "target_slug": "cost-minimization", "notes": "CHEERS 2022 is the reporting standard for cost-minimization analyses." }, { "relation_type": "see_also", "target_slug": "budget-impact", "notes": "Budget-impact analyses are adjacent but reported under the ISPOR BIA Good Practice, not CHEERS; CHEERS-style transparency on assumptions and perspective still applies." }, { "relation_type": "requires", "target_slug": "health-economic-modeling-methods-rwe", "notes": "Implements the CHEERS model structure, assumptions, and analytic-methods items." }, { "relation_type": "requires", "target_slug": "probabilistic-sensitivity-analysis-hea-rwe", "notes": "Implements the CHEERS characterization-of-uncertainty item (PSA and CEACs)." }, { "relation_type": "requires", "target_slug": "qaly-utility-mapping-rwe", "notes": "Implements the CHEERS health-state utilities / preference-measurement item for CUAs." }, { "relation_type": "requires", "target_slug": "discounting-costs-effects-rwe", "notes": "Implements the CHEERS discount-rate item for costs and effects." }, { "relation_type": "requires", "target_slug": "icer-net-monetary-benefit-rwe", "notes": "Implements the CHEERS results metrics (incremental cost-effectiveness ratio and net monetary/health benefit)." }, { "relation_type": "used_with", "target_slug": "survival-extrapolation-hta-rwe", "notes": "Survival extrapolation is a common CHEERS-reported analytic input when modeling beyond trial/RWE follow-up." }, { "relation_type": "used_with", "target_slug": "healthcare-costs-pppm-pppy-pmpm", "notes": "Real-world cost inputs (PPPM/PMPM) parameterize the cost arm CHEERS requires to be transparently sourced and valued." }, { "relation_type": "used_with", "target_slug": "target-trial-emulation", "notes": "When effectiveness inputs come from real-world data, the input study is reported to STROBE/RECORD-PE (not CHEERS); CHEERS reports how it was used in the economic model." }, { "relation_type": "used_with", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "Data-fitness of RWE inputs is reported alongside, not within, CHEERS; CHEERS still requires transparent sourcing of effectiveness, utility, and cost inputs." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "hta", "journal", "fda", "ema" ] }, { "slug": "cioms-rwd-rwe", "name": "CIOMS Working Group XIII: RWD and RWE in Regulatory Decision Making", "short_definition": "A global multi-stakeholder consensus framework that defines when and how real-world data and real-world evidence should enter regulatory decision making across the product lifecycle, setting expectations for research questions, fitness-for-purpose, study design, governance, and transparency.", "long_description": "**What it is** — *Real-World Data and Real-World Evidence in Regulatory Decision Making*, the\nconsensus report of the Council for International Organizations of Medical Sciences (CIOMS)\nWorking Group XIII (CIOMS WG XIII), published by CIOMS in 2024 with a peer-reviewed report\nsummary appearing in *Pharmacoepidemiology and Drug Safety* in 2025. CIOMS is a Geneva-based\ninternational NGO (established jointly by WHO and UNESCO) whose working groups produce global,\nmulti-stakeholder consensus guidance. Unlike a reporting checklist maintained by EQUATOR\n(STROBE, RECORD-PE) or a critical-appraisal instrument (ROBINS-I, Newcastle-Ottawa), WG XIII is\na *decision framework*: it harmonizes terminology and articulates the triggers, objectives,\nresearch questions, design features, governance, ethics, and timing that determine whether RWD\ncan generate RWE fit to support a regulatory action. It is descriptive and principles-based, not\na numbered scorecard, and it deliberately spans the full lifecycle from pre-authorization\nthrough post-marketing safety and effectiveness.\n\n**When to use** — Reach for CIOMS WG XIII at the *framing* stage of any non-interventional or\nhybrid study whose results are intended to inform a regulatory decision (FDA, EMA, or other\nnational agencies), and as a shared reference when an industry team, a regulator, and an academic\npartner must agree on whether an RWE question is answerable and worth pursuing. It is the\nappropriate lens when you are deciding *whether* RWD/RWE is the right evidence source for a given\nobjective (e.g., an external control arm, a label expansion, a safety characterization, or an\nimposed/voluntary PASS), and what design and data conditions would make that evidence credible.\nDecision rule for choosing WG XIII versus its siblings: use **WG XIII** to establish the\nregulatory rationale, evidence objective, and fitness expectations; switch to **HARPER** or\n**StaRT-RWE** when you move to writing the study protocol template; use the **ENCePP Guide on\nMethodological Standards** and the **ENCePP Checklist** for EU PASS methodological conduct; and\nuse **STROBE / RECORD-PE** at the manuscript-reporting stage. WG XIII complements, and does not\nreplace, agency-specific guidance such as the FDA RWE Framework or EMA/HMA RWE guidances.\n\n**What it requires** — The framework drives substantive expectations rather than checklist\nticks. It asks that teams (1) state a precise, decision-relevant research question and the\nregulatory objective it serves, with the estimand (target population, treatment strategies,\nintercurrent-event handling, summary measure) made explicit; (2) demonstrate *fitness-for-purpose*\nof the data source(s) — relevance (do the data capture the population, exposure, outcomes, and\ncovariates needed?) and reliability (provenance, completeness, accuracy, traceability, and data\nquality/QC) — before analysis; (3) validate phenotypes and operational definitions for exposure,\noutcomes, and key covariates (algorithm performance, e.g., PPV/sensitivity); (4) align time zero\nand follow-up to avoid immortal-time and related design biases; (5) pre-specify confounding\ncontrol and address measured and unmeasured confounding; (6) plan for attrition, missing data,\nand censoring; (7) pre-register protocol and analysis plans and commit to transparency and\nreproducibility (versioned code lists, study registration); and (8) carry quantitative bias and\nsensitivity analyses through to interpretation. Governance, patient privacy, ethics, and\nmulti-region regulatory acceptability run throughout.\n\n**When NOT to use — limitations and common misapplications** — WG XIII is not a reporting\nchecklist and not a risk-of-bias or quality-scoring instrument; completing or citing it does not\ndocument a study or grade its internal validity, and it cannot substitute for STROBE/RECORD-PE\nat write-up or for ROBINS-I at appraisal. It frames the *assessment* of fitness-for-purpose but\ndoes not perform it — invoking WG XIII does not certify that a database is fit; that judgment must\nbe evidenced. It is not a protocol template: do not use it where HARPER or StaRT-RWE is required\nto specify the design, and do not treat it as a replacement for the ENCePP Guide on Methodological\nStandards governing EU PASS conduct. The deepest failure mode is framework-as-theater: citing the\nconsensus to lend regulatory gloss to a study while leaving estimands vague, phenotypes\nunvalidated, time zero misaligned, and confounding unaddressed. Adopting its principles does not\nmake an observational comparison causal; the design, comparator, and analysis still have to earn\nthat. Finally, regional regulatory acceptance varies, so WG XIII consensus does not guarantee any\nsingle agency's acceptance of a given study.\n\n**How it maps to this catalog** — Each WG XIII expectation is implemented by concrete catalog\nconcepts. The estimand and intercurrent-event requirement is operationalized by\n`estimands-ate-att-intercurrent-events-rwe` and traced via `estimand-analysis-traceability-rwe`,\nwith `picots-framework-rwe` structuring the question. Fitness-for-purpose maps to\n`fit-for-purpose-data-assessment-rwe`, plus `database-feasibility-attrition-funnel-rwe` and\n`claims-analysis` for source-specific feasibility and provenance. Phenotype/algorithm validation\nis implemented by `diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe`,\n`claims-outcome-algorithm-ppv-sensitivity-rwe`, and `ehr-phenotyping-algorithms-rwe`. Time-zero\nalignment and the new-user/active-comparator design are realized through\n`time-zero-index-date-alignment-rwe`, `immortal-time-bias-handling`, and\n`active-comparator-new-user` (often within `target-trial-emulation`). Confounding control is\ndelivered by `high-dimensional-propensity-score-hdps-rwe` and\n`propensity-score-methods-psm-iptw`. Attrition, missing data, and sensitivity/quantitative bias\nanalysis are covered by `attrition-and-loss-to-follow-up-rwe`,\n`unmeasured-confounding-probabilistic-bias-analysis-rwe`, and `e-value-sensitivity-analysis`.\nApplied note: for a claims- or EHR-based external-control or PASS study, WG XIII is the upstream\nrationale and acceptability layer — establish the regulatory objective and data fitness with it,\nbuild the protocol with HARPER/StaRT-RWE, implement the design and analysis with the concepts\nabove, and report with RECORD-PE/STROBE.", "primary_category": "Guideline", "tags": [ "guideline", "framework", "rwe", "regulatory", "fitness-for-purpose" ], "aliases": [ "CIOMS WG XIII", "CIOMS Working Group XIII", "CIOMS RWD/RWE consensus report", "CIOMS-RWD-RWE" ], "applies_to_study_types": [ "cer_observational", "active_comparator_new_user", "claims_analysis", "ehr_study", "pass_imposed", "pass_voluntary", "drug_utilization", "single_arm_external_control" ], "data_sources": [ "claims", "ehr", "registry", "linked", "multi-database" ], "citations": [ { "role": "introduce", "doi": "10.1002/pds.70117", "url": "https://doi.org/10.1002/pds.70117", "citation_text": "Hennessy S, Atsuta Y, Hill S, Rägo L, Juhaeri J; CIOMS Working Group XIII. Real-World Data and Real-World Evidence in Regulatory Decision Making: Report Summary From the Council for International Organizations of Medical Sciences (CIOMS) Working Group XIII. Pharmacoepidemiology and Drug Safety. 2025;34(3):e70117.", "year": 2025, "authors_short": "Hennessy et al.", "notes": "Peer-reviewed report summary of the CIOMS WG XIII consensus, distilling the framework's triggers, objectives, fitness-for-purpose principles, and lifecycle scope. The citable statement for the guideline." }, { "role": "explain", "url": "https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11897686/", "citation_text": "Hennessy S, Atsuta Y, Hill S, Rägo L, Juhaeri J; CIOMS Working Group XIII. Report summary (open-access full text), Pharmacoepidemiology and Drug Safety, 2025. PMC11897686.", "year": 2025, "authors_short": "Hennessy et al.", "notes": "Open-access full text of the report summary; elaborates the consensus recommendations and how RWD/RWE map onto specific regulatory decision contexts." }, { "role": "use", "url": "https://cioms.ch/publications/product/real-world-data-and-real-world-evidence-in-regulatory-decision-making/", "citation_text": "Council for International Organizations of Medical Sciences (CIOMS). Real-World Data and Real-World Evidence in Regulatory Decision Making. Geneva: CIOMS, 2024.", "year": 2024, "authors_short": "CIOMS", "notes": "Official CIOMS publication landing page for the full consensus report (open-access PDF); the maintained primary source." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "cer-observational", "notes": "Provides the regulatory framing and fitness-for-purpose expectations for comparative observational effectiveness studies intended to inform regulatory decisions." }, { "relation_type": "applies_to", "target_slug": "pass-imposed", "notes": "Frames objectives, governance, and evidence expectations for imposed post-authorization safety studies." }, { "relation_type": "applies_to", "target_slug": "pass-voluntary", "notes": "Frames objectives and fitness-for-purpose for voluntary post-authorization safety studies." }, { "relation_type": "applies_to", "target_slug": "single-arm-external-control", "notes": "Articulates when RWD-derived external controls are an acceptable evidence source and the conditions for credibility." }, { "relation_type": "see_also", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "Implements the relevance-and-reliability fitness-for-purpose assessment that WG XIII requires but does not itself perform." }, { "relation_type": "see_also", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Operationalizes the explicit estimand and intercurrent-event handling the framework expects." }, { "relation_type": "see_also", "target_slug": "target-trial-emulation", "notes": "The design discipline WG XIII endorses for emulating the regulatory question with observational data." }, { "relation_type": "complements", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Implements the phenotype/algorithm validation WG XIII requires for exposures, outcomes, and covariates." }, { "relation_type": "complements", "target_slug": "claims-outcome-algorithm-ppv-sensitivity-rwe", "notes": "Supplies the outcome-algorithm performance (PPV/sensitivity) evidence the framework expects." }, { "relation_type": "complements", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "Implements the confounding-control expectations for routinely collected data." }, { "relation_type": "complements", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Implements the attrition, censoring, and missing-data accounting the framework requires." }, { "relation_type": "complements", "target_slug": "e-value-sensitivity-analysis", "notes": "One of the quantitative bias / sensitivity analyses WG XIII expects to carry through to interpretation." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "hta", "journal" ] }, { "slug": "claims-payer-observability-checklist-rwe", "name": "Claims Payer Observability Checklist", "short_definition": "A checklist for payer and enrollment variables that determine whether claims data can observe the exposure, outcome, cost, and covariate channels required for an RWE study.", "long_description": "**What it is** - This guideline is the checklist layer for payer and enrollment observability in\nclaims-based RWE. It keeps the checklist out of concept files and gives analysts a concrete gate\nbefore they build exposure, outcome, cost, adherence, or covariate variables from claims. The\ncentral question is not \"is the patient insured?\" but \"is the needed data channel observable for\nthis patient-month, plan, benefit, and vendor extract?\" Medicare Parts A, B, C, and D, LIS, dual\neligibility, entitlement reason, subscriber/member identifiers, ERISA funding status, and benefit\ncarve-outs all change what the dataset can see.\n\n**When to use** - Use it before claims-based cohort construction, especially when the study\ndepends on complete medical claims, Part D or PBM fills, specialty pharmacy, behavioral health,\nsite-of-care costs, cost sharing, socioeconomic proxies, or family/subscriber relationships. It\napplies to Medicare FFS, Medicare Advantage encounter data, Medicaid-linked Medicare, and\ncommercial claims from carriers, third-party administrators, employers, PBMs, and multi-source\naggregators. Use it again when defining analytic censoring rules, because a move into Medicare\nAdvantage, a pharmacy carve-out, or missing PBM feed can create unobservable person-time that\nlooks like no use or no event.\n\n**What it requires / checklist domains** - Build a month-level eligibility and benefit panel\nbefore interpreting claims. Confirm which channel is required for each endpoint or exposure:\nPart A/B FFS medical events, Part D retail fills, Part B administered drugs, commercial medical,\ncommercial pharmacy, specialty pharmacy, behavioral health, lab, or dental/vision ancillary\nfeeds. Separate member/person identifiers from subscriber/family identifiers. Identify plan\nfunding and administration where possible, including self-insured ERISA plans and third-party\nadministrators. For Medicare, explicitly carry A/B/C/D, LIS, dual eligibility, and reason for\nentitlement; Medicare has Parts A-D, and \"Part E\" should be treated as an informal/local term\nrequiring clarification, not as an official benefit.\n\n**When NOT to use - limitations and common misapplications** - Do not use this checklist as a\nsubstitute for outcome validation or confounding control; it only verifies observability. Do not\ninfer no medication use from absence of pharmacy claims when the pharmacy benefit is carved out.\nDo not infer no hospitalization from absent FFS claims during Medicare Advantage months unless MA\nencounter data are proven complete for the question. Do not de-duplicate people on subscriber ID\nalone, because a subscriber can cover multiple members. Do not treat LIS, dual status, or\nentitlement reason as static if month-level fields are available. Missing payer data are often a\nstructural data limitation, not a patient behavior.\n\n**How it maps to this catalog** - This guideline cross-references\n`medicare-entitlement-lis-dual-eligibility-rwe` for Medicare parts, LIS, dual, and entitlement\nfields; `subscriber-id-member-id-claims-rwe` for identifier grain; `erisa-self-insured-health-plans-rwe`\nfor commercial plan funding; `benefit-carve-outs-medical-pharmacy-rwe` for missing benefit\nchannels; `claims-analysis` for source mechanics; and\n`continuous-enrollment-observable-time-rwe` for the observable-time denominator. Use FDA's\nEHR/claims data guidance for the broader data reliability/relevance frame; use this checklist for\nthe payer-channel appendix.", "primary_category": "Guideline", "tags": [ "guideline", "checklist", "claims", "payer", "medicare", "commercial-claims", "observability", "carve-outs" ], "aliases": [ "payer data checklist", "claims observability checklist", "benefit completeness checklist", "enrollment checklist" ], "applies_to_study_types": [ "claims_analysis", "adherence", "comparative_effectiveness", "cost_analysis", "linkage" ], "data_sources": [ "claims" ], "citations": [ { "role": "introduce", "doi": null, "url": "https://www.fda.gov/regulatory-information/search-fda-guidance-documents/real-world-data-assessing-electronic-health-records-and-medical-claims-data-support-regulatory", "citation_text": "U.S. Food and Drug Administration. Real-World Data: Assessing Electronic Health Records and Medical Claims Data To Support Regulatory Decision-Making for Drug and Biological Products. Guidance for Industry. July 2024.", "year": 2024, "authors_short": "FDA", "notes": "Regulatory source for claims/EHR data relevance, reliability, provenance, and fitness-for-use." }, { "role": "explain", "url": "https://www.medicare.gov/basics/get-started-with-medicare/medicare-basics/parts-of-medicare", "citation_text": "Medicare.gov. Parts of Medicare.", "year": 2026, "authors_short": "Medicare.gov", "notes": "Official Part A/B/C/D definitions." }, { "role": "explain", "url": "https://resdac.org/cms-data/variables/monthly-medicare-medicaid-dual-eligibility-code-january", "citation_text": "Research Data Assistance Center. Monthly Medicare-Medicaid Dual Eligibility Code.", "year": 2026, "authors_short": "ResDAC", "notes": "Dual eligibility variable reference." }, { "role": "explain", "url": "https://www.govinfo.gov/content/pkg/USCODE-2024-title29/html/USCODE-2024-title29-chap18.htm", "citation_text": "United States Code. Title 29, Chapter 18: Employee Retirement Income Security Program.", "year": 2024, "authors_short": "U.S. Code", "notes": "ERISA statutory source." }, { "role": "explain", "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC10390975/", "citation_text": "Starner CI, et al. Carved-in and carved-out pharmacy benefits: implications for medication management and outcomes. Journal of Managed Care & Specialty Pharmacy. 2023.", "year": 2023, "authors_short": "Starner et al.", "notes": "Pharmacy benefit carve-in/carve-out source." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "medicare-entitlement-lis-dual-eligibility-rwe", "notes": "Medicare eligibility and benefit-state checklist." }, { "relation_type": "applies_to", "target_slug": "subscriber-id-member-id-claims-rwe", "notes": "Identifier-granularity checklist." }, { "relation_type": "applies_to", "target_slug": "erisa-self-insured-health-plans-rwe", "notes": "Commercial self-insured plan checklist." }, { "relation_type": "applies_to", "target_slug": "benefit-carve-outs-medical-pharmacy-rwe", "notes": "Benefit-channel completeness checklist." } ], "index_definitions": [], "checklist_items": [ "Build a month-level enrollment and benefit panel before exposure, outcome, or cost construction.", "Require the specific observable channel needed for the endpoint or exposure, such as Part D or commercial pharmacy for fills and A/B FFS for many Medicare medical events.", "Treat Medicare Advantage periods, missing PBM feeds, behavioral-health carve-outs, and specialty carve-outs as observability states, not no-use states.", "Separate subscriber/family identifiers from member/person identifiers; never de-duplicate patients solely on subscriber ID.", "Identify plan funding type, administrator, employer group, product, and benefit carve-outs when those fields are available.", "Include dual eligibility, LIS, disability, and entitlement reason when cost sharing, adherence, access, or socioeconomic confounding matters.", "Profile plan or employer groups for channel completeness before including them in pharmacy, behavioral-health, specialty-drug, or cost analyses.", "Document whether claims are from a carrier, third-party administrator, PBM, multi-source vendor, or linked extract." ], "regulatory_relevance": [ "hta", "cms" ] }, { "slug": "consort-ai", "name": "CONSORT-AI (CONSORT Extension for Trials of AI Interventions)", "short_definition": "The CONSORT 2010 reporting extension for randomized controlled trials evaluating interventions that include an artificial-intelligence or machine-learning component, adding 14 AI-specific reporting items to the parent checklist so that the algorithm, its inputs/outputs, its integration into the care pathway, and its failure analysis are reported transparently enough to appraise and reproduce.", "long_description": "**What it is.** CONSORT-AI is the official EQUATOR-registered extension of the CONSORT 2010 statement for reporting\nrandomized controlled trials (RCTs) in which the intervention contains an **artificial-intelligence / machine-learning\ncomponent** (Liu et al., *Nature Medicine* 2020; published in parallel in *BMJ* and *Lancet Digital Health*). It does\nnot replace CONSORT — it **layers AI-specific reporting items onto the parent checklist**. The extension adds new or\nelaborated items covering: a clear statement that the intervention involves AI and which version of the algorithm was\nevaluated; the intended use, clinical pathway, and intended user (clinician vs patient vs autonomous); the input data\nrequired (handling of poor-quality, missing, or out-of-distribution inputs); the AI output and how it feeds the\nclinical decision; the level of human–AI interaction and required skill; the setting and on-site integration; and a\npre-specified **analysis of performance errors / failure modes**. CONSORT-AI is the trial-*report* counterpart to\n**SPIRIT-AI** (Cruz Rivera et al. 2020), which governs the trial *protocol*; the two were developed together by the\nsame international consensus group (SPIRIT-AI and CONSORT-AI Steering Group). It is a **reporting** guideline — a\ntransparency checklist — not a critical-appraisal or risk-of-bias instrument and not a quality score.\n\n**When to use.** Use CONSORT-AI when you are reporting (or peer-reviewing, or registering the report of) a\n**randomized controlled trial whose intervention embeds an AI/ML model** — for example a deep-learning triage tool,\nan algorithmic treatment-recommendation system, or an ML-driven monitoring/alerting intervention tested head-to-head\nagainst standard care. The decision rules that distinguish CONSORT-AI from its siblings are sharp and worth stating\nexplicitly: (1) the design must be a **randomized trial** — if it is, use CONSORT-AI; the parent **CONSORT** still\ngoverns everything non-AI. (2) If the AI intervention is delivered in a **pragmatic, routine-care RCT**, combine\nCONSORT-AI with **CONSORT-Pragmatic** and characterize the explanatory–pragmatic position with **PRECIS-2** — this is\nthe principal point of contact with real-world evidence, since a pragmatic AI trial runs inside live EHR/claims\nworkflows. (3) If you are writing the **protocol**, use **SPIRIT-AI**, not CONSORT-AI. The relevant decision contexts\nspan regulatory submission (FDA/EMA software-as-a-medical-device and drug-device evidence), HTA/payer dossiers for\nAI-enabled technologies, peer-reviewed journals (most major journals require CONSORT-family adherence), and\ntrial-report registration.\n\n**What it requires.** As a reporting extension, CONSORT-AI enforces *complete transparent description* across the\ndomains that determine whether an AI trial can be appraised and acted on. Framed for evidence that touches real-world\ndata systems: **design and pre-specification** transparency (randomization, blinding where feasible, the specific\nalgorithm *version* — frozen vs continuously-learning); **data fitness-for-use of the AI inputs** (input data\nsources, formats, acquisition, and explicit handling of poor-quality, missing, or out-of-distribution inputs — the\nAI analogue of data-fitness-for-use); **algorithm / phenotype validation lineage** (how the model was developed and\nvalidated *before* the trial, and the performance claim it was trial-tested against); **integration and time-zero of\nthe intervention** (when and where in the care pathway the AI acts, the human–AI interaction level, and what the user\ndoes with the output); **outcomes, estimands, and intercurrent events** (CONSORT-AI inherits CONSORT/CONSORT-PRO\noutcome reporting; intercurrent events such as clinician override of the AI must be described); **attrition and\nmissing data**; and a distinctive requirement for **error/failure-mode analysis** (how algorithm performance errors\nwere identified, analyzed, and reported). Confounding control in the causal sense is handled by randomization itself\n— that is the point of using an RCT extension rather than an observational tool.\n\n**When NOT to use — limitations and common misapplications.** The single most important caveat in an RWE/HEOR\ncatalog: **CONSORT-AI is an RCT reporting extension and does not apply to non-randomized, observational evaluations\nof AI tools.** Concrete failure modes: (1) *Using CONSORT-AI for an observational performance study of a deployed\nalgorithm* — wrong tool; report model development/validation with **TRIPOD+AI** and early live-clinical evaluation\nwith **DECIDE-AI**. (2) *Using it for a diagnostic-accuracy study* — use **STARD-AI** (AI extension of STARD). (3)\n*Treating it as a risk-of-bias or quality instrument* — it is a reporting checklist; bias appraisal of AI prediction\nmodels is the job of **PROBAST/PROBAST-AI**, and trial risk of bias remains RoB 2. Completing the checklist does\n**not** confer internal validity, does not make a poorly-randomized trial sound, and does not make an observational\nAI evaluation causal. (4) *Checklist-as-theater* — pasting page numbers against items without the underlying\nreporting substance (e.g., declaring \"input data handling reported\" with no description of out-of-distribution\nbehavior) defeats the purpose. (5) *Wrong extension for the design* — using plain CONSORT (no AI items) for an AI\ntrial, or using CONSORT-AI to report the protocol instead of SPIRIT-AI.\n\n**How it maps to this catalog.** CONSORT-AI sits at the *reporting* end of the AI-trial lifecycle; several catalog\nconcepts implement the methodological substance it asks you to describe transparently. The **pragmatic-trial** and\n**target-trial-emulation** concepts implement the design-and-time-zero discipline the checklist reports, and target\ntrial thinking is the bridge when an AI intervention must be evaluated where randomization is infeasible (a setting\nCONSORT-AI itself does **not** cover — that is where TTE in observational data takes over). **estimands-ate-att-\nintercurrent-events-rwe** implements the estimand/intercurrent-event reporting (clinician override of the AI is a\ntextbook intercurrent event). **predictive-and-causal-ml-models-rwe** and **prediction-model-validation-\nrecalibration-rwe** implement the upstream model development/validation lineage that the trial report must reference.\n**attrition-and-loss-to-follow-up-rwe** implements the attrition reporting; **generalizability-transportability-\nexternal-validity-rwe** implements the \"intended setting / out-of-distribution input\" reasoning that makes an AI\nresult transportable. For the protocol-side counterpart, **study-protocol-or-sap-elements** carries the\npre-specification that SPIRIT-AI governs. **Applied RWE note:** the genuine real-world point of contact is the\n*pragmatic* AI trial embedded in routine EHR/claims workflows — here CONSORT-AI + CONSORT-Pragmatic + PRECIS-2 are\nused together, and the input-data items (missing/poor-quality/out-of-distribution EHR fields) and integration items\n(where in the live clinical workflow the algorithm fires) become the most consequential, because they are precisely\nthe elements that break silently when a model trained on one health system is deployed in another.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "artificial-intelligence", "machine-learning", "randomized-controlled-trial", "consort", "equator" ], "aliases": [ "CONSORT-AI", "CONSORT-Artificial Intelligence", "CONSORT extension for artificial intelligence", "CONSORT AI extension" ], "applies_to_study_types": [ "pragmatic_trial" ], "data_sources": [ "ehr", "claims", "registry", "linked" ], "citations": [ { "role": "introduce", "doi": "10.1038/s41591-020-1034-x", "url": "https://doi.org/10.1038/s41591-020-1034-x", "citation_text": "Liu X, Cruz Rivera S, Moher D, Calvert MJ, Denniston AK; SPIRIT-AI and CONSORT-AI Working Group. Reporting guidelines for clinical trial reports for interventions involving artificial intelligence: the CONSORT-AI extension. Nature Medicine. 2020;26(9):1364-1374.", "year": 2020, "authors_short": "Liu et al.", "notes": "Canonical statement paper; defines the 14 AI-specific extension items on the CONSORT 2010 checklist. Published in parallel in BMJ (m3164) and Lancet Digital Health." }, { "role": "explain", "doi": "10.1038/s41591-020-1037-7", "url": "https://doi.org/10.1038/s41591-020-1037-7", "citation_text": "Cruz Rivera S, Liu X, Chan AW, Denniston AK, Calvert MJ; SPIRIT-AI and CONSORT-AI Working Group. Guidelines for clinical trial protocols for interventions involving artificial intelligence: the SPIRIT-AI extension. Nature Medicine. 2020;26(9):1351-1363.", "year": 2020, "authors_short": "Cruz Rivera et al.", "notes": "Protocol counterpart developed by the same consensus group; CONSORT-AI governs the report, SPIRIT-AI the protocol." }, { "role": "explain", "doi": "10.1136/bmj.c332", "url": "https://doi.org/10.1136/bmj.c332", "citation_text": "Schulz KF, Altman DG, Moher D; CONSORT Group. CONSORT 2010 Statement: updated guidelines for reporting parallel group randomised trials. BMJ. 2010;340:c332.", "year": 2010, "authors_short": "Schulz et al.", "notes": "Parent statement that CONSORT-AI extends; all non-AI reporting items remain governed by CONSORT 2010." }, { "role": "use", "url": "https://www.equator-network.org/reporting-guidelines/consort-artificial-intelligence/", "citation_text": "CONSORT-AI extension — EQUATOR Network maintained reporting-guideline page (checklist, explanation, and links to the SPIRIT-AI/CONSORT-AI extensions).", "year": 2020, "authors_short": "EQUATOR Network", "notes": "Authoritative maintained landing page with the downloadable checklist." } ], "relations": [ { "relation_type": "part_of", "target_slug": "consort", "notes": "CONSORT-AI is an official extension of the CONSORT 2010 statement; the parent checklist still governs all non-AI reporting items." }, { "relation_type": "used_with", "target_slug": "consort-pragmatic", "notes": "For AI interventions evaluated in pragmatic routine-care RCTs, combine CONSORT-AI with the pragmatic extension." }, { "relation_type": "see_also", "target_slug": "spirit", "notes": "SPIRIT (and its SPIRIT-AI extension) governs the trial protocol; CONSORT-AI governs the trial report." }, { "relation_type": "see_also", "target_slug": "precis-2", "notes": "Use PRECIS-2 to characterize how explanatory vs pragmatic an AI trial is when it runs inside real-world workflows." }, { "relation_type": "see_also", "target_slug": "tripod", "notes": "TRIPOD (and TRIPOD+AI) is the correct guideline for reporting AI prediction-model development/validation — used instead of CONSORT-AI when there is no randomized trial." }, { "relation_type": "see_also", "target_slug": "stard", "notes": "STARD (and STARD-AI) is the correct guideline for diagnostic-accuracy studies of AI tools, not CONSORT-AI." }, { "relation_type": "applies_to", "target_slug": "pragmatic-trial", "notes": "The primary real-world point of contact — AI interventions tested in pragmatic trials embedded in routine care." }, { "relation_type": "see_also", "target_slug": "target-trial-emulation", "notes": "When randomizing an AI intervention is infeasible, target-trial emulation in observational data is the route — a setting CONSORT-AI explicitly does not cover." }, { "relation_type": "see_also", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Implements the estimand and intercurrent-event reporting CONSORT-AI requires; clinician override of the AI is a canonical intercurrent event." }, { "relation_type": "see_also", "target_slug": "predictive-and-causal-ml-models-rwe", "notes": "Implements the upstream model development that the trial report must reference (algorithm version, prior validation)." }, { "relation_type": "see_also", "target_slug": "prediction-model-validation-recalibration-rwe", "notes": "Implements the pre-trial validation lineage and recalibration the checklist asks authors to describe." }, { "relation_type": "see_also", "target_slug": "generalizability-transportability-external-validity-rwe", "notes": "Implements the intended-setting / out-of-distribution-input reasoning behind the AI input-handling items." }, { "relation_type": "see_also", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Implements the attrition / loss-to-follow-up reporting inherited from CONSORT." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "hta", "journal" ] }, { "slug": "consort-pragmatic", "name": "CONSORT Extension for Pragmatic Trials", "short_definition": "A reporting checklist (8 extended CONSORT items) for randomized pragmatic trials conducted in routine-care settings, prompting authors to describe how the trial's eligibility, intervention delivery, setting, outcomes, and analysis reflect real-world practice rather than an idealized explanatory design.", "long_description": "**What it is.** The **CONSORT extension for pragmatic trials** (Zwarenstein et al., BMJ\n2008) is a reporting guideline, not a design or risk-of-bias tool. Published against the\nthen-current CONSORT 2001 statement and used today alongside CONSORT 2010, it provides\nextended guidance on 8 checklist items (eligibility,\ninterventions, outcomes, sample size, blinding, participant flow, generalizability, and\ninterpretation) so that readers of a *pragmatic* randomized trial can judge how applicable\nthe results are to routine practice. It is curated through the EQUATOR Network and the\nCONSORT Group and is meant to be used alongside — never instead of — the parent CONSORT\nchecklist and the relevant trial registration. Its conceptual companion for *design* is the\nPRECIS-2 tool (Loudon et al., BMJ 2015), which scores how explanatory-versus-pragmatic a\ntrial is across nine domains; CONSORT-Pragmatic governs how that pragmatism is *reported*.\n\n**When to use.** Use it whenever you report a randomized trial that deliberately measures\neffectiveness under usual-care conditions — flexible delivery, broad eligibility, clinically\nmeaningful endpoints, minimal protocol-driven extra visits — and especially when the trial\nis embedded in real-world data infrastructure (a registry-based randomized trial, a\npragmatic trial with EHR/claims-ascertained outcomes, or a cluster-randomized\nimplementation trial). It is the right checklist for HTA/payer dossiers and journal\nsubmissions arguing external validity, and it supports FDA/EMA interest in pragmatic\neffectiveness evidence. Decision rule for choosing the right CONSORT family member: if the\nunit of randomization is the cluster, layer **CONSORT for cluster trials**; if you report\npatient-reported outcomes, harms, or non-inferiority, add those extensions; if your study is\n*non-randomized* real-world evidence, CONSORT does not apply at all — use **STROBE/RECORD-PE**\nfor reporting and **target-trial emulation** for design. Reach for CONSORT-Pragmatic\n(over plain CONSORT alone) precisely when the trial's value proposition is generalizability\nto routine care.\n\n**What it requires.** The extension forces authors to make pragmatism explicit and auditable.\nIts substantive domains, framed for trials that touch real-world data: (1) *eligibility and\nsetting* — describe the participants, practitioners, and care settings, and how closely they\nmirror the population the intervention targets in practice; (2) *intervention description* —\nstate how flexibly each intervention was delivered and the resources/expertise assumed,\nsince pragmatic trials permit clinician judgment rather than rigid protocols; (3) *outcomes*\n— justify endpoints as directly relevant to participants, clinicians, or payers, and report\nhow they were ascertained (including registry/EHR/claims-based capture, which raises\nthe same phenotype-validation and time-window questions as observational RWE); (4)\n*participant flow and follow-up* — a CONSORT flow diagram with explicit accounting of\nattrition and loss to follow-up, which in routine-care settings is often substantial and\ninformative; (5) *analysis and estimand* — pre-specify the analysis population (ITT under a\ntreatment-policy strategy is typical for pragmatic effectiveness) and how intercurrent\nevents (non-adherence, treatment switching, crossover) are handled; (6) *generalizability\nand interpretation* — discuss applicability to other populations, settings, and usual care.\nWhere outcomes or covariates are drawn from secondary data, the report should document\ndata fitness-for-use and algorithm definitions to the same standard expected of observational\nRWE.\n\n**When NOT to use — limitations and common misapplications.** (a) It is a *reporting*\nchecklist, so completing it improves transparency but does **not** reduce bias, raise study\nquality, or certify internal validity — a fully reported pragmatic trial can still be\nconfounded by poor allocation concealment or differential attrition; do not treat the\nchecklist as a risk-of-bias instrument (use RoB 2) or a quality score. (b) It applies only\nto *randomized* trials; using it to dress up a non-randomized database study is a category\nerror — that study needs RECORD-PE/STROBE, and randomization language will mislead reviewers.\n(c) Checklist-as-theater: ticking items in a submission appendix without the corresponding\ndetail in the manuscript defeats the purpose; page numbers must point to real content. (d)\nWrong family member: a cluster-randomized pragmatic trial reported with only the patient-level\nextension will under-report design effects and recruitment-after-randomization bias. (e)\nPragmatic ≠ low-rigor: the extension does not license loose endpoint ascertainment —\nEHR/claims-based outcomes still require validated phenotypes. (f) It does not, by itself,\nmake an effectiveness estimate causal or transportable; that depends on design and analysis,\nnot on the report.\n\n**How it maps to this catalog.** Several catalog concepts implement what the extension asks\nauthors to report. *Estimand and intercurrent-event handling* (treatment-policy ITT,\nswitching, non-adherence) is implemented by **estimands-ate-att-intercurrent-events-rwe**.\n*Outcome/exposure ascertainment from secondary data* — required whenever a pragmatic trial\nuses registry/EHR/claims endpoints — is implemented by\n**diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe** (algorithm definition, PPV\nvalidation) and **claims-analysis** (code lists, enrollment, data nuances). *Attrition and\nloss-to-follow-up reporting* maps to **attrition-and-loss-to-follow-up-rwe** and the CONSORT\nflow diagram. *Design framing and eligibility/time-zero discipline* is shared with\n**target-trial-emulation** and **active-comparator-new-user** (the new-user, time-zero logic\nthat pragmatic effectiveness designs borrow), while structured question specification uses\n**picots-framework-rwe**. *Data fitness-for-use* is implemented by\n**fit-for-purpose-data-assessment-rwe**, and *quantitative bias / sensitivity analysis* of\nresidual confounding (relevant when randomization is imperfect or outcomes are\ndatabase-derived) by **e-value-sensitivity-analysis**. *When confounding control is needed*\nfor embedded observational comparisons, see **high-dimensional-propensity-score-hdps-rwe**.\nApplied note for registry/EHR/claims-based pragmatic trials: report the outcome algorithm and\nits validation, continuous-enrollment/observation windows, and how loss to follow-up was\nhandled with the same rigor you would apply to a STROBE/RECORD-PE observational study — the\nrandomization protects the treatment contrast, but the *measurement* layer is pure RWE.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "pragmatic-trial", "consort", "external-validity", "rwe" ], "aliases": [ "CONSORT-Pragmatic", "CONSORT extension for pragmatic trials", "Pragmatic CONSORT" ], "applies_to_study_types": [ "pragmatic_trial" ], "data_sources": [ "claims", "ehr", "registry", "linked", "primary" ], "citations": [ { "role": "introduce", "doi": "10.1136/bmj.a2390", "url": "https://doi.org/10.1136/bmj.a2390", "citation_text": "Zwarenstein M, Treweek S, Gagnier JJ, et al. Improving the reporting of pragmatic trials: an extension of the CONSORT statement. BMJ. 2008;337:a2390.", "year": 2008, "authors_short": "Zwarenstein et al.", "notes": "Canonical statement paper defining the 8-item CONSORT extension for pragmatic trials; to be used alongside the parent CONSORT checklist." }, { "role": "explain", "doi": "10.1136/bmj.h2147", "url": "https://doi.org/10.1136/bmj.h2147", "citation_text": "Loudon K, Treweek S, Sullivan F, Donnan P, Thorpe KE, Zwarenstein M. The PRECIS-2 tool: designing trials that are fit for purpose. BMJ. 2015;350:h2147.", "year": 2015, "authors_short": "Loudon et al.", "notes": "Companion design tool that scores explanatory-versus-pragmatic intent across nine domains; clarifies what \"pragmatic\" means and therefore what the CONSORT extension asks authors to report." }, { "role": "use", "url": "https://www.equator-network.org/reporting-guidelines/improving-the-reporting-of-pragmatic-trials-an-extension-of-the-consort-statement/", "citation_text": "CONSORT extension for pragmatic trials — EQUATOR Network reporting-guideline library (maintained checklist and resources).", "year": 2008, "authors_short": "EQUATOR Network", "notes": "Stable, maintained landing page with the checklist and links to the CONSORT family." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "pragmatic-trial", "notes": "Reporting standard for randomized pragmatic trials conducted under routine-care conditions." }, { "relation_type": "used_with", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Implements the extension's analysis-and-estimand item — treatment-policy ITT and handling of intercurrent events (non-adherence, switching, crossover) typical of pragmatic effectiveness trials." }, { "relation_type": "used_with", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Implements outcome/exposure ascertainment when a pragmatic trial captures endpoints from registry/EHR/claims data; supplies algorithm definitions and PPV validation." }, { "relation_type": "used_with", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Implements the CONSORT participant-flow item — explicit accounting of attrition and loss to follow-up, often substantial in usual-care settings." }, { "relation_type": "used_with", "target_slug": "claims-analysis", "notes": "Operational detail for code lists, enrollment windows, and data nuances when pragmatic-trial outcomes or covariates come from claims." }, { "relation_type": "see_also", "target_slug": "target-trial-emulation", "notes": "Shares the eligibility/time-zero/estimand discipline; the design framework to use when the study is non-randomized rather than a pragmatic RCT." }, { "relation_type": "see_also", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "Documents data fitness-for-use when secondary data underpin endpoint ascertainment." }, { "relation_type": "see_also", "target_slug": "picots-framework-rwe", "notes": "Structured specification of population, intervention, comparator, outcomes, timing, and setting that the extension's eligibility and outcome items require." }, { "relation_type": "see_also", "target_slug": "e-value-sensitivity-analysis", "notes": "Quantitative bias analysis for residual confounding when randomization is imperfect or outcomes are database-derived." }, { "relation_type": "see_also", "target_slug": "active-comparator-new-user", "notes": "New-user, time-zero logic that pragmatic effectiveness designs borrow when framing the treatment contrast." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "hta", "journal" ] }, { "slug": "consort-pro", "name": "CONSORT-PRO", "short_definition": "Reporting extension to the CONSORT statement that specifies how patient-reported outcome (PRO) endpoints must be reported in randomized and pragmatic trials, covering PRO hypotheses, instrument selection and validation, the analysis of missing PRO data, and interpretation of PRO results.", "long_description": "**What it is.** CONSORT-PRO is the **Consolidated Standards of Reporting Trials extension for Patient-Reported\nOutcomes**, published by Calvert and colleagues (JAMA, 2013) as a formal extension to the CONSORT 2010 statement.\nIt is a **reporting checklist** — not a design protocol, a risk-of-bias tool, or a quality score — that tells authors,\nreviewers, and regulators what must appear in a trial report whenever a PRO is a primary or important secondary\nendpoint. It adds five PRO-specific elaborations to the core CONSORT items (and modifies several others): the PRO is\nidentified in the abstract; a PRO hypothesis is stated with relevant domains; the instrument's validity, reliability,\nand (where relevant) responsiveness are documented with citations; statistical approaches for dealing with **missing\nPRO data** are described; and PRO-specific limitations, generalizability, and clinical implications are discussed. It\nis maintained within the **EQUATOR Network** and is now supported by the **PROTEUS** (Patient-Reported Outcomes Tools:\nEngaging Users and Stakeholders) consortium, with a companion protocol extension, **SPIRIT-PRO** (Calvert et al., JAMA\n2018), governing the trial protocol stage.\n\n**When to use.** Apply CONSORT-PRO when reporting the *results* of any randomized or pragmatic trial in which a PRO —\nhealth-related quality of life, symptom burden, functioning, treatment satisfaction, or another patient-reported\nconstruct — is a primary endpoint, a key secondary endpoint, or used in the label or value story. It is the relevant\nguideline for peer-reviewed publication, for the PRO sections of an **HTA/payer dossier** (where HRQoL and utility\nevidence drive cost-utility models), and for **FDA/EMA** submissions where PRO endpoints support labeling claims.\nDecision rule for choosing the right member of the family: use **SPIRIT-PRO** at the *protocol* stage and **CONSORT-PRO**\nat the *trial-report* stage; use the parent **CONSORT 2010** when no PRO endpoint is in play; and if the PRO evidence\ncomes from a *non-randomized* real-world study, CONSORT-PRO does not govern the design — pair PRO-measurement rigor with\nSTROBE/RECORD-PE reporting and the RWE methods below. Within a registry-based or pragmatic randomized trial, CONSORT-PRO\nstill applies because randomization is present.\n\n**What it requires.** The substantive domains it enforces, framed for evidence that will face regulatory or HTA review:\n(1) **PRO endpoint definition and hypothesis** — the construct, the specific instrument and version, the recall period,\nthe mode and schedule of administration, and a pre-stated hypothesis with the responder/minimally important difference\nthreshold. (2) **Instrument fitness-for-purpose** — evidence of content validity, construct validity, reliability, and\nresponsiveness in the target population, with citations rather than assertions. (3) **Estimand and intercurrent-event\nthinking** — what the PRO contrast actually estimates and how death, treatment discontinuation, and disease progression\nare handled (terminal events that truncate PRO collection are estimand-defining, not nuisance missingness). (4)\n**Missing PRO data** — the expected and observed completion rates by arm and timepoint, the missingness mechanism\nassumed, the primary analysis model, and **sensitivity analyses** under alternative (e.g., not-at-random) assumptions.\n(5) **Multiplicity** across domains/timepoints and the analysis of repeated PRO measures. (6) **Interpretation** —\nclinical meaning anchored to a justified threshold, generalizability, and PRO-specific limitations. It does *not*\nrequire code, a numeric score, or a verdict on study validity.\n\n**When NOT to use — limitations and common misapplications.** CONSORT-PRO is a transparency instrument, so the\npredictable failures are: (a) **treating the checklist as a risk-of-bias or quality assessment** — a fully reported PRO\ntrial can still be biased; completeness of reporting is not internal validity, and CONSORT-PRO is not RoB 2 or\nROBINS-I. (b) **Checklist-as-theater** — pasting a completed checklist into supplementary material while the manuscript\nomits arm-specific completion rates or the assumed missingness mechanism. (c) **Wrong member of the family** — using\nCONSORT-PRO to appraise a single-arm or observational PRO study, where there is no randomization to report and STROBE\nplus PRO-measurement standards (and RECORD-PE for routinely collected data) are the correct frame. (d) **Confusing\nCONSORT-PRO with SPIRIT-PRO** — protocol content (sample-size justification for the PRO, data-collection plan, missing-\ndata strategy specified a priori) belongs in SPIRIT-PRO; reporting after the fact belongs in CONSORT-PRO. (e) **Ignoring\nterminal/intercurrent events** — analyzing PRO change only among survivors who completed assessments silently changes the\nestimand and inflates apparent benefit; this is the single most common substantive failure CONSORT-PRO is meant to surface.\n\n**How it maps to this catalog.** Each CONSORT-PRO requirement is implemented by a concept here. PRO instrument selection,\nvalidity, and responsiveness → **pro-validation**, **pro-development**, and **hrqol** (with **qaly-utility-mapping-rwe**\nwhen PRO/HRQoL feeds a cost-utility model for HTA). Estimand specification and intercurrent events (death, discontinuation,\nprogression) → **estimands-ate-att-intercurrent-events-rwe** and **estimand-analysis-traceability-rwe**. Missing-data\nreporting and analysis → **attrition-and-loss-to-follow-up-rwe**, **missing-data-pattern-table-rwe**,\n**multiple-imputation-longitudinal-rwe**, and **tipping-point-analysis-rwe** for the not-at-random sensitivity analyses\nCONSORT-PRO asks for. Repeated-measures and longitudinal PRO modeling → **mmrm-repeated-measures-rwe** and\n**longitudinal-outcomes-modeling-rwe**. Composite or multi-domain PRO endpoints → **composite-endpoint-construction-rwe**.\nBaseline comparability of the randomized arms → **baseline-characteristics-and-covariate-balance-rwe**.\n\n**Applied note for claims/EHR/registry RWE.** PROs are, by definition, collected directly from patients and are almost\nnever present in administrative claims; EHRs capture them only sporadically and non-systematically (free-text notes, ad\nhoc questionnaires). In a **registry-based or pragmatic randomized trial**, CONSORT-PRO applies in full — but the report\nmust be explicit that PRO ascertainment is **visit-driven**, so completion is tied to the same encounters that generate\nthe routinely collected data, making missingness informative and arm-differential. Report PRO completion as a funnel by\narm and timepoint, distinguish administrative censoring (disenrollment, end of registry follow-up) from non-response and\nfrom death, and pre-specify the estimand strategy for terminal events. Using a claims/EHR backbone to *recruit and\nfollow* a PRO cohort does not make claims a PRO source: the patient-facing instrument and its validation evidence still\ncarry the reporting burden.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "patient-reported-outcomes", "pro", "consort", "equator", "missing-data" ], "aliases": [ "CONSORT-PRO", "CONSORT PRO Extension", "CONSORT Patient-Reported Outcomes Extension", "CONSORT 2010 PRO extension" ], "applies_to_study_types": [ "pragmatic_trial", "registry_trial", "pro_rwe" ], "data_sources": [ "registry", "ehr", "linked" ], "citations": [ { "role": "introduce", "doi": "10.1001/jama.2013.879", "url": "https://doi.org/10.1001/jama.2013.879", "citation_text": "Calvert M, Blazeby J, Altman DG, Revicki DA, Moher D, Brundage MD; CONSORT PRO Group. Reporting of patient-reported outcomes in randomized trials: the CONSORT PRO extension. JAMA. 2013;309(8):814-822.", "year": 2013, "authors_short": "Calvert et al.", "notes": "Canonical statement paper defining the CONSORT extension for patient-reported outcomes and its five PRO-specific checklist elaborations." }, { "role": "explain", "doi": "10.1001/jama.2017.21903", "url": "https://doi.org/10.1001/jama.2017.21903", "citation_text": "Calvert M, Kyte D, Mercieca-Bebber R, et al. Guidelines for inclusion of patient-reported outcomes in clinical trial protocols: the SPIRIT-PRO extension. JAMA. 2018;319(5):483-494.", "year": 2018, "authors_short": "Calvert et al.", "notes": "Companion protocol-stage extension; pairs with CONSORT-PRO at the reporting stage and specifies a priori PRO sample-size, data-collection, and missing-data planning." }, { "role": "explain", "doi": "10.1136/bmj.c332", "url": "https://doi.org/10.1136/bmj.c332", "citation_text": "Schulz KF, Altman DG, Moher D; CONSORT Group. CONSORT 2010 statement: updated guidelines for reporting parallel group randomised trials. BMJ. 2010;340:c332.", "year": 2010, "authors_short": "Schulz et al.", "notes": "Parent statement that CONSORT-PRO extends; the base checklist applies in full and PRO items are layered on top." }, { "role": "use", "url": "https://www.equator-network.org/reporting-guidelines/consort-pro/", "citation_text": "CONSORT-PRO reporting guideline page, EQUATOR Network (maintained checklist, elaboration, and PROTEUS consortium resources).", "year": 2013, "authors_short": "EQUATOR Network", "notes": "Maintained checklist and supporting materials for day-to-day use." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "pragmatic-trial", "notes": "Governs reporting of PRO endpoints whenever a pragmatic randomized trial measures a patient-reported outcome." }, { "relation_type": "applies_to", "target_slug": "registry-trial", "notes": "Applies to registry-based randomized trials with PRO endpoints; the report must flag visit-driven, arm-differential PRO completion." }, { "relation_type": "applies_to", "target_slug": "pro-rwe", "notes": "Applies when PRO endpoints are collected and reported within a randomized or pragmatic real-world trial." }, { "relation_type": "used_with", "target_slug": "pro-validation", "notes": "Implements the instrument validity, reliability, and responsiveness evidence CONSORT-PRO requires authors to cite." }, { "relation_type": "used_with", "target_slug": "hrqol", "notes": "HRQoL is the most common PRO construct CONSORT-PRO governs; feeds HTA value arguments." }, { "relation_type": "used_with", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Implements the estimand and intercurrent-event reasoning (death, discontinuation, progression) that CONSORT-PRO asks PRO analyses to make explicit." }, { "relation_type": "used_with", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Implements the arm- and timepoint-specific PRO completion reporting that the checklist mandates." }, { "relation_type": "used_with", "target_slug": "multiple-imputation-longitudinal-rwe", "notes": "Implements the missing-PRO-data analysis and sensitivity strategy CONSORT-PRO requires authors to describe." }, { "relation_type": "used_with", "target_slug": "mmrm-repeated-measures-rwe", "notes": "Standard analytic approach for the repeated PRO measures CONSORT-PRO reporting covers." }, { "relation_type": "see_also", "target_slug": "tipping-point-analysis-rwe", "notes": "Supports the not-at-random missing-data sensitivity analyses CONSORT-PRO expects for PRO endpoints." }, { "relation_type": "see_also", "target_slug": "composite-endpoint-construction-rwe", "notes": "Relevant when the PRO endpoint is a multi-domain or composite score whose construction must be reported." }, { "relation_type": "see_also", "target_slug": "qaly-utility-mapping-rwe", "notes": "When PRO/HRQoL data are mapped to utilities for cost-utility analysis in an HTA dossier." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "hta", "journal" ] }, { "slug": "consort", "name": "CONSORT", "short_definition": "Consolidated Standards of Reporting Trials (CONSORT) — the EQUATOR-endorsed reporting guideline for parallel-group randomized controlled trials, defining a 25-item checklist plus participant flow diagram. In RWE it governs the subset of studies that actually randomize, chiefly pragmatic and registry-based randomized trials, with dedicated extensions for those designs.", "long_description": "**What it is.** The **Consolidated Standards of Reporting Trials (CONSORT) Statement** is the foundational reporting\nguideline for **randomized controlled trials (RCTs)**. It is maintained by the CONSORT Group and endorsed by the EQUATOR\nNetwork, and it is the single most widely adopted reporting standard in clinical research, required by hundreds of\nbiomedical journals (ICMJE-aligned), regulators, and HTA bodies. The core statement covers **parallel-group, two-arm,\nindividually randomized superiority trials** through a structured checklist (25 items in CONSORT 2010; an expanded item\nset in **CONSORT 2025**) and a **participant flow diagram** that traces enrollment, allocation, follow-up, and analysis.\nCONSORT is a *reporting* standard: it specifies what a finished trial report must contain so that readers can judge\nvalidity, applicability, and reproducibility. It sits alongside its protocol-stage sibling **SPIRIT** (used before the\ntrial) and a family of design- and content-specific extensions (pragmatic, AI, PRO, cluster, non-inferiority, harms).\n\n**When to use.** Use CONSORT whenever a study **actually randomizes** an intervention to participants (or clusters) and\nthe output is a trial report for a peer-reviewed journal, a regulatory submission (FDA/EMA), or an HTA/payer dossier.\nWithin the real-world evidence space the relevant cases are narrow but important: **pragmatic randomized trials** embedded\nin routine care and **registry-based randomized trials (RRCTs)** that randomize but draw eligibility, baseline data, and/or\nendpoints from claims, EHR, or disease registries. Decision rules for which CONSORT artifact applies:\n- **A unit was randomized → CONSORT (this guideline) is in scope.** No randomization → use **STROBE** (general\n observational) or **RECORD / RECORD-PE** (routinely-collected health data / pharmacoepidemiology) instead.\n- **Pragmatic trial in real-world settings → CONSORT extension for pragmatic trials (consort-pragmatic)**, paired with\n **PRECIS-2** to characterize where the trial sits on the explanatory–pragmatic continuum.\n- **PRO endpoint is primary or key-secondary → CONSORT-PRO (consort-pro)**.\n- **AI/ML-based intervention → CONSORT-AI (consort-ai)**.\n- **Writing the protocol, not the report → SPIRIT (spirit)**, not CONSORT.\n\n**What it requires.** CONSORT enforces transparent reporting across the trial lifecycle, and several items become load-bearing\nin registry/pragmatic RWE trials where data are routinely collected rather than purpose-measured:\n- **Trial design and eligibility** — explicit design (parallel, allocation ratio), pre-specified eligibility, and settings;\n in pragmatic/registry trials, how routine-care eligibility was operationalized in the data source.\n- **Randomization and allocation concealment** — sequence generation, type (block/stratified), and concealment mechanism;\n these distinguish a true RRCT from an observational comparison and are what license a causal interpretation.\n- **Outcomes and estimands** — completely defined pre-specified primary/secondary outcomes and, under CONSORT 2025, the\n **estimand** and handling of **intercurrent events**; in registry trials this requires that the routinely-collected\n outcome be a validated phenotype, not a raw code.\n- **Participant flow (the CONSORT diagram)** — numbers randomized, receiving intervention, lost to follow-up, excluded,\n and analyzed per arm. In registry/claims-based trials this is where attrition, disenrollment, and informative censoring\n must be made visible.\n- **Baseline data, numbers analyzed, and the analysis population** — ITT vs per-protocol, and denominators per arm.\n- **Results, harms, ancillary analyses, and limitations** — effect sizes with precision, pre-specified vs exploratory\n subgroups, and harms (CONSORT-Harms).\n- **Registration, protocol availability, and funding** — trial registration number, access to the full protocol/SAP, and\n declared funding/role of sponsor.\n\n**When NOT to use — limitations and common misapplications.**\n- **It is not for observational/non-randomized studies.** Applying CONSORT to a claims- or EHR-based cohort, case-control,\n self-controlled, or target-trial-*emulation* study is a category error; randomization is the trigger. Use **STROBE** or\n **RECORD/RECORD-PE** for secondary-data observational work. A target-trial *emulation* uses CONSORT-like discipline\n conceptually but is reported under STROBE/RECORD, not CONSORT.\n- **It is not a risk-of-bias instrument and not a quality score.** Completeness of reporting is not internal validity. Use\n **RoB 2** (for RCTs) or **ROBINS-I** (for non-randomized studies) to appraise bias; do not sum CONSORT items into a score.\n- **Using the wrong extension.** A pragmatic registry trial reported against the base statement without\n **CONSORT-Pragmatic** (or PRECIS-2 positioning) misses the items reviewers care about most — generalizability,\n routine-care delivery, and pragmatic outcome ascertainment. AI interventions need CONSORT-AI; PRO endpoints need\n CONSORT-PRO; protocols need SPIRIT.\n- **Checklist-as-theater.** Page-number ticks against each item without substantive transparency (e.g., \"estimand reported\"\n pointing to a vague sentence) satisfies the form and fails the function. The flow diagram is meaningless if attrition is\n rolled up rather than itemized by reason and arm.\n\n**How it maps to this catalog.** CONSORT items map onto concrete RWE methods that implement them in pragmatic and\nregistry-based randomized trials:\n- Participant-flow and attrition reporting → **attrition-and-loss-to-follow-up-rwe** (and, for the eligibility funnel from\n a routinely-collected source, database-feasibility-attrition-funnel-rwe).\n- Estimand and intercurrent-event reporting (CONSORT 2025) → **estimands-ate-att-intercurrent-events-rwe**.\n- Follow-up onset and avoidance of immortal time when randomization and index dates diverge in registry data →\n **time-zero-index-date-alignment-rwe**.\n- Validated registry/claims outcome and eligibility phenotypes underpinning \"completely defined outcomes\" →\n **diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe** and **claims-analysis**.\n- Whether the chosen registry/claims source can credibly support the trial's measurements →\n **fit-for-purpose-data-assessment-rwe**.\n\n**Applied note (registry-based randomized trial in claims/EHR).** Consider an RRCT that randomizes treatment but\nascertains the primary endpoint from linked claims. CONSORT compliance here is more than a checklist: the flow diagram\nmust separate *randomization* attrition from *data* attrition (disenrollment, loss of linkage), the \"completely defined\noutcome\" item demands a validated claims phenotype with reported PPV/sensitivity rather than a single ICD code, and the\nCONSORT 2025 estimand item should state the intercurrent-event strategy (e.g., treatment policy vs while-on-treatment)\ngiven switching and discontinuation observed in routine pharmacy data. Pair the report with consort-pragmatic and PRECIS-2\nso reviewers can locate the trial on the pragmatic–explanatory spectrum.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "randomized-controlled-trial", "pragmatic-trial", "equator" ], "aliases": [ "CONSORT", "CONSORT 2010", "CONSORT 2025", "Consolidated Standards of Reporting Trials" ], "applies_to_study_types": [ "pragmatic_trial", "registry_trial" ], "data_sources": [ "registry", "ehr", "claims", "linked" ], "citations": [ { "role": "introduce", "doi": "10.1136/bmj.c332", "url": "https://doi.org/10.1136/bmj.c332", "citation_text": "Schulz KF, Altman DG, Moher D, for the CONSORT Group. CONSORT 2010 Statement: updated guidelines for reporting parallel group randomised trials. BMJ. 2010;340:c332.", "year": 2010, "authors_short": "Schulz et al.", "notes": "Canonical CONSORT 2010 statement — the 25-item checklist and participant flow diagram that define the standard." }, { "role": "explain", "doi": "10.1136/bmj.c869", "url": "https://doi.org/10.1136/bmj.c869", "citation_text": "Moher D, Hopewell S, Schulz KF, et al. CONSORT 2010 Explanation and Elaboration: updated guidelines for reporting parallel group randomised trials. BMJ. 2010;340:c869.", "year": 2010, "authors_short": "Moher et al.", "notes": "Item-by-item explanation and elaboration giving the rationale and worked examples for each CONSORT 2010 item." }, { "role": "use", "doi": "10.1001/jama.2025.4347", "url": "https://doi.org/10.1001/jama.2025.4347", "citation_text": "Hopewell S, Chan A-W, Collins GS, et al. CONSORT 2025 Statement: Updated Guideline for Reporting Randomized Trials. JAMA. 2025.", "year": 2025, "authors_short": "Hopewell et al.", "notes": "Current major update; adds and clarifies items including estimands, intercurrent events, harms, and open-science reporting. Use this version for new trial reports." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "pragmatic-trial", "notes": "CONSORT (with the pragmatic-trials extension) governs reporting of pragmatic randomized trials conducted in routine-care settings." }, { "relation_type": "applies_to", "target_slug": "registry-trial", "notes": "CONSORT governs reporting of registry-based randomized trials that randomize but draw eligibility, baseline, or outcomes from registries/claims/EHR." }, { "relation_type": "see_also", "target_slug": "consort-pragmatic", "notes": "Required extension for pragmatic randomized trials embedded in real-world care." }, { "relation_type": "see_also", "target_slug": "consort-pro", "notes": "Extension for trials where a patient-reported outcome is the primary or key-secondary endpoint." }, { "relation_type": "see_also", "target_slug": "consort-ai", "notes": "Extension for trials evaluating AI/ML-based interventions." }, { "relation_type": "see_also", "target_slug": "precis-2", "notes": "Use to position a CONSORT-reported trial on the pragmatic–explanatory continuum." }, { "relation_type": "complements", "target_slug": "spirit", "notes": "Protocol-stage sibling; SPIRIT (not CONSORT) is used to report the trial protocol before conduct." }, { "relation_type": "alternative_to", "target_slug": "strobe", "notes": "For non-randomized observational studies use STROBE, not CONSORT; randomization is the discriminator." }, { "relation_type": "alternative_to", "target_slug": "record-pe", "notes": "For pharmacoepidemiologic studies using routinely-collected data without randomization, use RECORD-PE rather than CONSORT." }, { "relation_type": "complements", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Implements the CONSORT participant-flow diagram when registry/claims attrition and informative censoring must be made visible by arm." }, { "relation_type": "complements", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Operationalizes the CONSORT 2025 estimand and intercurrent-event reporting items in real-world randomized trials." }, { "relation_type": "complements", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Provides the validated phenotype behind a \"completely defined outcome\" when endpoints are ascertained from registry/claims data." }, { "relation_type": "complements", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "Establishes whether the registry/claims source can credibly support the trial's eligibility, baseline, and outcome measurements." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "hta", "journal" ] }, { "slug": "coreq", "name": "COREQ (Consolidated Criteria for Reporting Qualitative Research)", "short_definition": "A 32-item reporting checklist, organized in three domains, for transparent reporting of qualitative studies that collect data through individual interviews or focus groups.", "long_description": "**What it is** — COREQ (Consolidated Criteria for Reporting Qualitative Research) is a 32-item\nreporting checklist published by Tong, Sainsbury, and Craig in 2007 and maintained as an EQUATOR Network\nreporting guideline. It was developed by reviewing 22 prior qualitative checklists and consolidating their\nmost frequently recommended items into a single instrument for *interview- and focus-group-based* studies.\nThe 32 items sit in three domains: (1) **Research team and reflexivity** — interviewer/facilitator identity,\ncredentials, occupation, gender, training, and the relationship with participants (prior contact,\ninterviewer assumptions, reasons for the study); (2) **Study design** — theoretical/methodological\norientation (grounded theory, phenomenology, content analysis, etc.), participant selection and recruitment,\nsample size, non-participation, setting, presence of non-participants, and a description of the sample;\ndata collection covering the interview guide, pilot testing, repeat interviews, audio/visual recording, field\nnotes, duration, and whether data saturation was discussed; and (3) **Analysis and findings** — number of\ncoders, coding-tree derivation, software, participant checking (member checking), quotations with participant\nidentifiers, consistency between data and findings, presentation of major and minor themes. COREQ is a\n*reporting* tool: it specifies what must be disclosed so a reader can judge a study, not a verdict on whether\nthe study was done well.\n\n**When to use** — Reach for COREQ when you are reporting (or peer-reviewing, or registering a protocol for) a\nqualitative study whose primary data are **one-on-one interviews or focus groups**. In RWE/HEOR programs this\nis most often a qualitative arm rather than a standalone paper: concept-elicitation and cognitive-debriefing\ninterviews that establish content validity of a patient-reported outcome (PRO) instrument under the FDA PRO\nGuidance; the formative qualitative phase of a stated-preference / DCE study; patient, caregiver, clinician,\nor payer interview substudies embedded in a registry, observational, or mixed-methods program; and exit\ninterviews appended to pragmatic or single-arm studies. Decision rule for choosing COREQ over its siblings:\nif the data are interviews/focus groups, use COREQ; if the qualitative work spans broader methods\n(ethnography, document/textual analysis, observation), use **SRQR** (O'Brien 2014) instead; if you are\nsynthesizing primary qualitative studies, use **ENTREQ**, not COREQ. COREQ supports — it does not replace —\nthe design-level reporting standard for the quantitative components (STROBE/RECORD-PE for the observational\nanalysis, CONSORT-PRO/SPIRIT-PRO for trial-embedded PROs, CHEERS for the economic model).\n\n**What it requires** — The substantive disclosures COREQ enforces are specific to qualitative inquiry and do\nnot overlap with quantitative RWE reporting domains. It demands: **reflexivity** — who collected the data,\ntheir training and standpoint, and their relationship to participants, because the interviewer is the\nmeasurement instrument; **transparent sampling** — how participants were identified, approached, and selected,\nwho declined, and a sample description, so readers can judge transferability rather than statistical\nrepresentativeness; **instrument and procedure transparency** — the interview/topic guide, pilot testing,\nsetting, recording method, field notes, and an explicit statement on whether thematic saturation was reached;\nand **analytic auditability** — number of coders, how the coding framework was derived, analysis software,\nwhether participants reviewed findings, and verbatim quotations tied to identifiers that evidence each theme.\nThese map onto the catalog's *qualitative* fitness-for-purpose ideas (credibility, dependability, confirmability,\ntransferability) rather than phenotype validation, time-zero alignment, estimands, or confounding control,\nwhich are simply not in scope for an interview study.\n\n**When NOT to use — limitations and common misapplications** — (1) **It is a reporting checklist, not a\nrisk-of-bias or quality instrument.** A study that reports all 32 items is *fully described*, not *rigorous*;\nuse JBI's qualitative critical-appraisal checklist (or CASP) to appraise methodological soundness. Conflating\nCOREQ completeness with study quality is the single most common error. (2) **Wrong design.** Applying COREQ to\nethnography, observational fieldwork, document analysis, or grounded-theory work that is not interview/focus-\ngroup-based forces a mismatched instrument — use SRQR. (3) **Wrong activity.** COREQ does not govern qualitative\nevidence synthesis (use ENTREQ) and has nothing to say about the secondary-data, quantitative analyses in an\nRWE study — STROBE, RECORD-PE, HARPER, and CHEERS own those. Using COREQ where RECORD-PE/HARPER is required\nleaves the design-transparency gaps that regulators and HTA reviewers actually scrutinize. (4) **Checklist-as-\ntheater.** A submitted COREQ grid with page numbers that do not contain the claimed content, or items marked\n\"N/A\" without justification, is worse than none. (5) **Over-literal scoring.** Buus & Perron (2020) replicated\nCOREQ's development and documented inconsistent interpretation of several items; treat COREQ as a structured\nprompt for transparent narrative, not as a numeric score to be summed or thresholded.\n\n**How it maps to this catalog** — COREQ is the reporting layer for the catalog's qualitative and\nPRO-development concepts. It directly governs reporting for **qualitative-interview** (its core use case) and\nthe qualitative arm of **mixed-methods** programs. Its strongest RWE/HEOR anchor is **pro-development** and\n**pro-validation**: concept-elicitation and cognitive-debriefing interviews that establish PRO content validity\nshould be reported against COREQ so reviewers can audit saturation, sampling, and coding. It applies to the\nqualitative formative phase of a **preference-study** (the interviews that generate DCE attributes and levels).\nIt is *distinct from* **qualitative-synthesis**, which is reported under ENTREQ — a deliberate contrast, not an\noverlap. For ethnographic designs see **qualitative-ethnographic**, where SRQR rather than COREQ is the\nappropriate standard. Applied note for claims/EHR/registry RWE: COREQ does not govern the secondary-data\nanalysis itself — when a database study has no primary qualitative data collection, COREQ does not apply.\nIt becomes relevant only when a qualitative component is bolted onto the program (e.g., clinician interviews\nto validate an EHR phenotype's clinical face validity, patient interviews to interpret an unexpected\nregistry signal, or payer interviews in an HTA dossier). Report those qualitative substudies under COREQ and\nkeep the quantitative analysis under its own design-appropriate guideline.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "qualitative", "patient-reported-outcomes" ], "aliases": [ "COREQ", "Consolidated Criteria for Reporting Qualitative Research", "Tong COREQ checklist", "COREQ 32-item checklist" ], "applies_to_study_types": [ "qualitative_interview", "mixed_methods" ], "data_sources": [ "primary" ], "citations": [ { "role": "introduce", "doi": "10.1093/intqhc/mzm042", "url": "https://doi.org/10.1093/intqhc/mzm042", "citation_text": "Tong A, Sainsbury P, Craig J. Consolidated criteria for reporting qualitative research (COREQ): a 32-item checklist for interviews and focus groups. International Journal for Quality in Health Care. 2007;19(6):349-357.", "year": 2007, "authors_short": "Tong et al.", "notes": "Canonical statement paper; derives the 32 items by consolidating 22 prior qualitative checklists into three domains (research team/reflexivity, study design, analysis/findings)." }, { "role": "explain", "doi": "10.1016/j.ijnurstu.2019.103452", "url": "https://doi.org/10.1016/j.ijnurstu.2019.103452", "citation_text": "Buus N, Perron A. The quality of quality criteria: replicating the development of the Consolidated Criteria for Reporting Qualitative Research (COREQ). International Journal of Nursing Studies. 2020;102:103452.", "year": 2020, "authors_short": "Buus & Perron", "notes": "Critical appraisal that replicates COREQ's development; documents inconsistent interpretation of several items and cautions against treating COREQ as a quality score rather than a reporting aid." }, { "role": "use", "url": "https://www.equator-network.org/reporting-guidelines/coreq/", "citation_text": "Consolidated Criteria for Reporting Qualitative Research (COREQ) — EQUATOR Network maintained reporting-guideline page and downloadable 32-item checklist.", "year": 2007, "authors_short": "EQUATOR Network", "notes": "Authoritative maintained source for the current checklist and links to translations and exemplars." } ], "relations": [ { "relation_type": "used_with", "target_slug": "qualitative-interview", "notes": "COREQ is the primary reporting standard for interview- and focus-group-based qualitative studies." }, { "relation_type": "used_with", "target_slug": "pro-development", "notes": "Concept-elicitation and cognitive-debriefing interviews that establish PRO content validity should be reported against COREQ so reviewers can audit sampling, saturation, and coding." }, { "relation_type": "used_with", "target_slug": "pro-validation", "notes": "Cognitive-interview evidence supporting PRO validity is reported with COREQ alongside the psychometric results." }, { "relation_type": "used_with", "target_slug": "preference-study", "notes": "Apply COREQ to the formative qualitative phase (attribute/level-generating interviews) of a stated-preference or discrete-choice study." }, { "relation_type": "used_with", "target_slug": "mixed-methods", "notes": "In mixed-methods RWE, COREQ governs the qualitative arm while the quantitative arm follows its own design-appropriate reporting guideline." }, { "relation_type": "alternative_to", "target_slug": "qualitative-ethnographic", "notes": "For ethnographic and observational designs, SRQR (not COREQ) is the appropriate reporting standard." }, { "relation_type": "see_also", "target_slug": "qualitative-synthesis", "notes": "Synthesis of primary qualitative studies is reported under ENTREQ; COREQ covers the primary studies, not their synthesis." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "hta", "journal" ] }, { "slug": "cosmin-criteria", "name": "COSMIN Criteria for Good Measurement Properties", "short_definition": "COSMIN's consensus rating thresholds that classify each measurement property of a patient-reported outcome measure (PROM) as sufficient (+), insufficient (-), inconsistent (±), or indeterminate (?), used to judge whether an instrument is good enough for a given purpose.", "long_description": "**What it is** — The **COSMIN Criteria for Good Measurement Properties** are the consensus rating\nthresholds developed by the **COSMIN initiative** (COnsensus-based Standards for the selection of health\nMeasurement INstruments, hosted at Amsterdam UMC) for deciding whether the *result* of a measurement-property\nstudy on a **patient-reported outcome measure (PROM)** is acceptable. For each property — content validity,\nstructural validity, internal consistency, cross-cultural validity/measurement invariance, reliability,\nmeasurement error, criterion validity, construct validity (hypotheses testing), and responsiveness — the\ncriteria define when the estimate is rated **sufficient (+)**, **insufficient (−)**, **inconsistent (±)**, or\n**indeterminate (?)**. They are one of three distinct COSMIN tools and must not be confused with the other two:\nthe **COSMIN Risk of Bias (RoB) checklist** rates the *methodological quality* of the study that produced the\nestimate, and the **COSMIN reporting guideline** governs how a PROM study is written up. In a systematic review\nof PROMs the three are used in sequence (RoB → criteria for good measurement properties → modified GRADE for\nquality of evidence) to reach an evidence-based instrument recommendation.\n\n**When to use** — Apply the criteria whenever you must judge whether a PROM's measurement properties are good\nenough for a defined use: selecting an instrument for a trial, registry, or clinical program; appraising a newly\ndeveloped/validated PROM; or, most often, performing a **systematic review of PROMs**. In RWE/HEOR they apply\nwhen a PRO endpoint anchors the evidence — HRQoL trajectories in a disease registry, ePRO symptom capture in an\nEHR-linked cohort, or a preference/utility instrument feeding QALYs in an HTA dossier — and the instrument's\npsychometrics must be defended as fit-for-purpose for the population and mode of administration actually used.\nDecision rules for picking the right COSMIN tool (vs a sibling/extension): use **these criteria** to RATE a\nproperty estimate (is it good?); use the **COSMIN RoB checklist** to judge whether the study estimating it was\nwell conducted; use the **COSMIN reporting guideline** to REPORT a development/validation study; and use\n**PRISMA-COSMIN** to report a systematic review of PROMs. For diagnostic tests use STARD/QUADAS, and for\nprediction models use TRIPOD — COSMIN is for measurement instruments of latent constructs, not test accuracy.\n\n**What it requires** — The criteria operationalize each property against explicit thresholds, and the order\nmatters. **Content validity** (relevance, comprehensiveness, comprehensibility) is treated as the most important\nproperty and is rated first. **Structural validity** is judged on confirmatory factor-analysis fit (e.g.,\nCFI/TLI and RMSEA/SRMR within accepted bounds) or IRT/Rasch fit; **internal consistency** (Cronbach's alpha\n≥ 0.70) is only interpretable once at least low-quality evidence for sufficient structural validity exists.\n**Reliability** (ICC or weighted kappa ≥ 0.70) and **measurement error** (smallest detectable change versus the\nminimal important change) gauge stability. **Construct validity** and **responsiveness** require that ≥ 75% of\n*pre-specified* hypotheses about expected correlations or known-group/change differences are confirmed;\n**criterion validity** needs a genuine gold standard (correlation/AUC ≥ 0.70), which is rare for PRO constructs.\n**Cross-cultural validity/measurement invariance** requires no important differential item functioning across\nthe relevant subgroups, languages, or administration modes. For RWE this maps directly onto data-fitness-for-use\nand transportability: each property must hold in a population resembling the real-world cohort, at the recall\nperiod and mode (paper vs ePRO) actually deployed, with PRO missingness/attrition treated as potentially\ninformative rather than ignorable.\n\n**When NOT to use — limitations and common misapplications** — (1) The criteria are **not a risk-of-bias\ninstrument**: they rate whether a property *result* is good, not whether the *study* was sound — that is the RoB\nchecklist's job, and conflating the two is the single most common error. (2) They are **not a quality score**:\ndo not sum the per-property ratings into a composite \"COSMIN score\" or rank instruments by an arithmetic total;\nthe output is a property-by-property profile plus a GRADE-style certainty rating. (3) They were built for PROMs\n(and adaptable to clinician/observer-reported and performance-based outcomes), **not** for laboratory assays,\nimaging biomarkers, or diagnostic tests. (4) Passing all thresholds in one validation sample does **not**\ntransport — invariance must be re-examined whenever the RWE population, language, or administration mode differs\nfrom the validation study. (5) Applying defaults blindly: alpha ≥ 0.70 is meaningless without confirmed\nunidimensionality/structural validity first, and \"criterion validity\" claims collapse when no true gold standard\nexists. (6) **Checklist-as-theater**: labeling an instrument \"COSMIN-validated\" without stating which properties\nwere rated, in whom, and against which thresholds is a hallmark of misuse that reviewers will reject.\n\n**How it maps to this catalog** — The criteria are the appraisal layer over the catalog's PRO concepts. Content\nvalidity and instrument generation are implemented in **pro-development**; estimation of structural validity,\ninternal consistency, reliability, construct validity, and responsiveness lives in **pro-validation**; using PRO\nendpoints inside observational designs is **pro-rwe**; HRQoL constructs are **hrqol**; and preference/utility\ninstruments that feed cost-utility work are **qaly-utility-mapping-rwe**. Informative PRO missingness and\ncompletion attrition — central to whether measurement properties hold in real-world capture — are handled by\n**missing-data-pattern-table-rwe**, **multiple-imputation-longitudinal-rwe**, and **attrition-and-loss-to-follow-up-rwe**.\nWhether the chosen instrument remains valid in a population unlike its validation sample is a\n**generalizability-transportability-external-validity-rwe** question. *Applied claims/EHR/registry note:* PROMs\nalmost never appear in administrative claims; they enter RWE through disease/product registries, EHR-embedded\nePRO, or linked patient surveys. When a PRO endpoint anchors such a study, use the criteria to confirm the\ninstrument's properties are adequate at the administration mode and in a cohort resembling your real-world\npopulation — then document PRO missingness explicitly, because non-completion is rarely at random.", "primary_category": "Guideline", "tags": [ "guideline", "quality_assessment", "patient-reported-outcomes", "measurement-properties", "psychometrics", "prom", "cosmin" ], "aliases": [ "COSMIN", "COSMIN criteria for good measurement properties", "COSMIN quality criteria", "COSMIN measurement property criteria" ], "applies_to_study_types": [ "pro_development", "pro_validation" ], "data_sources": [ "registry", "ehr", "linked", "primary" ], "citations": [ { "role": "introduce", "doi": "10.1007/s11136-018-1798-3", "url": "https://doi.org/10.1007/s11136-018-1798-3", "citation_text": "Prinsen CAC, Mokkink LB, Bouter LM, et al. COSMIN guideline for systematic reviews of patient-reported outcome measures. Quality of Life Research. 2018;27(5):1147-1157.", "year": 2018, "authors_short": "Prinsen et al.", "notes": "Canonical COSMIN guideline that sets out the updated criteria for good measurement properties (the sufficient/insufficient/indeterminate rating thresholds) and their place in the RoB → criteria → modified GRADE workflow for reviews of PROMs." }, { "role": "explain", "doi": "10.1016/j.jclinepi.2006.03.012", "url": "https://doi.org/10.1016/j.jclinepi.2006.03.012", "citation_text": "Terwee CB, Bot SDM, de Boer MR, et al. Quality criteria were proposed for measurement properties of health status questionnaires. Journal of Clinical Epidemiology. 2007;60(1):34-42.", "year": 2007, "authors_short": "Terwee et al.", "notes": "Origin of the explicit quality-criteria thresholds (e.g., the 75%-of-hypotheses rule, alpha and ICC cut-offs) that the COSMIN criteria refined and standardized." }, { "role": "use", "url": "https://www.cosmin.nl/tools/guideline-conducting-systematic-review-outcome-measures/", "citation_text": "COSMIN initiative. Criteria for good measurement properties and user manual for systematic reviews of PROMs (maintained tools and resources).", "year": 2018, "authors_short": "COSMIN initiative", "notes": "Maintained source for the current criteria, rating tables, and accompanying user manual." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "pro-development", "notes": "Content-validity criteria (relevance, comprehensiveness, comprehensibility) appraise instruments produced during PROM development." }, { "relation_type": "applies_to", "target_slug": "pro-validation", "notes": "The criteria rate the structural, reliability, construct-validity, and responsiveness estimates produced by psychometric validation studies." }, { "relation_type": "used_with", "target_slug": "pro-validation", "notes": "The criteria are the appraisal layer applied to property estimates generated under the pro-validation concept." }, { "relation_type": "see_also", "target_slug": "pro-development", "notes": "Implements the content-validity requirements (item generation, cognitive interviews) that the criteria rate as most important." }, { "relation_type": "see_also", "target_slug": "pro-rwe", "notes": "Use the criteria to confirm a PRO endpoint's instrument is fit-for-purpose before using it in observational/real-world studies." }, { "relation_type": "see_also", "target_slug": "hrqol", "notes": "HRQoL instruments are the prototypical PROMs whose measurement properties the criteria evaluate." }, { "relation_type": "see_also", "target_slug": "qaly-utility-mapping-rwe", "notes": "Preference/utility instruments feeding QALYs should clear the relevant measurement-property criteria before their scores anchor cost-utility analyses." }, { "relation_type": "see_also", "target_slug": "generalizability-transportability-external-validity-rwe", "notes": "A PROM validated in one sample must show measurement invariance before its properties can be assumed in a different real-world population, mode, or language." }, { "relation_type": "see_also", "target_slug": "missing-data-pattern-table-rwe", "notes": "Real-world PRO completion is frequently non-random; informative missingness threatens the validity of property estimates and downstream analyses." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "hta", "journal" ] }, { "slug": "cosmin-reporting", "name": "COSMIN Reporting Guideline for Studies on Measurement Properties of PROMs", "short_definition": "A reporting guideline (EQUATOR-listed, maintained by the COSMIN initiative) specifying the minimum items authors must report for a single study that develops or evaluates a measurement property (reliability, validity, responsiveness, interpretability) of a patient-reported outcome measure.", "long_description": "**What it is.** The **COSMIN Reporting Guideline for Studies on Measurement Properties of Patient-Reported Outcome\nMeasures (COSMIN-RG)** is a 12-item, EQUATOR-Network-listed reporting checklist published by Gagnier, Lai, Mokkink, and\nTerwee (Gagnier et al., *Quality of Life Research*, 2021) and maintained by the **COSMIN** (COnsensus-based Standards for\nthe selection of health Measurement INstruments) initiative. Its purpose is narrow and specific: to make the *reporting* of\nan individual study on the measurement properties of a PROM complete and transparent, so that readers can judge what was\ndone and reuse or appraise the evidence. It sits inside the broader COSMIN methodology, which rests on a consensus taxonomy\nof measurement properties — content validity, structural validity, internal consistency, cross-cultural validity/measurement\ninvariance, reliability, measurement error, criterion validity, hypotheses-testing construct validity, and responsiveness\n(Mokkink et al., *J Clin Epidemiol*, 2010). COSMIN-RG is a *reporting* tool only. It is deliberately distinct from its two\nsiblings, with which it is constantly confused: the **COSMIN Risk of Bias (ROB) checklist** (a methodological-quality /\nrisk-of-bias appraisal tool) and the **COSMIN guideline for systematic reviews of PROMs** (Prinsen et al., *Quality of Life\nResearch*, 2018), which governs how to *synthesize and grade* PROM evidence across studies rather than how to report one.\n\n**When to use.** Apply COSMIN-RG whenever you author or referee a **single primary study whose object is a measurement\nproperty of a PROM** — instrument development, content-validity studies, factor-analytic/structural-validity studies,\ntest-retest reliability, measurement error, construct or criterion validity, cross-cultural adaptation, or responsiveness.\nThis is the relevant checklist for a journal manuscript reporting PROM psychometrics, for the measurement-property appendix\nof an FDA Patient-Focused Drug Development (PFDD) / Clinical Outcome Assessment (COA) submission, for an EMA dossier that\nrelies on a PRO endpoint, and for an HTA submission where the validity of the HRQoL/utility instrument underpins the value\nclaim. Decision rule for choosing the right COSMIN tool: reporting one measurement-property study → **COSMIN-RG (this\nguideline)**; appraising the risk of bias of such studies → **COSMIN ROB checklist**; conducting a systematic review that\nselects or recommends a PROM → **COSMIN-SR / Prinsen 2018**. If your manuscript is a *clinical* RWE study that merely uses a\nPROM as an endpoint (rather than studying the instrument itself), COSMIN-RG is not your primary reporting guideline —\nSTROBE/RECORD-PE govern the design, and COSMIN-evaluated evidence is what you *cite* to justify the instrument.\n\n**What it requires.** COSMIN-RG enforces complete reporting across the elements that let a reader reconstruct and appraise a\nmeasurement-property study: (1) the **construct and target population**, the PROM and its versions, and the intended use\n(discriminative, evaluative, predictive); (2) the **measurement property(ies) studied** and an a-priori statement of\nhypotheses for construct validity and responsiveness (expected direction and magnitude of correlations / known-groups\ndifferences), so that \"confirmation\" is not retrofitted; (3) **study design and sampling** — recruitment, setting, eligibility,\nadministration mode, and the time interval and stability assumption for test-retest reliability and responsiveness; (4)\n**sample size** and its justification; (5) **statistical methods** appropriate to each property (e.g., CFA/IRT fit indices\nfor structural validity, ICC model and form for reliability, smallest detectable change / limits of agreement for measurement\nerror, correlations and AUC for responsiveness); (6) **missing-data handling and attrition** across administrations, which is\ncentral because test-retest and responsiveness require repeated measurement and are vulnerable to informative loss; and (7)\nfull **results with uncertainty** plus interpretability anchors (minimal important change/difference, floor/ceiling effects).\nReframed for real-world data, COSMIN-RG is the discipline that makes a PROM \"fit for purpose\" before it is wired into a\nclaims/EHR/registry study: it documents the instrument's provenance, the population it was validated in (and therefore the\ntransportability of its scores), and the responsiveness evidence that justifies treating a score change as a real signal.\n\n**When NOT to use — limitations and common misapplications.**\n- **It is a reporting checklist, not a quality score or a risk-of-bias instrument.** Ticking all 12 items certifies that you\n *reported* the study completely; it says nothing about whether the design was sound. Risk of bias is the job of the separate\n **COSMIN ROB checklist** — do not present a completed COSMIN-RG as evidence the PROM is valid.\n- **Wrong sibling for the task.** Using COSMIN-RG to structure a *systematic review* of PROMs (where Prinsen 2018 / COSMIN-SR\n and the GRADE-style evidence synthesis apply), or to *appraise* primary studies, is a scope error that senior reviewers\n catch immediately.\n- **Wrong guideline for a clinical RWE study.** Applying COSMIN-RG to a comparative-effectiveness or safety study that simply\n *uses* a PROM endpoint is a category error; that study reports under STROBE/RECORD-PE, with COSMIN evidence cited to support\n instrument choice.\n- **Checklist-as-theater.** A complete checklist on an underpowered single-administration study, or one that declares\n construct-validity hypotheses *after* seeing the correlations, is transparent reporting of a weak study — not strong evidence.\n- **Construct/population drift.** A PROM validated in one population, language, or administration mode is not automatically valid\n in another; reporting cross-cultural validity / measurement invariance does not waive the need to confirm it in the RWE target\n population.\n\n**How it maps to this catalog.** COSMIN-RG governs the *measurement-property evidence layer* that RWE studies depend on; it\ndoes **not** implement causal-inference machinery (target-trial emulation, high-dimensional propensity scores, active-comparator\nnew-user designs), and claiming it does would itself be the \"checklist-as-theater\" failure mode above. Map it as follows.\nUpstream — the studies COSMIN-RG actually reports on — are implemented by **pro-development** and **pro-validation** (instrument\ncreation and psychometric evaluation) and **pro-rwe** (deploying PROs in real-world settings). Downstream — RWE analyses that\n*consume* a COSMIN-evaluated PROM — are where the measurement evidence pays off: **hrqol** and **qaly-utility-mapping-rwe**\n(HRQoL/utility instruments whose validity and responsiveness must be COSMIN-documented before scores drive cost-utility models),\n**estimands-ate-att-intercurrent-events-rwe** (PRO endpoints have intercurrent events — death, treatment switching,\nrescue therapy — that the estimand must address, and COSMIN's responsiveness/interpretability evidence defines what a\nmeaningful change is), and **missing-data-pattern-table-rwe** with **attrition-and-loss-to-follow-up-rwe** (repeated PROM\nadministration for test-retest and responsiveness makes informative missingness the dominant threat, which COSMIN-RG forces\nauthors to report and which these concepts handle analytically). For evidence synthesis, route to the sibling tools, not\nCOSMIN-RG. Applied note for claims/EHR/registry RWE: claims and most EHR data do **not** contain PROM scores at all, so the\npractical pattern is a linked or registry/primary-collection substrate; before a linked PROM endpoint is trusted, confirm via\nCOSMIN-RG-reported evidence that the instrument was validated in a comparable population and mode (paper vs ePRO vs telephone),\nthat a minimal important change is established, and that the responsiveness study's attrition would not bias the real-world\nsignal.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "prom", "measurement-properties", "patient-reported-outcomes", "psychometrics", "cosmin" ], "aliases": [ "COSMIN Reporting Guideline", "COSMIN-RG", "COSMIN PROM reporting guideline", "COSMIN reporting checklist for measurement-property studies" ], "applies_to_study_types": [ "pro_development", "pro_validation" ], "data_sources": [ "registry", "linked", "primary" ], "citations": [ { "role": "introduce", "doi": "10.1007/s11136-021-02822-4", "url": "https://doi.org/10.1007/s11136-021-02822-4", "citation_text": "Gagnier JJ, Lai J, Mokkink LB, Terwee CB. COSMIN reporting guideline for studies on measurement properties of patient-reported outcome measures. Quality of Life Research. 2021;30(8):2197-2218.", "year": 2021, "authors_short": "Gagnier et al.", "notes": "Canonical statement of the 12-item COSMIN reporting guideline (COSMIN-RG) for individual measurement-property studies; EQUATOR-listed." }, { "role": "explain", "doi": "10.1016/j.jclinepi.2010.02.006", "url": "https://doi.org/10.1016/j.jclinepi.2010.02.006", "citation_text": "Mokkink LB, Terwee CB, Patrick DL, et al. The COSMIN study reached international consensus on taxonomy, terminology, and definitions of measurement properties for health-related patient-reported outcomes. Journal of Clinical Epidemiology. 2010;63(7):737-745.", "year": 2010, "authors_short": "Mokkink et al.", "notes": "Consensus taxonomy and definitions of measurement properties that the reporting items are organized around (reliability, validity, responsiveness, interpretability)." }, { "role": "explain", "doi": "10.1007/s11136-018-1798-3", "url": "https://doi.org/10.1007/s11136-018-1798-3", "citation_text": "Prinsen CAC, Mokkink LB, Bouter LM, et al. COSMIN guideline for systematic reviews of patient-reported outcome measures. Quality of Life Research. 2018;27(5):1147-1157.", "year": 2018, "authors_short": "Prinsen et al.", "notes": "Sibling guideline for systematic reviews of PROMs — cite to distinguish scope; use this (not COSMIN-RG) when synthesizing/recommending a PROM across studies." }, { "role": "use", "url": "https://www.equator-network.org/reporting-guidelines/cosmin-reporting-guideline-for-studies-on-measurement-properties-of-patient-reported-outcome-measures/", "citation_text": "COSMIN. COSMIN Reporting Guideline for studies on measurement properties of PROMs — checklist and user manual. COSMIN initiative (also listed on the EQUATOR Network).", "year": 2021, "authors_short": "COSMIN initiative", "notes": "Maintained checklist, manual, and templates; primary source for the operational item list." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "pro-development", "notes": "Use when reporting a study that develops a PROM (item generation, content validity, initial structural validity)." }, { "relation_type": "applies_to", "target_slug": "pro-validation", "notes": "Use when reporting a study evaluating measurement properties (reliability, validity, responsiveness, measurement error) of an existing PROM." }, { "relation_type": "see_also", "target_slug": "pro-development", "notes": "Concept implementing the upstream instrument-development work that COSMIN-RG reports on." }, { "relation_type": "see_also", "target_slug": "pro-validation", "notes": "Concept implementing the psychometric evaluation (reliability/validity/responsiveness) whose reporting COSMIN-RG standardizes." }, { "relation_type": "see_also", "target_slug": "pro-rwe", "notes": "Real-world deployment of PROs; COSMIN-RG-reported evidence is the prerequisite for trusting a PROM as a real-world endpoint." }, { "relation_type": "used_with", "target_slug": "hrqol", "notes": "HRQoL instruments require COSMIN-documented validity and responsiveness before score changes are treated as meaningful." }, { "relation_type": "used_with", "target_slug": "qaly-utility-mapping-rwe", "notes": "Utility/HRQoL instruments feeding cost-utility models depend on the measurement-property evidence COSMIN-RG makes transparent." }, { "relation_type": "used_with", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "PRO endpoints carry intercurrent events (death, switching, rescue); COSMIN responsiveness/interpretability evidence defines the meaningful change the estimand targets." }, { "relation_type": "used_with", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Repeated PROM administration (test-retest, responsiveness) makes informative attrition the dominant threat that COSMIN-RG forces authors to report." }, { "relation_type": "see_also", "target_slug": "missing-data-pattern-table-rwe", "notes": "Missing-data documentation across PROM administrations is a core COSMIN-RG reporting requirement." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "hta", "journal" ] }, { "slug": "cosmin-rob", "name": "COSMIN Risk of Bias Checklist (PROMs)", "short_definition": "A standardized tool from the COSMIN initiative for rating the methodological quality (risk of bias) of individual studies on the measurement properties of patient-reported outcome measures, used to grade the evidence supporting a PROM within a systematic review.", "long_description": "**What it is** — The **COSMIN Risk of Bias checklist for systematic reviews of Patient-Reported Outcome Measures\n(PROMs)** is a consensus instrument from the **COSMIN initiative** (COnsensus-based Standards for the selection of health\nMeasurement INstruments; Amsterdam UMC — Mokkink, Terwee, Prinsen, de Vet and an international Delphi panel). Its purpose\nis narrow and specific: to rate the **methodological quality (risk of bias) of each individual study** that reports on a\nmeasurement property of a PROM, so that a systematic reviewer can decide how much to trust each measurement-property\nresult before pooling and grading the evidence. It is organized into **ten \"boxes,\" one per measurement property** — PROM\ndevelopment, content validity, structural validity, internal consistency, cross-cultural validity / measurement\ninvariance, reliability, measurement error, criterion validity, hypotheses testing for construct validity, and\nresponsiveness — each containing standards rated on a four-point scale (very good / adequate / doubtful / inadequate)\nunder a deliberate **\"worst-score-counts\"** rule. It is one piece of a larger COSMIN system: the **COSMIN guideline for\nsystematic reviews of PROMs** (Prinsen 2018) governs the full review workflow, and the **COSMIN methodology for content\nvalidity** (Terwee 2018) elaborates the content-validity box, which COSMIN treats as the single most important property.\n\n**When to use** — Use the COSMIN RoB checklist whenever you must **appraise the evidence base for a PRO instrument**: a\nsystematic review to select the best PROM for a construct and population; head-to-head comparison of candidate\ninstruments; or single-study critical appraisal of a validation paper. In HEOR/RWE the decision context is upstream of the\nstudy itself — it is how you **justify the PRO endpoint** you intend to use. Before a PRO supports a label claim (FDA PRO\nguidance), an HRQoL or symptom endpoint in an EMA submission, or a utility/HRQoL input in an HTA dossier, COSMIN RoB is\nthe defensible way to show the instrument's structural validity, reliability, measurement error, and responsiveness were\nestablished in low-risk-of-bias studies in a relevant population. **Decision rules for which COSMIN tool applies:** use the\n**RoB checklist** to rate risk of bias of each measurement-property study; use the **COSMIN guideline (Prinsen)** for the\nend-to-end review steps (search, eligibility, evidence synthesis, GRADE, recommendation); use the **content-validity\nmethodology (Terwee)** when content validity is the focus. If your task is *reporting* a PRO result rather than appraising\nan instrument, COSMIN is the wrong family — reach for CONSORT-PRO (trials) or ISOQOL reporting standards instead.\n\n**What it requires** — Within each of the ten boxes the checklist enforces **design requirements and preferred statistical\nmethods** appropriate to that property: e.g., confirmatory factor analysis or IRT/Rasch for structural validity; an\nadequate sample and internal-consistency statistic (Cronbach's alpha / KR-20) computed on a unidimensional set for\ninternal consistency; ICC or weighted kappa under stable conditions for reliability; standard error of measurement /\nsmallest detectable change / Bland-Altman limits for measurement error; **a priori, directional hypotheses** for construct\nvalidity and responsiveness (so that \"confirming\" post hoc expectations cannot be passed off as validity); and explicit\nhandling of missing PRO data and of attrition in the longitudinal designs used for responsiveness. PROM development and\ncontent validity are given primacy because a PROM that does not measure the right content cannot be rescued by good\npsychometrics downstream. Boxes are rated per standard and aggregated by **taking the lowest rating** in the box — there\nis **no summary numeric quality score**.\n\n**When NOT to use — limitations and common misapplications** — (1) It is a **risk-of-bias / methodological-quality**\ninstrument, **not a reporting checklist** and **not a measurement-property results-rating tool**. Conflating \"low risk of\nbias\" with \"good measurement properties\" is the cardinal error: how well a study was *done* is rated by COSMIN RoB, while\nwhether the *result* is good is judged against the separate \"criteria for good measurement properties\" and graded with\nGRADE. (2) It is **not a generic study-quality score** — there is no overall number, and the worst-score-counts logic is\nintentionally conservative. (3) It is **scoped to PROMs**; applying it to clinician-reported, performance-based, or\nlaboratory outcomes, or — most relevant to RWE — to a **claims- or EHR-derived outcome/phenotype algorithm**, is a\ncategory error. Validating a computable phenotype is PPV/sensitivity/chart-review work, not PROM measurement-property\nappraisal; COSMIN's boxes simply do not map onto an ICD-code algorithm. (4) **Wrong tool within the family** (using the\nRoB checklist when the question demands the full Prinsen guideline, or an outdated version of the content-validity\nstandards). (5) **Checklist-as-theater** — marking boxes without the design/statistics they require produces a confident\nrating that means nothing.\n\n**How it maps to this catalog** — The PRO-instrument concepts this checklist governs are **pro-development**,\n**pro-validation**, **pro-rwe**, and **hrqol** (and **qaly-utility-mapping-rwe** when a PROM feeds a utility value):\nCOSMIN RoB is the appraisal layer that tells you whether the instrument behind those endpoints earned its place. Its\n*analogue for non-PRO endpoints* — and the right tool when COSMIN does not apply — is endpoint validation:\n**claims-outcome-algorithm-ppv-sensitivity-rwe**, **algorithm-validation**, and **endpoint-adjudication-chart-review-rwe**\nimplement the equivalent \"is this measurement trustworthy?\" discipline for computable outcomes. The longitudinal designs\nbehind the responsiveness box invoke **attrition-and-loss-to-follow-up-rwe** and **missing-data-pattern-table-rwe**; for\nsurrogate or progression endpoints derived from PRO data, see **surrogate-endpoint-validation-rwe** and\n**real-world-progression-rwpfs-rwe**. **Applied note for claims/EHR/registry RWE:** PROs almost never exist in\nadministrative claims; in EHR and disease registries they appear as captured instrument scores. The practical use of\nCOSMIN RoB in a real-world study is therefore *pre-analytic* — before you attach a PROM to a real-world endpoint or a\ncost-utility model, use it to confirm the instrument has validated structural validity, reliability, measurement error,\nand responsiveness in a population resembling your data source; otherwise a methodologically clean RWE analysis rests on\nan unvalidated measurement.", "primary_category": "Guideline", "tags": [ "guideline", "risk-of-bias", "patient-reported-outcomes", "measurement-properties", "psychometrics", "systematic-review", "cosmin" ], "aliases": [ "COSMIN RoB", "COSMIN Risk of Bias checklist", "COSMIN Risk of Bias checklist for PROMs", "COSMIN risk of bias tool" ], "applies_to_study_types": [ "pro_development", "pro_validation", "systematic_review" ], "data_sources": [ "ehr", "registry", "primary" ], "citations": [ { "role": "introduce", "doi": "10.1007/s11136-017-1765-4", "url": "https://doi.org/10.1007/s11136-017-1765-4", "citation_text": "Mokkink LB, de Vet HCW, Prinsen CAC, et al. COSMIN Risk of Bias checklist for systematic reviews of Patient-Reported Outcome Measures. Quality of Life Research. 2018;27(5):1171-1179.", "year": 2018, "authors_short": "Mokkink et al.", "notes": "Canonical statement of the ten-box, worst-score-counts risk-of-bias checklist for individual measurement-property studies of PROMs." }, { "role": "explain", "doi": "10.1007/s11136-018-1798-3", "url": "https://doi.org/10.1007/s11136-018-1798-3", "citation_text": "Prinsen CAC, Mokkink LB, Bouter LM, et al. COSMIN guideline for systematic reviews of patient-reported outcome measures. Quality of Life Research. 2018;27(5):1147-1157.", "year": 2018, "authors_short": "Prinsen et al.", "notes": "Companion guideline for the full PROM systematic-review workflow; situates the RoB checklist alongside the criteria for good measurement properties and modified GRADE." }, { "role": "explain", "doi": "10.1007/s11136-018-1829-0", "url": "https://doi.org/10.1007/s11136-018-1829-0", "citation_text": "Terwee CB, Prinsen CAC, Chiarotto A, et al. COSMIN methodology for evaluating the content validity of patient-reported outcome measures: a Delphi study. Quality of Life Research. 2018;27(5):1159-1170.", "year": 2018, "authors_short": "Terwee et al.", "notes": "Elaborates the content-validity box, the property COSMIN treats as most important." }, { "role": "use", "url": "https://www.cosmin.nl/tools/checklists-assessing-methodological-study-qualities/", "citation_text": "COSMIN initiative — maintained checklists, manuals, and user guides for assessing the methodological quality of studies on PROM measurement properties.", "year": 2024, "authors_short": "COSMIN", "notes": "Maintained source for the current checklist versions, manuals, and worked examples." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "pro-validation", "notes": "Use to rate risk of bias of each study on a PROM measurement property when validating an instrument." }, { "relation_type": "applies_to", "target_slug": "pro-development", "notes": "The PROM-development box appraises how the instrument was created and its content derived." }, { "relation_type": "see_also", "target_slug": "pro-rwe", "notes": "COSMIN RoB justifies the PRO instrument before it is used as a real-world endpoint." }, { "relation_type": "see_also", "target_slug": "hrqol", "notes": "Appraises HRQoL instruments before they support symptom/HRQoL endpoints or label claims." }, { "relation_type": "used_with", "target_slug": "qaly-utility-mapping-rwe", "notes": "Confirms measurement-property validity of a PROM that feeds a utility value used in QALYs." }, { "relation_type": "alternative_to", "target_slug": "algorithm-validation", "notes": "For non-PRO computable endpoints, validity is established by algorithm validation (PPV/sensitivity), not by COSMIN's PROM boxes." }, { "relation_type": "alternative_to", "target_slug": "claims-outcome-algorithm-ppv-sensitivity-rwe", "notes": "The claims/EHR analogue of measurement-property appraisal for outcomes that are not patient-reported." }, { "relation_type": "see_also", "target_slug": "endpoint-adjudication-chart-review-rwe", "notes": "Chart-review adjudication is the trustworthiness check for clinically captured endpoints, parallel to COSMIN's role for PROMs." }, { "relation_type": "used_with", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "The responsiveness box relies on longitudinal designs where attrition and missing PRO data must be handled." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "hta", "journal" ] }, { "slug": "dag-framework", "name": "Directed Acyclic Graph (DAG) Framework for Causal Inference", "short_definition": "A graphical framework for encoding causal assumptions as a directed acyclic graph, then using the back-door criterion to derive a minimal sufficient adjustment set and to distinguish confounders from mediators and colliders before any analysis. It governs design transparency, not study quality.", "long_description": "**What it is** — The **directed acyclic graph (DAG) framework** is a non-parametric language for stating, sharing,\nand auditing the causal assumptions behind a study. A DAG is a graph whose nodes are variables (exposure, outcome,\nmeasured and *unmeasured* common causes, mediators, selection indicators) and whose directed edges encode assumed\ncause→effect relationships; \"acyclic\" means no variable can cause itself through a feedback loop, and arrows respect\ntime order. The framework was articulated for epidemiology by Greenland, Pearl, and Robins (1999), building on\nPearl's structural-causal-model theory. It is not a reporting checklist maintained by EQUATOR, Cochrane, ISPOR, or\na regulator — there is no single steward; the methodological community maintains it through textbooks (Hernán &\nRobins, *Causal Inference: What If*) and the **DAGitty** software/web tool (Textor et al.), which mechanizes the\ngraph-theoretic algorithms. Its purpose is narrow and powerful: make the identification step of a causal analysis\nexplicit and falsifiable-by-debate, so that the chosen covariate-adjustment set is a *derived consequence* of stated\nassumptions rather than an opaque modeling choice.\n\n**When to use** — Draw a DAG whenever the estimand is causal (comparative effectiveness/safety, treatment effects,\npolicy effects) and the design is non-interventional or a target-trial emulation in claims, EHR, registry, or linked\ndata — i.e., whenever confounding control is the central threat to validity. The DAG belongs at the *protocol/design*\nstage, before data are analyzed, and the resulting adjustment set should be pre-specified. It is the right tool when\n(a) you must justify *which* covariates to adjust for and, equally, which to leave alone; (b) reviewers (FDA/EMA, an\nHTA committee, a journal) will ask why your model is sufficient to remove confounding; (c) you need to expose\nunmeasured confounding honestly by drawing U-nodes; or (d) you must defend against subtle structural biases\n(collider stratification, M-bias, overadjustment, selection bias from informative censoring or differential\nenrollment). Use it *alongside*, not instead of, the reporting and protocol guidelines that actually govern the\nmanuscript (STROBE/RECORD-PE, HARPER, STaRT-RWE) — those tools increasingly *require* a DAG or equivalent causal\ndiagram as a design artifact.\n\n**What it requires** — Proper use of the framework — as opposed to drawing a decorative bubble chart — demands:\n(1) **Encode every common cause** of exposure and outcome, including variables you cannot measure; unmeasured\nconfounders are drawn explicitly as U-nodes — omitting them because they are unmeasured defeats the entire purpose.\n(2) **Respect time-ordering**: no arrow points backward in time; baseline covariates precede exposure, post-baseline\nvariables are flagged because adjusting for them can open bias paths. (3) **Subject-matter justification for every\nedge present and every edge absent** — a missing arrow is the strong assumption (no direct effect), and it must be\ndefensible. (4) **Apply the back-door criterion** (mechanized by DAGitty) to derive a *minimal sufficient adjustment\nset* that blocks all confounding paths without opening new ones. (5) **Classify each node** as confounder, mediator,\nor collider relative to the exposure–outcome contrast, because the correct action differs: adjust confounders, do\n*not* adjust mediators when the estimand is the total effect, never condition on colliders or their descendants.\n(6) **Pre-specify and version the DAG** as a study artifact (DAGitty exports a machine-readable model and the\ntestable implications), so the adjustment set is auditable and the assumptions are stated before results are seen.\n\n**When NOT to use — limitations and common misapplications** — A DAG encodes *assumptions*; it does not certify they\nare correct, and a completed graph never makes an observational study causal. Concrete failure modes:\n- **DAG-as-theater**: drawing the graph *after* choosing an adjustment set to retro-justify it. The DAG must precede\n and constrain the analysis, not decorate it.\n- **Treating the DAG as data-testable**: the core structural assumptions (which arrows exist) are not falsifiable\n from observed associations. DAGitty's testable implications check conditional independencies *implied by* the\n graph; they cannot confirm the graph is right.\n- **Conditioning on a collider** (or its descendant, or a selection variable) — this *induces* association where\n none existed (M-bias, selection bias), often worse than the confounding it was meant to fix.\n- **Adjusting for a mediator when the estimand is the total effect** — overadjustment that biases the estimate\n toward the null and can introduce collider bias at the mediator.\n- **Omitting unmeasured confounders** from the graph because they are inconvenient; the U-nodes are precisely what\n communicate residual confounding to a reviewer.\n- **Feedback loops / time-varying treatment-confounder feedback**: a standard DAG cannot represent cycles. For\n time-varying exposures with confounders affected by prior treatment, use time-expanded DAGs and g-methods, or\n single-world intervention graphs (SWIGs); a single cross-sectional DAG will mislead.\n- **Substituting the DAG for subject-matter expertise or for the reporting checklist** the venue actually requires\n (using a DAG where RECORD-PE/HARPER content is mandated, or vice versa).\n\n**How it maps to this catalog** — The DAG framework sits upstream of nearly every causal-inference concept here and\ntells the analyst *what to do*; the concepts tell them *how*:\n- **target-trial-emulation** — the emulated trial's eligibility, assignment, and time-zero structure is the design\n a DAG makes explicit; draw the DAG to justify the emulation's adjustment set.\n- **active-comparator-new-user** — the design choice that removes confounding by indication and immortal time *by\n structure*; the DAG shows why the active comparator blocks the back-door path that a non-user comparator leaves open.\n- **high-dimensional-propensity-score-hdps-rwe** and **propensity-score-methods-psm-iptw** — implement adjustment\n for the confounder set the DAG identifies; hdPS operationalizes proxies for the U-nodes you drew but cannot\n measure directly.\n- **estimands-ate-att-intercurrent-events-rwe** and **estimand-analysis-traceability-rwe** — the estimand\n (total vs. direct effect) determines whether a node is a mediator to leave alone or a confounder to adjust; the\n DAG enforces that distinction.\n- **diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe** — measurement-error and misclassification nodes belong\n on the DAG; phenotype validation quantifies the arrows into and out of the measured-outcome node.\n- **attrition-and-loss-to-follow-up-rwe**, **database-feasibility-attrition-funnel-rwe**, and\n **selection-bias-sensitivity-analysis-rwe** — censoring and enrollment are selection nodes; conditioning on them\n (or on their causes) is the collider/selection-bias failure mode the DAG is designed to catch.\n- **immortal-time-bias-handling** — a time-ordering violation the DAG exposes when follow-up starts before the\n exposure decision.\n- **empirical-calibration-negative-controls-rwe**, **negative-control-outcomes-rwe**,\n **negative-control-exposures-rwe**, and **e-value-sensitivity-analysis** — the quantitative bias analyses that\n probe the residual-confounding (U-node) and unblocked-path assumptions the DAG could not eliminate.\n- **claims-analysis** — the applied substrate: in claims/EHR/registry RWE, draw the DAG before extracting any\n covariate, derive the minimal sufficient adjustment set, then build only those baseline covariates (measured up\n to time zero) into the propensity model. Applied note: in administrative data the most consequential nodes are\n usually *unmeasured* (disease severity, frailty, lab values, over-the-counter use) — draw them as U-nodes,\n document which measured proxies stand in for them, and reserve a negative-control outcome and an E-value to\n bound the confounding the graph admits it cannot close.", "primary_category": "Guideline", "tags": [ "guideline", "framework", "causal-inference", "directed-acyclic-graph", "confounding", "collider-bias", "study-design" ], "aliases": [ "DAGs", "directed acyclic graphs", "causal diagrams", "causal DAG", "DAGitty framework" ], "applies_to_study_types": [ "cer_observational", "cohort_prospective", "cohort_retrospective", "case_control", "new_user", "active_comparator_new_user", "target_trial_emulation" ], "data_sources": [ "claims", "ehr", "registry", "linked", "primary" ], "citations": [ { "role": "introduce", "doi": "10.1097/00001648-199901000-00008", "url": "https://doi.org/10.1097/00001648-199901000-00008", "citation_text": "Greenland S, Pearl J, Robins JM. Causal diagrams for epidemiologic research. Epidemiology. 1999;10(1):37-48.", "year": 1999, "authors_short": "Greenland, Pearl & Robins", "notes": "Canonical statement that brought DAGs and the back-door criterion into epidemiology; defines confounders, mediators, and colliders graphically and shows when adjustment removes versus induces bias." }, { "role": "explain", "doi": "10.1016/j.jclinepi.2021.08.001", "url": "https://doi.org/10.1016/j.jclinepi.2021.08.001", "citation_text": "Digitale JC, Martin JN, Glymour MM. Tutorial on directed acyclic graphs. Journal of Clinical Epidemiology. 2022;142:264-267.", "year": 2022, "authors_short": "Digitale et al.", "notes": "Concise applied tutorial on building a DAG, reading paths, and choosing an adjustment set; good entry point for analysts new to the framework." }, { "role": "explain", "doi": "10.1093/ije/dyw341", "url": "https://doi.org/10.1093/ije/dyw341", "citation_text": "Textor J, van der Zander B, Gilthorpe MS, Liśkiewicz M, Ellison GTH. Robust causal inference using directed acyclic graphs: the R package 'dagitty'. International Journal of Epidemiology. 2016;45(6):1887-1894.", "year": 2016, "authors_short": "Textor et al.", "notes": "The practical software companion (DAGitty web tool and R package) that mechanizes adjustment-set derivation and the testable implications of a stated DAG." }, { "role": "use", "doi": "10.1093/ije/dyaa213", "url": "https://doi.org/10.1093/ije/dyaa213", "citation_text": "Tennant PWG, Murray EJ, Arnold KF, et al. Use of directed acyclic graphs (DAGs) to identify confounders in applied health research: review and recommendations. International Journal of Epidemiology. 2021;50(2):620-632.", "year": 2021, "authors_short": "Tennant et al.", "notes": "Review of how DAGs are (mis)used in practice with concrete reporting recommendations; the closest thing to a usage standard for drawing and reporting DAGs in applied studies." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "cer-observational", "notes": "A DAG is the design-stage tool for justifying confounding control in comparative observational studies." }, { "relation_type": "applies_to", "target_slug": "cohort-retrospective", "notes": "Encode the assumed confounding structure before selecting baseline covariates for a retrospective cohort." }, { "relation_type": "applies_to", "target_slug": "case-control", "notes": "DAGs expose selection (collider) bias that is especially easy to induce through case-control sampling and control selection." }, { "relation_type": "applies_to", "target_slug": "target-trial-emulation", "notes": "The emulated trial's structure is what the DAG makes explicit; draw it to justify the emulation's adjustment set and time-zero alignment." }, { "relation_type": "used_with", "target_slug": "active-comparator-new-user", "notes": "The DAG shows why an active comparator and new-user restriction block the back-door path that a non-user comparator leaves open." }, { "relation_type": "used_with", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "hdPS operationalizes adjustment for the confounder set the DAG identifies, using proxies for the U-nodes that cannot be measured directly." }, { "relation_type": "used_with", "target_slug": "propensity-score-methods-psm-iptw", "notes": "Propensity-score adjustment implements the minimal sufficient adjustment set derived from the DAG; it cannot fix an adjustment set that conditions on colliders or mediators." }, { "relation_type": "used_with", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "The estimand (total vs direct effect) decides whether a node is a mediator to leave alone or a confounder to adjust; the DAG enforces the distinction." }, { "relation_type": "see_also", "target_slug": "immortal-time-bias-handling", "notes": "A time-ordering violation the DAG exposes when follow-up begins before the exposure decision." }, { "relation_type": "see_also", "target_slug": "selection-bias-sensitivity-analysis-rwe", "notes": "Censoring and enrollment are selection nodes; conditioning on them or their causes is the collider/selection bias the DAG is built to detect." }, { "relation_type": "see_also", "target_slug": "e-value-sensitivity-analysis", "notes": "Quantifies the residual unmeasured confounding (U-nodes) the DAG admits it cannot eliminate." }, { "relation_type": "see_also", "target_slug": "empirical-calibration-negative-controls-rwe", "notes": "Negative controls empirically probe the unblocked-path and residual-confounding assumptions encoded in the DAG." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "hta", "journal" ] }, { "slug": "e-value", "name": "E-Value (Sensitivity Analysis for Unmeasured Confounding)", "short_definition": "A recommended quantitative-bias-analysis reporting standard for observational studies. The E-value is the minimum strength of association, on the risk-ratio scale, that an unmeasured confounder would need with both the exposure and the outcome to fully explain away an observed effect estimate (and, separately, to move its confidence limit to the null).", "long_description": "**What it is** — The **E-value** is a sensitivity-analysis metric introduced by VanderWeele and\nDing (Annals of Internal Medicine, 2017) for quantifying how robust an observed exposure-outcome\nassociation is to **unmeasured (residual) confounding**. It is the minimum strength of\nassociation — expressed as a risk ratio — that an unmeasured confounder would need to have with\n*both* the exposure and the outcome, over and above the measured covariates, to reduce the\nobserved point estimate to the null; a companion E-value is computed for the confidence-limit\nclosest to the null. Unlike a checklist with numbered items, the E-value is a single, transparent\nnumber that any reader can interpret without distributional assumptions about the unmeasured\nconfounder. It is not a regulatory checklist or a risk-of-bias instrument in the EQUATOR/Cochrane\nsense; rather it is a **methods standard** that reporting guidelines and agencies increasingly\nexpect within the sensitivity-analysis section of an observational study. ENCePP methodological\nstandards, HARPER/STaRT-RWE protocol templates, and high-impact journals all now treat a\nquantitative bias analysis for unmeasured confounding (of which the E-value is the most common\nform) as part of a complete, decision-grade report. There is no committee that \"maintains\" the\nE-value; it is maintained as a literature — the original statement paper, the technical-considerations\nfollow-up, clinician-facing explanations, and the freely available `EValue` R package and web\ncalculator (Mathur et al., 2018).\n\n**When to use** — Report an E-value whenever an observational (non-interventional) effect estimate\nis offered for a causal or comparative-effectiveness interpretation and unmeasured confounding is a\nlive threat: cohort, case-control, new-user, active-comparator new-user, and target-trial-emulation\nstudies in claims, EHR, registry, or linked data. It belongs in **FDA/EMA RWE submissions**, **HTA/payer\ndossiers**, and **peer-reviewed manuscripts** as the headline robustness statement accompanying the\nprimary adjusted estimate (and key subgroups). Decision rule for *which* sensitivity tool: use the\nE-value as the **default, communicable summary** when you have a single binary/rate/HR-type contrast\nand want a threshold readers can benchmark against known confounder strengths. Escalate beyond a bare\nE-value to fuller quantitative bias analysis — bias formulas with specified confounder prevalence,\nbounding factors, probabilistic/Monte Carlo bias analysis, or negative-control / empirical calibration —\nwhen you can credibly *parameterize* the suspected confounder, when multiple biases act jointly, or\nwhen a regulator/HTA reviewer asks for a quantified rather than threshold-style assessment. The E-value\nis the floor of good practice, not the ceiling.\n\n**What it requires** — As a reporting standard the E-value enforces, in the study's sensitivity\nanalysis, that authors: (1) compute the E-value for the **point estimate** *and* for the\n**confidence limit nearest the null**, and report both — a large point-estimate E-value with a\nconfidence limit E-value near 1 signals fragility; (2) compute it on the correct scale (risk ratio;\napproximate conversions are needed for odds ratios with common outcomes, hazard ratios, and\nstandardized mean differences, and the approximation should be acknowledged); (3) **interpret** the\nnumber against plausible real-world confounder strengths — i.e., name a measured confounder of\ncomparable strength and argue whether an *unmeasured* one of that magnitude could plausibly exist after\nthe design and adjustment already applied; and (4) tie the E-value back to the design choices that\nreduced confounding in the first place — active-comparator/new-user restriction, time-zero alignment,\nfit-for-purpose data, and the propensity-score/covariate strategy. In RWE specifically, the E-value\npresupposes that the *measured* confounding has already been handled credibly and that the estimand,\nintercurrent-event handling, and outcome/exposure phenotypes are sound; it speaks only to the\n*residual* gap.\n\n**When NOT to use — limitations and common misapplications** — (a) **It is not a quality score and not\na risk-of-bias tool.** A large E-value does not certify a study as unbiased; it only describes\nrobustness to one bias (unmeasured confounding) conditional on everything else being correct. (b) **It\ndoes not address other biases** — selection bias, immortal-time bias, misclassification of exposure or\noutcome, missing data, or measurement error are outside its scope; reporting an E-value while ignoring\nthese is checklist-as-theater. (c) **It is not a license to claim causation** — computing an E-value on\na hopelessly confounded claims comparison does not make the contrast causal. (d) **Misreading the\nthreshold:** a common error is reporting only the point-estimate E-value and omitting the\nconfidence-limit E-value, overstating robustness; another is treating the E-value as the *probability*\nof an explanatory confounder rather than the *strength* one would need. (e) **Scale errors:** applying\nthe risk-ratio formula to a hazard ratio or odds ratio from a common outcome without the appropriate\ntransformation inflates or deflates the value. (f) **Critique to disclose:** the E-value has been\ncriticized (notably by Ioannidis and colleagues) for being easy to report mechanically and for inviting\noverinterpretation when divorced from substantive argument about what the unmeasured confounder\nactually is; the defensible practice, set out in VanderWeele, Ding, and Mathur's technical-considerations\npaper, is to pair the number with a named, plausible confounder and the design that already constrains it.\nDo not substitute an E-value for the harder design work (negative controls, ACNU, hdPS); use it to\nsummarize what residual risk remains *after* that work.\n\n**How it maps to this catalog** — The computational mechanics, formulas, scale conversions, and the\n`EValue` R package live in the implementing concept **e-value-sensitivity-analysis**; this guideline\nentry is the *when/why/report-it* layer that points to it. Broader bias quantification —\nbounding factors, probabilistic bias analysis, multiple-bias modeling — is in\n**quantitative-bias-analysis-toolkit-rwe**, the natural escalation when a bare E-value is insufficient.\nThe E-value only earns its keep on top of credible measured-confounding control:\n**active-comparator-new-user** (removing confounding by indication and immortal time at the design\nstage) and **high-dimensional-propensity-score-hdps-rwe** (proxy-based adjustment) are the prerequisites\nit summarizes residual risk against, and **empirical-calibration-negative-controls-rwe** /\n**negative-control-outcomes-rwe** are complementary empirical checks that triangulate with it. The\nestimand the E-value qualifies should be defined via **estimands-ate-att-intercurrent-events-rwe**;\nthe exposure/outcome it depends on must be validly defined via\n**diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe**; the data must be appropriate per\n**fit-for-purpose-data-assessment-rwe**; cohort losses that could masquerade as confounding belong to\n**attrition-and-loss-to-follow-up-rwe**; and the question itself should be framed with\n**picots-framework-rwe**. For a claims/EHR/registry **applied note**: after building an ACNU cohort in\nclaims (see **claims-analysis**), balancing on an hdPS, and estimating, say, an adjusted HR of 0.70 for\na comparative-effectiveness outcome, report the E-value for the HR and for its upper confidence limit,\nthen state explicitly which *unmeasured* claims-invisible factor (frailty, disease severity, smoking,\nover-the-counter use) of that strength would be needed and whether the active-comparator design plus\nhdPS proxies plausibly already capture it.", "primary_category": "Guideline", "tags": [ "guideline", "sensitivity-analysis", "quantitative-bias-analysis", "unmeasured-confounding", "rwe", "reporting" ], "aliases": [ "E-value", "E value", "VanderWeele-Ding E-value", "E-value sensitivity analysis" ], "applies_to_study_types": [ "cer_observational", "cohort_prospective", "cohort_retrospective", "case_control", "new_user", "active_comparator_new_user", "target_trial_emulation" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "doi": "10.7326/M16-2607", "url": "https://doi.org/10.7326/M16-2607", "citation_text": "VanderWeele TJ, Ding P. Sensitivity analysis in observational research: introducing the E-value. Annals of Internal Medicine. 2017;167(4):268-274.", "year": 2017, "authors_short": "VanderWeele & Ding", "notes": "Canonical statement paper defining the E-value for the point estimate and the confidence limit nearest the null, with the bounding-factor derivation." }, { "role": "explain", "doi": "10.1515/jci-2018-0007", "url": "https://doi.org/10.1515/jci-2018-0007", "citation_text": "VanderWeele TJ, Ding P, Mathur M. Technical considerations in the use of the E-value. Journal of Causal Inference. 2019;7(2):20180007.", "year": 2019, "authors_short": "VanderWeele et al.", "notes": "Authors' rigorous follow-up clarifying scale conversions (HR, OR, risk difference), interpretation, and responses to early critique; the reference for defensible application." }, { "role": "explain", "doi": "10.1001/jama.2018.21554", "url": "https://doi.org/10.1001/jama.2018.21554", "citation_text": "Haneuse S, VanderWeele TJ, Arterburn D. Using the E-value to assess the potential effect of unmeasured confounding in observational studies. JAMA. 2019;321(6):602-603.", "year": 2019, "authors_short": "Haneuse et al.", "notes": "Concise clinician- and reviewer-facing explanation; useful for HTA/journal audiences who need the intuition without the derivation." }, { "role": "use", "doi": "10.1097/EDE.0000000000000864", "url": "https://doi.org/10.1097/EDE.0000000000000864", "citation_text": "Mathur MB, Ding P, Riddell CA, VanderWeele TJ. Web site and R package for computing E-values. Epidemiology. 2018;29(5):e45-e47.", "year": 2018, "authors_short": "Mathur et al.", "notes": "Documents the freely available EValue R package and online calculator implementing the metric; the practical how-to-compute pointer (mechanics live in the e-value-sensitivity-analysis concept)." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "cer-observational", "notes": "Report an E-value alongside the primary adjusted estimate in comparative observational studies." }, { "relation_type": "applies_to", "target_slug": "cohort-retrospective", "notes": "Standard robustness summary for adjusted cohort estimates in claims/EHR data." }, { "relation_type": "applies_to", "target_slug": "active-comparator-new-user", "notes": "Summarizes residual unmeasured confounding after the design has removed confounding by indication and immortal time." }, { "relation_type": "applies_to", "target_slug": "target-trial-emulation", "notes": "Report E-values for the emulated per-protocol / ITT contrasts as part of the sensitivity suite." }, { "relation_type": "requires", "target_slug": "e-value-sensitivity-analysis", "notes": "Implementing concept — formulas, scale conversions (HR/OR/RD), and the EValue R package live here." }, { "relation_type": "see_also", "target_slug": "quantitative-bias-analysis-toolkit-rwe", "notes": "Escalation path when a bare E-value is insufficient — bounding factors, probabilistic and multiple-bias analysis." }, { "relation_type": "used_with", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "The E-value quantifies residual confounding after measured confounding is controlled via hdPS." }, { "relation_type": "used_with", "target_slug": "active-comparator-new-user", "notes": "Design-stage confounding control that the E-value reports residual robustness against." }, { "relation_type": "used_with", "target_slug": "empirical-calibration-negative-controls-rwe", "notes": "Complementary empirical check; triangulate negative-control calibration with the E-value." }, { "relation_type": "see_also", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Define the estimand the E-value qualifies before computing it." }, { "relation_type": "see_also", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "A valid exposure/outcome phenotype is a precondition; the E-value addresses confounding, not misclassification." }, { "relation_type": "see_also", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "The E-value presupposes the data are fit for the causal question." }, { "relation_type": "see_also", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Differential attrition is selection bias, outside the E-value's scope — assess separately." }, { "relation_type": "see_also", "target_slug": "picots-framework-rwe", "notes": "Frame the question and estimand before reporting robustness." }, { "relation_type": "used_with", "target_slug": "claims-analysis", "notes": "Apply when the underlying contrast is built from administrative claims data." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "hta", "journal" ] }, { "slug": "encepp-checklist", "name": "ENCePP Checklist for Study Protocols", "short_definition": "A structured methodological checklist that prompts authors to document key design, data, bias-control, and analysis decisions in a pharmacoepidemiological or non-interventional study protocol; required as an annex for EMA-imposed post-authorisation safety studies and recommended for any ENCePP-badged study.", "long_description": "**What it is.** The **ENCePP Checklist for Study Protocols** (Revision 4, 2018) is a methodological protocol-completeness\nchecklist maintained by the **European Network of Centres for Pharmacoepidemiology and Pharmacovigilance (ENCePP)**, a\nnetwork coordinated by the European Medicines Agency (EMA). It is not a manuscript reporting checklist and not a\nrisk-of-bias score: it is a *pre-specification prompt list* that walks a protocol author through the epidemiological\ndecisions that determine whether a non-interventional study can yield credible causal or descriptive evidence — research\nquestion, study design, data source, exposure and outcome definitions, bias and confounding control, analysis, data\nmanagement, and quality assurance. Its purpose, in ENCePP's own words, is to encourage researchers to reflect on important\nepidemiological principles, to promote transparency about the methods actually used, and to keep protocols aligned with\ncontemporary methodological standards. The Checklist is the operational companion to the **ENCePP Guide on Methodological\nStandards in Pharmacoepidemiology** (the narrative standards reference): the Guide explains *how* to do each step well; the\nChecklist forces you to confirm, item by item, that the protocol addresses it.\n\n**When to use.** Apply the ENCePP Checklist at the **protocol stage**, before data access and before any analysis. It is\nmandatory as a signed annex for **EMA-imposed Post-Authorisation Safety Studies (PASS)** under the GVP Module VIII regime,\nand it is required for any study seeking the **ENCePP Study seal** registered in the EU PAS Register (HMA-EMA Catalogues).\nIt is strongly recommended for voluntary PASS, drug-utilisation studies, and comparative safety/effectiveness cohort or\ncase-control studies built on routinely collected European data. Decision rules versus siblings in this catalog: (1) Use\nthe **ENCePP Checklist** when you need a *completeness gate on the protocol* and a regulatory deliverable; use the\n**`encepp-guide`** (ENCePP Methodological Standards Guide) when you need the *substantive methodological reasoning* behind\na choice — they are paired, not interchangeable. (2) Use **`harper`** (HARmonized Protocol Template) or **`start-rwe`**\nwhen you need a *structured protocol template with prescribed tables and a causal-roadmap layout* — the ENCePP Checklist\nverifies coverage but does not supply the protocol skeleton, so the common pattern is HARPER/START-RWE to draft and the\nENCePP Checklist to certify. (3) Use **`record-pe`** or **`strobe`** for *final manuscript/report* reporting of a completed\npharmacoepidemiology study — those govern what you publish, the ENCePP Checklist governs what you plan. For a US-FDA\nsubmission, pair it with **`fda-rwe-framework`** / **`fda-rwe-noninterventional`** rather than treating it as sufficient on\nits own.\n\n**What it requires.** The Checklist enforces explicit protocol-level documentation across the domains that decide whether\nreal-world data can answer the question: (a) **research question and study design** stated as a clear objective with a\nnamed design (cohort, case-control, self-controlled, drug-utilisation) and, where causal, an explicit comparison and\nestimand; (b) **data source fitness-for-use** — provenance, coverage, relevant data quality dimensions, lag, and whether\nthe source can actually capture the exposure, outcome, and confounders required; (c) **population, exposure, and outcome\noperational definitions** — eligibility, washout/lookback, code lists, and validated phenotype/algorithm definitions with\nperformance metrics where available; (d) **time-zero/index-date alignment** and follow-up rules that avoid immortal time\nand prevalent-user bias; (e) **bias and confounding control** — confounders, the analytic strategy to address them\n(matching, propensity/disease-risk scores, restriction, design-based control), and residual/unmeasured confounding; (f)\n**statistical analysis** — primary and sensitivity/quantitative-bias analyses, handling of missing data, competing risks,\nand effect-measure modification; and (g) **data management, quality assurance, and reporting** — versioned code lists,\nvalidation, and a plan to report attrition transparently. Each \"no\" or \"not applicable\" must be justified, which is the\nmechanism that turns the list from a box-ticking exercise into a design critique.\n\n**When NOT to use — limitations and common misapplications.** The single most common error is treating the ENCePP\nChecklist as a **quality score or a risk-of-bias instrument** — it is neither. A fully completed checklist does **not**\ncertify that a study is low-bias, and it does **not** make an observational comparison causal; you can answer every item\naffirmatively and still have a fatally confounded estimate. Use **`robins-i`**/**`robins-e`** (or the relevant JBI/NOS\ntool) when you actually need a graded risk-of-bias appraisal, and a real causal design (target-trial emulation,\nactive-comparator new-user) when you need to defend a causal contrast. \"**Checklist-as-theater**\" — completing it for the\nbadge while the protocol remains vague on time-zero, comparator, or estimand — defeats its purpose; the justification\nfields exist precisely to expose that. Do not substitute the ENCePP Checklist for a **reporting** checklist\n(`record-pe`/`strobe`) at manuscript stage, and do not substitute it for the **protocol template** itself — it certifies a\nprotocol, it does not write one. Finally, it is a European pharmacoepidemiology instrument: for purely health-economic,\nsystematic-review, or US-only regulatory deliverables, the appropriate ISPOR/PRISMA/FDA guidance leads and the ENCePP\nChecklist plays at most a supporting role.\n\n**How it maps to this catalog.** Each ENCePP domain is implemented by concrete concepts here. *Research question / design /\nestimand* → **`picots-framework-rwe`**, **`target-trial-emulation`**, **`estimands-ate-att-intercurrent-events-rwe`**, and\n**`estimand-analysis-traceability-rwe`**. *Data source fitness* → **`fit-for-purpose-data-assessment-rwe`**,\n**`claims-analysis`**, and **`database-feasibility-attrition-funnel-rwe`**. *Phenotype / exposure / outcome definitions* →\n**`diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe`**, **`claims-outcome-algorithm-ppv-sensitivity-rwe`**, and\n**`washout-clean-lookback-period-rwe`**. *Time-zero and follow-up* → **`time-zero-index-date-alignment-rwe`** and\n**`continuous-enrollment-observable-time-rwe`**. *Confounding control* → **`active-comparator-new-user`**,\n**`high-dimensional-propensity-score-hdps-rwe`**, and **`propensity-score-methods-psm-iptw`**. *Analysis, attrition, and\nbias quantification* → **`competing-risks-cause-specific-fine-gray-rwe`**, **`attrition-and-loss-to-follow-up-rwe`**,\n**`e-value-sensitivity-analysis`**, and **`quantitative-bias-analysis-toolkit-rwe`**. **Applied note (claims/EHR/registry\nRWE):** in claims, the Checklist's data-fitness items should force a statement on enrollment continuity and\nMedicare-Advantage versus fee-for-service capture (see `medicare-ffs-ma-commercial-claims-differences-rwe`) before any\n\"no prior exposure\" washout is trusted as observed rather than missing; in EHR/registry work the same items should force a\nloss-to-follow-up and linkage-selection plan. The disciplined way to use this entry: draft the protocol with HARPER or\nSTART-RWE, build each component from the linked concept, then run the ENCePP Checklist as the final pre-submission gate.", "primary_category": "Guideline", "tags": [ "guideline", "methodological", "protocol", "pharmacoepidemiology", "pass", "encepp", "rwe" ], "aliases": [ "ENCePP Checklist", "ENCePP Checklist for Study Protocols", "ENCePP Methodological Checklist" ], "applies_to_study_types": [ "pass_imposed", "pass_voluntary", "cohort_prospective", "cohort_retrospective", "case_control", "drug_utilization", "active_comparator_new_user", "new_user" ], "data_sources": [ "claims", "ehr", "registry", "linked", "multi-database" ], "citations": [ { "role": "introduce", "doi": null, "url": "https://encepp.europa.eu/encepp-toolkit/encepp-checklist-study-protocols_en", "citation_text": "ENCePP Checklist for Study Protocols (Revision 4). European Network of Centres for Pharmacoepidemiology and Pharmacovigilance (ENCePP), European Medicines Agency, 2018.", "year": 2018, "authors_short": "ENCePP / EMA", "notes": "Canonical source. The ENCePP Checklist has no journal statement paper; the maintained ENCePP/EMA toolkit page is the authoritative reference. Revision 4 is the current version (October 2018)." }, { "role": "explain", "doi": null, "url": "https://encepp.europa.eu/encepp-toolkit/methodological-guide_en", "citation_text": "ENCePP Guide on Methodological Standards in Pharmacoepidemiology. European Network of Centres for Pharmacoepidemiology and Pharmacovigilance (ENCePP), European Medicines Agency.", "year": 2023, "authors_short": "ENCePP / EMA", "notes": "Companion narrative standards reference; the Checklist items map onto the methodological reasoning detailed in the Guide. Periodically revised." }, { "role": "use", "doi": null, "url": "https://www.ema.europa.eu/en/human-regulatory-overview/post-authorisation/pharmacovigilance-post-authorisation/good-pharmacovigilance-practices-gvp", "citation_text": "Guideline on good pharmacovigilance practices (GVP) Module VIII - Post-authorisation safety studies (Rev 3). European Medicines Agency.", "year": 2017, "authors_short": "EMA", "notes": "Regulatory basis requiring the ENCePP Checklist as an annex for imposed PASS protocols submitted under GVP Module VIII." } ], "relations": [ { "relation_type": "part_of", "target_slug": "pass-imposed", "notes": "The completed, signed ENCePP Checklist is a mandatory annex to an EMA-imposed PASS protocol." }, { "relation_type": "applies_to", "target_slug": "pass-voluntary", "notes": "Recommended completeness gate for voluntary post-authorisation safety studies and ENCePP-badged studies." }, { "relation_type": "applies_to", "target_slug": "cohort-retrospective", "notes": "Use at the protocol stage for retrospective cohort pharmacoepidemiology studies on routinely collected data." }, { "relation_type": "applies_to", "target_slug": "cohort-prospective", "notes": "Use at the protocol stage for prospective non-interventional cohort studies." }, { "relation_type": "applies_to", "target_slug": "case-control", "notes": "Use at the protocol stage for case-control and nested case-control pharmacoepidemiology designs." }, { "relation_type": "applies_to", "target_slug": "drug-utilization", "notes": "Use at the protocol stage for drug-utilisation studies." }, { "relation_type": "see_also", "target_slug": "harper", "notes": "HARPER supplies the structured protocol template; the ENCePP Checklist certifies coverage of it. Draft with HARPER, gate with ENCePP." }, { "relation_type": "see_also", "target_slug": "start-rwe", "notes": "START-RWE provides prescribed protocol tables and a causal-roadmap layout that the ENCePP Checklist then verifies." }, { "relation_type": "see_also", "target_slug": "record-pe", "notes": "RECORD-PE governs final reporting of a completed pharmacoepidemiology study; the ENCePP Checklist governs the protocol. Use both at their respective stages." }, { "relation_type": "complements", "target_slug": "target-trial-emulation", "notes": "Implements the design/estimand items - pre-specify the emulated trial before applying the checklist as a gate." }, { "relation_type": "complements", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "Implements the data-source fitness-for-use items the checklist requires." }, { "relation_type": "complements", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Implements the validated exposure/outcome phenotype-definition items." }, { "relation_type": "complements", "target_slug": "active-comparator-new-user", "notes": "Implements the confounding-control and time-zero alignment items for comparative designs." }, { "relation_type": "complements", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "Implements the analytic confounding-control strategy the checklist asks authors to specify." }, { "relation_type": "complements", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Implements the estimand and intercurrent-event specification the checklist expects for causal questions." }, { "relation_type": "complements", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Implements the transparent attrition/loss-to-follow-up reporting the checklist requires." }, { "relation_type": "complements", "target_slug": "e-value-sensitivity-analysis", "notes": "Implements the residual/unmeasured-confounding sensitivity analysis the checklist prompts for." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "ema", "fda", "hta" ] }, { "slug": "encepp-guide", "name": "ENCePP Guide on Methodological Standards in Pharmacoepidemiology", "short_definition": "A maintained, chapter-organized reference of methodological standards for the design, conduct, analysis, and reporting of pharmacoepidemiology and post-authorisation safety/efficacy studies, curated by the European Network of Centres for Pharmacoepidemiology and Pharmacovigilance (ENCePP) with the European Medicines Agency (EMA). Currently Revision 11 (July 2023).", "long_description": "**What it is** — The **ENCePP Guide on Methodological Standards in Pharmacoepidemiology** (commonly the \"ENCePP Guide\") is a living, web-published reference maintained by the **European Network of Centres for Pharmacoepidemiology and Pharmacovigilance (ENCePP)**, a scientific network coordinated by the **European Medicines Agency (EMA)**. First issued in 2010 and now at **Revision 11 (July 2023)**, it is organized into chapters that point to authoritative methodological references across the full study lifecycle — research-question formulation, study design, data sources and their fitness, exposure/outcome/covariate definition, statistical analysis, bias and confounding control, and reporting. Critically, it is a **curated standards-and-references guide, not an item-by-item checklist or scoring instrument**: it tells you where the methodological consensus sits and which tools to use, but the operational fill-in artifacts are its companions — the **ENCePP Checklist for Study Protocols** (a protocol completeness aid) and the **ENCePP Code of Conduct** (governance, transparency, and independence). It sits alongside, and explicitly cross-references, reporting guidelines (STROBE, RECORD/RECORD-PE) and protocol templates (STaRT-RWE, HARPER) rather than replacing them.\n\n**When to use** — Reach for the ENCePP Guide whenever you are **designing, conducting, or appraising a non-interventional or hybrid pharmacoepidemiology study** intended for an EMA regulatory pathway, an HTA/payer dossier, or a peer-reviewed journal — especially studies on routinely collected data (claims, EHR, registries, linked/multi-database). It is the **default European reference for post-authorisation safety studies (PASS)**, both imposed (a condition of marketing authorisation, governed by GVP Module VIII) and voluntary, and for drug-utilization, comparative-safety, and comparative-effectiveness work. Decision rule for which document to open: use the **Guide** to choose and justify methods; use the **ENCePP Checklist for Study Protocols** to confirm your *protocol* documents the required design choices; use **STaRT-RWE or HARPER** when you need a structured *template* to specify the study; and use **STROBE / RECORD-PE** when you reach the *reporting* stage. For systematic reviews/meta-analyses, the Guide is not the right home — use PRISMA-family guidance instead.\n\n**What it requires** — Although not a checklist, the Guide enforces a substantive set of methodological expectations that map directly onto real-world-data practice: (1) an explicit research question and **PICOTS/estimand** framing — target population, treatment strategies, intercurrent-event handling, and the causal contrast; (2) documented **data fitness-for-purpose** — provenance, capture, lag, completeness, and relevance of each source, with payer-specific caveats (e.g., Medicare fee-for-service vs Advantage capture); (3) validated **phenotype/algorithm definitions** for exposure, outcomes, and covariates, with reported operating characteristics (PPV/sensitivity); (4) **time-zero alignment** that emulates an eligibility-and-assignment moment and avoids immortal time; (5) rigorous **confounding control** (active-comparator/new-user design, high-dimensional propensity scores, negative controls); (6) transparent handling of **attrition, missing data, and competing risks**; and (7) pre-specified **sensitivity and quantitative bias analyses** (E-value, probabilistic bias analysis, empirical calibration) with versioned code lists and analysis specifications. Underlying all of it: transparency, registration (the EU PAS Register / HMA-EMA Catalogues), and scientific independence per the Code of Conduct.\n\n**When NOT to use — limitations and common misapplications** — The most frequent error is **treating the Guide as a fill-in template or a quality score**. It is neither. (a) It is **not a reporting checklist** — \"we followed the ENCePP Guide\" does not satisfy STROBE or RECORD-PE reporting requirements, and reviewers should ask for the actual reporting checklist. (b) It is **not a risk-of-bias instrument** — it does not yield a graded bias judgment the way ROBINS-I does; do not cite it as evidence that a study is low-bias. (c) **Confusing the three ENCePP artifacts** — the *Guide* (standards reference), the *Checklist for Study Protocols* (protocol completeness), and the *Code of Conduct* (governance) are distinct; submitting one when another is expected is a common regulatory friction point. (d) **Checklist-as-theater** — completing the protocol checklist or asserting Guide adherence does not make an observational study causal; the design and analysis still have to earn the causal interpretation. (e) **Wrong tool for the job** — using the Guide for systematic-review reporting (use PRISMA) or for interventional-trial reporting (use CONSORT) is a category error. (f) The Guide is **maintained and versioned** — citing a stale revision or a broken chapter link undermines an EMA submission; always confirm the current revision and the live URL.\n\n**How it maps to this catalog** — Each Guide expectation is implemented by a concept in this repository. Research question, estimand, and intercurrent events → `picots-framework-rwe` and `estimands-ate-att-intercurrent-events-rwe`; protocol/SAP structure → `study-protocol-or-sap-elements`. Data fitness → `fit-for-purpose-data-assessment-rwe`, with claims operational detail in `claims-analysis` and `continuous-enrollment-observable-time-rwe`. Phenotype/algorithm validation → `diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe`, `claims-outcome-algorithm-ppv-sensitivity-rwe`, and `algorithm-validation`. Time-zero and immortal time → `time-zero-index-date-alignment-rwe`, `washout-clean-lookback-period-rwe`, and `immortal-time-bias-handling`. Design and confounding control → `active-comparator-new-user`, `high-dimensional-propensity-score-hdps-rwe`, `target-trial-emulation`, and `empirical-calibration-negative-controls-rwe`. Attrition, competing risks, and sensitivity → `attrition-and-loss-to-follow-up-rwe`, `competing-risks-cause-specific-fine-gray-rwe`, `e-value-sensitivity-analysis`, and `quantitative-bias-analysis-toolkit-rwe`. Regulatory packaging for PASS → `regulatory-readiness-rwe`, `pass-imposed`, and `pass-voluntary`. **Applied note (claims/EHR/registry RWE):** for a claims-based imposed PASS, treat the Guide as the spine of the protocol — pre-register on the EU PAS Register, specify the active-comparator new-user cohort with a 365-day continuous-enrollment washout, validate the outcome phenotype (report PPV), align time zero at first qualifying fill, build a high-dimensional propensity score from the baseline window, and pre-specify a CONSORT-style attrition flow plus E-value/negative-control sensitivity analyses — then report against RECORD-PE.", "primary_category": "Guideline", "tags": [ "guideline", "pharmacoepidemiology", "methodological-standards", "pass", "ema", "encepp", "real-world-evidence" ], "aliases": [ "ENCePP Guide", "ENCePP Methodological Guide", "Guide on Methodological Standards in Pharmacoepidemiology", "ENCePP Guide on Methodological Standards in Pharmacoepidemiology" ], "applies_to_study_types": [ "new_user", "active_comparator_new_user", "self_controlled_case_series", "case_crossover", "drug_utilization", "pass_imposed", "pass_voluntary", "cohort_prospective", "cohort_retrospective", "case_control", "claims_analysis", "ehr_study" ], "data_sources": [ "claims", "ehr", "registry", "linked", "multi-database" ], "citations": [ { "role": "introduce", "doi": null, "url": "https://encepp.europa.eu/encepp-toolkit/methodological-guide_en", "citation_text": "European Network of Centres for Pharmacoepidemiology and Pharmacovigilance (ENCePP). Guide on Methodological Standards in Pharmacoepidemiology (Revision 11). EMA/95098/2010 Rev.11. European Medicines Agency; July 2023.", "year": 2023, "authors_short": "ENCePP / EMA", "notes": "Canonical, maintained source document. No journal DOI exists; the Guide is a living web publication updated by structured review (current Revision 11, July 2023). Confirm the live revision before citing in a submission." }, { "role": "explain", "doi": null, "url": "https://encepp.europa.eu/encepp-toolkit/encepp-checklist-study-protocols_en", "citation_text": "ENCePP Checklist for Study Protocols. European Network of Centres for Pharmacoepidemiology and Pharmacovigilance (ENCePP), European Medicines Agency.", "year": 2023, "authors_short": "ENCePP / EMA", "notes": "Companion protocol-completeness aid to the Guide; the operational checklist a protocol is mapped against, distinct from the standards Guide itself." }, { "role": "use", "doi": "10.1136/bmj.m4856", "url": "https://doi.org/10.1136/bmj.m4856", "citation_text": "Wang SV, Pinheiro S, Hua W, et al. STaRT-RWE: structured template for planning and reporting on the implementation of real world evidence studies. BMJ. 2021;372:m4856.", "year": 2021, "authors_short": "Wang et al.", "notes": "Structured template the Guide recommends for transparently specifying and reporting RWE study implementation; pairs with the Guide at the protocol stage." }, { "role": "use", "doi": "10.1002/pds.5507", "url": "https://doi.org/10.1002/pds.5507", "citation_text": "Wang SV, Pottegård A, Crown W, et al. HARmonized Protocol Template to Enhance Reproducibility of hypothesis evaluating real-world evidence studies on treatment effects: A good practices report of a joint ISPE/ISPOR task force. Pharmacoepidemiology and Drug Safety. 2023;32(1):44-55.", "year": 2023, "authors_short": "Wang et al.", "notes": "HARPER protocol template for hypothesis-evaluating treatment-effect RWE; an endorsed protocol artifact that operationalizes the Guide's design expectations." } ], "relations": [ { "relation_type": "see_also", "target_slug": "target-trial-emulation", "notes": "The Guide endorses target-trial thinking to make the causal question and eligibility/assignment structure explicit before emulation in observational data." }, { "relation_type": "see_also", "target_slug": "active-comparator-new-user", "notes": "The active-comparator new-user design is the Guide's default analytic core for comparative safety/effectiveness, controlling confounding by indication." }, { "relation_type": "see_also", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "hdPS implements the Guide's expectation of rigorous confounding control with proxy adjustment in high-dimensional claims/EHR data." }, { "relation_type": "see_also", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Implements the Guide's requirement to pre-specify the estimand, target of inference, and intercurrent-event strategy." }, { "relation_type": "see_also", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Implements the Guide's expectation of validated, transparently defined phenotype/outcome algorithms with reported operating characteristics." }, { "relation_type": "see_also", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "Implements the Guide's data-fitness-for-purpose assessment of provenance, capture, and relevance for each data source." }, { "relation_type": "see_also", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Implements the Guide's expectation of transparent attrition reporting and handling of informative loss to follow-up." }, { "relation_type": "see_also", "target_slug": "quantitative-bias-analysis-toolkit-rwe", "notes": "Implements the Guide's call for pre-specified quantitative bias and sensitivity analyses for unmeasured confounding and misclassification." }, { "relation_type": "see_also", "target_slug": "regulatory-readiness-rwe", "notes": "Operationalizes packaging of an ENCePP/PASS study for EMA submission and the EU PAS Register." }, { "relation_type": "used_with", "target_slug": "claims-analysis", "notes": "The Guide's standards apply directly to claims-based PASS and comparative studies; see this concept for claims operational detail." }, { "relation_type": "used_with", "target_slug": "study-protocol-or-sap-elements", "notes": "The Guide governs the methodological content of the protocol/SAP that this concept structures." }, { "relation_type": "applies_to", "target_slug": "pass-imposed", "notes": "Default European methodological reference for imposed post-authorisation safety studies (GVP Module VIII)." }, { "relation_type": "applies_to", "target_slug": "pass-voluntary", "notes": "Applies to voluntary post-authorisation safety/efficacy studies on routinely collected data." }, { "relation_type": "applies_to", "target_slug": "drug-utilization", "notes": "Provides methodological standards for drug-utilization studies." }, { "relation_type": "applies_to", "target_slug": "active-comparator-new-user", "notes": "Methodological standards for comparative safety/effectiveness new-user cohort studies." }, { "relation_type": "applies_to", "target_slug": "claims-analysis", "notes": "Standards for studies built on administrative claims data." }, { "relation_type": "applies_to", "target_slug": "ehr-study", "notes": "Standards for studies built on electronic health record data." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "ema", "fda", "hta", "journal" ] }, { "slug": "entreq", "name": "ENTREQ", "short_definition": "Enhancing Transparency in Reporting the Synthesis of Qualitative Research — a 21-item, five-domain reporting statement for qualitative evidence syntheses (meta-ethnography, thematic synthesis, framework synthesis, etc.), not a critical-appraisal or risk-of-bias tool.", "long_description": "**What it is.** ENTREQ (Enhancing Transparency in Reporting the Synthesis of Qualitative Research) is a reporting\nguideline — a 21-item checklist organized into **five domains** (introduction, methods and methodology, literature\nsearch and selection, appraisal, and synthesis of findings) — published by Tong, Flemming, McInnes, Oliver, and Craig\nin *BMC Medical Research Methodology* in 2012 and indexed on the EQUATOR Network. Its purpose is to make the *process*\nof synthesizing primary qualitative studies transparent and reproducible: how the review question was framed, how\nqualitative studies were searched and selected, how (and whether) their quality was appraised, and — the part most\noften left opaque — how primary findings were coded, compared, and transformed into higher-order interpretive themes.\nENTREQ is method-agnostic across the qualitative synthesis family: it applies whether the synthesis approach is\nmeta-ethnography, thematic synthesis, framework synthesis, meta-aggregation, grounded-theory synthesis, or critical\ninterpretive synthesis. It is maintained as a community reporting standard rather than by a regulatory agency.\n\n**When to use.** Use ENTREQ when the deliverable is a **synthesis of primary qualitative research** and you need a\ntransparent reporting backbone — a peer-reviewed journal manuscript, a protocol, or the qualitative-evidence section of\nan HTA patient-input submission. In HEOR/RWE this is the qualitative arm of the evidence base: syntheses of patient and\ncaregiver experience of treatment burden, illness and symptom experience, barriers and facilitators to adherence,\nacceptability of an intervention, or the lived-experience evidence that anchors PRO instrument content and value\nframeworks. Decision rule for which guideline applies: if you are synthesizing *qualitative* studies, ENTREQ is the\numbrella reporting standard; if the synthesis method is specifically **meta-ethnography** (Noblit & Hare), the\npurpose-built extension **eMERGe** (France et al., 2019) is more granular and is the right choice. If you are reporting\na synthesis of *quantitative* (RCT or observational) studies, ENTREQ is the wrong family — use PRISMA 2020 and, for\nprotocols, PRISMA-P. For mixed-methods reviews, ENTREQ governs the qualitative strand while a PRISMA-family statement\ngoverns the quantitative strand.\n\n**What it requires.** ENTREQ enforces explicit reporting across its five domains. (1) *Introduction* — a clear\nqualitative review aim. (2) *Methods/methodology* — the synthesis methodology and the epistemological/theoretical\napproach stated up front, plus the reviewers' role and reflexivity. (3) *Literature search and selection* — search\nstrategy, sources, the sampling approach (exhaustive vs purposive/theoretical sampling, which is legitimate and even\npreferred in qualitative synthesis), inclusion/exclusion rationale, and a documented selection process. (4) *Appraisal*\n— whether and how the quality of included studies was assessed and what was done with that assessment (note: ENTREQ\nreports *that* appraisal happened; it does not itself appraise). (5) *Synthesis of findings* — how data were extracted,\nthe software used, the coding and comparison process, how themes were derived (data-driven vs theory-driven), how\nprimary-author quotations are distinguished from reviewers' interpretations, and the audit trail from raw quotations to\nthird-order constructs. The throughline is **interpretive traceability**: a reader should be able to follow each\nsynthesized theme back to the primary data that produced it.\n\n**When NOT to use — limitations and common misapplications.** ENTREQ is a **reporting** checklist; it is *not* a\nrisk-of-bias instrument and *not* a quality score. Completing it tells the reader what you did, not whether the\nsynthesis was done well — for critical appraisal of the included qualitative studies use CASP, JBI, or an equivalent\ntool, and do not confuse a fully ticked ENTREQ checklist with methodological rigor (checklist-as-theater). Concrete\nfailure modes: (a) **wrong extension for the design** — applying generic ENTREQ when the method is meta-ethnography and\nthe more specific eMERGe guidance exists, or — conversely — forcing ENTREQ onto a quantitative review where PRISMA is\nrequired; (b) treating it as a *conduct* guide — ENTREQ does not tell you how to perform meta-ethnography, only how to\nreport it; (c) using ENTREQ to lend false rigor to a thin synthesis with few primary studies or no audit trail; (d)\napplying it to a single primary qualitative study (where COREQ or SRQR is the relevant standard) rather than to a\n*synthesis* of studies. ENTREQ also predates GRADE-CERQual, which assesses confidence in qualitative synthesis findings\n— ENTREQ reports the synthesis; CERQual rates how much to trust each finding, and the two are complementary, not\ninterchangeable.\n\n**How it maps to this catalog.** ENTREQ is the reporting envelope for the catalog's qualitative and patient-experience\nconcepts. The synthesis itself is implemented by `qualitative-synthesis`; the primary studies it ingests are described\nby `qualitative-interview` and `qualitative-ethnographic`; when the qualitative strand sits inside a broader design,\n`mixed-methods` is the integrating frame. ENTREQ's domains directly support the patient-experience evidence used in\n`pro-development`, `pro-validation`, and `pro-rwe` (a transparent synthesis of lived-experience studies is how PRO item\ncontent and conceptual frameworks are justified for FDA PFDD and for HTA value cases). For where ENTREQ sits relative to\nits quantitative siblings, see `systematic-review` and `scoping-review` (and PRISMA/PRISMA-P in the guideline set).\n**Applied note for claims/EHR/registry RWE:** ENTREQ does not touch phenotype algorithms, time-zero, confounding\ncontrol, or attrition — those belong to the quantitative concepts. Its role in a claims/EHR/registry program is the\n*qualitative complement*: when a mixed-methods HEOR study pairs a claims-based treatment-pattern or cost analysis with a\nsynthesis of patient-experience studies (e.g., to explain why adherence breaks down, or to ground a treatment-burden\nendpoint), ENTREQ is the standard that keeps the qualitative half auditable while the quantitative half follows\nRECORD-PE/STROBE and the relevant design concepts.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "qualitative-synthesis", "meta-ethnography", "patient-experience", "equator", "mixed-methods" ], "aliases": [ "ENTREQ statement", "Enhancing Transparency in Reporting the Synthesis of Qualitative Research" ], "applies_to_study_types": [ "qualitative_synthesis", "mixed_methods" ], "data_sources": [ "primary" ], "citations": [ { "role": "introduce", "doi": "10.1186/1471-2288-12-181", "url": "https://doi.org/10.1186/1471-2288-12-181", "citation_text": "Tong A, Flemming K, McInnes E, Oliver S, Craig J. Enhancing transparency in reporting the synthesis of qualitative research: ENTREQ. BMC Medical Research Methodology. 2012;12:181.", "year": 2012, "authors_short": "Tong et al.", "notes": "The canonical ENTREQ statement — derives the 21 items across five domains and is the primary citation for any qualitative evidence synthesis using the framework." }, { "role": "use", "doi": "10.1186/s12874-018-0600-0", "url": "https://doi.org/10.1186/s12874-018-0600-0", "citation_text": "France EF, Cunningham M, Ring N, et al. Improving reporting of meta-ethnography: the eMERGe reporting guidance. BMC Medical Research Methodology. 2019;19:25.", "year": 2019, "authors_short": "France et al.", "notes": "Method-specific sibling extension. When the synthesis approach is meta-ethnography specifically, eMERGe is the more granular reporting standard to use alongside or instead of generic ENTREQ." }, { "role": "use", "url": "https://www.equator-network.org/reporting-guidelines/entreq/", "citation_text": "ENTREQ — Enhancing Transparency in Reporting the Synthesis of Qualitative Research. EQUATOR Network reporting-guidelines library.", "year": 2012, "authors_short": "EQUATOR Network", "notes": "Maintained landing page with the checklist and supporting resources." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "qualitative-synthesis", "notes": "ENTREQ is the reporting standard for qualitative evidence syntheses regardless of synthesis method." }, { "relation_type": "applies_to", "target_slug": "mixed-methods", "notes": "Governs the qualitative strand of a mixed-methods review; a PRISMA-family statement governs the quantitative strand." }, { "relation_type": "used_with", "target_slug": "qualitative-synthesis", "notes": "The implementing concept — meta-ethnography, thematic synthesis, framework synthesis, and meta-aggregation are the methods ENTREQ asks you to report transparently." }, { "relation_type": "used_with", "target_slug": "qualitative-interview", "notes": "Interview studies are typical primary inputs whose findings ENTREQ traces from raw quotation to synthesized theme." }, { "relation_type": "see_also", "target_slug": "qualitative-ethnographic", "notes": "Ethnographic primary studies are common source material for meta-ethnographic synthesis under ENTREQ." }, { "relation_type": "see_also", "target_slug": "pro-development", "notes": "Transparent qualitative synthesis of lived-experience studies justifies PRO item content and conceptual frameworks (FDA PFDD; HTA patient input)." }, { "relation_type": "see_also", "target_slug": "pro-rwe", "notes": "ENTREQ-reported syntheses supply the patient-experience evidence that complements quantitative PRO-RWE endpoints." }, { "relation_type": "see_also", "target_slug": "systematic-review", "notes": "ENTREQ is the qualitative counterpart to PRISMA-governed reviews of quantitative studies." }, { "relation_type": "see_also", "target_slug": "scoping-review", "notes": "Distinguish a qualitative evidence synthesis (ENTREQ) from a scoping review (PRISMA-ScR) when scoping the question." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "journal", "hta", "fda" ] }, { "slug": "eu-hta-jca", "name": "EU HTA Regulation — Joint Clinical Assessment (JCA)", "short_definition": "EU Regulation 2021/2282 on HTA established a mandatory Joint Clinical Assessment (JCA) process, effective January 2025 for oncology medicines and ATMPs (advanced therapy medicinal products). Conducted by EUnetHTA 21 member agencies, the JCA produces a PICO-driven relative effectiveness assessment of a new medicine vs. a defined comparator, which EU member states must use as a basis — not a mandatory conclusion — for their national HTA decisions.", "long_description": "**What it is** — **EU Regulation 2021/2282 on Health Technology Assessment**, commonly called the\n**EU HTA Regulation**, established a new pan-European mechanism for coordinated clinical assessment\nof medicines and medical devices. Its centrepiece is the **Joint Clinical Assessment (JCA)**: a\nstructured relative-effectiveness evaluation conducted by the **EUnetHTA 21** network of national\nHTA bodies (coordinated by the European Commission, with AIFA, HAS, and others serving as\nco-author agencies). The JCA is **not** a pricing, reimbursement, or full HTA decision — it assesses\nclinical evidence only (effectiveness and safety relative to a defined comparator), leaving economic\nevaluation, pricing negotiation, and final reimbursement to member states. The JCA follows a\n**PICO-driven scope** (Population, Intervention, Comparator(s), Outcomes) agreed between the\nmanufacturer and the assessment team via a **scoping process**; manufacturers submit a **Joint\nSubmission Dossier (JSD)** structured per EUnetHTA 21 templates (analogous to, but distinct from,\nAMCP Format for the US). The regulation entered into force 12 January 2022; **JCA became mandatory\nfor oncology medicines and ATMPs on 12 January 2025**, with mandatory extension to orphan products\nin 2028 and all other medicinal products in 2030.\n\n**When to use** — Apply the EU HTA JCA framework when: preparing a regulatory or market-access\nstrategy for a medicine seeking European authorisation in oncology or ATMP indications from 2025\nonward; designing a global evidence generation plan that must support both the JCA and national HTA\ndecisions in Germany (AMNOG), France (HAS), UK (NICE), and others simultaneously; or when developing\nan indirect treatment comparison, network meta-analysis, or RWE package to supplement pivotal trial\ndata for the JSD. Specific design implications: (1) **Comparators** — the JCA scope process\ndetermines which comparators are assessed; manufacturers must anticipate requests for multiple\ncountry-specific comparators that span the member-state landscape, often requiring multi-comparator\nindirect treatment comparisons. (2) **Outcomes scope** — the PICO agreement includes a pre-specified\noutcome set; evidence packages must address each outcome domain (overall survival, progression-free\nsurvival, PROs, safety) per the agreed PICO rather than the manufacturer's preferred endpoints.\n(3) **RWE role** — while pivotal trial data remain primary, EUnetHTA 21 methods guidance acknowledges\nRWE for supplementary safety, long-term outcomes, and real-world comparative effectiveness when\ntrial data are immature or missing. Decision rule: the JCA applies to **EMA-centralised procedure\nproducts in oncology/ATMPs** from January 2025; other products remain subject to their current\nnational HTA processes until 2028–2030.\n\n**What it requires (checklist domains)** — The JCA dossier and process enforce these substantive\nelements. *Scoping*: an agreed PICO for each member-state sub-population; manufacturer and assessment\nteam exchange positions on comparators, populations, and outcomes in a structured pre-submission\ndialogue. *Joint Submission Dossier (JSD)*: (a) **Product information** — EMA label and approved\nindication; (b) **Clinical trial data** — full clinical study reports for pivotal trials, including\nsubgroup data for the agreed PICO populations; (c) **Indirect treatment comparisons / NMA** — when no\nhead-to-head evidence exists against the required comparators, a systematic review plus indirect\ncomparison following EUnetHTA 21 methods guidance (building on ISPOR and NICE DSU technical support);\n(d) **Real-world evidence** — where pivotal trial follow-up is immature or lacks an agreed comparator,\nobservational studies or registry data may supplement; EUnetHTA 21 has published methods guidance on\nRWE acceptability standards; (e) **Patient-reported outcomes** — PRO evidence using validated\ninstruments for the agreed outcome domains; (f) **Safety data** — pooled and individual-study safety\nprofiles. *Assessment report*: the co-author agencies produce a joint report with a structured\nrelative effectiveness conclusion (added benefit, no added benefit, insufficient evidence) by outcome\ndomain; member states must consult this report before national HTA decisions. *Transparency*:\nJCA reports are publicly available; the JSD is submitted to the IT platform maintained by the\nEuropean Commission.\n\n**When NOT to use — limitations and common misapplications** — (1) **Confusing JCA with\nreimbursement decisions** — the JCA assesses clinical evidence only; national bodies retain full\nauthority over pricing, cost-effectiveness thresholds, and reimbursement decisions. A positive JCA\ndoes not guarantee reimbursement in any member state. (2) **Assuming JCA replaces national HTAs** —\nmember states must *use* the JCA as a basis but are not bound by its conclusions; Germany's AMNOG\nprocess and France's HAS will continue their national processes, informed by but not replaced by\nthe JCA. (3) **Using pre-JCA submission templates** — the Joint Submission Dossier format (EUnetHTA\n21 templates) is distinct from older EUnetHTA Core Model submissions; using outdated templates or\nAMCP-structured dossiers without adaptation creates non-conforming submissions. (4) **Ignoring the\nmulti-comparator challenge** — a single JCA scope may include 5–10 comparators reflecting different\nmember-state standard-of-care practices; a manufacturer who designs evidence around a single\npreferred comparator may face requests for indirect comparisons that their evidence package cannot\nsupport. (5) **Assuming RWE is accepted as primary evidence** — EUnetHTA 21 methods guidance\npositions RWE as supplementary; the JCA hierarchy places randomised evidence highest, and RWE must\nmeet explicit quality criteria (design rigor, population representativeness, outcome validity) to\nbe given weight. (6) **Neglecting the timeline** — mandatory JCA for oncology/ATMPs began January\n2025; evidence generation plans for products in late-stage development must already account for JCA\nrequirements alongside pivotal trial design.\n\n**How it maps to this catalog** — The JCA dossier draws on the same methods-catalog concepts that\nunderpin rigorous global evidence packages. *Indirect treatment comparisons and NMA* — the most\ncommon technical challenge in JCA submissions — are governed by **ispor-indirect** and **prisma-nma**\n(the systematic review and reporting standard for the evidence base). *RWE supplementary packages*\nsubmitted under EUnetHTA 21 methods guidance rely on **target-trial-emulation** (rigorous\nobservational design), **propensity-score-methods-psm-iptw** (confounding control), and\n**fit-for-purpose-data-assessment-rwe** (demonstrating that the European registry or EHR data source\nis adequate for the PICO population). *Outcome validity* for PROs relies on **cosmin-criteria** and\n**isoqol-standards**; for claims/registry outcomes on **claims-outcome-algorithm-ppv-sensitivity-rwe**.\n*Evidence certainty* across the JCA dossier is structured with **grade** (GRADE certainty of\nevidence), which EUnetHTA 21 has incorporated into its methods guidance for rating confidence in\nrelative effectiveness conclusions. *Patient-centricity and estimands* — the JCA PICO scope process\nincreasingly aligns with ICH E9(R1) estimand thinking (**ich-e9-r1** in this catalog), especially\nfor oncology endpoints where treatment switching and intercurrent events complicate OS estimation.\n*Systematic review underpinning the clinical evidence section* should follow **prisma-2020**. The\nJCA also intersects with **cheers-2022** (CHEERS economic reporting) for member states that request\ncost-effectiveness analysis alongside the clinical JCA.", "primary_category": "Guideline", "tags": [ "guideline", "eu-hta", "joint-clinical-assessment", "market-access", "relative-effectiveness", "indirect-treatment-comparison", "real-world-evidence", "european-regulation", "pico" ], "aliases": [ "EU HTA JCA", "Joint Clinical Assessment", "EU HTA Regulation", "EU Regulation 2021/2282", "EUnetHTA 21", "JCA", "Joint Submission Dossier", "JSD" ], "applies_to_study_types": [ "cer_observational", "rct", "nma", "systematic_review", "registry" ], "data_sources": [ "claims", "ehr", "registry", "linked", "primary" ], "citations": [ { "role": "introduce", "doi": "10.3390/jmahp12020008", "url": "https://doi.org/10.3390/jmahp12020008", "citation_text": "Schuster V, Ecker S, Wild C. EU HTA Regulation and Joint Clinical Assessment — Threat or Opportunity? Journal of Market Access & Health Policy. 2024;12(2):8.", "year": 2024, "authors_short": "Schuster et al.", "notes": "Peer-reviewed commentary in Journal of Market Access & Health Policy assessing the implications of EU Regulation 2021/2282 and the JCA process for manufacturers, HTA bodies, and health systems; published ahead of the January 2025 mandatory go-live for oncology/ATMPs." }, { "role": "explain", "url": "https://health.ec.europa.eu/health-technology-assessment_en", "citation_text": "European Commission. Health Technology Assessment — EU HTA Regulation 2021/2282 and Joint Clinical Assessment. European Commission Health website (maintained). Brussels: EC; 2025.", "year": 2025, "authors_short": "European Commission", "notes": "Official EU Commission page for the HTA Regulation, JCA process documentation, IT platform access, and EUnetHTA 21 methods deliverables." }, { "role": "use", "doi": "10.1017/s026646232610364x", "url": "https://doi.org/10.1017/s026646232610364x", "citation_text": "Meregaglia M, Costa F, Cavallaro L, et al. Implementing the EU HTA Regulation and Joint Clinical Assessment: a multi-stakeholder perspective from Italy. International Journal of Technology Assessment in Health Care. 2026;42:e12.", "year": 2026, "authors_short": "Meregaglia et al.", "notes": "Multi-stakeholder perspective on JCA implementation experience in a major EU member state; illustrates practical submission, scoping, and evidence-gap challenges for manufacturers." } ], "relations": [ { "relation_type": "used_with", "target_slug": "ispor-indirect", "notes": "Multi-comparator ITCs and NMAs are among the most common technical challenges in JCA dossiers; ISPOR Indirect Treatment Comparisons guidance governs their conduct and reporting." }, { "relation_type": "used_with", "target_slug": "grade", "notes": "EUnetHTA 21 uses GRADE to rate certainty of relative effectiveness evidence in JCA reports; manufacturers should anticipate GRADE downgrades for indirect comparisons or immature RWE." }, { "relation_type": "used_with", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "RWE submitted as JCA supplementary evidence must demonstrate that the data source is adequate for the PICO population; EUnetHTA 21 methods guidance specifies acceptability criteria." }, { "relation_type": "used_with", "target_slug": "target-trial-emulation", "notes": "The target-trial emulation framework is the design discipline that produces JCA-acceptable observational RWE by making the design transparent and aligned with the agreed PICO." }, { "relation_type": "used_with", "target_slug": "cosmin-criteria", "notes": "PRO evidence in the JCA dossier must use validated instruments; COSMIN criteria govern measurement-property adequacy." }, { "relation_type": "see_also", "target_slug": "prisma-nma", "notes": "Systematic reviews and NMAs underpinning JCA indirect comparisons should follow PRISMA-NMA reporting standards." }, { "relation_type": "see_also", "target_slug": "ich-e9-r1", "notes": "ICH E9(R1) estimand framework is increasingly relevant for JCA oncology submissions where treatment switching and intercurrent events affect overall survival estimation." }, { "relation_type": "see_also", "target_slug": "cheers-2022", "notes": "CHEERS governs economic evaluation reporting that may accompany national HTA submissions that build on the JCA clinical evidence base." }, { "relation_type": "see_also", "target_slug": "amcp-format", "notes": "The AMCP Format is the US payer dossier equivalent; global evidence packages must serve both JCA and AMCP audiences simultaneously, requiring careful comparator and outcome planning." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "ema", "hta", "journal" ] }, { "slug": "eu-pas-register", "name": "EU PAS Register / HMA-EMA Catalogue of Real-World Data Studies", "short_definition": "Public registration platform for non-interventional and real-world data studies on authorised medicines in the EU. Launched 2010 as the EU PAS Register by ENCePP under EMA coordination; superseded in February 2024 by the HMA-EMA Catalogue of real-world data studies, the study-records companion to the data-sources catalogue.", "long_description": "**What it is** — The **EU PAS Register** (EU Post-Authorisation Study Register) is a free, publicly searchable\nregistration platform for non-interventional post-authorisation studies of authorised medicines, launched in November\n2010 by the European Network of Centres for Pharmacoepidemiology and Pharmacovigilance (**ENCePP**) under European\nMedicines Agency (**EMA**) coordination. In **February 2024** it was decommissioned and replaced by the **HMA-EMA\nCatalogues of real-world data sources and studies** — a paired system in which the *studies* catalogue (the EU PAS\nRegister's successor) holds study records and links each one to entries in the companion *data sources* catalogue.\nThroughout, the platform is a **transparency container**: it records that a study exists, its objectives, design, data\nsources, milestones, protocol, amendments, and results — it does not appraise, score, or certify methodological\nquality. It is governed alongside the ENCePP Code of Conduct and the ENCePP Guide on Methodological Standards in\nPharmacoepidemiology, which supply the methods substance the register itself does not.\n\n**When to use** — Register a study here when it is a **non-interventional / observational** study of an authorised\nmedicine in the EU regulatory ecosystem. Registration is a **legal obligation for imposed PASS** (post-authorisation\nsafety studies required as a condition of marketing authorisation, under Good Pharmacovigilance Practices Module VIII)\nand is strongly expected for **voluntary PASS, drug-utilisation studies, and effectiveness/safety RWE** intended for\nEU regulatory interaction, ENCePP study status, the ENCePP Seal, or transparency-driven publication. Decision rules for\npicking the *right* registry: interventional clinical trials go to **CTIS** (EU Clinical Trials Regulation; legacy\nEudraCT) or ClinicalTrials.gov, **not** here; systematic reviews / meta-analyses go to **PROSPERO**; this platform is\nfor primary observational and real-world data studies. Post-February 2024, register new studies in the **HMA-EMA\nCatalogue of real-world data studies** and link the study record to the database's entry in the data-sources catalogue;\ndo not attempt to register in the retired EU PAS Register or cite dead `encepp.eu/encepp/studiesDatabase.jsp` links.\n\n**What it requires** — The platform enforces **fields and timing**, not analytic content. At registration it captures\nstudy identification, research question and objectives, design, study population, exposures/outcomes of interest, the\ndata source(s), planned methods, sponsor/funding, and milestones; for imposed PASS the **agreed protocol must be posted\nbefore data collection starts**, with substantial amendments, progress, and the final study report or results posted as\nthey occur. The post-2024 catalogue adds the requirement to **link the study to a registered real-world data source**,\nreinforcing data-provenance transparency. Framed for real-world data work, the register operationalises **design\ntransparency** (prospective, time-stamped posting of objectives, design, and analysis intentions before results are\nknown) and **data-fitness traceability** (an auditable link from the study to the database it used). It does **not**\nitself mandate phenotype/algorithm validation, time-zero alignment, estimand specification, confounding control, or\nquantitative bias analysis — those are the substance the registrant supplies, governed by the ENCePP Guide, GVP Module\nVIII, HARPER, and STaRT-RWE, and implemented through the catalog concepts mapped below.\n\n**When NOT to use — limitations and common misapplications** — The single most consequential error is treating\nregistration as a **quality or validity guarantee**: the register is not a reporting checklist (use STROBE / RECORD-PE),\nnot a protocol template (use HARPER / STaRT-RWE), and not a risk-of-bias instrument (use ROBINS-I / the ENCePP\nChecklist). A registered protocol can still be a poorly designed, confounded, immortal-time-ridden observational study —\n**registration does not make a study causal.** Other failure modes: (1) **confusing registration with the ENCePP Seal**\n— the Seal signals adherence to the ENCePP Code of Conduct and is a separate, additional step; (2) **retrospective\nregistration** after results are known, which defeats the anti-selective-reporting purpose and still permits outcome\nswitching; (3) **wrong registry** — entering an RCT (belongs in CTIS) or a systematic review (belongs in PROSPERO);\n(4) **registration-as-theater** — posting a thin record to satisfy an obligation while the substantive protocol,\namendments, or results are never completed or updated; (5) **stale-platform citation** — referencing the retired EU PAS\nRegister or its old ENCePP URLs after the February 2024 migration to the HMA-EMA Catalogue.\n\n**How it maps to this catalog** — The register is the transparency wrapper; these concepts implement what actually goes\ninside each field. Objectives / PICOTS and protocol scaffolding -> `picots-framework-rwe`, `study-protocol-or-sap-elements`.\nDesign, time-zero, and a defensible comparator -> `target-trial-emulation`, `active-comparator-new-user`,\n`time-zero-index-date-alignment-rwe`. Data-source fitness (the link to the data-sources catalogue) ->\n`fit-for-purpose-data-assessment-rwe`, `claims-analysis`, `regulatory-readiness-rwe`. Outcome/exposure operational\ndefinitions and phenotype validation -> `diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe`,\n`claims-outcome-algorithm-ppv-sensitivity-rwe`. Confounding control -> `high-dimensional-propensity-score-hdps-rwe`,\n`propensity-score-methods-psm-iptw`. Estimands and intercurrent events -> `estimands-ate-att-intercurrent-events-rwe`.\nAttrition and results reporting -> `attrition-and-loss-to-follow-up-rwe`, `database-feasibility-attrition-funnel-rwe`.\nSensitivity / quantitative bias analysis -> `e-value-sensitivity-analysis`, `quantitative-bias-analysis-toolkit-rwe`.\n\n**Applied note (claims / EHR / registry RWE).** For a US-claims or EHR study registered for EU regulatory or\ntransparency purposes, create the study record and link it to the database's entry in the HMA-EMA *data-sources*\ncatalogue; where a US source (Medicare FFS, commercial, or an EHR network) has sparse or absent catalogue metadata,\ncarry provenance explicitly through a `fit-for-purpose-data-assessment-rwe` write-up rather than leaning on the\ncatalogue entry. Pair registration with HARPER or STaRT-RWE for the protocol you post, and RECORD-PE for the eventual\nreport — the register fixes *that* and *when*, those guidelines fix *what* and *how well*.", "primary_category": "Guideline", "tags": [ "guideline", "registration", "transparency", "pharmacoepidemiology", "rwe", "pass", "encepp" ], "aliases": [ "EU PAS Register", "EU Post-Authorisation Study Register", "EUPAS", "HMA-EMA Catalogue of Real-World Data Studies", "HMA-EMA Catalogues of real-world data sources and studies", "ENCePP EU PAS Register" ], "applies_to_study_types": [ "pass_imposed", "pass_voluntary", "drug_utilization", "cohort_prospective", "cohort_retrospective" ], "data_sources": [ "claims", "ehr", "registry", "linked", "multi-database" ], "citations": [ { "role": "introduce", "doi": null, "url": "https://catalogues.ema.europa.eu/", "citation_text": "HMA-EMA Catalogues of real-world data sources and studies. Heads of Medicines Agencies (HMA) and European Medicines Agency (EMA), 2024 (successor to the EU PAS Register, launched by ENCePP/EMA in 2010).", "year": 2024, "authors_short": "HMA / EMA", "notes": "Canonical, maintained agency platform. No journal statement paper exists for the register itself; this stable URL is the authoritative reference. The EU PAS Register was retired and migrated here in February 2024." }, { "role": "explain", "doi": "10.1002/pds.5413", "url": "https://doi.org/10.1002/pds.5413", "citation_text": "Sultana J, Crisafulli S, Almas M, et al. Overview of the European post-authorisation study register post-authorization studies performed in Europe from September 2010 to December 2018. Pharmacoepidemiology and Drug Safety. 2022;31(8):823-835.", "year": 2022, "authors_short": "Sultana et al.", "notes": "Empirical characterisation of the studies actually registered in the EU PAS Register (scope, design, data collection, transparency), the best peer-reviewed description of the platform's content and use." }, { "role": "explain", "doi": null, "url": "https://encepp.europa.eu/encepp-toolkit/methodological-guide_en", "citation_text": "ENCePP Guide on Methodological Standards in Pharmacoepidemiology (Revision 11). European Network of Centres for Pharmacoepidemiology and Pharmacovigilance, 2023.", "year": 2023, "authors_short": "ENCePP", "notes": "The methods companion to the register; supplies the methodological substance (design, data, analysis standards) that registration alone does not enforce." }, { "role": "use", "doi": "10.1002/pds.5507", "url": "https://doi.org/10.1002/pds.5507", "citation_text": "Wang SV, Pottegård A, Crown W, et al. HARmonized Protocol Template to Enhance Reproducibility of hypothesis evaluating real-world evidence studies on treatment effects: A good practices report of a joint ISPE/ISPOR task force (HARPER). Pharmacoepidemiology and Drug Safety. 2023;32(1):44-55.", "year": 2023, "authors_short": "Wang et al.", "notes": "Structured protocol template to populate the register's protocol field for hypothesis-evaluating treatment-effect RWE." }, { "role": "use", "doi": "10.1136/bmj.m4856", "url": "https://doi.org/10.1136/bmj.m4856", "citation_text": "Wang SV, Pinheiro S, Hua W, et al. STaRT-RWE: structured template for planning and reporting on the implementation of real world evidence studies. BMJ. 2021;372:m4856.", "year": 2021, "authors_short": "Wang et al.", "notes": "Structured planning/reporting template that pairs with registration to make the registered protocol and final report reproducible." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "pass-imposed", "notes": "Registration in the HMA-EMA Catalogue (formerly EU PAS Register) is a legal obligation for imposed PASS under GVP Module VIII; the protocol must be posted before data collection." }, { "relation_type": "applies_to", "target_slug": "pass-voluntary", "notes": "Voluntary PASS are registered to support transparency and ENCePP study status / Seal." }, { "relation_type": "applies_to", "target_slug": "drug-utilization", "notes": "Drug-utilisation studies are a core registered study type on the platform." }, { "relation_type": "applies_to", "target_slug": "cohort-prospective", "notes": "Prospective non-interventional cohorts are registered with protocol, milestones, and results." }, { "relation_type": "applies_to", "target_slug": "cohort-retrospective", "notes": "Retrospective database cohorts are registered with their data source linked to the data-sources catalogue." }, { "relation_type": "see_also", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "Supplies the data-fitness substance behind the catalogue's data-source link; required when the catalogue's source metadata is thin (e.g., US claims/EHR)." }, { "relation_type": "see_also", "target_slug": "regulatory-readiness-rwe", "notes": "Registration is one element of a regulatory-ready RWE package alongside protocol, validation, and reporting artifacts." }, { "relation_type": "see_also", "target_slug": "picots-framework-rwe", "notes": "Implements the objectives/population/comparator fields the register captures." }, { "relation_type": "see_also", "target_slug": "study-protocol-or-sap-elements", "notes": "Defines the protocol/SAP content that populates the posted protocol field." }, { "relation_type": "complements", "target_slug": "target-trial-emulation", "notes": "Pre-specifying the emulated trial before registration enforces the prospective design transparency the register is built to capture." }, { "relation_type": "complements", "target_slug": "active-comparator-new-user", "notes": "Supplies the defensible comparator and time-zero design registered under study design." }, { "relation_type": "complements", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Implements the validated outcome/exposure phenotypes the registered protocol must specify." }, { "relation_type": "complements", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "Implements the confounding-control plan declared in the registered analysis methods." }, { "relation_type": "complements", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Specifies the estimand and intercurrent-event handling the protocol field should state." }, { "relation_type": "complements", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Implements the attrition/results reporting posted as study results." }, { "relation_type": "complements", "target_slug": "claims-analysis", "notes": "Operational detail for registering and reporting claims-based real-world data studies." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "ema", "hta", "journal" ] }, { "slug": "fda-pragmatic-trials", "name": "FDA Draft Guidance: Integrating Randomized Controlled Trials into Routine Clinical Practice", "short_definition": "FDA draft guidance (September 2024, CDER/CBER; Docket FDA-2024-D-2052) describing how to design and conduct randomized controlled trials with streamlined protocols embedded in routine care, using real-world data infrastructure (EHR, registries, claims) for enrollment, endpoint ascertainment, and follow-up.", "long_description": "**What it is** — *Integrating Randomized Controlled Trials for Drug and Biological Products Into\nRoutine Clinical Practice* is a draft guidance for industry issued by the U.S. FDA (CDER and CBER)\nin September 2024 (Docket FDA-2024-D-2052; comment period closed December 2024). It is part of FDA's\nReal-World Evidence (RWE) program and sits alongside the Agency's framework and the RWD-source\nguidances. Its purpose is to describe how sponsors can run *randomized* controlled trials with\nsimplified, streamlined protocols and procedures focused on essential data, embedding the trial in\nordinary clinical care and leveraging real-world data infrastructure (EHR, disease and product\nregistries, claims) for recruitment, baseline characterization, endpoint capture, and follow-up.\nThese designs are also called pragmatic trials, point-of-care trials, or large simple trials. The\ndefining feature that separates this guidance from the rest of FDA's RWE suite is that **randomization\nis retained** — the real-world component is the *data and care setting*, not the causal contrast.\nIt is a regulatory framework (what makes such a trial acceptable as substantial evidence), not a\nreporting checklist and not a risk-of-bias instrument.\n\n**When to use** — Reach for this guidance when you are planning or defending a randomized,\ninterventional trial that will be conducted under real-world conditions for an FDA submission:\nbroad eligibility, care delivered by treating clinicians rather than dedicated research staff,\nobjective endpoints (death, hospitalization, major clinical events) ascertained from routinely\ncollected data, and risk-based monitoring. Decision rule for picking the right document: if the\nstudy **randomizes** treatment and uses RWD mainly for data capture and follow-up, this guidance\ngoverns; if the study is **non-interventional** (treatment decisions made in routine care, no\nrandomization, confounding controlled by design and analysis), use `fda-rwe-noninterventional`\nand `fda-rwd-ehr-claims` instead. PRECIS-2 helps you *position* a design on the\nexplanatory–pragmatic continuum and CONSORT-Pragmatic governs *journal reporting*; this FDA\nguidance addresses *regulatory acceptability* — they are complementary, not interchangeable.\nRegistry-based randomized trials (RRTs) that randomize within an existing registry fall squarely\nin scope.\n\n**What it requires** — The substantive expectations cluster into: (1) **trial design integrity** —\npre-specified, simplified protocol; preserved randomization and allocation concealment; broadened\nbut defensible eligibility; equipoise. (2) **Fitness-for-use of the RWD that supports the trial** —\nthe EHR, registry, or claims source used for endpoints, covariates, and follow-up must be relevant\nand reliable (provenance, completeness, accuracy, traceability), held to the same standard as in\nFDA's RWD guidances. (3) **Endpoint ascertainment from routine data** — endpoints must be\nobjective and reliably captured from the data stream; validated computable phenotypes / outcome\nalgorithms, adjudication where needed, and a defensible mortality source. (4) **Estimands and\nintercurrent events** — an explicit estimand with treatment-policy (ITT-like) framing for the\ncomparative effect, and pre-specified handling of non-adherence, treatment switching, and\ncrossover that are common when patients are treated in routine care. (5) **Retention, missing data,\nand follow-up** — attrition and loss to follow-up minimized and characterized; missing-data\nmechanisms and analytic handling pre-specified. (6) **Human-subject protection and pragmatic\nconduct** — informed consent, safety reporting, and risk-based monitoring adapted to the care\nsetting without compromising participant protection or data integrity.\n\n**When NOT to use — limitations and common misapplications** — (a) Do **not** apply this guidance\nto non-interventional observational database studies; randomization is its premise, and an\nobservational target-trial emulation needs `fda-rwe-noninterventional` / `fda-rwd-ehr-claims`\nplus design-based confounding control, not this document. (b) Do **not** treat it as a reporting\nchecklist or a quality score — it confers no \"compliance badge,\" and a streamlined protocol that\nticks its themes is not automatically pragmatic, valid, or generalizable. (c) Do **not** conflate\nit with PRECIS-2 (a design-positioning tool) or CONSORT-Pragmatic / SPIRIT (reporting and protocol\ntemplates); using a reporting extension where regulatory design guidance is required (or vice\nversa) is the classic wrong-tool error. (d) \"Streamlining\" is not license to weaken the parts that\ncarry the evidence: randomization integrity, endpoint ascertainability from routine data, adequate\npower, and equipoise still bind, and endpoints that cannot be reliably captured from the available\nRWD (e.g., subjective or imaging-adjudicated outcomes absent in the data) are a fatal design flaw.\n(e) It does not relax human-subject protections or pharmacovigilance obligations.\n\n**How it maps to this catalog** — The randomization premise changes which catalog concepts\nimplement which requirement, and an honest mapping says so. Confounding-control concepts such as\n`high-dimensional-propensity-score-hdps-rwe` and `active-comparator-new-user` are **not** the\nworkhorses here — randomization, not propensity adjustment, balances confounders; those concepts\nbelong to the observational alternative, `target-trial-emulation`, which you would reach for only\nwhen a randomized pragmatic trial is infeasible. What *does* implement this guidance:\n`fit-for-purpose-data-assessment-rwe` (relevance/reliability of the supporting RWD source);\n`diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe`, `outcome-algorithm-construction-rwe`, and\n`endpoint-adjudication-chart-review-rwe` (endpoint ascertainment from EHR/claims, with validation);\n`mortality-source-hierarchy-rwe` (death capture and censoring); `composite-endpoint-construction-rwe`\nand `estimands-ate-att-intercurrent-events-rwe` (estimand definition and intercurrent-event\nhandling); `attrition-and-loss-to-follow-up-rwe` and `missing-data-pattern-table-rwe` (retention and\nmissing data); `generalizability-transportability-external-validity-rwe` (the representativeness\npayoff that motivates pragmatic designs); `picots-framework-rwe` and `sample-size-power-precision-rwe`\n(scoping and adequacy); and `claims-analysis` for the operational mechanics of the supporting data.\nApplied note for claims/EHR/registry RWE: the trial randomizes, but every endpoint, covariate, and\nfollow-up flag is still read from a real-world feed — so the same phenotype validation, continuous-\nobservability, and mortality-source discipline used in observational studies apply to the trial's\ndata layer, and a weak data source undermines a randomized pragmatic trial just as surely as it\nundermines an observational one.", "primary_category": "Guideline", "tags": [ "guideline", "regulatory-framework", "fda", "pragmatic-trial", "point-of-care-trial", "large-simple-trial", "randomized", "rwe" ], "aliases": [ "FDA Pragmatic Trials", "Integrating RCTs into Routine Clinical Practice", "Point-of-Care Trials", "Large Simple Trials", "Registry-Based Randomized Trials", "FDA-2024-D-2052" ], "applies_to_study_types": [ "pragmatic_trial", "registry_trial" ], "data_sources": [ "ehr", "registry", "claims", "linked" ], "citations": [ { "role": "introduce", "doi": null, "url": "https://www.fda.gov/regulatory-information/search-fda-guidance-documents/integrating-randomized-controlled-trials-drug-and-biological-products-routine-clinical-practice", "citation_text": "U.S. Food and Drug Administration. Integrating Randomized Controlled Trials for Drug and Biological Products Into Routine Clinical Practice: Draft Guidance for Industry. CDER/CBER, September 2024 (Docket FDA-2024-D-2052).", "year": 2024, "authors_short": "FDA (CDER/CBER)", "notes": "Canonical statement document. Agency draft guidance with no journal DOI; the FDA guidance-documents permalink is the stable reference." }, { "role": "explain", "doi": "10.1056/NEJMra1510059", "url": "https://doi.org/10.1056/NEJMra1510059", "citation_text": "Ford I, Norrie J. Pragmatic Trials. New England Journal of Medicine. 2016;375(5):454-463.", "year": 2016, "authors_short": "Ford & Norrie", "notes": "Authoritative review of the design principles (streamlined protocols, routine-care conduct, objective endpoints) that the FDA guidance operationalizes for regulatory use." }, { "role": "explain", "doi": "10.1177/1740774515598334", "url": "https://doi.org/10.1177/1740774515598334", "citation_text": "Califf RM, Sugarman J. Exploring the ethical and regulatory issues in pragmatic clinical trials. Clinical Trials. 2015;12(5):436-441.", "year": 2015, "authors_short": "Califf & Sugarman", "notes": "Frames the ethics, consent, and regulatory questions raised when randomized trials are embedded in routine practice." }, { "role": "use", "doi": null, "url": "https://www.federalregister.gov/documents/2024/09/18/2024-21077/integrating-randomized-controlled-trials-for-drug-and-biological-products-into-routine-clinical", "citation_text": "Federal Register. Integrating Randomized Controlled Trials for Drug and Biological Products Into Routine Clinical Practice; Draft Guidance for Industry; Availability. 89 FR, September 18, 2024.", "year": 2024, "authors_short": "Federal Register", "notes": "Official availability notice and docket of record for submitting comments." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "pragmatic-trial", "notes": "Primary scope — randomized pragmatic / point-of-care / large simple trials conducted in routine clinical practice for FDA submission." }, { "relation_type": "applies_to", "target_slug": "registry-trial", "notes": "Covers registry-based randomized trials that randomize within an existing registry and use registry follow-up for endpoints." }, { "relation_type": "see_also", "target_slug": "consort-pragmatic", "notes": "CONSORT extension for reporting pragmatic trials in journals; complements (does not replace) the FDA design/regulatory expectations." }, { "relation_type": "see_also", "target_slug": "precis-2", "notes": "Design-positioning tool for placing a trial on the explanatory–pragmatic continuum; use at planning to justify pragmatic choices the guidance expects." }, { "relation_type": "see_also", "target_slug": "fda-rwe-framework", "notes": "Parent FDA RWE program framework within which this guidance sits." }, { "relation_type": "alternative_to", "target_slug": "target-trial", "notes": "When randomization is infeasible, the observational target-trial-emulation pathway (and fda-rwe-noninterventional) replaces this guidance and confounding must be controlled by design/analysis rather than randomization." }, { "relation_type": "used_with", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "Implements the relevance/reliability assessment of the RWD source that supports enrollment, endpoints, and follow-up." }, { "relation_type": "used_with", "target_slug": "outcome-algorithm-construction-rwe", "notes": "Implements objective endpoint ascertainment from EHR/claims, with validation." }, { "relation_type": "used_with", "target_slug": "endpoint-adjudication-chart-review-rwe", "notes": "Implements adjudication of endpoints that routine data cannot reliably classify on their own." }, { "relation_type": "used_with", "target_slug": "mortality-source-hierarchy-rwe", "notes": "Implements reliable death ascertainment and censoring for time-to-event endpoints." }, { "relation_type": "used_with", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Implements the estimand and intercurrent-event handling (non-adherence, switching, crossover) the guidance requires." }, { "relation_type": "used_with", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Implements characterization and minimization of attrition under routine-care follow-up." }, { "relation_type": "used_with", "target_slug": "generalizability-transportability-external-validity-rwe", "notes": "Implements the representativeness/generalizability gains that motivate pragmatic designs." }, { "relation_type": "used_with", "target_slug": "claims-analysis", "notes": "Operational mechanics of the claims/EHR data layer that supports the randomized trial." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda" ] }, { "slug": "fda-rwd-ehr-claims", "name": "FDA RWD Guidance: Assessing EHR and Medical Claims Data", "short_definition": "FDA final guidance (July 2024) on assessing the relevance and reliability of electronic health record (EHR) and medical claims data proposed to support effectiveness or safety conclusions in drug and biologic regulatory submissions.", "long_description": "**What it is** — *Real-World Data: Assessing Electronic Health Records and Medical Claims Data\nTo Support Regulatory Decision-Making for Drug and Biological Products* is a final FDA guidance\nfor industry (CDER/CBER, July 2024; finalizing the September 2021 draft of the same title) issued\nunder the 21st Century Cures Act Real-World Evidence (RWE) Program. It is not a journal reporting\nchecklist maintained by EQUATOR; it is a regulator's statement of the considerations sponsors\nshould address when they propose to use EHR or medical claims data — alone or linked — as the\ndata source for a clinical study intended to support a regulatory determination of effectiveness\nor safety. Its organizing logic is the *fitness-for-use* assessment: whether a given data source\nis **relevant** (does it capture the exposures, outcomes, covariates, and population needed to\nanswer the question) and **reliable** (are the data accrued, curated, and quality-assured such\nthat the captured values can be trusted), and whether those properties can be documented and\naudited. It sits alongside FDA's parent RWE Framework and its companion guidances on study designs\nusing RWD and on regulatory submission of RWD/RWE.\n\n**When to use** — Apply this guidance whenever an EHR- or claims-based non-interventional study\n(or the external-control or hybrid arm of a trial) is being designed, conducted, or documented\nwith the intent of submitting it to FDA to support an effectiveness or safety claim — an IND,\nNDA, BLA, sNDA/sBLA, or a required post-marketing study. Use it from the protocol stage, not\nretrospectively: the guidance expects the data-source assessment, study design, and analysis plan\nto be pre-specified before analytic results are seen, and it expects sponsors to engage the Agency\nearly. Decision rules for *this* document versus its siblings: use **this guidance** when the\ncentral question is *can this EHR/claims data source support the study* (data provenance, linkage,\nvalidation, accrual, quality control); use FDA's **\"Considerations for the Use of RWD and RWE\nto Support Regulatory Decision-Making\"** guidance when the question is broader design/analysis\nconsiderations across RWD types; use the **\"Data Standards for Drug and Biological Product\nSubmissions Containing RWD\"** guidance for the submission/format mechanics. For an HTA or payer\ndossier, or a peer-reviewed manuscript, this guidance is a strong reference standard but the\n*reporting* vehicle is typically STaRT-RWE, STROBE/RECORD-PE, or HARPER — use those for the report\nand this guidance to justify the data source. EMA/ENCePP work is governed by its own GVP and ENCePP\ninstruments; this guidance carries no statutory force outside FDA but its fitness-for-use\nexpectations are broadly concordant.\n\n**What it requires** — The guidance enforces documentation across several substantive domains.\n(1) *Data source provenance and selection*: why this EHR/claims source, who curates it, how raw\nrecords become the analytic dataset, and the full set of data-management/transformation steps,\nincluding any extraction-transformation-load conversions to a common data model. (2)\n*Fitness-for-use — relevance*: availability of the key exposures, outcomes, covariates, and a\npopulation that maps to the target; adequate follow-up and the temporality needed for the estimand.\n(3) *Fitness-for-use — reliability*: data accrual and **lag**, completeness, plausibility,\nconformance, and provenance/audit traceability back to source records. (4) *Definition and\nvalidation of study variables*: pre-specified operational definitions for exposure, outcomes, and\ncovariates, and **validation of the computable phenotype/algorithm** (e.g., PPV, sensitivity against\na reference standard such as chart review or adjudication), with the validation population\nrepresentative of the study population. (5) *Linkage*: when EHR, claims, registries, or mortality\nfiles are linked, the linkage method, match rate, and the error/selection it introduces. (6)\n*Design integrity*: clear time-zero/index definition that avoids immortal time, appropriate\ncomparator, and covariate assessment windows. (7) *Quality assurance and governance*: a data\nquality plan, study monitoring, and access for FDA inspection/audit of source data. Although FDA\ndoes not publish a numbered checklist, sponsors are expected to provide this evidence prospectively\nand to be able to reproduce the analytic dataset from source.\n\n**When NOT to use — limitations and common misapplications** — This is a regulatory framework, not\na risk-of-bias instrument and not a numeric quality score: there is no item count, no threshold,\nand \"addressing the guidance\" produces no grade. The most damaging misapplications: (a) treating\nthe guidance as a **scorecard or checklist-as-theater** — listing that each domain was \"considered\"\nwithout the underlying validation, lineage, and quality evidence; (b) assuming that satisfying the\ndata-fitness expectations confers **causal validity** — the guidance governs whether the *data* can\nsupport the question, not whether the *design* identifies a causal effect, which still requires\ncomparator choice, confounding control, and bias analysis; (c) conflating **data accrual lag** with\nfitness-for-use, or ignoring lag entirely so that recent outcomes are differentially undercaptured;\n(d) using an unvalidated computable phenotype, or borrowing a PPV from a different database/era as\nif it transports; (e) using **EHR problem-list or claims service dates as time-zero** without\nconfirming the actual index event, which manufactures immortal time and misclassifies exposure;\n(f) declining negative-control or quantitative bias diagnostics on the grounds that \"the guidance\ndoes not require them\" — they are how reliability claims are stress-tested. It is also the wrong\nprimary instrument for prospective registries collected to protocol, for primary-data-collection\nstudies, or for non-US submissions where ENCePP/GVP govern; and it does not replace the design or\ndata-standards guidances in its own family.\n\n**How it maps to this catalog** — Each requirement is implemented by a concrete concept here.\n*Fitness-for-use (relevance + reliability)* → `fit-for-purpose-data-assessment-rwe`, with payer\nstructure and coding-intensity nuances in `medicare-ffs-ma-commercial-claims-differences-rwe` and\ndata-source mechanics in `claims-analysis`. *Phenotype/algorithm definition and validation* →\n`diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe`. *Design integrity, comparator, and\ntime-zero* → `target-trial-emulation` and `active-comparator-new-user`, with immortal-time pitfalls\nin `immortal-time-bias-handling`. *Estimand and intercurrent events* →\n`estimands-ate-att-intercurrent-events-rwe`. *Confounding control* →\n`high-dimensional-propensity-score-hdps-rwe` and `propensity-score-methods-psm-iptw`. *Attrition,\naccrual lag, and missingness* → `attrition-and-loss-to-follow-up-rwe` and\n`database-feasibility-attrition-funnel-rwe`. *Reliability stress-testing and sensitivity* →\n`empirical-calibration-negative-controls-rwe`, `e-value-sensitivity-analysis`, and\n`quantitative-bias-analysis-toolkit-rwe`.\n\n*Applied note (claims/EHR/registry RWE).* In a Medicare FFS + commercial claims effectiveness\nstudy, satisfying this guidance means: documenting the licensor, refresh cadence, and claims\n**adjudication lag** (often 3–6 months, longer for some settings) and excluding immature\nperson-time; requiring continuous medical+pharmacy enrollment so absence of a code is true-negative\nrather than missing; restricting or flagging Medicare Advantage person-time where FFS claims are\nabsent; pre-specifying and validating the outcome phenotype (e.g., a 1-inpatient-or-2-outpatient\nrule with PPV from chart-confirmed cases in *this* source); fixing time-zero at the first qualifying\nfill, not the diagnosis date; and pre-registering negative-control outcomes and an E-value so the\nreliability of the comparison is demonstrated, not asserted. Linked EHR adds severity and lab detail\nbut its match rate and linkage selection must be reported.", "primary_category": "Guideline", "tags": [ "guideline", "regulatory", "fda", "rwe", "fitness-for-use", "data-quality" ], "aliases": [ "FDA RWD EHR/Claims guidance", "FDA EHR and claims data guidance", "FDA 2024 RWD EHR/claims guidance", "Assessing EHR and Medical Claims Data" ], "applies_to_study_types": [ "claims_analysis", "ehr_study", "linked_data" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "doi": null, "url": "https://www.fda.gov/regulatory-information/search-fda-guidance-documents/real-world-data-assessing-electronic-health-records-and-medical-claims-data-support-regulatory", "citation_text": "U.S. Food and Drug Administration. Real-World Data: Assessing Electronic Health Records and Medical Claims Data To Support Regulatory Decision-Making for Drug and Biological Products. Guidance for Industry (Final). CDER/CBER; July 2024.", "year": 2024, "authors_short": "FDA (CDER/CBER)", "notes": "Canonical statement of the guidance. Agency guidance has no journal DOI; this is the stable FDA guidance-documents landing page. Finalizes the September 2021 draft of the same title." }, { "role": "explain", "doi": "10.1136/bmj.m4856", "url": "https://doi.org/10.1136/bmj.m4856", "citation_text": "Wang SV, Pinheiro S, Hua W, et al. STaRT-RWE: structured template for planning and reporting on the implementation of real world evidence studies. BMJ. 2021;372:m4856.", "year": 2021, "authors_short": "Wang et al.", "notes": "Structured template that operationalizes the same design and data-source parameters the FDA guidance expects sponsors to pre-specify (data source, exposure/outcome/covariate definitions, time-zero, follow-up, analysis); a practical reporting vehicle for compliance." }, { "role": "explain", "doi": "10.1002/pds.4297", "url": "https://doi.org/10.1002/pds.4297", "citation_text": "Berger ML, Sox H, Willke RJ, et al. Good practices for real-world data studies of treatment and/or comparative effectiveness: Recommendations from the joint ISPOR-ISPE Special Task Force on real-world evidence in health care decision making. Pharmacoepidemiol Drug Saf. 2017;26(9):1033-1039.", "year": 2017, "authors_short": "Berger et al.", "notes": "Cross-society good-practice recommendations on RWD relevance, reliability, transparency, and a-priori specification that prefigure FDA's fitness-for-use expectations for EHR/claims data." }, { "role": "use", "doi": null, "url": "https://www.fda.gov/media/152503/download", "citation_text": "U.S. Food and Drug Administration. Real-World Data: Assessing Electronic Health Records and Medical Claims Data To Support Regulatory Decision-Making for Drug and Biological Products (full guidance PDF). July 2024.", "year": 2024, "authors_short": "FDA", "notes": "Direct PDF of the final guidance text for working reference and protocol citation." }, { "role": "use", "doi": null, "url": "https://www.fda.gov/science-research/science-and-research-special-topics/real-world-evidence", "citation_text": "U.S. Food and Drug Administration. Real-World Evidence Program (parent framework and companion guidances). FDA Science & Research.", "year": 2024, "authors_short": "FDA", "notes": "Parent RWE Program page situating this guidance among the design and data-standards companion guidances." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "claims-analysis", "notes": "Governs the data-fitness, validation, and documentation expectations when a claims-based study is intended to support an FDA effectiveness or safety determination." }, { "relation_type": "applies_to", "target_slug": "ehr-study", "notes": "Governs relevance/reliability assessment, phenotype validation, and time-zero integrity for EHR-based regulatory studies." }, { "relation_type": "applies_to", "target_slug": "linked-data", "notes": "Adds linkage-method, match-rate, and linkage-selection documentation when EHR, claims, registry, or mortality sources are linked." }, { "relation_type": "requires", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "Implements the guidance's core relevance-and-reliability fitness-for-use assessment." }, { "relation_type": "requires", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Implements the required definition and validation of computable phenotypes/algorithms (e.g., PPV against a reference standard)." }, { "relation_type": "used_with", "target_slug": "target-trial-emulation", "notes": "Provides the design discipline (eligibility, treatment strategies, time-zero) the guidance expects sponsors to pre-specify." }, { "relation_type": "used_with", "target_slug": "active-comparator-new-user", "notes": "Supplies a defensible comparator and incident-user, immortal-time-free time-zero structure consistent with the guidance's design-integrity expectations." }, { "relation_type": "used_with", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "Implements confounding control over the documented covariates in claims/EHR." }, { "relation_type": "used_with", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Defines the estimand and intercurrent-event handling the data source must support." }, { "relation_type": "used_with", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Documents attrition, follow-up adequacy, and the accrual-lag effects central to reliability." }, { "relation_type": "see_also", "target_slug": "medicare-ffs-ma-commercial-claims-differences-rwe", "notes": "Payer-structure and coding-intensity differences that drive relevance/reliability and Medicare Advantage person-time handling." }, { "relation_type": "see_also", "target_slug": "empirical-calibration-negative-controls-rwe", "notes": "Negative-control diagnostics that stress-test reliability claims beyond the guidance's explicit asks." }, { "relation_type": "see_also", "target_slug": "e-value-sensitivity-analysis", "notes": "Quantitative bias analysis to demonstrate, not assert, robustness of the comparison." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "hta", "journal" ] }, { "slug": "fda-rwe-devices", "name": "FDA RWE for Medical Devices", "short_definition": "FDA's December 2025 final guidance on how real-world evidence is evaluated and used to support medical-device regulatory decisions across the total product life cycle; it centers on real-world data relevance and reliability (\"fit-for-purpose\"). Issued by CDRH and CBER under FDORA section 3629; supersedes the 2017 first edition.", "long_description": "**What it is** — *Use of Real-World Evidence to Support Regulatory Decision-Making for Medical\nDevices* is an FDA guidance for industry and FDA staff, issued by the Center for Devices and\nRadiological Health (CDRH) with the Center for Biologics Evaluation and Research (CBER). The\nfirst edition was finalized in August 2017; a revised final version was issued in December 2025\n(Federal Register notice of availability, 18 December 2025) pursuant to section 3629 of the Food\nand Drug Omnibus Reform Act of 2022 (FDORA), which directed FDA to update its real-world data\n(RWD) / real-world evidence (RWE) device guidance. It is a **regulatory decision framework, not\na reporting checklist or a risk-of-bias instrument**: it describes how FDA evaluates whether RWD\nare of sufficient **relevance** and **reliability** to generate RWE that can support a\ndevice regulatory decision, and it is maintained by FDA (CDRH).\n\n**When to use** — Use this guidance whenever a device sponsor proposes to generate or submit RWE\nto support a CDRH/CBER regulatory action: a marketing submission (510(k), De Novo, PMA, PMA\nsupplement, HDE), an expanded or modified indication, label changes, a condition-of-approval\npost-approval study (PAS), a section 522 postmarket surveillance order, active safety\nsurveillance / signal evaluation through programs such as NEST, or construction of an\nexternal/historical control for a single-arm device study. **Decision rule:** apply *this*\ndevice guidance — rather than the drug/biologic-oriented `fda-rwe-framework` and its companions\n(`fda-rwe-noninterventional`, `fda-rwd-ehr-claims`) — whenever the regulated product is a medical\ndevice or device-led combination product reviewed by CDRH/CBER. Device-specific realities (UDI-\nrather than NDC-based exposure ascertainment, operator/site learning-curve effects, iterative\ndesign changes within a product family, and registry-centric data) make the device guidance\ncontrolling. Pair it with study-conduct and reporting tools (HARPER/STaRT-RWE for protocols,\nSTROBE/RECORD-PE for reporting) — those are complementary, not substitutes.\n\n**What it requires** — The substantive backbone is **fit-for-purpose** assessment of RWD against\nthe specific regulatory question. *Relevance:* the data must capture the device exposure (unique\ndevice identifier or device/procedure codes), the target population and indication, and the\noutcomes at adequate granularity, with sufficient follow-up and sample size. *Reliability:* data\n**accrual** (provenance, completeness, timeliness, representativeness) and **data quality\nassurance/control** (accuracy, conformance, transformation/transcription integrity, auditability).\nBeyond data fitness, the guidance expects the methodological discipline of a credible\nnon-randomized study: a pre-specified protocol and statistical analysis plan; transparent,\nvalidated operational definitions for device exposure and outcome phenotypes; correct time-zero/\nindex-date alignment to avoid immortal-time and other time-related biases; an explicit estimand\nwith pre-stated handling of intercurrent events; rigorous confounding control (active comparator,\npropensity-score and high-dimensional methods) given the absence of randomization; rigor for\nexternal/historical controls where used; accounting for attrition and missing data; and\nsensitivity / quantitative bias analyses sized to the decision's risk. It also addresses curating\nand linking device registries, the use of data collected under Emergency Use Authorization, and\nearly engagement with FDA (pre-submission/Q-Submission) to align on data and design before lock.\n\n**When NOT to use — limitations and common misapplications** — (1) It is **not** a reporting\nchecklist or quality score; you cannot \"complete\" it the way you tick STROBE or RECORD items —\nit demands substantive evidence of relevance, reliability, and design validity. (2) It does **not**\nlower the evidentiary bar: clean, high-quality data do not make an observational comparison\ncausal; confounding, selection, and time-related biases must still be designed out. (3)\nWrong-document error: applying it to a drug/biologic (use the `fda-rwe-framework` family) or\napplying a drug-oriented RWE guidance to a device. (4) Treating registry participation as\nautomatic fitness-for-use — many device registries lack the linkage, comparator, or outcome\nascertainment needed for the question at hand. (5) Assuming RWE substitutes for a trial when the\nquestion (novel device, no adequate comparator, high residual-bias risk) genuinely requires a\nrandomized design or a single-arm study with a rigorously justified external control. (6)\nChecklist-as-theater: asserting \"fit-for-purpose\" without the phenotype validation, balance\ndiagnostics, and bias analysis that substantiate it.\n\n**How it maps to this catalog** — Each requirement is implemented by a concept entry here.\n`fit-for-purpose-data-assessment-rwe` operationalizes the relevance/reliability core.\n`diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe` and `claims-analysis` implement\ndevice-exposure and outcome operational definitions and their validation (UDI, procedure/CPT/\nICD-10-PCS coding). `time-zero-index-date-alignment-rwe` implements index/time-zero discipline.\n`active-comparator-new-user` and `high-dimensional-propensity-score-hdps-rwe` implement\nnon-randomized confounding control, while `target-trial-emulation` supplies the overarching\ndesign discipline. `estimands-ate-att-intercurrent-events-rwe` implements estimand and\nintercurrent-event specification. `attrition-and-loss-to-follow-up-rwe` implements attrition and\nmissing-data accounting; `e-value-sensitivity-analysis` and `quantitative-bias-analysis-toolkit-rwe`\nimplement sensitivity and quantitative bias analysis. `single-arm-external-control` and\n`rare-disease-external-controls-rwe` implement the external/historical control arms that are\ncommon in device evaluation; `generalizability-transportability-external-validity-rwe` and\n`regulatory-readiness-rwe` support transportability and submission readiness. **Applied note for\nregistry/claims/EHR device RWE:** unlike drugs, device exposure has no NDC — anchor exposure on\nthe UDI, device-specific procedure codes, and curated device registries (e.g., cardiovascular\nimplant registries), reconcile operator/site learning-curve effects and iterative device\nversions across a product family, and link registry data to claims and a death index to complete\nfollow-up and capture mortality.", "primary_category": "Guideline", "tags": [ "guideline", "regulatory", "fda", "medical-devices", "rwe", "fit-for-purpose", "framework" ], "aliases": [ "FDA RWE Devices", "FDA Medical Device RWE Guidance", "CDRH RWE Guidance", "Use of Real-World Evidence to Support Regulatory Decision-Making for Medical Devices" ], "applies_to_study_types": [ "disease_registry", "product_registry", "cer_observational", "single_arm_external_control" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "doi": null, "url": "https://www.fda.gov/regulatory-information/search-fda-guidance-documents/use-real-world-evidence-support-regulatory-decision-making-medical-devices", "citation_text": "U.S. Food and Drug Administration (CDRH/CBER). Use of Real-World Evidence to Support Regulatory Decision-Making for Medical Devices: Guidance for Industry and Food and Drug Administration Staff. Final guidance, December 2025.", "year": 2025, "authors_short": "FDA CDRH/CBER", "notes": "Canonical agency document. Issued under FDORA section 3629; the December 2025 final version supersedes the original August 2017 guidance. No journal DOI exists for an FDA guidance, so the stable FDA landing page is the introduce reference." }, { "role": "explain", "doi": "10.1056/NEJMsb1609216", "url": "https://doi.org/10.1056/NEJMsb1609216", "citation_text": "Sherman RE, Anderson SA, Dal Pan GJ, et al. Real-World Evidence — What Is It and What Can It Tell Us? New England Journal of Medicine. 2016;375(23):2293-2297.", "year": 2016, "authors_short": "Sherman et al.", "notes": "FDA-authored conceptual framing of RWD/RWE and the relevance-and-reliability logic that underpins the agency's device fit-for-purpose evaluation." }, { "role": "use", "doi": null, "url": "https://www.federalregister.gov/documents/2025/12/18/2025-23252/use-of-real-world-evidence-to-support-regulatory-decision-making-for-medical-devices-guidance-for", "citation_text": "Federal Register. Use of Real-World Evidence To Support Regulatory Decision-Making for Medical Devices; Guidance for Industry and Food and Drug Administration Staff; Availability. 18 December 2025.", "year": 2025, "authors_short": "FDA / Federal Register", "notes": "Official notice of availability for the final guidance; the authoritative record of the December 2025 finalization." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "disease-registry", "notes": "Governs RWD relevance/reliability and design expectations when a disease registry supports a device regulatory decision." }, { "relation_type": "applies_to", "target_slug": "product-registry", "notes": "Device/product registries are a central RWD source under this guidance; participation alone does not establish fitness-for-use." }, { "relation_type": "applies_to", "target_slug": "cer-observational", "notes": "Applies to observational comparative-effectiveness/safety studies of devices proposed for regulatory use." }, { "relation_type": "applies_to", "target_slug": "single-arm-external-control", "notes": "Single-arm device studies with an external/historical control must meet the guidance's relevance, reliability, and comparability expectations." }, { "relation_type": "complements", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "Implements the relevance-and-reliability core (data accrual + quality assurance/control) that the guidance requires." }, { "relation_type": "complements", "target_slug": "target-trial-emulation", "notes": "Supplies the design discipline (explicit protocol, eligibility, assignment, time zero) the guidance expects of a non-randomized device study." }, { "relation_type": "complements", "target_slug": "active-comparator-new-user", "notes": "Implements confounding control by indication for comparative device studies lacking randomization." }, { "relation_type": "complements", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "Implements proxy-based confounding adjustment when key confounders are unmeasured in device RWD." }, { "relation_type": "complements", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Implements explicit estimand definition and intercurrent-event handling required for the analysis plan." }, { "relation_type": "complements", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Implements validated exposure/outcome operational definitions (here anchored on UDI and procedure codes) and PPV/validation evidence." }, { "relation_type": "complements", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Implements attrition and missing-data accounting the guidance expects in transparent reporting." }, { "relation_type": "see_also", "target_slug": "quantitative-bias-analysis-toolkit-rwe", "notes": "Sensitivity and quantitative bias analyses substantiate fitness-for-use claims and address residual confounding/misclassification." }, { "relation_type": "see_also", "target_slug": "rare-disease-external-controls-rwe", "notes": "External/historical controls are common in device evaluation (rare conditions, novel devices) and carry specific comparability requirements." }, { "relation_type": "see_also", "target_slug": "fda-rwe-framework", "notes": "Sibling guidance for drugs/biologics; use the device guidance instead when the regulated product is a device or device-led combination product." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda" ] }, { "slug": "fda-rwe-framework", "name": "Framework for FDA's Real-World Evidence Program", "short_definition": "FDA's December 2018 programmatic framework, issued under 21st Century Cures Act Section 3022, that sets out how the agency will evaluate the potential use of real-world data and real-world evidence to support regulatory decisions on effectiveness for approved drugs and biologics. It is a strategic evaluation lens (data fit-for-use, study design, study conduct) — not a reporting checklist or a risk-of-bias instrument.", "long_description": "**What it is** — The *Framework for FDA's Real-World Evidence Program* is a strategic, programmatic\ndocument published by the U.S. Food and Drug Administration (Center for Drug Evaluation and Research\nand Center for Biologics Evaluation and Research) in December 2018, in response to Section 3022 of\nthe 21st Century Cures Act. It articulates the principles FDA will use when assessing whether\nreal-world data (RWD — data on health status and care delivery routinely collected from sources such\nas claims, EHRs, product and disease registries, and digital health technologies) can generate\nreal-world evidence (RWE — clinical evidence about a product's use and potential benefits or risks\nderived from analysis of RWD) of sufficient quality to support a new indication for an approved drug\nor to satisfy post-approval study requirements. The Framework is organized around three evaluation\npillars that the agency applies to any proposed RWE submission: (1) whether the **RWD are fit for use**\n— relevance and reliability of the data; (2) whether the **study design** used to generate RWE\nprovides adequate scientific evidence to answer the regulatory question; and (3) whether the **study\nconduct** meets FDA regulatory requirements (e.g., monitoring, data integrity). It is maintained by\nFDA, which has since operationalized the Framework through a series of more specific RWE guidances\n(2021–2024). The Framework is the umbrella; the operational expectations live in the sibling\nguidances.\n\n**When to use** — Consult the Framework when scoping a regulatory RWE strategy for an approved drug\nor biologic: deciding whether a non-interventional study, a hybrid/pragmatic design, an externally\ncontrolled trial, or a registry-based study could plausibly support a label expansion or fulfill a\npost-marketing requirement, and when preparing for a sponsor-FDA meeting on RWE. It is the right\nreference for the *programmatic, decision-grade* question — \"will FDA consider this kind of RWE\ncredible for this kind of decision?\" — and for orienting a team to FDA's three-pillar evaluation\nlogic before any protocol is drafted. **Decision rule for which document applies:** use THIS Framework\nfor high-level strategy and the agency's evaluation lens; switch to the operational FDA guidances for\ndesign and submission detail — `fda-rwe-noninterventional` (non-interventional study design and\nanalysis expectations), `fda-rwd-ehr-claims` (assessing EHR and medical-claims data for\nfitness-for-use), and `fda-rwe-devices` (RWE for medical devices, a separate CDRH pathway). For\nEU/EMA strategy use the ENCePP and GVP Module VIII references; for HTA/payer dossiers, RWE strategy is\ngoverned by HTA-body methods guidance (e.g., NICE RWE framework), not by this FDA document. For\nprotocol templating and reporting — the steps the Framework expects but does not itself specify — use\nHARPER/StaRT-RWE (protocol) and STROBE/RECORD-PE (reporting).\n\n**What it requires** — The Framework does not impose a numbered checklist; it sets evaluation\nexpectations across its three pillars, each of which maps onto concrete RWD methods.\n*Data relevance and reliability (fit-for-use):* the data must capture the exposures, outcomes, and\nkey covariates needed for the question, with adequate accuracy, completeness, provenance, and quality\ncontrols — including validated phenotype/outcome algorithms and a documented data-curation/linkage\ntrail. *Study design rigor:* FDA explicitly endorses applying clinical-trial design principles to\nobservational data — pre-specification, a clear causal estimand, an appropriate comparator, correct\ntime-zero/index alignment to avoid immortal-time and selection bias, and rigorous confounding\ncontrol. *Study conduct:* transparency, pre-registration where applicable, data integrity, and\npre-specified sensitivity and quantitative bias analyses so that the evidence is reproducible and\ndefensible. In practice this means a submission must document data-source fitness, phenotype\nvalidation (PPV/sensitivity), exposure and outcome definitions, estimands and intercurrent-event\nhandling, attrition, and bias analyses — the substantive domains a reviewer will probe.\n\n**When NOT to use — limitations and common misapplications** — The single most common error is\ntreating the Framework as a *checklist to complete*. It is a strategic/programmatic document, not a\nreporting checklist, not a critical-appraisal or risk-of-bias instrument, and not a quality score;\n\"satisfying the Framework\" is not a meaningful claim, and an analyst who needs item-level reporting\nstructure should reach for STaRT-RWE, HARPER, RECORD-PE, or STROBE instead. A second error is using\nthe 2018 Framework where the *operational* FDA guidance is required — it is the umbrella, and the\ndesign- and submission-level expectations live in the 2021–2024 guidances (`fda-rwe-noninterventional`,\n`fda-rwd-ehr-claims`, `fda-rwe-devices`); citing the Framework when a reviewer expects the specific\nguidance is the \"wrong-document\" trap. The Framework also does not, by itself, make an observational\nstudy causal: invoking it does not substitute for the actual design and analytic work (active\ncomparator, time-zero alignment, confounding control, sensitivity analysis) that earns a causal\ninterpretation. It is U.S.-FDA and drug/biologic specific — it does not govern EU/EMA submissions,\nHTA decisions, or (directly) devices — and it addresses *effectiveness* for approved products, not\ninitial approval or safety surveillance per se. Finally, it is a framework, not a standard: it\nsignals what FDA values but defers the methodological specifics to other guidances and to the\nscientific literature, so it should never be the sole cited authority for a design choice.\n\n**How it maps to this catalog** — The three pillars map directly onto implementing concepts in this\nrepository. *Data fit-for-use* is implemented by `fit-for-purpose-data-assessment-rwe` (relevance and\nreliability assessment), `claims-analysis` (claims structure and limitations),\n`medicare-ffs-ma-commercial-claims-differences-rwe` (payer-driven data-capture differences), and the\nphenotype concepts `diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe`,\n`ehr-phenotyping-algorithms-rwe`, and `claims-outcome-algorithm-ppv-sensitivity-rwe` (algorithm\ndevelopment and PPV/sensitivity validation). *Study design rigor* is implemented by\n`target-trial-emulation` (pre-specify the hypothetical trial before emulating it),\n`active-comparator-new-user` (comparator choice and incident-user restriction),\n`time-zero-index-date-alignment-rwe` and `immortal-time-bias-handling` (correct index alignment),\n`estimands-ate-att-intercurrent-events-rwe` (causal estimand and intercurrent events), and\n`high-dimensional-propensity-score-hdps-rwe` and `propensity-score-methods-psm-iptw` (confounding\ncontrol). *Study conduct, transparency, and robustness* is implemented by\n`study-protocol-or-sap-elements`, `attrition-and-loss-to-follow-up-rwe` and\n`database-feasibility-attrition-funnel-rwe` (attrition reporting),\n`e-value-sensitivity-analysis` and `quantitative-bias-analysis-toolkit-rwe` (quantitative bias\nanalysis), and `regulatory-readiness-rwe` (assembling the submission package). A reviewer evaluating\na claims- or EHR-based submission under this Framework will, in effect, walk the chain from\n`fit-for-purpose-data-assessment-rwe` and `diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe`\n(is the data fit and are the phenotypes validated?) through `active-comparator-new-user` and\n`time-zero-index-date-alignment-rwe` (is the design trial-emulating and free of immortal time?) to\n`e-value-sensitivity-analysis` (how robust is the result to unmeasured confounding?). Treat the\nFramework as the lens and these concepts as the lenses' implementation.", "primary_category": "Guideline", "tags": [ "guideline", "regulatory-framework", "fda", "rwe", "rwd", "21st-century-cures" ], "aliases": [ "FDA RWE Framework", "Framework for FDA's Real-World Evidence Program", "FDA RWE Program Framework", "21st Century Cures RWE Framework" ], "applies_to_study_types": [ "cer_observational", "new_user", "active_comparator_new_user", "claims_analysis", "ehr_study", "disease_registry", "product_registry", "single_arm_external_control" ], "data_sources": [ "claims", "ehr", "registry", "linked", "multi-database" ], "citations": [ { "role": "introduce", "doi": null, "url": "https://www.fda.gov/media/120060/download", "citation_text": "U.S. Food and Drug Administration. Framework for FDA's Real-World Evidence Program. Silver Spring, MD; December 2018.", "year": 2018, "authors_short": "FDA", "notes": "Canonical agency document, issued under 21st Century Cures Act Section 3022; no journal DOI exists. Stable FDA-hosted PDF (40 pp.) verified to resolve and to carry the December 2018 title page." }, { "role": "explain", "doi": "10.1056/NEJMsb1609216", "url": "https://doi.org/10.1056/NEJMsb1609216", "citation_text": "Sherman RE, Anderson SA, Dal Pan GJ, et al. Real-World Evidence — What Is It and What Can It Tell Us? New England Journal of Medicine. 2016;375(23):2293-2297.", "year": 2016, "authors_short": "Sherman et al.", "notes": "FDA-authored conceptual foundation distinguishing RWD from RWE and motivating the program the 2018 Framework formalizes. DOI verified via Crossref (title and first author Sherman, NEJM, 2016)." }, { "role": "use", "doi": "10.1056/NEJMp2200089", "url": "https://doi.org/10.1056/NEJMp2200089", "citation_text": "Concato J, Corrigan-Curay J. Real-World Evidence — Where Are We Now? New England Journal of Medicine. 2022;386(18):1680-1682.", "year": 2022, "authors_short": "Concato & Corrigan-Curay", "notes": "FDA RWE-program leadership update on how the Framework's pillars are being operationalized across the 2021+ guidance series. DOI verified via Crossref (title and first author Concato, NEJM, 2022)." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "cer-observational", "notes": "The Framework governs how FDA evaluates comparative observational evidence proposed to support an effectiveness decision." }, { "relation_type": "applies_to", "target_slug": "new-user-design", "notes": "New-user (incident-user) restriction is part of the trial-emulating design rigor the Framework expects under its study-design pillar." }, { "relation_type": "applies_to", "target_slug": "active-comparator-new-user", "notes": "An active-comparator, new-user design is the canonical way to meet the Framework's design-rigor expectations (comparator choice, time-zero alignment, confounding control)." }, { "relation_type": "applies_to", "target_slug": "claims-analysis", "notes": "Claims studies must satisfy the data fit-for-use pillar; see fda-rwd-ehr-claims for the operational fitness criteria." }, { "relation_type": "applies_to", "target_slug": "single-arm-external-control", "notes": "Externally controlled designs are a frequent RWE submission type evaluated under all three pillars." }, { "relation_type": "see_also", "target_slug": "fda-rwe-noninterventional", "notes": "Operational sibling guidance specifying non-interventional study design and analysis expectations — use this for design-level detail the Framework defers." }, { "relation_type": "see_also", "target_slug": "fda-rwd-ehr-claims", "notes": "Operational sibling guidance on assessing EHR and medical-claims data for fitness-for-use (Framework pillar 1)." }, { "relation_type": "see_also", "target_slug": "fda-rwe-devices", "notes": "Parallel CDRH framework for RWE on medical devices; cite this, not the drug/biologic Framework, for device questions." }, { "relation_type": "complements", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "Implements the data relevance-and-reliability pillar." }, { "relation_type": "complements", "target_slug": "target-trial-emulation", "notes": "Implements the study-design pillar — pre-specify the hypothetical trial before emulating it." }, { "relation_type": "complements", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Implements phenotype/algorithm validation required for data reliability." }, { "relation_type": "complements", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "Implements the confounding-control expectation under study-design rigor." }, { "relation_type": "complements", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Implements the pre-specified causal estimand and intercurrent-event handling." }, { "relation_type": "complements", "target_slug": "e-value-sensitivity-analysis", "notes": "Implements the quantitative bias / sensitivity analysis expectation under study conduct." }, { "relation_type": "complements", "target_slug": "regulatory-readiness-rwe", "notes": "Assembles the data-fitness, design, and conduct evidence into a submission package." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "hta" ] }, { "slug": "fda-rwe-noninterventional", "name": "FDA RWE: Considerations Regarding Non-Interventional Studies", "short_definition": "FDA draft guidance (March 2024) advising sponsors on planning, designing, and reporting non-interventional (observational) studies of drugs and biologics intended to contribute to substantial evidence of effectiveness and/or evidence of safety, with an emphasis on data reliability/relevance, design transparency, and pre-specification.", "long_description": "**What it is** — *Real-World Evidence: Considerations Regarding Non-Interventional Studies for Drug and Biological Products* is a draft guidance for industry issued by the U.S. FDA (CDER and CBER) on 20 March 2024 (Federal Register notice 2024-05969, published 21 March 2024; comment period closed 18 June 2024). It is one deliverable of FDA's Real-World Evidence Program, mandated by the 21st Century Cures Act and the 2016 FDA Reauthorization Act. Unlike an EQUATOR reporting checklist (STROBE, RECORD-PE) or a Cochrane/ISPOR appraisal instrument, this is an agency expectations document: it tells sponsors who intend to rely on a non-interventional study to support a regulatory decision *what FDA will look for* when judging whether the study can contribute to substantial evidence of effectiveness or to evidence of safety. It is maintained by FDA, not by a methods-guideline consortium, and it sits alongside two sibling FDA RWE guidances it must not be conflated with: the data-focused guidance on *Assessing Electronic Health Records and Medical Claims Data To Support Regulatory Decision-Making* (data reliability/relevance) and the *Considerations for the Use of Real-World Data and Real-World Evidence To Support Regulatory Decision-Making for Drug and Biological Products* framework. This guidance is the design-and-conduct layer for the specifically *non-interventional* case.\n\n**When to use** — Apply it whenever a sponsor (or a partner generating evidence for a sponsor) is planning, conducting, or documenting a non-interventional study — a cohort, case-control, or comparative-effectiveness study in claims, EHR, registry, or linked data — that is intended to contribute to a U.S. regulatory submission for a drug or biologic, whether for effectiveness or for safety. The decision rule for *which* guidance applies: (1) if the question is whether a given **data source** is fit for regulatory use, the EHR/claims data guidance governs; (2) if the study is **interventional or a hybrid/pragmatic trial** with a protocol-driven intervention, this guidance does not apply and the RCT-with-RWD or hybrid-design pathway does; (3) if the study is a **non-interventional comparison of treatments as used in routine care**, this guidance is the operative one. It is not a journal reporting checklist and not an HTA reference case — for EMA/ENCePP submissions or PASS, pair it with the ENCePP Guide and Methodological Standards; for an HTA dossier, pair it with the relevant HTA reference case; for journal publication, pair it with STROBE/RECORD-PE. Early engagement with FDA (e.g., a meeting before the protocol is finalized) is an explicit expectation, not an option, for studies meant to support effectiveness.\n\n**What it requires** — The guidance organizes its expectations around making a non-interventional study as transparent and pre-specified as the trial it stands in for. The substantive domains it enforces for real-world data are: (1) **Data fitness-for-use** — documented reliability (accrual, completeness, provenance, data-quality checks, conformance, plausibility) and relevance (availability of the exposure, outcome, covariates, and population needed to answer the question), with the supporting data guidance invoked by reference; (2) **Design transparency and pre-specification** — a finalized protocol and statistical analysis plan with date stamps and version control submitted *before* analysis, ideally before any analytic data access, so that design choices are not data-driven; (3) **Study design framed against a target trial** — explicit specification of eligibility, treatment strategies, assignment, time zero/index, follow-up, outcome, and the causal contrast, so the observational analysis emulates a hypothetical pragmatic trial; (4) **Time-zero alignment** — index-date definition that avoids immortal-time bias and prohibits using post-baseline information to define baseline; (5) **Estimands and intercurrent events** — a clearly stated estimand (target population, treatment condition, outcome, summary measure) and a pre-specified strategy for intercurrent events such as treatment switching, discontinuation, and death; (6) **Exposure, outcome, and covariate definitions** — validated operational/phenotype algorithms with reported performance (e.g., PPV) where outcomes are derived from coded data; (7) **Confounding control** — pre-specified covariates and methods (propensity-score or high-dimensional approaches, active-comparator/new-user framing) with diagnostics for balance and positivity; (8) **Missing data and attrition** — characterization of loss to follow-up and missingness with pre-specified handling; (9) **Sensitivity and quantitative bias analysis** — robustness checks and bias quantification (negative controls, E-values) to probe residual confounding and design assumptions; and (10) **Transparency and reproducibility** — versioned code lists, algorithms, and analytic programs that would allow FDA to reproduce the analysis.\n\n**When NOT to use — limitations and common misapplications** — This is a regulatory expectations document, not a scoring tool, and several failure modes recur. (1) *Treating it as a fit-for-purpose data assessment.* The guidance repeatedly defers to the EHR/claims data guidance for reliability and relevance; checking design boxes here does not establish that the underlying data can answer the question. (2) *Conflating it with its sibling guidances.* Using this guidance to justify a data source (that is the data guidance's job), or invoking it for an interventional hybrid or single-arm externally controlled trial (a different pathway), is a category error — name and route to the correct guidance. (3) *Checklist-as-theater.* A complete, well-formatted protocol does not make an observational comparison causal; unmeasured confounding, channeling, and an indefensible comparator are not cured by documentation. (4) *Assuming method endorsement.* The guidance does not bless any particular estimator (PS matching, IPTW, g-methods) — it requires justification and diagnostics, not adoption of a favorite method. (5) *Retrofitting pre-specification.* Submitting a protocol after the analysis has been run, or after iterating on results, defeats the central safeguard the guidance is built around; the date-stamped, pre-analysis protocol is the point. (6) *Using it where a different guideline is required.* For an EMA submission rely on the ENCePP standards; for journal reporting use STROBE/RECORD-PE; for an HTA dossier use the payer reference case — this FDA guidance is necessary but not sufficient for those audiences.\n\n**How it maps to this catalog** — Each requirement above is implemented by a concept in this repository. Data fitness-for-use → `fit-for-purpose-data-assessment-rwe` and `database-feasibility-attrition-funnel-rwe`, with payer-specific nuances in `medicare-ffs-ma-commercial-claims-differences-rwe` and source mechanics in `claims-analysis`. Target-trial framing → `target-trial-emulation`, with `clone-censor-weight-per-protocol` when the estimand is a sustained per-protocol strategy. Time-zero alignment → `time-zero-index-date-alignment-rwe` and `immortal-time-bias-handling`. Estimands and intercurrent events → `estimands-ate-att-intercurrent-events-rwe` and `estimand-analysis-traceability-rwe`. Phenotype/algorithm validation → `diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe`, `claims-outcome-algorithm-ppv-sensitivity-rwe`, and `ehr-phenotyping-algorithms-rwe`. Confounding control → `active-comparator-new-user`, `high-dimensional-propensity-score-hdps-rwe`, `propensity-score-methods-psm-iptw`, and `dags-backdoor-criterion-drug-studies`. Attrition and missing data → `attrition-and-loss-to-follow-up-rwe` and `missing-data-pattern-table-rwe`. Sensitivity and quantitative bias analysis → `e-value-sensitivity-analysis`, `negative-control-outcomes-rwe`, and `quantitative-bias-analysis-toolkit-rwe`. Question framing → `picots-framework-rwe`; overall regulatory packaging → `regulatory-readiness-rwe` and `study-protocol-or-sap-elements`. **Applied note (claims/EHR/registry):** in claims, the highest-yield FDA-facing artifacts are a versioned attrition funnel from source population to analytic cohort, a validated outcome algorithm with reported PPV, an active-comparator new-user design with continuous-enrollment requirements over the full lookback (excluding Medicare Advantage person-time where fee-for-service claims are absent), and a pre-specified estimand with an intercurrent-event strategy — all locked in a date-stamped protocol before analytic data access.", "primary_category": "Guideline", "tags": [ "guideline", "regulatory", "fda", "rwe", "non-interventional", "observational", "effectiveness", "safety" ], "aliases": [ "FDA RWE Non-Interventional", "FDA Non-Interventional Studies Guidance", "FDA Observational Studies RWE Guidance", "Real-World Evidence: Considerations Regarding Non-Interventional Studies" ], "applies_to_study_types": [ "cohort_retrospective", "cer_observational", "new_user", "active_comparator_new_user", "claims_analysis", "ehr_study" ], "data_sources": [ "claims", "ehr", "registry", "linked", "multi-database" ], "citations": [ { "role": "introduce", "doi": null, "url": "https://www.fda.gov/regulatory-information/search-fda-guidance-documents/real-world-evidence-considerations-regarding-non-interventional-studies-drug-and-biological-products", "citation_text": "U.S. Food and Drug Administration. Real-World Evidence: Considerations Regarding Non-Interventional Studies for Drug and Biological Products. Draft Guidance for Industry. CDER/CBER; March 2024.", "year": 2024, "authors_short": "FDA (CDER/CBER)", "notes": "Canonical agency source. Draft guidance issued 20 March 2024; Federal Register notice 2024-05969 (89 FR, published 21 March 2024); comment period closed 18 June 2024. No journal DOI exists for the guidance itself; the FDA guidance-document landing page is the stable reference." }, { "role": "explain", "doi": "10.1002/cpt.857", "url": "https://doi.org/10.1002/cpt.857", "citation_text": "Franklin JM, Schneeweiss S. When and How Can Real World Data Analyses Substitute for Randomized Controlled Trials? Clinical Pharmacology & Therapeutics. 2017;102(6):924-933.", "year": 2017, "authors_short": "Franklin & Schneeweiss", "notes": "Frames when non-interventional analyses can support effectiveness conclusions and the design/transparency conditions required — directly relevant to the substantial-evidence bar the guidance addresses." }, { "role": "use", "doi": "10.1093/aje/kwv254", "url": "https://doi.org/10.1093/aje/kwv254", "citation_text": "Hernán MA, Robins JM. Using Big Data to Emulate a Target Trial When a Randomized Trial Is Not Available. American Journal of Epidemiology. 2016;183(8):758-764.", "year": 2016, "authors_short": "Hernán & Robins", "notes": "The target-trial emulation framework that operationalizes the guidance's design-and-pre-specification expectations (eligibility, time zero, treatment strategies, estimand) for routinely collected data." } ], "relations": [ { "relation_type": "see_also", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "The guidance defers data reliability/relevance to FDA's EHR/claims data guidance; this concept implements that fitness-for-use assessment." }, { "relation_type": "used_with", "target_slug": "target-trial-emulation", "notes": "Implements the guidance's expectation that a non-interventional study be designed against an explicit hypothetical trial (eligibility, time zero, strategies, estimand)." }, { "relation_type": "used_with", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Implements the required statement of estimand and a pre-specified strategy for intercurrent events (switching, discontinuation, death)." }, { "relation_type": "used_with", "target_slug": "time-zero-index-date-alignment-rwe", "notes": "Implements time-zero/index-date definition that avoids immortal-time bias and post-baseline leakage into baseline." }, { "relation_type": "used_with", "target_slug": "active-comparator-new-user", "notes": "Standard design framing for the comparative non-interventional studies this guidance addresses." }, { "relation_type": "used_with", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "Confounding-control method with diagnostics for balance and positivity, as the guidance expects design choices to be justified rather than assumed." }, { "relation_type": "used_with", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Implements validated outcome/phenotype algorithms with reported performance (e.g., PPV) for coded data." }, { "relation_type": "used_with", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Implements characterization and pre-specified handling of attrition and missing data the guidance requires." }, { "relation_type": "used_with", "target_slug": "e-value-sensitivity-analysis", "notes": "Implements quantitative bias analysis / sensitivity checks for residual confounding." }, { "relation_type": "used_with", "target_slug": "claims-analysis", "notes": "Source-level mechanics (enrollment, NDC/diagnosis logic, payer nuances) underlying non-interventional claims studies under this guidance." }, { "relation_type": "see_also", "target_slug": "regulatory-readiness-rwe", "notes": "Overall packaging of a non-interventional study for regulatory review, including transparency and reproducibility expectations." }, { "relation_type": "see_also", "target_slug": "picots-framework-rwe", "notes": "Structures the research question (population, intervention, comparator, outcomes, timing, setting) the guidance expects to be pre-specified." }, { "relation_type": "applies_to", "target_slug": "cer-observational", "notes": "Comparative-effectiveness observational studies are the core use case for this guidance." }, { "relation_type": "applies_to", "target_slug": "cohort-retrospective", "notes": "Retrospective cohort designs in routinely collected data fall squarely under the guidance's scope." }, { "relation_type": "applies_to", "target_slug": "claims-analysis", "notes": "Claims-based non-interventional studies are a primary target of the guidance." }, { "relation_type": "applies_to", "target_slug": "ehr-study", "notes": "EHR-based non-interventional studies are a primary target of the guidance." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "hta" ] }, { "slug": "gpp", "name": "Guidelines for Good Pharmacoepidemiology Practices (GPP)", "short_definition": "ISPE's overarching good-practice guidance for the planning, conduct, analysis, archiving, and communication of pharmacoepidemiologic and non-interventional real-world studies. It is substantive practice guidance, not a line-by-line reporting checklist or a risk-of-bias instrument.", "long_description": "**What it is** — The **Guidelines for Good Pharmacoepidemiology Practices (GPP)** are the\nprofession-defining good-practice statement maintained by the **International Society for\nPharmacoepidemiology (ISPE)** through its Public Policy Committee. First issued in 1996 and\nrevised in 2004, 2007, and 2015 (the current \"fourth version\"), GPP is *substantive practice\nguidance* spanning the entire study lifecycle: protocol development, responsibilities and\nqualifications of personnel, study conduct, data-quality assurance, statistical analysis,\ndocumentation and **archiving**, privacy/ethics, conflict-of-interest disclosure, adverse-event\nreporting, and communication of results. It is deliberately design-agnostic — it governs cohort,\ncase-control, self-controlled, drug-utilization, and database studies alike — and sits alongside\nENCePP's *Guide on Methodological Standards* as one of the two foundational good-practice\nreferences for the field. GPP is *not* a reporting checklist (that is STROBE/RECORD-PE) and *not*\na risk-of-bias tool (that is ROBINS-I); it is the operating standard for *how the work is done*.\n\n**When to use** — Treat GPP as the default standard of conduct for any non-interventional or\nhybrid pharmacoepidemiologic study, regardless of data source or submission target: FDA or EMA\nregulatory submissions and post-authorization safety/effectiveness studies (PASS), HTA/payer\ndossiers, and peer-reviewed publication. Apply it from the moment a question is framed —\nbefore data access — to discipline protocol pre-specification, personnel qualifications, and a\ndata-management/archiving plan. **Decision rule for siblings:** use GPP for *overarching conduct\nand quality-system expectations*; reach for **ENCePP's Guide/Checklist** when the work is an\nEU-regulated PASS or you need the agency-facing methodological-standards mapping; reach for\n**HARPER / StaRT-RWE** when you need a structured *protocol template* with study-design diagrams\nand pre-specified parameter tables; and reach for **STROBE/RECORD-PE** at the *reporting* stage\nfor the manuscript checklist and attrition flow diagram. These are complementary, not\ninterchangeable: GPP tells you to pre-specify, archive, and quality-assure; the templates tell\nyou exactly which fields to fill; the reporting checklists tell you what to disclose in print.\n\n**What it requires** — GPP enforces good practice across domains that, for real-world data, map\ndirectly onto the hardest design decisions: (1) **design transparency** — an a-priori written\nprotocol stating objectives, the causal/descriptive question, design, and analysis plan, amended\nwith version control rather than rewritten; (2) **data fitness-for-use** — documented assessment\nof whether the source (claims, EHR, registry, linked) can capture the exposures, outcomes, and\ncovariates required, including provenance, completeness, lags, and validation; (3)\n**operational definitions and phenotype/algorithm validation** — explicit, reproducible exposure\nand outcome algorithms with reported performance (e.g., PPV) where feasible; (4) **time-zero\nalignment and follow-up rules** that avoid immortal time; (5) **estimands and the analytic\ncontrast**, with pre-specified handling of intercurrent events (switching, discontinuation,\ndeath); (6) **confounding control** strategy declared a priori; (7) **attrition and\nmissing-data** handling with a transparent flow from source to analytic cohort; (8)\n**sensitivity and quantitative bias analysis** to probe key assumptions; and (9) **documentation,\ncode/algorithm versioning, and long-term archiving** so the study is auditable and reproducible.\nGPP frames these as quality-system obligations — who is responsible, what is documented, and what\nis retained — rather than as a manuscript checklist.\n\n**When NOT to use — limitations and common misapplications** — The dominant reviewer-facing error\nis **category confusion**. GPP is *not* a reporting checklist: you cannot \"complete GPP\" item by\nitem to satisfy a journal's reporting requirements — that is STROBE or RECORD-PE, and submitting\na \"GPP checklist\" in place of a RECORD-PE flow diagram will be rejected. GPP is also *not* a\n**risk-of-bias / critical-appraisal instrument**: it does not score study quality the way\nROBINS-I rates confounding, selection, and measurement bias, so it cannot be cited as the\nappraisal tool in an evidence synthesis. Nor is GPP a **protocol template** — pointing to GPP\ndoes not relieve you of producing the structured design and parameter tables that HARPER /\nStaRT-RWE (or the ENCePP protocol shell for EU PASS) provide. **Adherence to GPP does not make a\nstudy causal or unbiased**: a beautifully documented, fully archived study can still be wrecked\nby confounding by indication, immortal-time bias, or an unvalidated outcome algorithm — GPP\ngoverns *process and transparency*, not the *validity* of any single design choice. Finally,\navoid **GPP-as-theater**: a generic statement that the study \"followed GPP\" with no protocol\nversion, no data-fitness assessment, no phenotype validation, and no archiving plan is the\nfailure mode senior regulatory and HTA reviewers flag first.\n\n**How it maps to this catalog** — GPP's good-practice domains are *implemented* by specific\nconcepts in this repository. Design transparency and pre-specification → **picots-framework-rwe**\nand **estimands-ate-att-intercurrent-events-rwe** (the estimand and intercurrent-event handling\nGPP demands a priori). Data fitness-for-use → **fit-for-purpose-data-assessment-rwe**, with\nsource-specific nuance in **claims-analysis** and **medicare-ffs-ma-commercial-claims-differences-rwe**.\nPhenotype/algorithm validation → **diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe**.\nTime-zero alignment that avoids immortal time → **time-zero-index-date-alignment-rwe** and the\ndrug-free lookback in **washout-clean-lookback-period-rwe**. Confounding control → the\n**active-comparator-new-user** design and **high-dimensional-propensity-score-hdps-rwe**, with\nthe trial-emulation scaffold in **target-trial-emulation**. Attrition/missing data → the\nsource-to-analytic flow in **attrition-and-loss-to-follow-up-rwe**. Sensitivity / quantitative\nbias analysis → **e-value-sensitivity-analysis** and **quantitative-bias-analysis-toolkit-rwe**.\nIn practice for a claims/EHR/registry study, satisfying GPP means: write the protocol and lock\nPICOTS and the estimand before pulling data; run a fit-for-purpose assessment of the source;\ndefine and (where possible) validate exposure and outcome phenotypes; set time zero at first\nqualifying fill after a documented washout; pre-specify the confounding-control approach; report\nthe full attrition funnel; run pre-planned sensitivity and bias analyses; and archive code,\ncode lists, and protocol versions so an auditor can reproduce the result.", "primary_category": "Guideline", "tags": [ "guideline", "good-practice", "pharmacoepidemiology", "rwe", "conduct-standard", "ispe" ], "aliases": [ "GPP", "ISPE GPP", "Good Pharmacoepidemiology Practices", "Guidelines for Good Pharmacoepidemiology Practice" ], "applies_to_study_types": [ "new_user", "active_comparator_new_user", "self_controlled_case_series", "case_crossover", "drug_utilization", "cohort_prospective", "cohort_retrospective", "case_control", "nested_case_control" ], "data_sources": [ "claims", "ehr", "registry", "linked", "multi-database", "primary" ], "citations": [ { "role": "introduce", "doi": "10.1002/pds.3891", "url": "https://doi.org/10.1002/pds.3891", "citation_text": "Public Policy Committee, International Society of Pharmacoepidemiology. Guidelines for good pharmacoepidemiology practice (GPP). Pharmacoepidemiology and Drug Safety. 2016;25(1):2-10.", "year": 2016, "authors_short": "ISPE Public Policy Committee", "notes": "Canonical statement of the current (2015 revision / fourth version) GPP, published by ISPE in Pharmacoepidemiology and Drug Safety. The authoritative reference defining the good-practice expectations across the study lifecycle." }, { "role": "use", "url": "https://www.pharmacoepi.org/resources/policies/guidelines-08027/", "citation_text": "International Society for Pharmacoepidemiology. Guidelines for Good Pharmacoepidemiology Practices (GPP), Revision 3 (June 2015; fourth version). ISPE.", "year": 2015, "authors_short": "ISPE", "notes": "Official ISPE landing page with the full text, revision history (1996, 2004, 2007, 2015), and the maintained good-practice document. Use as the stable primary reference." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "cohort-retrospective", "notes": "GPP governs conduct, documentation, and archiving for retrospective cohort studies in routinely collected data." }, { "relation_type": "applies_to", "target_slug": "cohort-prospective", "notes": "GPP governs conduct and quality-system expectations for prospective non-interventional cohort studies." }, { "relation_type": "applies_to", "target_slug": "active-comparator-new-user", "notes": "GPP frames the a-priori protocol and documentation expectations; the design itself is implemented in the active-comparator-new-user concept." }, { "relation_type": "applies_to", "target_slug": "case-control", "notes": "GPP applies to case-control (and nested case-control) studies as part of its design-agnostic conduct standard." }, { "relation_type": "applies_to", "target_slug": "drug-utilization", "notes": "GPP covers drug-utilization studies, including data-fitness and operational-definition expectations." }, { "relation_type": "applies_to", "target_slug": "self-controlled-case-series", "notes": "GPP's conduct and documentation standards apply to self-controlled designs." }, { "relation_type": "complements", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "Implements GPP's data fitness-for-use requirement — assessing whether the source can capture the needed exposures, outcomes, and covariates." }, { "relation_type": "complements", "target_slug": "picots-framework-rwe", "notes": "Implements GPP's a-priori specification of the question (population, intervention, comparator, outcomes, timing, setting)." }, { "relation_type": "complements", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Implements GPP's requirement to pre-specify the analytic contrast and the handling of intercurrent events." }, { "relation_type": "complements", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Implements GPP's operational-definition and phenotype/algorithm-validation requirement." }, { "relation_type": "complements", "target_slug": "time-zero-index-date-alignment-rwe", "notes": "Implements GPP's expectation of a justified time zero that avoids immortal-time bias." }, { "relation_type": "complements", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "Implements GPP's confounding-control expectations for high-dimensional claims/EHR data." }, { "relation_type": "complements", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Implements GPP's transparent source-to-analytic attrition reporting and loss-to-follow-up handling." }, { "relation_type": "complements", "target_slug": "quantitative-bias-analysis-toolkit-rwe", "notes": "Implements GPP's expectation to quantify the impact of key biases through sensitivity and quantitative bias analysis." }, { "relation_type": "see_also", "target_slug": "target-trial-emulation", "notes": "A trial-emulation scaffold operationalizes GPP's pre-specification discipline for causal real-world questions." }, { "relation_type": "see_also", "target_slug": "active-comparator-new-user", "notes": "The standard confounding-control design GPP-compliant comparative studies typically adopt." }, { "relation_type": "see_also", "target_slug": "claims-analysis", "notes": "Source-specific operational depth for executing a GPP-compliant claims study." }, { "relation_type": "see_also", "target_slug": "medicare-ffs-ma-commercial-claims-differences-rwe", "notes": "Payer-specific data-fitness nuances relevant to GPP's fit-for-use assessment." }, { "relation_type": "see_also", "target_slug": "e-value-sensitivity-analysis", "notes": "A concrete sensitivity/quantitative-bias tool supporting GPP's robustness expectations." }, { "relation_type": "see_also", "target_slug": "visualizations-pharmacoepidemiology-rwe", "notes": "Lifecycle and attrition visualizations supporting GPP-compliant documentation and reporting." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "hta", "journal" ] }, { "slug": "grace", "name": "GRACE (Good ReseArch for Comparative Effectiveness)", "short_definition": "A validated 11-item instrument for critically appraising the quality of non-randomized (observational) studies of comparative effectiveness, scoring whether the data and methods were good enough to support a comparative conclusion. It is an appraisal tool, not a reporting checklist and not a numeric quality score.", "long_description": "**What it is.** GRACE — **Good ReseArch for Comparative Effectiveness** — is a critical-appraisal\ninstrument for non-interventional (observational) comparative effectiveness research (CER). It was\ndeveloped by the GRACE Initiative (an academic-industry collaboration led by Nancy Dreyer and\ncolleagues, with roots at the Comparative Effectiveness Research Collaborative Initiative) to give\ndecision-makers a structured, *empirically validated* way to judge whether an observational CER study\nwas conducted and reported well enough to inform a treatment comparison. The current instrument is the\n**GRACE Checklist**: 11 yes/no/unclear items in two domains — **6 items on data quality** (were the\ndata adequate to study the treatments, outcomes, and key confounders?) and **5 items on methods** (was\nthe design and analysis sound enough to support a comparative claim?). What distinguishes GRACE from\nmost appraisal tools is that its items were tested against expert global-quality ratings and refined to\nretain only discriminating questions; the validation study reported roughly 71% sensitivity and 81%\nspecificity for separating higher- from lower-quality studies. It is publicly maintained (checklist,\nelements documents, and the GRACE Principles) and is freely available.\n\n**When to use.** Reach for GRACE when you must *appraise* a published or proposed observational\ncomparative-effectiveness or comparative-safety study — typically claims-, EHR-, registry-, or\nlinked-data-based — and need a defensible, reproducible quality judgment rather than a gut reaction.\nConcrete contexts: an HTA or payer evidence team grading the non-RCT evidence base in a dossier or\nvalue assessment; a systematic reviewer or guideline panel weighting observational CER studies; an\ninternal evidence-quality gate before a real-world study is cited in a regulatory or reimbursement\nsubmission; a peer reviewer or methods editor assessing a CER manuscript. Decision rule for picking the\nright tool: use **GRACE** when the question is \"is this *comparative-effectiveness observational* study\ngood enough to believe?\"; use **ROBINS-I** when you need a formal, signalling-question risk-of-bias\nassessment mapped to a target trial (e.g., for a Cochrane review or GRADE certainty downgrade); use\n**STROBE / RECORD / RECORD-PE / HARPER** when the task is *reporting completeness* of an observational\nor routinely-collected-data study, not appraisal. GRACE is deliberately lightweight and CER-specific,\nwhich is its strength for fast, comparable triage and its limit for fine-grained bias attribution.\n\n**What it requires.** The two GRACE domains, read through a real-world-data lens, enforce the\nsubstantive questions a senior reviewer asks. *Data domain:* whether the data were adequate to capture\nthe **treatments/exposures** with enough detail (timing, dose, switching) — in claims this is fill- and\nNDC-level exposure construction with continuous-enrollment observability; whether **outcomes** were\nmeasured with acceptable validity — i.e., whether the **phenotype/algorithm was validated** (PPV,\nsensitivity) rather than assumed; and whether the data captured the **key confounders and effect\nmodifiers** needed for the comparison. *Methods domain:* whether **comparison groups** were concurrent\nand appropriately defined (an active-comparator, new-user structure with aligned **time zero** rather\nthan prevalent users or immortal time); whether the **analysis controlled confounding** credibly\n(propensity-score or multivariable methods, with attention to unmeasured confounding); whether\n**classification of exposure/outcome was independent** of the comparison; whether **attrition and\nfollow-up** were handled and reported; and whether the investigators ran **sensitivity / quantitative\nbias analyses** to test robustness. Implicit throughout is fitness-for-use of the data source and a\npre-specified, transparent design — the same disciplines a regulator or HTA reviewer expects.\n\n**When NOT to use — limitations and common misapplications.** GRACE is a *quality-appraisal* tool, not\na reporting checklist: do not hand authors GRACE as a writing template (use STROBE/RECORD for that),\nand do not treat a completed GRACE form as evidence the study was *reported* completely. It is also\n**not a numeric quality score**: the items are diagnostic prompts, and tallying \"yes\" answers into a\ncut-off or pooling them as a weight in a meta-analysis is exactly the kind of quality-scoring practice\nmethodologists warn against — GRACE supports a structured *judgment*, not arithmetic. It is scoped to\n**comparative-effectiveness observational designs**; applying it to a single-arm descriptive study, a\ndiagnostic-accuracy study, an RCT, or a systematic review is a category error (use ROBINS-I, QUADAS-2,\nRoB 2, or AMSTAR 2 respectively). A high GRACE rating does **not** certify causality — a study can pass\nevery item and still be confounded by an unmeasured factor; GRACE checks that the right defenses were\nattempted, not that bias was eliminated. Other failure modes: \"checklist-as-theater,\" where boxes are\nticked without engaging the underlying methods; using GRACE where a formal target-trial / ROBINS-I\nbias assessment is required for GRADE certainty rating; and over-reliance on its modest sensitivity —\nGRACE triages, it does not adjudicate, and borderline studies still need expert methods review.\n\n**How it maps to this catalog.** Each GRACE item points to a concept here that *implements* the thing\nGRACE only asks about. Comparison-group adequacy and aligned time zero are operationalized by\n[active-comparator-new-user](active-comparator-new-user) and\n[time-zero-index-date-alignment-rwe](time-zero-index-date-alignment-rwe), with\n[immortal-time-bias-handling](immortal-time-bias-handling) for the classic follow-up trap. Outcome and\nexposure validity — the heart of the GRACE data domain in claims/EHR — are implemented by\n[diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe](diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe)\nand [claims-outcome-algorithm-ppv-sensitivity-rwe](claims-outcome-algorithm-ppv-sensitivity-rwe).\nConfounding control maps to [propensity-score-methods-psm-iptw](propensity-score-methods-psm-iptw) and\n[high-dimensional-propensity-score-hdps-rwe](high-dimensional-propensity-score-hdps-rwe), with\n[e-value-sensitivity-analysis](e-value-sensitivity-analysis) and\n[quantitative-bias-analysis-toolkit-rwe](quantitative-bias-analysis-toolkit-rwe) covering the\nsensitivity / quantitative-bias-analysis item. The comparative estimand the study is actually\ndefending is made explicit by\n[estimands-ate-att-intercurrent-events-rwe](estimands-ate-att-intercurrent-events-rwe);\nattrition/follow-up by [attrition-and-loss-to-follow-up-rwe](attrition-and-loss-to-follow-up-rwe); and\ndata-source fitness by [fit-for-purpose-data-assessment-rwe](fit-for-purpose-data-assessment-rwe) and\n[claims-analysis](claims-analysis). The aspirational benchmark behind a strong GRACE rating is a\nwell-specified [target-trial-emulation](target-trial-emulation). **Applied note (claims/EHR/registry\nRWE):** when appraising a claims study, do not accept a \"yes\" on the outcome-data item unless the\nauthors cite a validated algorithm with PPV/sensitivity in a comparable population; treat an\nunvalidated diagnosis-code definition, a prevalent-user comparison, or an absent negative-control /\nsensitivity analysis as fatal weaknesses even if every other box is ticked.", "primary_category": "Guideline", "tags": [ "guideline", "critical-appraisal", "quality-assessment", "comparative-effectiveness", "observational-studies", "real-world-evidence" ], "aliases": [ "GRACE", "GRACE Checklist", "GRACE Principles", "Good ReseArch for Comparative Effectiveness", "Good Research for Comparative Effectiveness Checklist" ], "applies_to_study_types": [ "cer_observational", "active_comparator_new_user", "new_user", "cohort_retrospective" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "doi": "10.18553/jmcp.2014.20.3.301", "url": "https://doi.org/10.18553/jmcp.2014.20.3.301", "citation_text": "Dreyer NA, Velentgas P, Westrich K, Dubois R. The GRACE Checklist for Rating the Quality of Observational Studies of Comparative Effectiveness: A Tale of Hope and Caution. Journal of Managed Care Pharmacy. 2014;20(3):301-308.", "year": 2014, "authors_short": "Dreyer et al.", "notes": "Statement paper introducing the 11-item GRACE Checklist (6 data-quality + 5 methods items) and its intended use for appraising observational comparative-effectiveness studies." }, { "role": "explain", "doi": "10.18553/jmcp.2016.22.10.1107", "url": "https://doi.org/10.18553/jmcp.2016.22.10.1107", "citation_text": "Dreyer NA, Bryant A, Velentgas P. The GRACE Checklist: A Validated Assessment Tool for High Quality Observational Studies of Comparative Effectiveness. Journal of Managed Care & Specialty Pharmacy. 2016;22(10):1107-1113.", "year": 2016, "authors_short": "Dreyer et al.", "notes": "Empirical validation against expert global-quality ratings; reports the discriminating items and the approximately 71% sensitivity / 81% specificity performance that underpins the checklist." }, { "role": "use", "url": "https://www.latitudes-network.org/tool/grace-checklist/", "citation_text": "GRACE Checklist — maintained library entry and tool resources (Latitudes Network), linking the GRACE Checklist for appraising observational comparative-effectiveness studies.", "year": 2016, "authors_short": "Latitudes Network", "notes": "Maintained public reference for the checklist; confirms GRACE is a quality-appraisal tool, not a domain-by-domain risk-of-bias instrument." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "cer-observational", "notes": "Core scope — GRACE appraises non-randomized comparative-effectiveness studies." }, { "relation_type": "applies_to", "target_slug": "active-comparator-new-user", "notes": "The comparison-group and time-zero items reward exactly the active-comparator, new-user structure this concept implements." }, { "relation_type": "see_also", "target_slug": "target-trial-emulation", "notes": "A well-specified target-trial emulation is the design that most reliably earns strong marks on the GRACE methods domain." }, { "relation_type": "used_with", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Implements the GRACE outcome-/exposure-data-quality items; a validated phenotype with PPV is what a \"yes\" on those items should require." }, { "relation_type": "used_with", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "Implements the GRACE confounding-control methods item for claims/EHR data." }, { "relation_type": "used_with", "target_slug": "quantitative-bias-analysis-toolkit-rwe", "notes": "Implements the GRACE sensitivity/robustness item via quantitative bias analysis and negative-control checks." }, { "relation_type": "see_also", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Supports the GRACE follow-up/attrition item." }, { "relation_type": "see_also", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Makes explicit the comparative estimand a high-quality CER study must define and defend." }, { "relation_type": "see_also", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "Underpins the GRACE data-domain question of whether the source is adequate to study the treatments, outcomes, and confounders." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "hta", "journal" ] }, { "slug": "grade", "name": "GRADE", "short_definition": "A structured framework for rating the certainty (quality) of a body of evidence per outcome and the strength of the resulting recommendations, used by guideline developers and HTA bodies. Evidence is rated High, Moderate, Low, or Very Low; recommendations are rated Strong or Conditional.", "long_description": "**What it is.** **GRADE — Grading of Recommendations Assessment, Development and Evaluation** is the dominant\nframework for two distinct tasks in evidence-based guideline development: (1) rating the **certainty** (formerly\n\"quality\") of a *body of evidence* for each outcome of interest, and (2) moving from that evidence to a graded\n**strength of recommendation**. It is maintained by the **GRADE Working Group**, an open international collaboration\nof guideline developers, methodologists, and clinicians, and is endorsed by Cochrane, the WHO, NICE, and dozens of\nprofessional societies and journals. GRADE rates certainty on four ordered levels — **High, Moderate, Low, Very\nLow** — reflecting confidence that the estimated effect is close to the true effect. Recommendations are rated as\n**Strong** or **Conditional (weak)**. Critically, GRADE does **not** rate individual studies; it rates the\n*aggregate* evidence per outcome, consuming study-level risk-of-bias judgments (from ROBINS-I, RoB 2, etc.) as\ninputs. The framework's two reference artifacts are the **Evidence Profile / Summary-of-Findings (SoF) table** and,\nfor recommendations, the **Evidence-to-Decision (EtD) framework**.\n\n**When to use.** Use GRADE when you are synthesizing a *body* of evidence to inform a clinical practice guideline,\na Cochrane or other systematic review, an HTA appraisal, or a payer/formulary recommendation — i.e., decision\ncontexts where the unit of judgment is \"how confident are we in the effect on this outcome, across all the\nevidence?\" It applies to systematic reviews and meta-analyses (of RCTs, of observational studies, or mixed),\nnetwork meta-analyses (via the NMA-specific GRADE guidance), and to the recommendations that guideline panels\nderive from them. Decision rule for *which* GRADE you apply: standard GRADE covers intervention effects on\npatient-important outcomes; questions of **diagnostic test accuracy** require the GRADE diagnostic extension, and\n**prognosis** questions require the prognosis adaptation — using core GRADE on these without the extension is a\nmisapplication. GRADE is the right tool for HTA dossiers and journal-mandated certainty assessment; it is **not** a\nregulatory submission instrument (FDA and EMA do not adjudicate marketing decisions on GRADE), and it is not a\nreporting checklist (use PRISMA for the synthesis report, CHEERS for economic evaluations).\n\n**What it requires.** GRADE enforces an explicit, domain-by-domain, auditable rating for each outcome:\n- **Initial certainty by design.** A body of RCT evidence starts at **High**; a body of observational (including\n real-world data) evidence starts at **Low**.\n- **Five downgrading domains.** Rate down for (1) **risk of bias** (limitations in the contributing studies — for\n RWE this is where confounding control, time-zero alignment, and exposure/outcome misclassification enter), (2)\n **inconsistency** (unexplained heterogeneity across studies), (3) **indirectness** (mismatch of population,\n intervention, comparator, or outcome — including surrogate vs patient-important endpoints — to the decision\n question), (4) **imprecision** (wide confidence intervals around the pooled estimate; optimal-information-size\n considerations), and (5) **publication bias**.\n- **Three upgrading domains (observational evidence only).** Rate up for a **large magnitude of effect**, a\n **dose-response gradient**, or **plausible residual confounding that would bias toward the null** (i.e.,\n confounding that, if present, would shrink rather than inflate the observed effect).\n- **Final certainty** is one of High / Moderate / Low / Very Low, with each up- or down-grade documented and\n justified in the Evidence Profile / SoF table.\n- **Strength of recommendation** is determined *separately* through the **EtD framework**, weighing the balance of\n desirable and undesirable effects, the certainty of evidence, patients' values and preferences, resource use and\n cost-effectiveness, equity, acceptability, and feasibility. Strong vs Conditional reflects how confident the\n panel is that the desirable consequences outweigh the undesirable ones for most patients.\n\n**When NOT to use — limitations and common misapplications.**\n- **Rating a single study.** GRADE rates a body of evidence per outcome. Applying the four certainty levels to one\n cohort study is a category error; for a single study's internal validity, use a risk-of-bias tool (ROBINS-I for\n non-randomized, RoB 2 for randomized).\n- **GRADE is not a risk-of-bias instrument.** Study-level bias is an *input* (the \"risk of bias\" downgrade domain),\n assessed with ROBINS-I/RoB 2; GRADE does not itself appraise individual studies.\n- **Conflating certainty with strength of recommendation.** They are separate axes and can legitimately diverge — a\n *strong* recommendation on *low-certainty* evidence is valid in defined \"discordant\" situations (e.g., life-saving\n intervention with little downside). Treating certainty as if it dictates recommendation strength misreads the\n framework.\n- **Wrong question type without the extension.** Diagnostic-accuracy and prognosis questions need the dedicated\n GRADE adaptations; core GRADE on these produces invalid ratings.\n- **Checklist theater.** Assigning \"Low certainty — observational\" by default, or ticking domains without explicit,\n transparent justification for each up/down move, defeats the purpose. The Evidence Profile must show the reasoning.\n- **Treating GRADE as a reporting or RWD-design guideline.** Phenotype validation, immortal-time avoidance, and\n estimand specification are study-conduct matters; they *feed* GRADE's risk-of-bias and indirectness judgments but\n are governed by STROBE/RECORD-PE, HARPER/STaRT-RWE, and the like — not by GRADE.\n\n**How it maps to this catalog.** In an RWE evidence synthesis, the catalog's design and analysis concepts are not\n*implemented by* GRADE — they are the **inputs** that move GRADE's domains for an observational evidence body:\n- **Risk-of-bias domain.** A rigorous **active-comparator, new-user** design (active-comparator-new-user),\n **target-trial emulation** (target-trial-emulation), **high-dimensional propensity scores**\n (high-dimensional-propensity-score-hdps-rwe), and explicit immortal-time handling materially reduce the risk-of-bias\n concern — strong execution can justify *not* downgrading further, and in rare cases supports the\n \"plausible-confounding-toward-the-null\" upgrade.\n- **Indirectness domain.** **Phenotype/algorithm validation** (diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe)\n and **fit-for-purpose data assessment** determine whether the measured construct matches the PICO outcome;\n **estimands and intercurrent-event handling** (estimands-ate-att-intercurrent-events-rwe) determine whether the\n quantity estimated matches the decision question. Misalignment here is an indirectness downgrade.\n- **Plausible-confounding upgrade / sensitivity.** **Quantitative bias analysis and the E-value** speak directly to\n the \"plausible residual confounding would bias toward the null\" upgrade criterion and to how seriously to take the\n risk-of-bias downgrade.\n- **Source-data context.** Differences across claims, EHR, and registry data (e.g., **claims-analysis**, Medicare\n FFS vs MA coding intensity) affect outcome capture and transportability, again feeding the risk-of-bias and\n indirectness judgments.\n\n**Applied note (claims/EHR/registry RWE).** When a guideline panel grades an outcome supported only by claims- or\nEHR-based comparative studies, the body starts at **Low**. Document, per study, the confounding-control strategy and\nvalidation of the exposure/outcome phenotypes; decide the **risk-of-bias** downgrade on that basis (a well-executed\nACNU target-trial emulation with validated phenotypes and negative-control diagnostics is a different animal from a\nprevalent-user, drug-vs-non-user claims analysis). Use a **surrogate vs patient-important outcome** check to set the\n**indirectness** rating, and bring an **E-value or formal bias analysis** to bear on whether residual confounding\ncould plausibly explain — or, conversely, could only attenuate — the effect. Record every move in the Evidence\nProfile so the rating is reproducible and defensible to an HTA reviewer.", "primary_category": "Guideline", "tags": [ "guideline", "grade", "certainty-of-evidence", "strength-of-recommendation", "evidence-synthesis", "hta", "guideline-development" ], "aliases": [ "GRADE", "Grading of Recommendations Assessment, Development and Evaluation", "GRADE approach" ], "applies_to_study_types": [ "systematic_review", "meta_analysis_rct", "meta_analysis_obs", "network_meta_analysis" ], "data_sources": [ "claims", "ehr", "registry", "linked", "primary" ], "citations": [ { "role": "introduce", "doi": "10.1136/bmj.39489.470347.AD", "url": "https://doi.org/10.1136/bmj.39489.470347.AD", "citation_text": "Guyatt GH, Oxman AD, Vist GE, et al. GRADE: an emerging consensus on rating quality of evidence and strength of recommendations. BMJ. 2008;336(7650):924-926.", "year": 2008, "authors_short": "Guyatt et al.", "notes": "Canonical consensus statement introducing the GRADE approach — four certainty levels and the separation of evidence certainty from recommendation strength." }, { "role": "explain", "doi": "10.1016/j.jclinepi.2010.04.026", "url": "https://doi.org/10.1016/j.jclinepi.2010.04.026", "citation_text": "Guyatt GH, Oxman AD, Schünemann HJ, et al. GRADE guidelines: 1. Introduction — GRADE evidence profiles and summary of findings tables. Journal of Clinical Epidemiology. 2011;64(4):383-394.", "year": 2011, "authors_short": "Guyatt et al.", "notes": "Opening article of the 20+ part GRADE guidelines series; defines evidence profiles, summary-of-findings tables, and the operational mechanics of the up/down-grading domains." }, { "role": "use", "doi": "10.1136/bmj.i2016", "url": "https://doi.org/10.1136/bmj.i2016", "citation_text": "Alonso-Coello P, Schünemann HJ, Moberg J, et al. GRADE Evidence to Decision (EtD) frameworks: a systematic and transparent approach to making well informed healthcare choices. 1: Introduction. BMJ. 2016;353:i2016.", "year": 2016, "authors_short": "Alonso-Coello et al.", "notes": "Defines the Evidence-to-Decision framework used to move from certainty of evidence to a Strong or Conditional recommendation in clinical guidelines and HTA/coverage decisions." }, { "role": "use", "url": "https://www.gradeworkinggroup.org/", "citation_text": "GRADE Working Group — official handbook, GRADEpro GDT software, and maintained guidance series.", "year": 2024, "authors_short": "GRADE Working Group", "notes": "Authoritative maintained resource, including the GRADE Handbook and extensions for diagnosis, prognosis, and network meta-analysis." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "systematic-review", "notes": "GRADE rates the certainty of evidence per outcome within a systematic review and feeds the summary-of-findings table." }, { "relation_type": "applies_to", "target_slug": "meta-analysis-rct", "notes": "A pooled RCT body starts at High certainty; the five downgrade domains are applied to the meta-analytic estimate." }, { "relation_type": "applies_to", "target_slug": "meta-analysis-obs", "notes": "A pooled observational body starts at Low certainty; both downgrade and the three observational-only upgrade domains apply." }, { "relation_type": "applies_to", "target_slug": "network-meta-analysis", "notes": "GRADE has dedicated NMA guidance for rating certainty of direct, indirect, and network estimates." }, { "relation_type": "see_also", "target_slug": "target-trial-emulation", "notes": "A well-specified target-trial emulation strengthens the risk-of-bias judgment for an observational evidence body, reducing the GRADE downgrade." }, { "relation_type": "see_also", "target_slug": "active-comparator-new-user", "notes": "ACNU design choices (incident users, active comparator, time-zero alignment) are inputs to GRADE's risk-of-bias domain for RWE." }, { "relation_type": "see_also", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "Rigorous confounding control informs the risk-of-bias judgment and can support not downgrading further." }, { "relation_type": "see_also", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Whether the estimand matches the decision question feeds GRADE's indirectness domain." }, { "relation_type": "see_also", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Phenotype/algorithm validity (PPV, sensitivity) informs the indirectness and risk-of-bias domains for the measured outcome." }, { "relation_type": "see_also", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Differential attrition in contributing RWE studies is a risk-of-bias consideration when grading the body of evidence." }, { "relation_type": "see_also", "target_slug": "claims-analysis", "notes": "Source-data characteristics (capture, coding intensity, transportability) feed the risk-of-bias and indirectness judgments for claims-based evidence." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "hta", "journal" ] }, { "slug": "gramms", "name": "GRAMMS (Good Reporting of A Mixed Methods Study)", "short_definition": "A six-item reporting checklist for mixed-methods health-services research, requiring authors to justify the mixed-methods approach, describe the design and each strand's methods, and report how the quantitative and qualitative strands were integrated, their limitations, and the insights that integration produced.", "long_description": "**What it is.** **GRAMMS — Good Reporting of A Mixed Methods Study** — is a six-item reporting checklist introduced by\nO'Cathain, Murphy, and Nicholl (J Health Serv Res Policy, 2008) and listed on the EQUATOR Network. It was derived from a\nstructured review of mixed-methods studies in health services research that found their reporting was frequently incomplete,\nparticularly around *why* the methods were mixed and *how* the strands were brought together. GRAMMS asks authors to (1)\njustify the mixed-methods design (what the combination adds over a single method); (2) describe the design — its priority\n(which strand dominates), sequence (concurrent vs sequential), and purpose; (3) describe each method (sampling, data\ncollection, and analysis) for the quantitative and qualitative strands separately; (4) describe where, how, and by whom\n**integration** occurred; (5) describe any limitation associated with one component arising from the *presence* of the other\n(e.g., a reduced sample because of dual data collection); and (6) describe the insights gained by mixing or integrating the\nmethods. It is a *reporting* tool — a transparency checklist for the manuscript — not a critical-appraisal instrument, a\nrisk-of-bias tool, or a numeric quality score. It is maintained as a community resource via EQUATOR rather than through a\nformal versioned update process.\n\n**When to use.** Use GRAMMS when the deliverable is a **mixed-methods study report or protocol** that genuinely combines a\nquantitative and a qualitative strand — for example, a real-world effectiveness or health-services evaluation paired with\npatient or clinician interviews, a survey nested in a registry analysis, or a process evaluation accompanying an\nobservational comparative-effectiveness study. The natural decision contexts are **peer-reviewed journal submission** and,\nsecondarily, an **HTA/payer dossier** where qualitative patient-experience or implementation evidence is integrated with\nquantitative outcomes to inform value. Decision rule for picking GRAMMS over a sibling guideline: if the study is *purely*\nobservational quantitative RWE (claims/EHR cohort, comparative effectiveness), reach for STROBE/RECORD-PE or HARPER, not\nGRAMMS; if the study is *purely* qualitative, reach for COREQ or SRQR; choose GRAMMS only when integration of two strands is\nitself a reported object of the study. GRAMMS is best used **alongside** the strand-specific guideline for each component\n(e.g., GRAMMS + RECORD-PE for the quantitative arm + COREQ for the interviews), because GRAMMS deliberately does not specify\nhow to report the internals of either strand in depth.\n\n**What it requires.** GRAMMS enforces six reporting domains, and only some have analogs in quantitative RWE practice:\n- **Design transparency** — the justification, priority, sequence, and purpose of the mixed-methods design must be explicit.\n This is the domain that overlaps most with RWE design-transparency expectations (pre-specified design, clear research\n question).\n- **Per-strand methods** — for the *quantitative* strand specifically, the catalog's substantive RWE standards apply: data\n fitness-for-use, phenotype/algorithm definition, time-zero alignment, the estimand and intercurrent events, confounding\n control, and attrition/missing data. GRAMMS itself states only that each strand's methods be described; it does not\n enumerate these RWE-specific items, so a credible report borrows them from the appropriate quantitative guideline.\n- **Integration** — the distinctive GRAMMS requirement: report where (design, methods, or interpretation level), how\n (e.g., triangulation, following a thread, a joint display), and by whom integration was performed. This domain has **no\n counterpart** among the quantitative methods in this catalog.\n- **Strand-interaction limitations and integration insights** — what one strand cost or contributed to the other, and what\n was learned that neither strand could have produced alone.\n\n**When NOT to use — limitations and common misapplications.**\n- **Treating GRAMMS as a risk-of-bias or quality-scoring tool.** It is a reporting checklist; a fully GRAMMS-compliant\n paper can still describe a biased, underpowered, or poorly integrated study. Completeness of reporting is not validity.\n For appraisal of the quantitative strand use ROBINS-I; do not tally GRAMMS items into a \"quality score.\"\n- **Believing the checklist makes the study causal or rigorous.** Reporting that integration occurred does not make the\n quantitative strand confounding-controlled; the design must earn that separately.\n- **Wrong guideline for the design.** Using GRAMMS for a single-method observational study (where STROBE/RECORD-PE belongs),\n or using STROBE alone for a study whose central claim rests on integrating qualitative and quantitative findings (where\n the integration items GRAMMS polices would otherwise go unreported).\n- **Checklist-as-theater.** The most common failure GRAMMS was designed to catch is naming a study \"mixed methods\" while\n reporting two parallel strands that are never actually integrated — item 4 (integration) and item 6 (insights from mixing)\n exist precisely to expose this, and submitting a checklist with those items hand-waved defeats the purpose.\n\n**How it maps to this catalog.** GRAMMS sits one level above the quantitative methods catalog: it governs the *report\nstructure* of a mixed-methods study, and its quantitative strand must still satisfy this catalog's design standards. Map\nonly the items that genuinely overlap, and treat the integration/qualitative items as out of catalog scope (a finding in\nitself): the **per-strand methods** requirement for the quantitative arm is implemented by `active-comparator-new-user`\n(defensible design and time-zero alignment), `diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe` (phenotype/algorithm\nvalidation), `estimands-ate-att-intercurrent-events-rwe` (estimand and intercurrent events), `high-dimensional-propensity-score-hdps-rwe`\n(confounding control), and `attrition-and-loss-to-follow-up-rwe` (attrition/missing data); the **design-transparency** item\nis reinforced by `target-trial-emulation` (pre-specification discipline); and `claims-analysis` carries the data-fitness\nconsiderations for a claims/EHR/registry quantitative strand. GRAMMS' **integration**, **strand-interaction limitation**,\nand **integration-insight** items have no implementing concept here because the catalog scopes quantitative RWE only —\nauthors should pair GRAMMS with COREQ/SRQR for the qualitative strand and report integration per GRAMMS directly.\n\n**Applied note (claims/EHR/registry RWE).** A typical use is a sequential explanatory design: a claims-based active-comparator\nnew-user cohort estimates a comparative safety effect, then clinician interviews explain an unexpected channeling pattern.\nGRAMMS forces the report to state that the quantitative finding *drove* the qualitative sampling (integration at the methods\nlevel), to acknowledge that recruiting interviewees from the cohort narrowed the quantitative window (strand-interaction\nlimitation), and to articulate the insight (e.g., prescribing rationale that residual-confounding diagnostics could not\nsurface). The cohort itself is still held to `active-comparator-new-user` and `attrition-and-loss-to-follow-up-rwe`\nstandards — GRAMMS does not relax them.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "mixed-methods", "equator", "health-services-research", "integration" ], "aliases": [ "GRAMMS", "Good Reporting of A Mixed Methods Study" ], "applies_to_study_types": [ "mixed_methods" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "doi": "10.1258/jhsrp.2007.007074", "url": "https://doi.org/10.1258/jhsrp.2007.007074", "citation_text": "O'Cathain A, Murphy E, Nicholl J. The quality of mixed methods studies in health services research. Journal of Health Services Research & Policy. 2008;13(2):92-98.", "year": 2008, "authors_short": "O'Cathain et al.", "notes": "Canonical paper introducing the six-item GRAMMS reporting checklist, derived from a structured review of mixed-methods health-services research." }, { "role": "explain", "doi": "10.1136/bmj.c4587", "url": "https://doi.org/10.1136/bmj.c4587", "citation_text": "O'Cathain A, Murphy E, Nicholl J. Three techniques for integrating data in mixed methods studies. BMJ. 2010;341:c4587.", "year": 2010, "authors_short": "O'Cathain et al.", "notes": "Elaborates the integration item at the heart of GRAMMS (triangulation, following a thread, mixed-methods matrix/joint displays), the domain authors most often under-report." }, { "role": "use", "url": "https://www.equator-network.org/reporting-guidelines/the-quality-of-mixed-methods-studies-in-health-services-research/", "citation_text": "GRAMMS (Good Reporting of A Mixed Methods Study), EQUATOR Network — maintained checklist landing page.", "year": 2008, "authors_short": "EQUATOR Network", "notes": "Stable EQUATOR reference for the checklist and related mixed-methods reporting resources." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "mixed-methods", "notes": "Use when reporting or protocolizing a study that integrates a quantitative and a qualitative strand." }, { "relation_type": "see_also", "target_slug": "target-trial-emulation", "notes": "Pre-specification and design-transparency discipline supports the GRAMMS design-justification item for the quantitative strand." }, { "relation_type": "used_with", "target_slug": "active-comparator-new-user", "notes": "When the quantitative strand is a comparative cohort, GRAMMS' per-strand methods item is satisfied by a defensible ACNU design with aligned time zero." }, { "relation_type": "used_with", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Implements phenotype/algorithm definition and validation for the quantitative strand's outcomes and covariates." }, { "relation_type": "used_with", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Specifies the estimand and intercurrent-event handling that the quantitative strand must report under GRAMMS' methods item." }, { "relation_type": "used_with", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "Confounding control for the quantitative strand; GRAMMS reports that it was done, not how to do it." }, { "relation_type": "used_with", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Attrition and missing-data reporting for the quantitative strand and for the strand-interaction limitation item." }, { "relation_type": "used_with", "target_slug": "claims-analysis", "notes": "Data fitness-for-use for a claims/EHR/registry quantitative strand within a mixed-methods study." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "journal", "hta" ] }, { "slug": "gvp-module-viii", "name": "GVP Module VIII: Post-Authorisation Safety Studies", "short_definition": "EMA Good Pharmacovigilance Practices Module VIII (Rev. 3) — the regulatory framework governing the design, protocol format, registration, conduct, and PRAC reporting of post-authorisation safety studies (PASS), including non-interventional studies built on routinely collected real-world data.", "long_description": "**What it is.** GVP Module VIII — *Guideline on good pharmacovigilance practices (GVP): Module VIII —\nPost-authorisation safety studies (Rev. 3)* — is the European Medicines Agency (EMA) framework that governs\npost-authorisation safety studies (PASS). It is a binding *regulatory* guideline maintained by EMA in\ncollaboration with national competent authorities, operating under the EU pharmacovigilance legislation\n(Directive 2001/83/EC and Regulation (EC) No 726/2004). Revision 3 took legal effect on 13 October 2017 and is\naccompanied by *Addendum I* (Rev. 3, 2020), which sets the technical requirements for submitting information on\nnon-interventional PASS. Module VIII is a *process-and-conduct* framework, not a journal reporting checklist: it\ndefines what a PASS is, when it is imposed versus voluntary, what the protocol must contain (using the ENCePP\nprotocol structure), how and when protocols and results are submitted to the Pharmacovigilance Risk Assessment\nCommittee (PRAC), the role of the EU PAS Register, and the oversight, abstract, progress-report, and final-report\nobligations across the study lifecycle.\n\n**When to use.** Apply Module VIII whenever a study's *primary aim* is to evaluate the safety of an authorised\nmedicinal product in the EU — quantifying a risk, characterising a known risk, confirming the safety profile, or\nmeasuring the effectiveness of a risk-minimisation measure — and the study is conducted, sponsored, or imposed in\nthe EU regulatory context. The decision rules that select Module VIII (versus a sibling guideline) are: (1)\n*Imposed PASS* — a category-1 or category-2 obligation of the marketing authorisation (Art. 9, 21a, 22a, 104(2)) — falls fully under\nModule VIII with mandatory PRAC protocol endorsement before the study starts; (2) *voluntary PASS* initiated by\nthe marketing authorisation holder follows Module VIII conduct and EU PAS Register registration but with lighter\nsubmission obligations; (3) any *non-interventional PASS* using secondary data additionally invokes the ENCePP\nCode of Conduct and the ENCePP Checklist for study protocols. Module VIII is the *regulatory wrapper*; it does not\nreplace journal reporting guidelines (STROBE/RECORD-PE) or protocol-templating tools (HARPER, STaRT-RWE), which\nare used *inside* a Module VIII submission. If the study is interventional (a clinical trial), Module VIII does\nnot apply — the Clinical Trials Regulation governs instead. For a parallel FDA submission, the FDA RWE/RWD\nguidances run alongside, not instead of, Module VIII.\n\n**What it requires.** Module VIII enforces a structured protocol and lifecycle. The protocol — following the\nENCePP template — must pre-specify the research question and objectives; the study design and rationale; the data\nsource(s) and an explicit assessment of their *fitness for use* (coverage, completeness, validity, lag, recording\npractices); the study population with eligibility, *time-zero/index-date* definition, and follow-up; the exposure\ndefinition; the outcome definitions and the *operational algorithms/phenotypes* used to ascertain them, with their\nvalidation (e.g., positive predictive value); covariates and the strategy for *confounding control*; the\nstatistical analysis plan including the target estimand and handling of *intercurrent events*, missing data,\n*attrition/loss to follow-up*, and competing risks; planned *sensitivity and quantitative bias analyses*; data\nmanagement and quality control; and dissemination. Lifecycle requirements include EU PAS Register registration\n*before data collection starts*, PRAC protocol assessment for imposed studies, a study-progress/interim mechanism,\na study abstract, and a final study report submitted to the competent authority within 12 months of data-collection\nend. The standing principle is transparency and pre-specification: design choices are fixed and documented before\nthe data are touched, and any amendment is versioned and justified.\n\n**When NOT to use — limitations and common misapplications.** Module VIII is a regulatory *conduct-and-submission*\nframework, **not** a risk-of-bias instrument and **not** a quality score — endorsing a protocol does not certify\nthe study free of confounding or selection bias; that judgement comes from methodological appraisal (e.g., ROBINS-I)\nand the analytic safeguards the protocol commits to. Completing the Module VIII process does **not** make an\nobservational safety study causal; an under-specified estimand, a poorly validated outcome phenotype, immortal-time\nbias from misaligned time-zero, or residual confounding will survive a fully compliant submission. Common failure\nmodes: treating the ENCePP template as box-ticking *theatre* while leaving the data-fitness assessment and\nphenotype validation hollow; using Module VIII as if it were a journal reporting checklist (use STROBE/RECORD-PE\nfor the manuscript) or a protocol-quality template (use HARPER/STaRT-RWE for the design tables); applying it to\ninterventional trials (wrong regulatory regime); confusing imposed and voluntary obligations and thereby missing\nmandatory PRAC endorsement before study start; and registering on the EU PAS Register *after* data collection has\nbegun, which defeats the pre-specification it exists to enforce. Module VIII is EU-specific — it does not satisfy\nFDA expectations on its own.\n\n**How it maps to this catalog.** Module VIII names requirements; the following concepts implement them. Pre-specify\nthe eligibility–exposure–outcome–follow-up structure and a defensible estimand by emulating the trial you cannot\nrun with **target-trial-emulation**, and build the comparative analytic engine with the **active-comparator-new-user**\ndesign (incident-user washout + active comparator + time-zero alignment to kill confounding by indication and\nimmortal time). State and defend the causal contrast — including intercurrent-event handling — with\n**estimands-ate-att-intercurrent-events-rwe**. Operationalise outcome and covariate ascertainment, with validation\nmetrics, via **diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe**. Control measured and proxy confounding\nusing **high-dimensional-propensity-score-hdps-rwe**. Address the data-fitness, exposure-windowing, and recording\nrealities of the underlying data with **claims-analysis**. Document and analyse dropout with\n**attrition-and-loss-to-follow-up-rwe**, and quantify robustness to unmeasured confounding with sensitivity and\nquantitative-bias analysis. *Applied note (claims/EHR/registry RWE):* for a claims-based imposed PASS, the\nModule VIII data-fitness section must state coverage and lag, exclude person-time where the data source cannot\nobserve the relevant claims, define the outcome phenotype with its validated PPV, fix time-zero at the qualifying\nexposure event (no immortal time), and pre-register the protocol on the EU PAS Register before pull — each of which\nis implemented by the catalog concepts above and reported through STROBE/RECORD-PE in the final study report.", "primary_category": "Guideline", "tags": [ "guideline", "regulatory", "pharmacovigilance", "pass", "post-authorisation-safety-study", "ema", "encepp", "rwe" ], "aliases": [ "GVP Module VIII", "Good Pharmacovigilance Practices Module VIII", "Module VIII PASS", "EMA PASS guideline", "post-authorisation safety study guideline" ], "applies_to_study_types": [ "pass_imposed", "pass_voluntary", "drug_utilization", "signal_detection" ], "data_sources": [ "claims", "ehr", "registry", "linked", "primary" ], "citations": [ { "role": "introduce", "doi": null, "url": "https://www.ema.europa.eu/en/human-regulatory-overview/post-authorisation/pharmacovigilance-post-authorisation/good-pharmacovigilance-practices-gvp", "citation_text": "European Medicines Agency. Guideline on good pharmacovigilance practices (GVP) — Module VIII: Post-authorisation safety studies (Rev. 3). EMA/813938/2011 Rev 3. Effective 13 October 2017.", "year": 2017, "authors_short": "EMA", "notes": "Canonical EMA GVP page linking Module VIII and related PASS guidance; use this maintained landing page because individual EMA document URLs move." }, { "role": "explain", "doi": null, "url": "https://www.ema.europa.eu/en/human-regulatory-overview/post-authorisation/pharmacovigilance-post-authorisation/good-pharmacovigilance-practices-gvp", "citation_text": "European Medicines Agency. Guideline on good pharmacovigilance practices (GVP) — Module VIII Addendum I: Requirements and recommendations for the submission of information on non-interventional post-authorisation safety studies (Rev. 3). 2020.", "year": 2020, "authors_short": "EMA", "notes": "Maintained EMA GVP page linking Module VIII Addendum I and other PASS submission requirements; cite the page rather than brittle direct document URLs." }, { "role": "use", "doi": "10.1136/bmj.m4856", "url": "https://doi.org/10.1136/bmj.m4856", "citation_text": "Wang SV, Pinheiro S, Hua W, et al. STaRT-RWE: structured template for planning and reporting on the implementation of real world evidence studies. BMJ. 2021;372:m4856.", "year": 2021, "authors_short": "Wang et al.", "notes": "Structured template for fixing design parameters (time-zero, exposure/outcome windows, analysis) that operationalises the pre-specification a Module VIII protocol demands for RWD studies." }, { "role": "use", "doi": "10.1002/pds.5507", "url": "https://doi.org/10.1002/pds.5507", "citation_text": "Wang SV, Pottegård A, Crown W, et al. HARmonized Protocol Template to Enhance Reproducibility of hypothesis evaluating real-world evidence studies on treatment effects: A good practices report of a joint ISPE/ISPOR task force. Pharmacoepidemiology and Drug Safety. 2023;32(1):44-55.", "year": 2023, "authors_short": "Wang et al.", "notes": "ISPE/ISPOR harmonised protocol template for hypothesis-evaluating RWE; used inside a Module VIII protocol to specify design, estimand, and analysis with reproducibility." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "pass-imposed", "notes": "Module VIII fully governs imposed PASS, including mandatory PRAC protocol endorsement before study start and the statutory reporting timeline." }, { "relation_type": "applies_to", "target_slug": "pass-voluntary", "notes": "Module VIII governs the conduct and EU PAS Register registration of voluntary, MAH-initiated PASS with lighter submission obligations." }, { "relation_type": "applies_to", "target_slug": "drug-utilization", "notes": "Drug-utilisation studies conducted for safety/risk-minimisation purposes fall within the PASS framework defined by Module VIII." }, { "relation_type": "applies_to", "target_slug": "signal-detection", "notes": "Signal-evaluation studies that become formal safety studies are conducted under Module VIII conduct and submission rules." }, { "relation_type": "used_with", "target_slug": "target-trial-emulation", "notes": "Use to pre-specify the hypothetical trial (eligibility, exposure, time-zero, estimand) that a Module VIII PASS protocol must articulate before emulation." }, { "relation_type": "used_with", "target_slug": "active-comparator-new-user", "notes": "Implements the comparative design and time-zero alignment Module VIII protocols require to control confounding by indication and immortal time." }, { "relation_type": "used_with", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "Implements the confounding-control strategy the protocol must declare for secondary-data PASS." }, { "relation_type": "used_with", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Implements the outcome/covariate phenotype definitions and validation (PPV) that the data-fitness and ascertainment sections require." }, { "relation_type": "used_with", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Implements the target estimand and intercurrent-event handling the statistical analysis plan must specify." }, { "relation_type": "used_with", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Implements the documentation and analysis of dropout that the conduct and reporting requirements demand." }, { "relation_type": "used_with", "target_slug": "claims-analysis", "notes": "Implements the data-source fitness-for-use assessment, exposure windowing, and recording-practice handling central to secondary-data PASS." }, { "relation_type": "see_also", "target_slug": "prisma-p", "notes": "Sibling protocol-discipline guideline for evidence synthesis; Module VIII governs primary safety studies, PRISMA-P governs systematic-review protocols." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "ema", "fda", "hta" ] }, { "slug": "harper", "name": "HARPER", "short_definition": "A harmonized protocol template (text, tabular, and visual) for hypothesis-evaluating real-world evidence studies of treatment effects, jointly developed by an ISPE/ISPOR task force to force complete, unambiguous pre-specification of every design and analysis decision before data are analyzed.", "long_description": "**What it is.** HARPER — the **HARmonized Protocol Template to Enhance Reproducibility** of hypothesis-evaluating\nreal-world evidence (RWE) studies on treatment effects — is a structured protocol template developed by a\njoint **ISPE/ISPOR** (International Society for Pharmacoepidemiology / International Society for Pharmacoeconomics\nand Outcomes Research) task force and published as a Good Practices report in 2022 (concurrently in\n*Pharmacoepidemiology and Drug Safety* and *Value in Health*). It is not a brief reporting checklist; it is a\nfill-in protocol skeleton combining narrative text, standardized tables, and a study-design diagram that compels\nauthors to state every parameter a reader needs to reproduce the study: design choice, eligibility, exposure and\ncomparator operationalization, outcome algorithms, time-zero (index) alignment, covariate windows, the estimand\nand its intercurrent-event handling, the analysis specification, and the sensitivity analyses. HARPER is the\n*protocol-stage* successor to and harmonization of **STaRT-RWE** (the Structured Template for plAnning and\nReporting on the implementation of RWE studies, Wang et al., BMJ 2021); the two share the same DNA and the task\nforce intends HARPER to be the single template for both planning and reporting transparency. It is maintained by\nthe ISPE/ISPOR task force authorship and promoted through both societies and, in the US policy context, referenced\nalongside CMS/FDA expectations for transparent RWE.\n\n**When to use.** Use HARPER whenever you are designing or registering a **hypothesis-evaluating** (confirmatory,\ncomparative effectiveness or safety) non-interventional study of a treatment effect using routinely collected data\n— claims, EHR, registry, or linked sources — and the study is destined for a regulatory submission (FDA RWE\nprogram, EMA, an imposed or voluntary PASS), an HTA/payer dossier, or a high-impact peer-reviewed journal. It is\nthe right instrument at the **protocol stage**, ideally before data access and certainly before any outcome-dependent\nanalytic choices are made; registering a completed HARPER protocol is the strongest available defense against\naccusations of data-driven specification. Decision rules for choosing HARPER over a sibling: use HARPER (or its\npredecessor STaRT-RWE) when the study estimates a *treatment effect* and you need a full protocol template;\nuse **STROBE/RECORD/RECORD-PE** instead when your task is the final *reporting* of an observational study in a\nmanuscript (those are reporting checklists, not protocol templates); use the **ENCePP Checklist for Study Protocols**\nin parallel when an EU PASS or ENCePP seal is in scope (HARPER organizes the science, ENCePP confirms regulatory\ngovernance); use **SPIRIT** for interventional trial protocols and **CHEERS** for economic evaluations — HARPER is\nfor descriptive-of-effect, non-interventional designs, not RCTs or cost-effectiveness models. HARPER is generally\nnot the tool for purely descriptive epidemiology (incidence, prevalence, utilization), where a lighter design\ndescription suffices.\n\n**What it requires.** HARPER's tables and narrative force substantive content, mapped here to real-world-data\nrealities: (1) **Design transparency** — explicit design label, a study-design diagram on a calendar timeline\n(assessment, washout, baseline, and follow-up windows drawn relative to time zero), and a PICOTS-style framing of\nthe question. (2) **Data fitness-for-use** — naming the data source(s), provenance, capture mechanism, known\ncoverage gaps, and a justification that the data can actually measure the exposure, outcome, and confounders\nrequired. (3) **Exposure/comparator operationalization** — code lists, grace periods, stockpiling rules, and (for\ncomparative work) a defensible active comparator. (4) **Outcome phenotype/algorithm validation** — the operational\ndefinition and its validation metrics (PPV/sensitivity) or a plan to obtain them. (5) **Time-zero alignment** —\nindex-date definition that aligns eligibility, treatment assignment, and start of follow-up to avoid immortal time.\n(6) **Estimand and intercurrent events** — the target population, treatment strategies being contrasted, and the\npre-specified strategy for intercurrent events (treatment switching, discontinuation, death). (7) **Confounding\ncontrol** — covariate measurement windows and the adjustment method (e.g., propensity or high-dimensional\npropensity scores). (8) **Attrition and missing data** — an attrition (CONSORT-style) accounting and a missing-data\nplan. (9) **Sensitivity / quantitative bias analysis** — pre-specified robustness checks (washout/grace-period\nvariants, negative controls, E-value or other bias analysis) and versioned code lists so the analysis can be\nreproduced exactly.\n\n**When NOT to use — limitations and common misapplications.** HARPER is a *transparency and reproducibility*\ninstrument, **not** a risk-of-bias tool and **not** a quality score: a fully completed HARPER protocol can describe a\nbadly confounded study with great clarity. Do not treat a filled-in template as evidence the study is valid, nor as\na substitute for a formal bias assessment (ROBINS-I) or for the design thinking of target-trial emulation —\ncompleting the template does not make an observational estimate causal; it only makes the (possibly biased) design\nlegible. Common failure modes: **template-as-theater** — pasting boilerplate into the cells without genuine\npre-specification, then changing analyses after seeing results; using a *reporting* checklist (STROBE) where a\n*protocol template* (HARPER/STaRT-RWE) was needed, or vice-versa; using HARPER for an interventional trial (use\nSPIRIT) or an economic model (use CHEERS); and omitting the data-fitness and phenotype-validation cells, which are\nexactly the parts regulators and HTA reviewers scrutinize. HARPER also does not itself govern EU PASS regulatory\nprocess — pair it with the ENCePP checklist rather than assuming HARPER satisfies that requirement.\n\n**How it maps to this catalog.** HARPER is an organizing frame; the concepts in this catalog implement each cell.\nIts design-and-estimand spine is implemented by **target-trial-emulation** (pre-specify the hypothetical trial\nbefore emulating it), **active-comparator-new-user** (the new-user + active-comparator + time-zero structure that\nfills the eligibility/exposure/index tables), and **estimands-ate-att-intercurrent-events-rwe** (the estimand and\nintercurrent-event cells). The data-fitness cell is implemented by **fit-for-purpose-data-assessment-rwe** and, for\nUS claims nuance, **claims-analysis** and **medicare-ffs-ma-commercial-claims-differences-rwe** (FFS vs MA capture\nand coding-intensity differences that determine whether \"no prior claim\" is a true washout). Outcome/exposure\noperationalization maps to **diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe** (1-inpatient / 2-outpatient\nrules, time windows, PPV) and **time-zero-index-date-alignment-rwe**. Confounding control maps to\n**high-dimensional-propensity-score-hdps-rwe** and **propensity-score-methods-psm-iptw**. Attrition and robustness\nmap to **attrition-and-loss-to-follow-up-rwe**, **e-value-sensitivity-analysis**, and\n**quantitative-bias-analysis-toolkit-rwe**. For claims/EHR/registry RWE specifically: a defensible HARPER protocol\npre-registers the diagnosis/outcome algorithm *with* its validation metrics, requires continuous-enrollment and\ndata-capture conditions so absence-of-fill is observed rather than missing, draws the calendar-timeline design\ndiagram so immortal time is visibly excluded, and versions every code list — so the eventual STROBE/RECORD-PE\nmanuscript can be checked back against a protocol that was fixed before the analyst saw an effect estimate.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "protocol-template", "rwe", "pharmacoepidemiology", "transparency", "reproducibility", "ispe", "ispor" ], "aliases": [ "HARPER", "HARmonized Protocol Template to Enhance Reproducibility", "ISPE/ISPOR HARPER template" ], "applies_to_study_types": [ "new_user", "active_comparator_new_user", "cer_observational", "comparative_effectiveness", "target_trial_emulation", "claims_analysis", "ehr_study", "cohort_retrospective", "cohort_prospective", "pass_imposed", "pass_voluntary" ], "data_sources": [ "claims", "ehr", "registry", "linked", "multi-database" ], "citations": [ { "role": "introduce", "doi": "10.1002/pds.5507", "url": "https://doi.org/10.1002/pds.5507", "citation_text": "Wang SV, Pottegård A, Crown W, et al. HARmonized Protocol Template to Enhance Reproducibility of hypothesis evaluating real-world evidence studies on treatment effects: a good practices report of a joint ISPE/ISPOR task force. Pharmacoepidemiology and Drug Safety. 2023;32(1):44-55.", "year": 2023, "authors_short": "Wang et al.", "notes": "Canonical ISPE/ISPOR Good Practices statement defining the HARPER template, its tables, and the study-design diagram. Published concurrently in Value in Health." }, { "role": "explain", "doi": "10.1016/j.jval.2022.09.001", "url": "https://doi.org/10.1016/j.jval.2022.09.001", "citation_text": "Wang SV, Pottegård A, Crown W, et al. HARmonized Protocol Template to Enhance Reproducibility of Hypothesis Evaluating Real-World Evidence Studies on Treatment Effects: A Good Practices Report of a Joint ISPE/ISPOR Task Force. Value in Health. 2022;25(10):1663-1672.", "year": 2022, "authors_short": "Wang et al.", "notes": "Co-publication of the same task-force report in the ISPOR journal; identical template, HTA/health-economics audience." }, { "role": "explain", "doi": "10.1136/bmj.m4856", "url": "https://doi.org/10.1136/bmj.m4856", "citation_text": "Wang SV, Pinheiro S, Hua W, et al. STaRT-RWE: structured template for planning and reporting on the implementation of real world evidence studies. BMJ. 2021;372:m4856.", "year": 2021, "authors_short": "Wang et al.", "notes": "Predecessor structured template that HARPER harmonizes and extends; useful for the reporting-stage tables and as the conceptual basis for HARPER's design parameters." } ], "relations": [ { "relation_type": "see_also", "target_slug": "target-trial-emulation", "notes": "HARPER's design and estimand tables are best populated by explicitly pre-specifying the hypothetical target trial the observational study emulates." }, { "relation_type": "see_also", "target_slug": "active-comparator-new-user", "notes": "The new-user + active-comparator + time-zero structure fills HARPER's eligibility, exposure/comparator, and index-date cells for comparative drug studies." }, { "relation_type": "see_also", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Implements HARPER's requirement to state the target estimand and the pre-specified handling of intercurrent events (switching, discontinuation, death)." }, { "relation_type": "see_also", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "Implements HARPER's data fitness-for-use cell - whether the source can actually measure the exposure, outcome, and confounders required." }, { "relation_type": "see_also", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Implements HARPER's outcome operationalization and phenotype-validation requirement (algorithm rules, time windows, PPV/sensitivity)." }, { "relation_type": "see_also", "target_slug": "time-zero-index-date-alignment-rwe", "notes": "Implements HARPER's time-zero alignment so eligibility, assignment, and follow-up start together and immortal time is excluded." }, { "relation_type": "see_also", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "Implements HARPER's confounding-control specification for claims/EHR data with many proxy covariates." }, { "relation_type": "see_also", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Implements HARPER's CONSORT-style attrition accounting and loss-to-follow-up handling." }, { "relation_type": "see_also", "target_slug": "e-value-sensitivity-analysis", "notes": "Implements HARPER's quantitative-bias/sensitivity cell for residual unmeasured confounding." }, { "relation_type": "see_also", "target_slug": "claims-analysis", "notes": "Operational detail for the claims-specific data-fitness, enrollment, and capture conditions HARPER asks authors to document." }, { "relation_type": "complements", "target_slug": "study-protocol-or-sap-elements", "notes": "HARPER organizes the protocol-level transparency that a fuller SAP then expands into analytic detail." }, { "relation_type": "applies_to", "target_slug": "target-trial-emulation", "notes": "HARPER provides the protocol template and transparency expectations for reporting a target-trial-emulation study." }, { "relation_type": "applies_to", "target_slug": "active-comparator-new-user", "notes": "HARPER provides the protocol template and transparency expectations for an active-comparator new-user study." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "hta", "journal" ] }, { "slug": "ich-e6-r2", "name": "ICH E6(R2) Good Clinical Practice", "short_definition": "ICH Harmonised Guideline establishing the international ethical and scientific quality standard for designing, conducting, recording, and reporting trials that involve human subjects. In RWE it governs only the prospective/interventional component of hybrid designs (pragmatic and registry-based randomized trials, decentralized trials, single-arm trials whose treated arm is run prospectively) — not retrospective database studies.", "long_description": "**What it is** — *Integrated Addendum to ICH E6(R1): Guideline for Good Clinical Practice E6(R2)* (Step 4, 9 November 2016),\nthe international Good Clinical Practice (GCP) standard maintained by the **International Council for Harmonisation of\nTechnical Requirements for Pharmaceuticals for Human Use (ICH)** and adopted into law/guidance by FDA, EMA, PMDA, and other\nICH regulators. GCP is an **ethical and scientific quality standard for the conduct of clinical trials in human subjects** —\nit defines the responsibilities of sponsors, investigators, and IRBs/IECs; informed consent; the Investigator's Brochure and\nprotocol; trial monitoring; source documentation and data integrity (ALCOA); essential documents; and reporting of safety.\nE6(R2) added a risk-based quality management framework (quality-by-design, risk-based monitoring, sponsor oversight of\nelectronic systems and vendors). It is **not** a reporting checklist, **not** a risk-of-bias instrument, and **not** an RWE\ndesign framework — it is a *conduct* standard for prospective research. Note that **ICH E6(R3)** (Step 4, 6 January 2025)\nsupersedes R2 for new protocols as ICH members transition; R3 is explicitly media-neutral and builds in real-world data\ncapture and decentralized elements, so for any newly designed trial confirm whether your region has adopted R3 before\ndefaulting to R2.\n\n**When to use** — Apply GCP when a study has a **prospective, interventional component conducted under a protocol in human\nsubjects**, even when that study also leans on real-world data. The RWE-relevant cases are narrow and specific: pragmatic\n(point-of-care) randomized trials; registry-based randomized trials that randomize within an existing registry; decentralized\nor hybrid trials that collect protocol endpoints from EHRs, wearables, or claims; and single-arm trials whose *treated arm is\nprospectively enrolled and dosed* (even if the comparator is an external real-world control). Decision rule: if a patient is\n**assigned to an intervention by the investigator/protocol and followed prospectively**, GCP applies to that arm and its data\ncollection. If the study is **secondary use of already-existing routinely collected data** (a retrospective claims or EHR\ncohort, a non-interventional PASS, an HTA dossier built from observational evidence), GCP does **not** apply — the governing\ndocuments are RECORD/RECORD-PE and STROBE for reporting, HARPER/STaRT-RWE for protocol templating, ENCePP for PASS conduct,\nand the FDA/EMA RWE frameworks for regulatory fitness. For new prospective protocols, prefer **ICH E6(R3)** where adopted;\nreserve E6(R2) for trials initiated under R2 or in regions still transitioning.\n\n**What it requires** — For the prospective component it governs, GCP enforces: (1) a written, IRB/IEC-approved **protocol**\nand a current **Investigator's Brochure**; (2) documented **informed consent** appropriate to the data and specimens\ncollected (including secondary use and linkage permissions when EHR/registry data feed trial endpoints); (3) **investigator\nand site qualification** and delegation of duties; (4) **sponsor oversight** with a risk-based quality management plan —\nprospective identification of critical-to-quality factors, risk-based monitoring, and defined data-integrity controls\n(E6(R2)'s central addition); (5) **source data and source documents** that are attributable, legible, contemporaneous,\noriginal, and accurate (ALCOA), with a credible audit trail when endpoints are extracted from EHRs or devices; (6)\n**traceability from source to analysis** — when a real-world data source (registry, EHR, claims feed) serves as source data,\nthe sponsor must validate the system, control access, and document data provenance and any transformations; (7) **safety\nreporting** and essential-document retention. In hybrid designs the load-bearing GCP question is *data integrity of\nroutinely collected data used as trial source data*: who validated the extract, can the audit trail reconstruct the value,\nand was consent adequate for the use.\n\n**When NOT to use — limitations and common misapplications** — The dominant failure mode is **forcing GCP onto a\nretrospective database study**. A claims-only or EHR-only secondary-use cohort, a non-interventional PASS, or an HTA RWE\nanalysis is not a clinical trial; its participants were not assigned an intervention by a protocol, so GCP's machinery (IRB\napproval of an intervention, individual informed consent, the Investigator's Brochure, on-site monitoring, source data\nverification against a CRF) does not map. Pasting \"conducted in accordance with ICH E6(R2)\" onto such a study is\n**checklist-as-theater** and a reviewer will reject it. Equally, GCP is **not a substitute for reporting or bias\nguidance**: completing GCP conduct requirements says nothing about whether confounding was controlled, whether the\nphenotype was validated, or whether the study was reported transparently — those are RECORD-PE/STROBE/HARPER concerns. Do\nnot treat GCP as a *quality score* (it is binary conduct compliance, audited, not rated), and do not cite E6(R2) when the\nprotocol was designed under, and should follow, **E6(R3)**. Finally, GCP applies to the **interventional arm**, not to an\nexternal real-world control: the control data are governed by RWE data-fitness and reporting standards, not by GCP.\n\n**How it maps to this catalog** — GCP intersects the catalog only where a study is prospective or hybrid. Its conduct\nstandards attach to: **pragmatic-trial** and **registry-trial** (the protocol, consent, IRB, monitoring, and data-integrity\nrequirements for point-of-care and registry-based randomized trials); **single-arm-external-control** (GCP governs the\nprospectively enrolled treated arm; the external control is governed by RWD-fitness concepts, not GCP);\n**endpoint-adjudication-chart-review-rwe** (GCP source-data and audit-trail requirements when trial endpoints are\nadjudicated from EHR/chart data); **regulatory-readiness-rwe** and **estimand-analysis-traceability-rwe** (source-to-analysis\ntraceability and pre-specification expected of regulatory submissions); and **study-protocol-or-sap-elements** /\n**picots-framework-rwe** (the protocol discipline GCP requires). For **pass-imposed** and **pass-voluntary**, GCP applies\n*only if the PASS is interventional* (rare) — most PASS are non-interventional and fall under ENCePP, not GCP. Concepts that\nbelong to **retrospective** RWE — claims-analysis, high-dimensional-propensity-score-hdps-rwe,\ndiagnosis-phenotype-algorithm-1ip-2op-time-window-rwe, time-zero-index-date-alignment-rwe,\nattrition-and-loss-to-follow-up-rwe, fit-for-purpose-data-assessment-rwe, and target-trial-emulation — are explicitly **out\nof GCP scope**: target-trial emulation deliberately *substitutes design logic for the trial GCP would have governed* and is\nreported under STROBE/RECORD-PE, not E6.\n\n*Applied note (hybrid claims/EHR/registry RWE).* When a registry-based or decentralized trial uses an EHR or claims feed as\n**source data** for a protocol endpoint, GCP compliance hinges on system validation, a reconstructable audit trail, access\ncontrol, and consent that covers the secondary use and any linkage — document these in the protocol and the data-management\nplan. The epidemiologic quality of that same data (phenotype PPV, completeness, transportability) is a separate question\nanswered by the RWE concepts above; GCP and RWE-fitness standards are complementary, not interchangeable.", "primary_category": "Guideline", "tags": [ "guideline", "good-clinical-practice", "regulatory", "clinical-trial-conduct", "pragmatic-trial", "registry-trial", "data-integrity" ], "aliases": [ "ICH E6(R2)", "ICH E6(R2)/GCP", "GCP", "Good Clinical Practice", "Integrated Addendum to ICH E6(R1)", "ICH-E6-R2" ], "applies_to_study_types": [ "pragmatic_trial", "registry_trial" ], "data_sources": [ "registry", "ehr", "linked", "primary" ], "citations": [ { "role": "introduce", "doi": null, "url": "https://database.ich.org/sites/default/files/E6_R2_Addendum.pdf", "citation_text": "International Council for Harmonisation. Integrated Addendum to ICH E6(R1): Guideline for Good Clinical Practice E6(R2). Current Step 4 version, 9 November 2016.", "year": 2016, "authors_short": "ICH", "notes": "Canonical guideline text. No journal DOI exists; the ICH stable PDF on database.ich.org is the authoritative source. Adopted Step 4 by the Regulatory Members of the ICH Assembly on 9 November 2016 as an integrated addendum to E6(R1) (1996)." }, { "role": "explain", "doi": null, "url": "https://database.ich.org/sites/default/files/ICH_E6%28R3%29_Step4_FinalGuideline_2025_0106.pdf", "citation_text": "International Council for Harmonisation. Guideline for Good Clinical Practice E6(R3). Final version, adopted 6 January 2025 (Step 4).", "year": 2025, "authors_short": "ICH", "notes": "Successor revision that supersedes E6(R2) as ICH members transition. Media-neutral and explicitly accommodating real-world data capture and decentralized/hybrid designs; prefer for newly designed protocols where adopted." }, { "role": "use", "doi": null, "url": "https://database.ich.org/sites/default/files/E6_R2_Addendum.pdf", "citation_text": "International Council for Harmonisation. Integrated Addendum to ICH E6(R1): Guideline for Good Clinical Practice E6(R2). Current Step 4 version, 9 November 2016.", "year": 2016, "authors_short": "ICH", "notes": "Stable official ICH database PDF for the E6(R2) text; local regulator adoption notices may be cited in submissions, but this is the authoritative guideline text." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "pragmatic-trial", "notes": "GCP governs the conduct, consent, IRB oversight, monitoring, and source-data integrity of pragmatic (point-of-care) randomized trials, including any routinely collected data used as trial source data." }, { "relation_type": "applies_to", "target_slug": "registry-trial", "notes": "GCP governs registry-based randomized trials — the prospective randomization and the integrity/validation of registry data used as trial source data." }, { "relation_type": "used_with", "target_slug": "single-arm-external-control", "notes": "GCP applies to the prospectively enrolled treated arm; the external real-world control is governed by RWD fitness-for-use and reporting standards, not by GCP." }, { "relation_type": "used_with", "target_slug": "endpoint-adjudication-chart-review-rwe", "notes": "GCP source-data (ALCOA) and audit-trail requirements apply when trial endpoints are adjudicated from EHR or chart data." }, { "relation_type": "see_also", "target_slug": "regulatory-readiness-rwe", "notes": "GCP's source-to-analysis traceability and sponsor-oversight expectations align with regulatory readiness for prospective and hybrid submissions." }, { "relation_type": "see_also", "target_slug": "study-protocol-or-sap-elements", "notes": "GCP requires a written, IRB-approved protocol; the protocol/SAP elements concept details the substantive content GCP presumes." }, { "relation_type": "see_also", "target_slug": "target-trial-emulation", "notes": "Out of GCP scope — target-trial emulation substitutes explicit design logic for the prospective trial GCP would have governed, and is reported under STROBE/RECORD-PE rather than ICH E6." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "journal" ] }, { "slug": "ich-e9-r1", "name": "ICH E9(R1) Estimands and Sensitivity Analysis", "short_definition": "ICH E9(R1) is the regulatory addendum to ICH E9 that defines the estimand framework — five attributes (population, treatment, endpoint, intercurrent events, population-level summary) that together specify precisely what treatment effect a study aims to estimate — and requires aligned sensitivity analyses to probe the assumptions behind it.", "long_description": "**What it is** — ICH E9(R1), *Addendum on Estimands and Sensitivity Analysis to the Guideline on Statistical Principles for Clinical Trials*, is a harmonised regulatory guidance adopted at ICH Step 4 in November 2019. It is maintained by the International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use (ICH) and implemented by member regulators including FDA and EMA (EMA/CHMP/ICH/436221/2017). It is not a reporting checklist and not a critical-appraisal tool; it is a *thinking framework* that forces a study to state, before any analysis, exactly which treatment effect (\"estimand\") it is targeting. An estimand is defined by five attributes: (1) the **population** (target patients), (2) the **treatment** condition(s) being compared, (3) the **endpoint** (variable), (4) how **intercurrent events** (ICEs — events after treatment initiation that alter the existence or interpretation of the outcome, such as treatment discontinuation, switching, rescue medication, or death) are handled, and (5) the **population-level summary** (e.g., difference in means, risk ratio, hazard ratio). The addendum specifies five strategies for ICEs — treatment-policy, hypothetical, composite, while-on-treatment, and principal-stratum — and requires that the chosen estimator and any sensitivity analyses be aligned to the named estimand rather than chosen by default.\n\n**When to use** — Apply ICH E9(R1) whenever a study estimates a treatment effect and that effect must be defended to a regulator, an HTA body, or a critical journal — i.e., the design and analysis stage of pragmatic trials, target-trial emulations, comparative-effectiveness and safety studies in routinely collected data, and post-authorisation safety/efficacy studies (PASS/PAES). It is the governing framework for FDA and EMA submissions where a primary or secondary treatment effect is claimed, and increasingly for HTA dossiers where the relevant decision question (e.g., effectiveness under real-world adherence vs. efficacy under perfect adherence) hinges on how intercurrent events are treated. Decision rule for which tool applies: use **ICH E9(R1)** to *define the question and the effect* (what are we estimating, and under what handling of discontinuation/switching/death?); use a reporting guideline such as STROBE/RECORD-PE or a protocol template such as HARPER/STaRT-RWE to *document the completed study*; use ROBINS-I to *appraise its risk of bias*. These are complementary, not interchangeable. In RWE specifically, E9(R1) is most powerful when paired with target-trial emulation, which gives the otherwise-abstract intercurrent-event strategies concrete operational meaning (e.g., a treatment-policy estimand maps to intention-to-treat from time zero; a hypothetical \"no switching\" estimand maps to per-protocol with censoring and inverse-probability-of-censoring weighting).\n\n**What it requires** — The addendum requires that a study, at the protocol stage, (i) name the **target population** and align eligibility and time zero to it (in RWE, the new-user/active-comparator structure and an immortal-time-free index date); (ii) specify the **treatment conditions** as explicit strategies, including duration, switching, and concomitant therapy; (iii) define the **endpoint** with its measurement and assessment window; (iv) enumerate the **intercurrent events** that can occur and assign each an explicit strategy (treatment-policy, hypothetical, composite, while-on-treatment, or principal-stratum) — the analytic heart of the framework, because the same data can yield very different effects depending on how discontinuation, switching, rescue, and especially death are handled; (v) state the **population-level summary**; and (vi) pre-specify **sensitivity analyses** that vary the assumptions specific to the chosen estimand (e.g., missing-data mechanisms under a hypothetical strategy, grace periods and censoring rules under a per-protocol estimand) and are distinguished from supplementary analyses that target a different estimand. In real-world data the framework forces several disciplines the catalog implements concretely: data fitness-for-use must support the chosen ICE strategy (e.g., reliable death capture for a composite or while-on-treatment estimand), phenotype/outcome algorithms must be validated, time-zero must be aligned to the population and treatment definitions, attrition and informative censoring must be modeled rather than ignored, and confounding control must target the named estimand (ATT vs. ATE) rather than whatever the default estimator produces.\n\n**When NOT to use — limitations and common misapplications** — The dominant failure mode is **estimand-as-documentation**: writing a five-attribute table to satisfy a template while the analysis is unchanged and the intercurrent-event strategy is never actually operationalized. A second is **defaulting to a single treatment-policy estimand** for every question without asking whether discontinuation and switching are part of the effect of interest or a nuisance to be removed — the choice is substantive, not administrative, and the wrong default can answer a question no decision-maker asked. A third is **confusing the estimand with the estimator**: E9(R1) defines *what* is being estimated; it does not endorse any particular model, and a beautifully specified estimand fitted with a biased estimator (e.g., adjusting for a post-baseline mediator, or ignoring informative censoring under a hypothetical strategy) is still wrong. A fourth, specific to RWE, is **importing the framework without the target-trial scaffolding** — intercurrent events such as switching and rescue have no operational meaning until eligibility, treatment strategies, and time zero are defined, so applying E9(R1) to a prevalent-user or immortal-time-contaminated cohort produces a precisely-worded estimand for a biased number. Finally, E9(R1) is **not a risk-of-bias instrument and not a quality score**: a clearly specified estimand does not make an observational comparison causal; residual confounding, positivity violations, and measurement error must still be addressed with the relevant design and analysis methods and appraised with ROBINS-I.\n\n**How it maps to this catalog** — Each E9(R1) requirement is implemented by one or more concepts in this repository. The **population/treatment/time-zero** attributes are operationalized by **target-trial-emulation**, **active-comparator-new-user**, and **immortal-time-bias-handling**. The **intercurrent-events and population-level-summary** attributes are implemented by **estimands-ate-att-intercurrent-events-rwe** (the five ICE strategies and ATE/ATT contrasts) and made auditable by **estimand-analysis-traceability-rwe** (estimand-to-estimator traceability). **Confounding control** aligned to the chosen estimand is implemented by **high-dimensional-propensity-score-hdps-rwe** and **propensity-score-methods-psm-iptw**. **Endpoint/phenotype validity** is implemented by **diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe** and **claims-outcome-algorithm-ppv-sensitivity-rwe**; competing events (notably death as an intercurrent event) by **competing-risks-cause-specific-fine-gray-rwe**. **Data fitness** for the chosen ICE strategy is implemented by **fit-for-purpose-data-assessment-rwe**; **attrition and informative censoring** by **attrition-and-loss-to-follow-up-rwe**; the **sensitivity-analysis** requirement by **e-value-sensitivity-analysis**; and the **decision-question framing** by **picots-framework-rwe**. Applied note for claims/EHR/registry RWE: choosing an intercurrent-event strategy is also a data-fitness decision — a *while-on-treatment* or *composite-with-death* estimand demands reliable mortality and discontinuation capture (link to a death index; reconcile MA-only person-time and days-supply gaps), whereas a *hypothetical* \"if patients had not switched\" estimand demands defensible censoring-weight models. Write the estimand first, then verify the database can actually support it; the order is what separates decision-grade RWE from an estimand table.", "primary_category": "Guideline", "tags": [ "guideline", "methodological", "estimand", "intercurrent-events", "sensitivity-analysis", "rwe", "framework" ], "aliases": [ "ICH E9(R1)", "ICH E9 R1", "ICH-E9-R1", "Estimand addendum", "ICH E9 addendum on estimands" ], "applies_to_study_types": [ "pragmatic_trial", "cer_observational", "target_trial_emulation", "new_user", "active_comparator_new_user" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "doi": null, "url": "https://database.ich.org/sites/default/files/E9-R1_Step4_Guideline_2019_1203.pdf", "citation_text": "International Council for Harmonisation. ICH E9(R1): Addendum on Estimands and Sensitivity Analysis in Clinical Trials to the Guideline on Statistical Principles for Clinical Trials. ICH Harmonised Guideline, Step 4, 2019.", "year": 2019, "authors_short": "ICH", "notes": "The canonical regulatory guidance defining the five-attribute estimand framework and the five intercurrent-event strategies; adopted at ICH Step 4 (Nov 2019) and implemented by FDA and EMA. No journal DOI exists for the guidance itself; this is the official ICH stable document." }, { "role": "explain", "doi": "10.1177/1740774516633115", "url": "https://doi.org/10.1177/1740774516633115", "citation_text": "Mehrotra DV, Hemmings RJ, Russek-Cohen E, on behalf of the ICH E9/R1 Expert Working Group. Seeking harmony: estimands and sensitivity analyses for confirmatory clinical trials. Clinical Trials. 2016;13(4):456-458.", "year": 2016, "authors_short": "Mehrotra et al.", "notes": "Foundational statement from members of the ICH E9/R1 Expert Working Group articulating the rationale for the estimand framework and the alignment of sensitivity analyses to a named estimand." }, { "role": "use", "doi": "10.1007/978-3-031-26328-6_9", "url": "https://doi.org/10.1007/978-3-031-26328-6_9", "citation_text": "Wu Y. Estimand in real-world evidence study: from frameworks to application. In: Real-World Evidence in Medical Product Development. Springer; 2023:181-200.", "year": 2023, "authors_short": "Wu", "notes": "Extends the ICH E9(R1) estimand framework to non-interventional real-world evidence studies, including how intercurrent events and the population-level summary translate to observational data and target-trial emulation." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "pragmatic-trial", "notes": "E9(R1) governs estimand specification and aligned sensitivity analyses for pragmatic trials, where intercurrent events such as non-adherence and switching are common and the treatment-policy vs. hypothetical choice is consequential." }, { "relation_type": "applies_to", "target_slug": "cer-observational", "notes": "Comparative-effectiveness observational studies should name the estimand (population, treatment strategies, endpoint, intercurrent-event handling, summary) before analysis." }, { "relation_type": "applies_to", "target_slug": "target-trial-emulation", "notes": "Target-trial emulation gives the abstract intercurrent-event strategies concrete operational meaning; E9(R1) supplies the vocabulary the emulated protocol must specify." }, { "relation_type": "applies_to", "target_slug": "active-comparator-new-user", "notes": "The new-user/active-comparator structure operationalizes the population and treatment attributes and an immortal-time-free time zero." }, { "relation_type": "complements", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Implements the intercurrent-events and population-level-summary attributes (five ICE strategies; ATE vs. ATT contrasts) for real-world data." }, { "relation_type": "complements", "target_slug": "estimand-analysis-traceability-rwe", "notes": "Implements the requirement that the estimator and sensitivity analyses be traceable to the named estimand rather than chosen by default." }, { "relation_type": "complements", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "Confounding control aligned to the chosen estimand (ATT vs. ATE); hdPS supplies proxy adjustment when key confounders are unmeasured." }, { "relation_type": "complements", "target_slug": "competing-risks-cause-specific-fine-gray-rwe", "notes": "Death as an intercurrent or competing event is handled here; the choice of cause-specific vs. subdistribution hazard must match the estimand." }, { "relation_type": "complements", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "The chosen intercurrent-event strategy is also a data-fitness decision; the database must reliably capture the events (e.g., death, discontinuation) the estimand depends on." }, { "relation_type": "complements", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Informative censoring and attrition must be modeled to estimate hypothetical and per-protocol estimands without bias." }, { "relation_type": "see_also", "target_slug": "picots-framework-rwe", "notes": "PICOTS frames the decision question that the estimand then makes precise and analyzable." }, { "relation_type": "see_also", "target_slug": "immortal-time-bias-handling", "notes": "Time-zero alignment for the population and treatment attributes prevents immortal time that would corrupt any estimand." }, { "relation_type": "see_also", "target_slug": "e-value-sensitivity-analysis", "notes": "Quantitative-bias and sensitivity analyses probe the unverifiable assumptions behind the named estimand." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "hta", "journal" ] }, { "slug": "inahta-checklist", "name": "INAHTA HTA Checklist", "short_definition": "A 17-item, 5-domain transparency checklist developed by the International Network of Agencies for Health Technology Assessment (INAHTA) to promote consistent, transparent reporting of health technology assessment reports — explicitly a reporting/transparency aid, not a quality scorecard.", "long_description": "**What it is** — The **INAHTA HTA Checklist** (formally the *Checklist for HTA Reports*, Hailey 2003) is a 17-item instrument, grouped into five domains, developed and maintained by the **International Network of Agencies for Health Technology Assessment (INAHTA)** — the global network of publicly funded HTA bodies (NICE, CADTH, IQWiG, HAS, and ~50 others). It was built by summarizing the key elements of HTA reports, drawing on agency experience and existing HTA guidelines, then circulating for consensus among member agencies. Its purpose is narrow and explicit: to further a **consistent and transparent** approach to *reporting* HTA, so that a reader can tell what question was asked, how the assessment was done, what was found, and what it implies. The five domains are: (1) **preliminary information** (authorship, contact, review status, links to other reports); (2) **why and how the assessment was prepared** (the policy/research question, scope, methods for retrieving and appraising evidence, sources of data); (3) **the results of the assessment** (findings on effectiveness, safety, and economic considerations, with stated assumptions and uncertainty); (4) **implications and limitations** (medico-legal, ethical, social, and organizational implications; limitations of the report); and (5) **conclusions and recommendations** (clearly distinguished from results). It is a short, agency-level reporting/transparency tool — not a measurement method and not a critical-appraisal/risk-of-bias instrument.\n\n**When to use** — Use the INAHTA checklist when **producing or reading a full HTA report** intended for a coverage, reimbursement, or health-system decision — i.e., a document that *integrates* clinical effectiveness, safety, and economic evidence (cost-effectiveness/cost-utility, budget impact) plus organizational/ethical/social implications, rather than a single primary study. It is the natural transparency backbone for an **HTA/payer dossier** and for agency-issued assessments. Decision rules for choosing *this* tool versus siblings: if you are reporting the *economic model* itself, **CHEERS 2022** is the correct reporting guideline and INAHTA sits above it as the report-level wrapper; if you are reporting the *systematic-review component*, use **PRISMA 2020** for the synthesis and INAHTA for the overall report; if you are appraising the *methodological quality/risk of bias* of the included evidence, INAHTA is the wrong instrument entirely (use ROBINS-I, RoB 2, AMSTAR 2, or GRADE). INAHTA governs how the *assessment as a whole* is reported and made transparent, not how any one piece of evidence was generated or graded.\n\n**What it requires** — Framed for real-world-evidence-heavy HTA submissions, the checklist's domains demand explicit documentation of: the **policy and research question and scope** (population, intervention, comparator, outcomes, setting — the PICOTS that anchors the assessment); the **methods for identifying, selecting, and synthesizing evidence**, including which **data sources** were used and why (claims, EHR, registry, linked data) and their **fitness for the decision question**; the **design and analytic choices** behind any real-world comparative analysis (time-zero alignment, comparator selection, confounding control, attrition/missing data, and the estimand actually targeted); the **results with assumptions and uncertainty made visible** (effectiveness, safety, and economic findings, with sensitivity/scenario analyses); the **implications** (transferability/generalizability to the decision context, plus ethical, legal, social, and organizational consequences); and **conclusions kept distinct from results**. For RWE used in HTA, the checklist effectively forces a reader to see whether the real-world evidence is fit for purpose, whether the comparison is causally interpretable, and where residual uncertainty lies — without itself prescribing the methods.\n\n**When NOT to use — limitations and common misapplications** — The single most important caveat is INAHTA's own: the checklist **\"is not intended to be viewed or used as a scorecard to rate HTA reports,\"** and reports \"may be valid and useful without meeting all the criteria.\" Treating item counts as a quality score is the canonical misuse — it is a *reporting/transparency* aid, not a risk-of-bias tool and not a quality rating. Other failure modes: (1) **Completing the checklist does not make the underlying evidence sound** — a fully transparent report can rest on a confounded, immortal-time-biased real-world analysis; transparency reveals problems, it does not cure them. (2) **Wrong instrument for the layer of work** — using INAHTA to appraise an individual study (use ROBINS-I/RoB 2), to grade certainty of a body of evidence (use GRADE), to appraise a systematic review's conduct (use AMSTAR 2), or to report an economic evaluation in detail (use CHEERS 2022). (3) **Checklist-as-theater** — pasting a completed checklist into an appendix while the body of the report still omits the comparator rationale, data-source fitness, or sensitivity analyses; the checklist is satisfied in form but defeats its own purpose. (4) **Over-reach onto primary research** — it was designed for agency-style HTA *reports*, not for a stand-alone pharmacoepidemiology manuscript, where STROBE/RECORD-PE/HARPER apply.\n\n**How it maps to this catalog** — Each checklist domain points to the concept(s) in this repo that actually implement the requirement. *Scope / research question:* **picots-framework-rwe** structures the population-intervention-comparator-outcome-timing-setting frame the checklist's \"why and how\" domain demands. *Data identification and fitness:* **fit-for-purpose-data-assessment-rwe** and **claims-analysis** establish whether the claims/EHR/registry source can answer the decision question. *Design and causal interpretability of any RWE comparison:* **target-trial-emulation**, **active-comparator-new-user**, **time-zero-index-date-alignment-rwe**, **diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe** (case-finding/phenotype validity), **high-dimensional-propensity-score-hdps-rwe** (confounding control), **estimands-ate-att-intercurrent-events-rwe** (the estimand actually reported), and **attrition-and-loss-to-follow-up-rwe** (follow-up completeness). *Results, uncertainty, and sensitivity:* **quantitative-bias-analysis-toolkit-rwe** and **e-value-sensitivity-analysis** make residual-confounding uncertainty explicit; the economic findings draw on **cost-effectiveness**, **cost-utility**, **icer-net-monetary-benefit-rwe**, **budget-impact**, and (for long-run modeling) **survival-extrapolation-hta-rwe**. *Implications / transferability:* **generalizability-transportability-external-validity-rwe** addresses whether the evidence carries to the decision context. *Applied note (claims/EHR/registry RWE):* when an HTA dossier leans on a claims-based comparative analysis, the checklist's transparency domains are satisfied only if the report names the database and its FFS/MA/commercial coverage limits, defines the phenotype/outcome algorithms and their validation, fixes time zero, states the comparator and confounding-adjustment strategy, reports attrition through the analytic funnel, and shows sensitivity/quantitative-bias analyses — i.e., the checklist is a transparency wrapper that the listed concepts fill in.", "primary_category": "Guideline", "tags": [ "guideline", "hta", "reporting", "transparency", "health-technology-assessment" ], "aliases": [ "INAHTA Checklist", "INAHTA HTA Checklist", "INAHTA Checklist for HTA Reports", "Checklist for HTA Reports", "Hailey Checklist" ], "applies_to_study_types": [ "cost_effectiveness", "cost_utility", "budget_impact", "systematic_review", "meta_analysis_rct", "meta_analysis_obs", "cer_observational" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "doi": "10.1017/S0266462303000011", "url": "https://doi.org/10.1017/S0266462303000011", "citation_text": "Hailey D. Toward transparency in health technology assessment: a checklist for HTA reports. International Journal of Technology Assessment in Health Care. 2003;19(1):1-7.", "year": 2003, "authors_short": "Hailey", "notes": "Canonical INAHTA-consensus statement describing the 17-item, 5-domain checklist and its intent as a transparency aid (explicitly not a scorecard)." }, { "role": "use", "url": "https://www.inahta.org/hta-tools-resources/briefs/", "citation_text": "INAHTA. INAHTA Briefs, Checklists & Impact Framework — Checklist for HTA Reports (maintained PDF and resources). International Network of Agencies for Health Technology Assessment.", "year": 2003, "authors_short": "INAHTA", "notes": "Maintained agency landing page hosting the current checklist PDF and related HTA reporting resources." }, { "role": "demonstrate", "doi": "10.1017/S0266462317004512", "url": "https://doi.org/10.1017/S0266462317004512", "citation_text": "Vukovic V, Favaretti C, Ricciardi W, de Waure C. Health technology assessment evidence on e-health/m-health technologies: evaluating the transparency and thoroughness. International Journal of Technology Assessment in Health Care. 2018;34(1):87-96.", "year": 2018, "authors_short": "Vukovic et al.", "notes": "Applied use of the INAHTA checklist to appraise the transparency of a body of HTA reports, illustrating how the domains surface reporting gaps." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "cost-effectiveness", "notes": "INAHTA wraps the report-level transparency around an HTA that includes a cost-effectiveness analysis; CHEERS governs the economic-evaluation reporting itself." }, { "relation_type": "applies_to", "target_slug": "cost-utility", "notes": "Same report-level transparency wrapper for cost-utility (QALY-based) analyses within an HTA." }, { "relation_type": "applies_to", "target_slug": "budget-impact", "notes": "HTA reports routinely integrate a budget-impact analysis; the checklist requires its assumptions and uncertainty be made transparent." }, { "relation_type": "applies_to", "target_slug": "systematic-review", "notes": "The evidence-synthesis component of an HTA; report the synthesis with PRISMA and wrap the overall report with INAHTA." }, { "relation_type": "see_also", "target_slug": "picots-framework-rwe", "notes": "Implements the checklist's scope/research-question domain (population, intervention, comparator, outcomes, timing, setting)." }, { "relation_type": "see_also", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "Implements the data-fitness expectation behind the checklist's methods/data-sources domain for claims/EHR/registry evidence." }, { "relation_type": "see_also", "target_slug": "target-trial-emulation", "notes": "Provides the causal scaffold a transparent HTA needs when real-world comparative effectiveness drives the assessment." }, { "relation_type": "see_also", "target_slug": "active-comparator-new-user", "notes": "Concrete design that makes a real-world comparison in an HTA causally interpretable (confounding by indication, time-zero alignment)." }, { "relation_type": "see_also", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Forces the report to state the estimand actually targeted and how intercurrent events were handled." }, { "relation_type": "see_also", "target_slug": "quantitative-bias-analysis-toolkit-rwe", "notes": "Makes residual-confounding and bias uncertainty explicit in the results/uncertainty domain." }, { "relation_type": "see_also", "target_slug": "survival-extrapolation-hta-rwe", "notes": "HTA-specific long-run modeling whose assumptions the checklist requires to be transparent." }, { "relation_type": "used_with", "target_slug": "claims-analysis", "notes": "When an HTA dossier rests on a claims-based analysis, the checklist's transparency domains apply directly to database choice, phenotypes, and sensitivity analyses." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "hta", "journal" ] }, { "slug": "index-score-operationalization-checklist-rwe", "name": "Index and Score Operationalization Checklist for RWE", "short_definition": "A reporting and implementation checklist for source-backed indices, scores, and code groupers used in RWE, including Charlson/Quan-Charlson, Elixhauser, CMS-HCC, NCI comorbidity, AHRQ CCS/CCSR, and ECOG.", "long_description": "**What it is** - This guideline is the checklist layer for source-backed indices, scores, and\ncode groupers used in RWE. Concepts explain what Charlson, Quan-Charlson, Elixhauser, CMS-HCC,\nNCI comorbidity, AHRQ CCS/CCSR, ECOG, and related scores are; this guideline states what must\nbe specified, archived, and reported when one of those derived variables is used in a protocol,\nSAP, manuscript, payer dossier, or regulatory appendix. It is intentionally separate from the\nconcept files because checklists belong in guideline records, while concept records carry the\noperational method.\n\n**When to use** - Use it whenever an index, score, risk-adjustment model, code grouper, or\nclinical scale becomes an analytic variable rather than a topic label. Typical uses include a\nQuan-Charlson baseline comorbidity score in claims, a CMS-HCC risk score in Medicare Advantage\nor payer work, an AHRQ CCS/CCSR grouping used to collapse diagnoses, an NCI comorbidity score in\nSEER-Medicare oncology cohorts, or ECOG performance status abstracted from EHR/registry data.\nApply it before data derivation starts, because version, lookback, source-field, and hierarchy\nchoices change the numeric output and can change effect estimates.\n\n**What it requires / checklist domains** - The common failure mode is under-specification:\n\"Charlson,\" \"HCC,\" \"CCSR,\" or \"ECOG\" is not enough. A reproducible report names the exact\nindex or grouper family, version or release, source data fields, code maps, diagnosis/procedure\npositions, assessment window, hierarchy rules, coefficient or point set, missingness handling,\nand intended analytic role. The checklist also requires source-backed definitions for each\ncomponent or category when a table of index definitions is shown. For payment or administrative\nmodels such as CMS-HCC, report the model year and do not mix risk-adjustment and causal\nconfounding language. For clinical scores such as ECOG, report timing relative to index date and\nwhether the value came from structured fields, abstraction, NLP, registry capture, or imputation.\n\n**When NOT to use - limitations and common misapplications** - Do not use this checklist as a\nclaim that the score is clinically valid in the current population. It verifies operational\ntransparency, not predictive performance, calibration, construct validity, or causal adequacy.\nDo not substitute a payment model for a clinical severity measure without labeling the\nlimitation. Do not compare scores across studies if versions, lookback windows, code systems, or\ncomponent hierarchies differ. Do not present area, payer, or registry proxies as individual\nclinical facts. It is actively misleading to say \"Quan-Charlson adjusted\" if the implementation\nused a different ICD map, omitted hierarchy rules, mixed ICD-9 and ICD-10 without a transition\nplan, or derived comorbidities during follow-up.\n\n**How it maps to this catalog** - This guideline cross-references the implementing concepts:\n`charlson-comorbidity-index-rwe` for Charlson and Quan administrative maps,\n`elixhauser-comorbidity-index-rwe` for Elixhauser indicators, `cms-hcc-risk-adjustment-rwe` for\nCMS-HCC model-year handling, `nci-comorbidity-index-seer-medicare-rwe` for SEER-Medicare\noncology comorbidity, `ahrq-ccs-ccsr-clinical-classifications-rwe` for AHRQ diagnosis grouping,\nand `ecog-performance-status-score-rwe` for oncology performance status. Use STaRT-RWE or\nRECORD-PE for the broader study report; use this checklist for the index-specific derivation\ntable and reproducibility appendix.", "primary_category": "Guideline", "tags": [ "guideline", "checklist", "indices", "scores", "comorbidity", "ecog", "hcc", "ccsr" ], "aliases": [ "index checklist", "score checklist", "comorbidity index checklist", "ECOG checklist", "HCC checklist", "CCSR checklist" ], "applies_to_study_types": [ "claims_analysis", "ehr_study", "registry_linkage", "comparative_effectiveness", "oncology_rwe" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "doi": "10.1136/bmj.m4856", "url": "https://doi.org/10.1136/bmj.m4856", "citation_text": "Wang SV, Pinheiro S, Hua W, et al. STaRT-RWE: structured template for planning and reporting on the implementation of real world evidence studies. BMJ. 2021;372:m4856.", "year": 2021, "authors_short": "Wang et al.", "notes": "Structured RWE planning/reporting template that supports explicit operational definitions, windows, code lists, and analysis variables." }, { "role": "explain", "url": "https://pubmed.ncbi.nlm.nih.gov/3558716/", "citation_text": "Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies. J Chronic Dis. 1987;40(5):373-383.", "year": 1987, "authors_short": "Charlson et al.", "notes": "Original Charlson score source." }, { "role": "explain", "doi": "10.1097/01.mlr.0000182534.19832.83", "url": "https://doi.org/10.1097/01.mlr.0000182534.19832.83", "citation_text": "Quan H, Sundararajan V, Halfon P, et al. Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data. Medical Care. 2005;43(11):1130-1139.", "year": 2005, "authors_short": "Quan et al.", "notes": "ICD-9/ICD-10 comorbidity code maps." }, { "role": "explain", "url": "https://www.cms.gov/medicare/payment/medicare-advantage-rates-statistics/risk-adjustment", "citation_text": "Centers for Medicare & Medicaid Services. Risk Adjustment.", "year": 2026, "authors_short": "CMS", "notes": "CMS-HCC model source materials." }, { "role": "explain", "url": "https://hcup-us.ahrq.gov/toolssoftware/ccsr/ccs_refined.jsp", "citation_text": "Agency for Healthcare Research and Quality. Clinical Classifications Software Refined (CCSR). Healthcare Cost and Utilization Project.", "year": 2026, "authors_short": "AHRQ", "notes": "AHRQ CCS/CCSR source materials." }, { "role": "explain", "url": "https://ecog-acrin.org/resources/ecog-performance-status/", "citation_text": "ECOG-ACRIN Cancer Research Group. ECOG Performance Status Scale.", "year": 2026, "authors_short": "ECOG-ACRIN", "notes": "ECOG score definitions." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "charlson-comorbidity-index-rwe", "notes": "Reporting checklist for Charlson / Quan-Charlson implementations." }, { "relation_type": "applies_to", "target_slug": "elixhauser-comorbidity-index-rwe", "notes": "Reporting checklist for Elixhauser indicators and scores." }, { "relation_type": "applies_to", "target_slug": "cms-hcc-risk-adjustment-rwe", "notes": "Reporting checklist for CMS-HCC model-year and coding-intensity use." }, { "relation_type": "applies_to", "target_slug": "nci-comorbidity-index-seer-medicare-rwe", "notes": "Reporting checklist for NCI/SEER-Medicare comorbidity implementation." }, { "relation_type": "applies_to", "target_slug": "ahrq-ccs-ccsr-clinical-classifications-rwe", "notes": "Reporting checklist for AHRQ CCS/CCSR grouping." }, { "relation_type": "applies_to", "target_slug": "ecog-performance-status-score-rwe", "notes": "Reporting checklist for ECOG measurement timing and source." } ], "index_definitions": [], "checklist_items": [ "Name the exact index, score, or grouper family, not only the shorthand label.", "Pin the version, release, model year, macro, mapping file, or coefficient set used.", "Define the source data fields, claim types, diagnosis/procedure positions, structured EHR fields, registry fields, and extraction method.", "Define the baseline, lookback, assessment, or capture window relative to index date.", "Apply and report hierarchy, severity-collapse, code grouping, or score-weight rules before analysis.", "State whether the output enters analysis as a continuous score, categorical score, component flags, grouped code features, or matching/stratification variable.", "Report missingness, unobservable periods, and source-specific capture problems by cohort arm or data source.", "Archive the code, mappings, input fields, and versioned source files needed to reproduce the derived score.", "Do not substitute a payment model, code grouper, or proxy score for a validated clinical construct without labeling the limitation." ], "regulatory_relevance": [ "fda", "hta", "cms" ] }, { "slug": "ipcw-reporting-diagnostics-checklist-rwe", "name": "IPCW Reporting and Diagnostics Checklist", "short_definition": "A checklist for reporting inverse probability of censoring weighting, including censoring definitions, weight models, positivity diagnostics, truncation, and variance estimation.", "long_description": "**What it is** - This guideline is the checklist layer for inverse probability of censoring\nweighting (IPCW). The concept explains the estimator; this guideline states what a study must\nreport when censoring weights are used to address informative censoring, treatment switching,\nartificial censoring, loss to follow-up, or protocol deviations. The goal is to make the\nidentifying assumptions, weight models, diagnostics, truncation, and variance estimation visible\nenough for a reviewer to judge whether the weighted analysis is credible.\n\n**When to use** - Use it for longitudinal cohort, target-trial emulation, per-protocol,\nas-treated, sustained-treatment, or dynamic-strategy analyses when censoring is plausibly related\nto future outcomes. Common triggers include treatment discontinuation, switching, add-on therapy,\ndisenrollment, site leakage, missing follow-up assessments, and artificial censoring introduced by\nclone-censor-weight designs. Apply the checklist before fitting the model, because the numerator,\ndenominator, time scale, covariate history, and truncation rules should be pre-specified.\n\n**What it requires / checklist domains** - Define each censoring process separately:\nadministrative study end, loss of observability, treatment switch, protocol deviation, death,\ncare outside the system, and missing assessment. State whether death is censoring, an endpoint,\na competing event, or part of a composite estimand. Specify numerator and denominator models for\nstabilized weights, the time grid, covariate history, and time-varying predictors of both\ncensoring and outcome. Report weight mean, range, percentiles, effective sample size, and\npractical positivity violations. Pre-specify truncation or winsorization thresholds and report\nthe fraction of records affected. Use robust or bootstrap variance because weights are estimated.\n\n**When NOT to use - limitations and common misapplications** - Do not use IPCW as a cosmetic fix\nfor unmeasured reasons for loss to follow-up; the exchangeability assumption still depends on\nmeasured predictors. Do not censor death as if it were routine loss to follow-up when the estimand\nrequires a competing-risk, composite, or principal-stratum strategy. Do not hide extreme weights,\npositivity failure, or truncation behind a single adjusted estimate. Do not fit censoring models\nusing post-censoring information or variables affected by future treatment unless the estimand and\ntime ordering justify it. IPCW can reduce bias from informative censoring, but unstable weights can\nincrease variance and sensitivity to model misspecification.\n\n**How it maps to this catalog** - This guideline cross-references\n`inverse-probability-of-censoring-weighting-rwe` for the estimator,\n`censoring-mechanisms-rwe` for censoring taxonomy, `clone-censor-weight-per-protocol` for the\ncommon target-trial use case, `attrition-and-loss-to-follow-up-rwe` for follow-up diagnostics,\n`missing-data-pattern-table-rwe` for assessment missingness, `estimands-ate-att-intercurrent-events-rwe`\nfor the estimand decision, and `positivity-overlap` concepts through the weighting and propensity\nscore family where relevant. Use this checklist as the reporting gate around the IPCW concept.", "primary_category": "Guideline", "tags": [ "guideline", "checklist", "ipcw", "censoring", "weights", "positivity" ], "aliases": [ "IPCW checklist", "censoring weights checklist", "stabilized weights checklist" ], "applies_to_study_types": [ "cohort_retrospective", "cohort_prospective", "target_trial_emulation", "comparative_effectiveness" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "doi": null, "url": "https://pubmed.ncbi.nlm.nih.gov/10985216/", "citation_text": "Robins JM, Finkelstein DM. Correcting for noncompliance and dependent censoring in an AIDS clinical trial with inverse probability of censoring weighted (IPCW) log-rank tests. Biometrics. 2000;56(3):779-788.", "year": 2000, "authors_short": "Robins and Finkelstein", "notes": "Canonical IPCW source." }, { "role": "explain", "doi": "10.1093/aje/kwn164", "url": "https://doi.org/10.1093/aje/kwn164", "citation_text": "Cole SR, Hernan MA. Constructing inverse probability weights for marginal structural models. American Journal of Epidemiology. 2008;168(6):656-664.", "year": 2008, "authors_short": "Cole and Hernan", "notes": "Weight construction and diagnostics." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "inverse-probability-of-censoring-weighting-rwe", "notes": "Checklist for IPCW implementation and reporting." } ], "index_definitions": [], "checklist_items": [ "Define censoring reasons separately, including administrative end, disenrollment, treatment switch, protocol deviation, care leakage, loss to follow-up, and death.", "Do not treat death as ordinary censoring when the estimand requires a competing-risk or composite-event strategy.", "Specify numerator and denominator models for stabilized censoring weights.", "Include measured predictors of both censoring and outcome, including time-varying disease severity and utilization when available.", "Inspect and report weight mean, range, percentiles, effective sample size, and practical positivity violations.", "Pre-specify truncation or winsorization thresholds and report the proportion of records affected.", "Use robust or bootstrap variance because censoring weights are estimated." ], "regulatory_relevance": [ "fda", "ema", "hta" ] }, { "slug": "isoqol-standards", "name": "ISOQOL Minimum Standards for PRO Measures", "short_definition": "A consensus standard from the International Society for Quality of Life Research specifying the minimum evidence a patient-reported outcome (PRO) measure must demonstrate — a documented conceptual and measurement model, reliability, content and construct validity, responsiveness, interpretability of scores, and acceptable respondent burden — before it is fit to support patient-centered and comparative-effectiveness research.", "long_description": "**What it is.** The **ISOQOL Minimum Standards for PRO Measures** are a consensus statement issued by the **International\nSociety for Quality of Life Research (ISOQOL)** (Reeve et al., *Quality of Life Research*, 2013) that defines the floor of\nmeasurement evidence a patient-reported outcome (PRO) instrument must meet to be credible in patient-centered outcomes\nresearch (PCOR) and comparative-effectiveness research (CER). It is a *measure-appraisal* standard — it tells you whether a\nquestionnaire is good enough to use — and is the PRO-science complement to instrument-property frameworks such as **COSMIN**\nand to regulatory PRO guidance from the **FDA** and **EMA**. It is not a study-reporting checklist and not a risk-of-bias\ntool for a completed analysis. ISOQOL maintains the standard and the broader PRO methods literature through its task forces;\nthe standard is operationalized alongside the COSMIN taxonomy of measurement properties (Mokkink et al., 2010) and the\nISPOR good-practice report for translation and cultural adaptation (Wild et al., 2005).\n\n**When to use.** Apply ISOQOL Minimum Standards whenever a PRO, HRQoL, symptom, or functional-status instrument carries\nevidentiary weight in a decision: selecting or pre-specifying an endpoint instrument for a non-interventional or hybrid RWE\nstudy, an HTA/payer dossier, or a peer-reviewed submission; appraising whether an existing measure embedded in EHR-, claims-\nlinked-, or registry-collected data is fit for the intended population and use; or documenting instrument quality in a study\nprotocol. Decision rule for which standard governs: use **ISOQOL Minimum Standards** to judge *the measure itself*; use\n**COSMIN** when you need a granular, per-property methodological-quality rating of the validation *studies* behind a measure;\nuse **FDA/EMA PRO guidance** when the PRO is a labeling or registration endpoint (those guidances impose the additional bar\nof fit-for-purpose qualification and content validity in the *target* context of use). ISOQOL is the right tool when the\nquestion is \"is this instrument scientifically adequate for my population and decision,\" not \"did I report my trial\ncorrectly\" (that is a CONSORT-PRO / SPIRIT-PRO question) and not \"is my observational analysis unbiased\" (that is\nRECORD/STROBE plus a risk-of-bias instrument).\n\n**What it requires.** The standard enforces a small set of substantive domains, each of which must be documented for the\n*specific population and application* at hand, not inherited from the instrument's original development sample:\n- **Conceptual and measurement model** — an explicit statement of the construct measured, its dimensionality, the items\n mapped to each domain, scoring/aggregation rules, and the recall period. This is the PRO analogue of design transparency:\n a measure with no documented model cannot be defended.\n- **Reliability** — internal consistency and, where scores are used over time, test–retest reliability; for clinician- or\n observer-completed instruments, inter-rater reliability.\n- **Content validity** — evidence (typically qualitative, from the intended patient population) that items are relevant,\n comprehensible, and comprehensive for the concept and context of use. ISOQOL and the regulators treat this as the\n foundational property; a psychometrically clean but conceptually irrelevant measure still fails.\n- **Construct validity** — convergent/divergent and known-groups evidence that scores behave as theory predicts.\n- **Responsiveness** — ability to detect change over time when the underlying construct changes, essential for any\n longitudinal RWE endpoint.\n- **Interpretability of scores** — meaning attached to scores and to score changes, including a defensible threshold for\n *meaningful within-patient change* (the construct that has largely supplanted distribution-only MCID heuristics).\n- **Translation and cultural validity** — for any non-source-language administration, documented translation and\n cross-cultural adaptation (the ISPOR Wild et al. process) and measurement-invariance evidence so scores are comparable\n across language groups.\n- **Acceptable respondent and administrative burden** — completion time, missing-data behavior, and mode-of-administration\n equivalence (paper vs ePRO) appropriate to the data-collection setting.\n\n**When NOT to use — limitations and common misapplications.**\n- **It is not a risk-of-bias instrument and not a quality *score*.** The standards are pass/fail minima for a *measure*;\n they do not grade the internal validity of the *study* that used the measure, and \"8 of 8 standards met\" is not a numeric\n quality rating to rank instruments.\n- **Meeting the standards does not make an observational study causal.** A perfectly validated PRO endpoint analyzed in a\n confounded, immortal-time-ridden cohort still yields a biased treatment effect. ISOQOL adequacy is necessary, never\n sufficient.\n- **Validity is contextual, not a property stamped on the instrument.** A measure validated in trial populations may fail\n in a frailer, multimorbid, or differently literate real-world population; reusing a measure off-the-shelf because it \"is\n validated\" — without checking the conceptual model, language, and responsiveness in *your* population — is the most common\n misapplication.\n- **Wrong tool for the job:** using ISOQOL to appraise a clinician-rated efficacy outcome that is not patient-reported;\n using it where the decision actually needs FDA fit-for-purpose qualification for a labeling claim; or substituting it for\n CONSORT-PRO/SPIRIT-PRO reporting completeness. Checklist-as-theater — citing the standard without producing the\n population-specific content-validity and responsiveness evidence it demands — is rejected by knowledgeable reviewers.\n- **Single-source reliance.** A standard cannot rescue a PRO that is structurally missing in the data source (e.g.,\n administrative claims have essentially no PROs); applicability is gated by whether the construct was actually captured.\n\n**How it maps to this catalog.** ISOQOL governs the *endpoint*; the catalog concepts govern the *design and analysis* that\nmust be in place around an adequate PRO for the evidence to be credible:\n- **time-zero alignment, eligibility, and incident exposure** → `active-comparator-new-user` and `target-trial-emulation`:\n a validated, responsive PRO is only interpretable if measured from a clean, common baseline with comparable arms.\n- **estimand and intercurrent events** → `estimands-ate-att-intercurrent-events-rwe`: PRO endpoints are acutely sensitive\n to intercurrent events (death, treatment discontinuation, rescue therapy) and to terminal-event/missing-data handling;\n the estimand must be specified before the PRO contrast means anything.\n- **attrition and missing data** → `attrition-and-loss-to-follow-up-rwe`: longitudinal PROs are dominated by informative\n dropout; the ISOQOL responsiveness and burden domains presuppose that missingness is addressed analytically.\n- **confounding control** → `high-dimensional-propensity-score-hdps-rwe`: an unbiased PRO comparison still needs baseline\n confounding controlled like any other outcome.\n- **identifying the eligible population in coded data** → `diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe`: the\n condition cohort in which a PRO is fielded is typically defined by a coded phenotype, which must be validated.\n- **data fitness for use** → `claims-analysis`: a reminder that the dominant RWE data source carries no PROs, so ISOQOL\n applicability depends on EHR-, registry-, or de-novo-collected instruments, often via record linkage.\n\n**Applied note (claims/EHR/registry RWE).** Claims data contain no PROs, so an ISOQOL appraisal applies only where PROs are\nactually collected — disease registries, EHR-embedded questionnaires (e.g., PROMIS short forms in oncology or rheumatology),\nor prospectively fielded ePRO. In these settings the binding questions are: (1) does the instrument's documented conceptual\nmodel and content validity hold in *this* real-world population, which is usually older, more comorbid, and more variably\nliterate than the development sample; (2) is the language/translation validated for the administered population\n(Wild/ISPOR); (3) is mode-of-administration equivalence established if collection mixed paper and ePRO; and (4) is the\nmeasure responsive over the actual observation cadence, given that registry/EHR PRO capture is irregular and\nvisit-driven — irregular timing interacts directly with the attrition and estimand concerns above and can masquerade as a\nresponsiveness failure.", "primary_category": "Guideline", "tags": [ "guideline", "patient-reported-outcomes", "hrqol", "measure-validation", "psychometrics", "endpoints" ], "aliases": [ "ISOQOL Standards", "ISOQOL Minimum Standards", "ISOQOL minimum standards for patient-reported outcome measures", "Reeve 2013 PRO minimum standards" ], "applies_to_study_types": [ "pro_development", "pro_validation", "pro_rwe", "hrqol" ], "data_sources": [ "ehr", "registry", "linked", "primary" ], "citations": [ { "role": "introduce", "doi": "10.1007/s11136-012-0344-y", "url": "https://doi.org/10.1007/s11136-012-0344-y", "citation_text": "Reeve BB, Wyrwich KW, Wu AW, et al. ISOQOL recommends minimum standards for patient-reported outcome measures used in patient-centered outcomes and comparative effectiveness research. Quality of Life Research. 2013;22(8):1889-1905.", "year": 2013, "authors_short": "Reeve et al.", "notes": "Canonical ISOQOL consensus statement defining the minimum measurement standards (conceptual/measurement model, reliability, validity, responsiveness, interpretability, translation, burden) for PRO measures in PCOR/CER." }, { "role": "explain", "doi": "10.1007/s11136-010-9606-8", "url": "https://doi.org/10.1007/s11136-010-9606-8", "citation_text": "Mokkink LB, Terwee CB, Patrick DL, et al. The COSMIN checklist for assessing the methodological quality of studies on measurement properties of health status measurement instruments: an international Delphi study. Quality of Life Research. 2010;19(4):539-549.", "year": 2010, "authors_short": "Mokkink et al.", "notes": "Provides the standardized taxonomy and definitions of measurement properties that operationalize and granularly rate the validation studies behind a measure judged against the ISOQOL minima." }, { "role": "use", "doi": "10.1111/j.1524-4733.2005.04054.x", "url": "https://doi.org/10.1111/j.1524-4733.2005.04054.x", "citation_text": "Wild D, Grove A, Martin M, et al. Principles of good practice for the translation and cultural adaptation process for patient-reported outcomes (PRO) measures: report of the ISPOR Task Force for Translation and Cultural Adaptation. Value in Health. 2005;8(2):94-104.", "year": 2005, "authors_short": "Wild et al.", "notes": "ISPOR good-practice process that operationalizes the ISOQOL translation/cross-cultural-validity standard for any non-source-language administration in real-world data." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "pro-development", "notes": "Sets the minimum evidence a newly developed PRO measure must accumulate (conceptual model, content validity, reliability) before use." }, { "relation_type": "applies_to", "target_slug": "pro-validation", "notes": "Defines the appraisal floor when validating or re-validating a PRO measure in a new population or context of use." }, { "relation_type": "applies_to", "target_slug": "pro-rwe", "notes": "Governs whether a PRO endpoint fielded in registry/EHR-based real-world studies is fit for purpose." }, { "relation_type": "applies_to", "target_slug": "hrqol", "notes": "Applies to HRQoL instruments used as endpoints or descriptors in observational and comparative-effectiveness work." }, { "relation_type": "used_with", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "PRO endpoints require an explicit estimand and intercurrent-event strategy (death, discontinuation, rescue) before a between-arm PRO contrast is interpretable." }, { "relation_type": "used_with", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Longitudinal PROs are dominated by informative dropout; the responsiveness and burden domains presuppose principled missing-data handling." }, { "relation_type": "see_also", "target_slug": "target-trial-emulation", "notes": "A validated, responsive PRO is only interpretable when measured from a common, well-defined time zero with comparable arms." }, { "relation_type": "see_also", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "An adequate PRO endpoint still needs baseline confounding controlled like any other outcome." }, { "relation_type": "see_also", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "The condition cohort in which a PRO is fielded is usually defined by a validated coded phenotype." }, { "relation_type": "see_also", "target_slug": "claims-analysis", "notes": "Administrative claims carry no PROs, so ISOQOL applicability depends on EHR-, registry-, or de-novo-collected instruments, often via linkage." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "hta", "journal" ] }, { "slug": "ispe-scope", "name": "ISPE Self-Controlled Designs Guidance (Cadarette 2021)", "short_definition": "ISPE-endorsed methodological guidance on the design, conduct, and reporting of self-controlled study designs in pharmacoepidemiology — the self-controlled case series (SCCS), case-crossover, case-time-control, and case-case-time-control designs — in which each person serves as their own control so that all time-invariant confounding is removed by construction.", "long_description": "**What it is.** \"Control yourself: ISPE-endorsed guidance in the application of self-controlled study designs in\npharmacoepidemiology\" (Cadarette et al., *Pharmacoepidemiology and Drug Safety*, 2021), endorsed by the\nInternational Society for Pharmacoepidemiology (ISPE), is a methodological reference — not a one-page reporting\nchecklist — that establishes common language, a structured worksheet, and best-practice recommendations for the\nfamily of **self-controlled (within-person) designs**: the self-controlled case series (SCCS), the case-crossover\n(CCO) design, the case-time-control (CTC) design, and the case-case-time-control (CCTC) design. Its organizing\ninsight is that these designs make the *case its own control*, so every **time-invariant confounder** — measured or\nunmeasured, including stable genetics, chronic comorbidity, sex, and durable lifestyle — is eliminated by the design\nitself rather than by adjustment. The guidance walks the analyst through the assumptions each design depends on\n(transient exposure, abrupt outcome, exposure independent of the event, event independent of observation) and how\nthe choice among the four designs follows from which of those assumptions hold. It is maintained by ISPE as part of\nits methods-guidance output and is widely cited as the field's reference statement on when and how to use\nself-controlled designs. (This catalog refers to it descriptively as the ISPE self-controlled-designs guidance; \"SCOPE\" is not an official name or acronym for it and is avoided to prevent confusion with the EU SCOPE Joint Action in pharmacovigilance.)\n\n**When to use.** Reach for this guidance whenever the scientific question concerns the **transient effect of an\nintermittent or short-duration exposure on the risk of an abrupt, well-dated acute event** — vaccine safety\n(seizures, intussusception, myocarditis after a dated dose), acute drug toxicity (NSAIDs and GI bleed, antibiotics\nand arrhythmia), or any \"did exposure raise risk in the hours-to-weeks after it occurred\" question — using claims,\nEHR, registry, or linked routinely collected data destined for an FDA/EMA submission, an imposed or voluntary PASS,\nan HTA/payer safety narrative, or a peer-reviewed journal. The within-person comparison is most valuable precisely\nwhen **between-person confounding by indication or frailty is severe and hard to measure**, because it sidesteps the\nneed for a comparator cohort entirely. Decision rules among the siblings: use **SCCS** when complete exposure and\nevent histories are observable across a defined window and the rate of events is the estimand; use **case-crossover**\nwhen you have cases only and want to compare each case's exposure in a hazard window against earlier referent\nwindow(s); add **case-time-control** when an underlying **time trend in exposure prevalence** would bias the\ncase-crossover (a separate control series estimates and removes that trend); and use **case-case-time-control** when\nthat exposure-trend control itself risks **overadjustment** because the future-case control series shares the cases'\ncharacteristics. This is the design family to consult *instead of* a cohort/active-comparator guidance when the\ncomparison is within-person.\n\n**What it requires.** The guidance enforces the substantive elements that determine whether a self-controlled\nanalysis is valid, framed for real-world data: (1) a clear, dated, intermittent **exposure definition** with explicit\nstart/stop, and an **induction/latency and risk (hazard) window** justified on pharmacology rather than convenience;\n(2) an **abrupt, accurately dated outcome** with a validated phenotype/algorithm and a defined deduplication rule so\nrecurrent or carried-forward codes are not double-counted; (3) explicit treatment of the **observation period** and\nof person-time partitioned into risk versus control windows; (4) a stated and defended position on the two design-\nbreaking dependencies — **event-dependent exposure** (does the event itself change future exposure probability, as a\nstroke stops anticoagulation?) and **event-dependent observation** (is the event fatal or does it truncate\nfollow-up?), invoking the appropriate SCCS extension when either holds; (5) handling of **time-varying confounders\nwithin the person** (age, calendar season, disease progression) by modeling or by escalating to CTC/CCTC; and (6)\npre-specified **sensitivity and bias diagnostics** — alternative risk-window lengths, negative-control exposures,\nand washout/clean-period definitions. It is fundamentally a transparency-and-assumptions instrument: it forces the\nanalyst to state which design was chosen and why, and to demonstrate that the design's assumptions are met.\n\n**When NOT to use — limitations and common misapplications.** This is methodological guidance for within-person\ndesigns, not a universal RWE quality score, and it does not certify causality; meeting its recommendations does not\nrescue a study whose design assumptions are violated. Concrete failure modes: (a) **the outcome is cumulative or\nchronic** (cancer, atherosclerosis, fibrosis) rather than abrupt — the transient-effect logic does not hold and a\nself-controlled design is the wrong tool; (b) **event-dependent exposure** — the acute event changes the chance of\nfuture exposure, biasing a standard SCCS unless the event-dependent-exposure extension is used; (c) **event-dependent\nobservation** — the outcome is fatal or terminates observation, requiring Farrington's extension or a different\ndesign; (d) **uncontrolled within-person time trends** in exposure or background risk, which a plain case-crossover\nwill absorb as spurious association unless escalated to case-time-control / case-case-time-control; (e) **induction\nor latency longer than the observation window**, so the relevant exposure cannot be captured; and (f) **using a self-\ncontrolled design when the real question is a between-person comparison** (drug A vs drug B effectiveness, absolute\nrisk in the population) — for that, a cohort/active-comparator design is required and this guidance does not apply.\nPicking the wrong sibling (e.g., case-crossover where a time trend mandates CTC) is itself a misapplication this\nguidance is written to prevent.\n\n**How it maps to this catalog.** The four designs this guidance governs are implemented directly by `self-controlled-case-series`,\n`case-crossover`, `case-time-control`, and `case-case-time-control` — consult those for estimators, code, and worked\nexamples. The central risk-window requirement is implemented by `exposure-lag-induction-latency-window-rwe` (induction\nand latency) and `as-treated-risk-window-construction-rwe` (hazard-window construction). Event handling maps to\n`acute-event-deduplication-window-rwe` (one-event-per-episode rules) and `recurrent-events-analysis-rwe`. The dated,\nvalidated outcome that self-controlled designs depend on is delivered by `diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe`\nand `claims-outcome-algorithm-ppv-sensitivity-rwe`. Bias diagnostics map to `negative-control-exposures-rwe`, and the\nclean-period logic to `washout-clean-lookback-period-rwe`. For the operational substrate, `claims-analysis` covers the\nroutinely collected data these designs run on. **Applied note (claims/EHR/registry RWE):** in claims, the exposure\ndate is the pharmacy fill or administration date and the hazard window is built in `days_supply` space, so deduplicate\nacute outcome codes (e.g., a single seizure may generate multiple inpatient + outpatient claims within days) before\npartitioning person-time; in vaccine-safety EHR/registry work, the dated dose makes SCCS especially clean, but confirm\nthe event did not stop or alter subsequent dosing (event-dependent exposure) and that fatal events are handled with the\nappropriate extension rather than silently dropped.", "primary_category": "Guideline", "tags": [ "guideline", "methodological", "self-controlled-designs", "pharmacoepidemiology", "sccs", "case-crossover", "rwe" ], "aliases": [ "ISPE self-controlled designs guidance", "Control yourself (Cadarette 2021)", "Cadarette 2021 self-controlled designs" ], "applies_to_study_types": [ "self_controlled_case_series", "case_crossover", "case_time_control", "case_case_time_control" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "doi": "10.1002/pds.5227", "url": "https://doi.org/10.1002/pds.5227", "citation_text": "Cadarette SM, Maclure M, Delaney JAC, et al. Control yourself: ISPE-endorsed guidance in the application of self-controlled study designs in pharmacoepidemiology. Pharmacoepidemiology and Drug Safety. 2021;30(6):671-684.", "year": 2021, "authors_short": "Cadarette et al.", "notes": "Canonical ISPE-endorsed statement establishing common language, a worksheet, and assumption-driven design selection across the SCCS, case-crossover, case-time-control, and case-case-time-control designs." }, { "role": "explain", "doi": "10.1002/sim.2302", "url": "https://doi.org/10.1002/sim.2302", "citation_text": "Whitaker HJ, Farrington CP, Spiessens B, Musonda P. Tutorial in biostatistics: the self-controlled case series method. Statistics in Medicine. 2005;24(10):1563-1588.", "year": 2005, "authors_short": "Whitaker et al.", "notes": "Foundational tutorial on the SCCS estimator and its assumptions (transient exposure, event independent of exposure and of observation) that the guidance operationalizes." }, { "role": "explain", "doi": "10.1093/oxfordjournals.aje.a115853", "url": "https://doi.org/10.1093/oxfordjournals.aje.a115853", "citation_text": "Maclure M. The case-crossover design: a method for studying transient effects on the risk of acute events. American Journal of Epidemiology. 1991;133(2):144-153.", "year": 1991, "authors_short": "Maclure", "notes": "Original statement of the case-crossover design and the transient-exposure / abrupt-outcome logic that motivates the case-time-control and case-case-time-control refinements covered by the guidance." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "self-controlled-case-series", "notes": "Primary design governed by the guidance; consult the SCCS concept for the estimator, assumptions, and extensions for event-dependent exposure/observation." }, { "relation_type": "applies_to", "target_slug": "case-crossover", "notes": "Cases-only within-person design for transient effects on abrupt events; the guidance specifies referent-window and time-trend considerations." }, { "relation_type": "applies_to", "target_slug": "case-time-control", "notes": "Adds a control series to remove a time trend in exposure prevalence that would bias a case-crossover." }, { "relation_type": "applies_to", "target_slug": "case-case-time-control", "notes": "Refines case-time-control to avoid overadjustment when the control series shares case characteristics." }, { "relation_type": "requires", "target_slug": "exposure-lag-induction-latency-window-rwe", "notes": "Self-controlled validity hinges on a pharmacologically justified induction/latency window; this concept implements that specification." }, { "relation_type": "requires", "target_slug": "as-treated-risk-window-construction-rwe", "notes": "Implements the hazard (risk) window construction that partitions within-person time into risk vs control periods." }, { "relation_type": "used_with", "target_slug": "acute-event-deduplication-window-rwe", "notes": "Abrupt outcomes must be deduplicated to one event per episode before person-time is partitioned." }, { "relation_type": "used_with", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Supplies the dated, validated outcome phenotype the self-controlled comparison depends on." }, { "relation_type": "used_with", "target_slug": "claims-outcome-algorithm-ppv-sensitivity-rwe", "notes": "Quantifies outcome-algorithm PPV/sensitivity, since outcome misclassification biases self-controlled estimates." }, { "relation_type": "used_with", "target_slug": "negative-control-exposures-rwe", "notes": "Negative-control exposures are the standard bias diagnostic for residual within-person confounding." }, { "relation_type": "see_also", "target_slug": "recurrent-events-analysis-rwe", "notes": "Relevant when individuals contribute multiple events within the observation window." }, { "relation_type": "see_also", "target_slug": "claims-analysis", "notes": "Operational substrate (fill/administration dates, days_supply, code capture) on which these designs run." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "hta", "journal" ] }, { "slug": "ispor-bia", "name": "ISPOR Budget Impact Analysis Good Practices", "short_definition": "ISPOR Good Practices task force report on the principles, structure, and reporting of budget impact analyses (BIA) — the affordability counterpart to cost-effectiveness analysis, estimating the financial consequences for a specific budget holder of adopting a new health technology over a short (typically 1–5 year), undiscounted time horizon.", "long_description": "**What it is** — The **ISPOR Budget Impact Analysis (BIA) Good Practices** reports are consensus\nmethods guidance issued by the *International Society for Pharmacoeconomics and Outcomes Research\n(ISPOR)* through its Good Practices for Outcomes Research task forces. The current canonical statement is\nthe **2012 Budget Impact Analysis Good Practice II** report (Sullivan et al., *Value in Health*, 2014),\nwhich updates the original 2007 task force report (Mauskopf et al.). It defines what a budget impact\nanalysis is, the analytic framework it should follow, the inputs it requires, and how it must be\nreported. A BIA estimates the **financial consequences of adopting and diffusing a new intervention**\nwithin a specific population and a specific budget holder's accounts — typically a health plan, national\npayer, hospital, or integrated delivery system. It answers \"**can we afford this, and what will it do to\nour budget over the next few years?**\", which is a fundamentally different question from the\nvalue-for-money question answered by cost-effectiveness analysis (CEA). Most HTA submissions and payer\ndossiers require a BIA alongside (and distinct from) the economic evaluation.\n\n**When to use** — Use a BIA whenever a **decision maker who holds a budget** needs to understand the\nnear-term, total financial impact of a coverage, formulary, or adoption decision: payer P&T / formulary\nreviews, HTA affordability assessments (e.g., NICE budget impact test, ICER's potential-budget-impact\nanalysis), hospital pharmacy and value-analysis committees, and the budget-impact module of an AMCP-style\nor country-specific reimbursement dossier. The defining decision rules: (1) the perspective is the\n**specific budget holder's** (payer / health-system), not societal; (2) the time horizon is **short**\n(commonly 1–5 years) and reported **year by year, undiscounted**; (3) the comparator is the **current\ntreatment mix (\"world without\")** versus the **anticipated mix after adoption (\"world with\")**, not a\nsingle head-to-head comparator. Choose a **BIA** when the question is affordability and cash-flow\nplanning. Choose its sibling **cost-effectiveness / cost-utility analysis** when the question is whether\nthe technology is worth its price per unit of health gained — the two are complementary and a complete\ndossier usually contains both, but they are not interchangeable.\n\n**What it requires** — The guideline enforces a specific structure and set of reportable elements: (1) a\nclearly stated **perspective and budget holder**, with the matching cost inventory; (2) an explicit\n**eligible-population size** built from epidemiology (prevalence/incidence), diagnosis and treatment\nrates, and any eligibility restrictions — sized for the **open or closed** population as appropriate; (3)\nthe **current and projected treatment mix**, including realistic **uptake / market-diffusion curves**\nrather than instantaneous 100% switching; (4) **per-patient cost streams** disaggregated into\nintervention/drug acquisition costs, other medical costs, and **cost offsets** (e.g., avoided\nhospitalizations, displaced therapies); (5) a transparent **model structure** (a simple cost calculator,\nor a Markov/state-transition or discrete-event model when condition dynamics matter); (6) **scenario and\none-way sensitivity analyses** on the dominant drivers (population size, uptake, market share, costs),\nwith probabilistic sensitivity analysis optional rather than required; and (7) **disaggregated annual\nresults** — total and per-member-per-month (PMPM) where relevant — so the budget holder can trace the\ncost components. Crucially, costs and outcomes are **undiscounted** and the horizon is short, which\ndistinguishes BIA reporting from CEA.\n\n**When NOT to use — limitations and common misapplications** — A BIA is **not** a value assessment and\nmust never be used to make a value-for-money claim (\"our drug is cost-saving\" / \"cost-effective\") — that\nis the job of CEA/cost-utility analysis, and conflating the two is the single most common\nmisapplication. Specific failure modes: (a) applying a **lifetime horizon or discounting** — if you are\ndiscounting future costs you have likely drifted into building a CEA, not a BIA; (b) adopting a\n**societal perspective** and the wrong cost inventory instead of the budget holder's accounts; (c)\nassuming **instantaneous full uptake** instead of a realistic diffusion curve, which overstates\nearly-year impact; (d) reporting only a single **net cost** figure without disaggregated drug, medical,\nand offset line items, defeating the payer's need to interrogate the components; (e) ignoring that\n**eligible-population sizing and market-share uncertainty usually dominate** the result far more than\nunit-cost precision, yet are often left unexplored in sensitivity analysis; and (f) treating the BIA as a\nmarketing artifact rather than a budget-planning input. A BIA is also the **wrong tool** when the\ndecision turns on health outcomes per dollar (use CEA) or when no defined budget holder exists.\n\n**How it maps to this catalog** — The BIA framework is implemented through several concepts in this\ncatalog. Overall structure and model choice are covered by **health-economic-modeling-methods-rwe**, with\ndynamic structures in **markov-transition-probabilities-rwe** and **discrete-event-simulation-rwe** when\ncondition progression must be modeled. The cost inputs are operationalized through\n**healthcare-costs-pppm-pppy-pmpm** (per-patient and per-member cost streams, including the PMPM\nreporting BIA favors), **hcru-healthcare-resource-utilization** (resource-use parameters), and\n**all-cause-vs-attributable-costs-rwe** (choosing the correct cost basis for offsets). The\nshort-horizon, undiscounted convention is the key contrast captured in **discounting-costs-effects-rwe**\n— note that, unlike CEA, a BIA deliberately does **not** discount. Population transportability and uptake\nrealism map to **generalizability-transportability-external-validity-rwe**, and uncertainty handling maps\nto **probabilistic-sensitivity-analysis-hea-rwe** (with the caveat that one-way/scenario analysis, not\nfull PSA, is the BIA norm). The sibling-versus-distinct relationship is anchored to **cost-effectiveness**,\nand the decision-context concept is **budget-impact**. *Applied note for claims/EHR/registry RWE:* a BIA\nis a modeling exercise, but its most influential parameters — eligible-population size, current treatment\npatterns and market share, HCRU, and per-patient costs — are typically sourced from real-world claims,\nEHR, and registry data. The quality of a BIA therefore hinges on the fitness of those RWE inputs, even\nthough the BIA framework itself prescribes no study design.", "primary_category": "Guideline", "tags": [ "guideline", "budget-impact-analysis", "health-economics", "affordability", "payer", "hta", "ispor" ], "aliases": [ "ISPOR BIA", "ISPOR-BIA", "Budget Impact Analysis Good Practice II", "ISPOR BIA Good Practices", "BIA Good Practice II" ], "applies_to_study_types": [ "budget_impact" ], "data_sources": [ "claims", "ehr", "registry" ], "citations": [ { "role": "introduce", "doi": "10.1016/j.jval.2013.08.2291", "url": "https://doi.org/10.1016/j.jval.2013.08.2291", "citation_text": "Sullivan SD, Mauskopf JA, Augustovski F, et al. Budget impact analysis—principles of good practice: report of the ISPOR 2012 Budget Impact Analysis Good Practice II Task Force. Value in Health. 2014;17(1):5-14.", "year": 2014, "authors_short": "Sullivan et al.", "notes": "Current canonical ISPOR statement on BIA. Defines perspective, budget holder, short undiscounted horizon, population sizing, treatment-mix scenarios, disaggregated cost streams, and required sensitivity analyses; supersedes and refines the 2007 report." }, { "role": "explain", "doi": "10.1111/j.1524-4733.2007.00187.x", "url": "https://doi.org/10.1111/j.1524-4733.2007.00187.x", "citation_text": "Mauskopf JA, Sullivan SD, Annemans L, et al. Principles of good practice for budget impact analysis: report of the ISPOR Task Force on good research practices—budget impact analysis. Value in Health. 2007;10(5):336-347.", "year": 2007, "authors_short": "Mauskopf et al.", "notes": "Original 2007 task force report that the 2014 Good Practice II report updates; establishes the BIA versus cost-effectiveness distinction and the affordability framing." }, { "role": "use", "url": "https://www.ispor.org/heor-resources/good-practices-for-outcomes-research/report", "citation_text": "ISPOR Good Practices for Outcomes Research Reports — Budget Impact Analysis. ISPOR (International Society for Pharmacoeconomics and Outcomes Research).", "year": 2014, "authors_short": "ISPOR", "notes": "Maintained ISPOR landing page for the Good Practices task force reports, including the BIA guidance and supplementary materials." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "budget-impact", "notes": "The ISPOR BIA Good Practices report defines the structure and reporting expectations for any budget impact analysis; see the budget-impact concept for implementation detail." }, { "relation_type": "alternative_to", "target_slug": "cost-effectiveness", "notes": "BIA and cost-effectiveness analysis are complementary but distinct — BIA answers affordability over a short undiscounted horizon for a specific budget holder; CEA answers value for money per unit of health gained. Do not substitute one for the other." }, { "relation_type": "used_with", "target_slug": "health-economic-modeling-methods-rwe", "notes": "Provides the model structures (cost calculators, state-transition, simulation) that implement the BIA framework." }, { "relation_type": "used_with", "target_slug": "healthcare-costs-pppm-pppy-pmpm", "notes": "Supplies the per-patient and per-member cost streams (including the PMPM reporting BIA favors) that feed the budget impact calculation." }, { "relation_type": "used_with", "target_slug": "hcru-healthcare-resource-utilization", "notes": "Source of the resource-use parameters underlying medical-cost streams and offsets in a BIA." }, { "relation_type": "used_with", "target_slug": "all-cause-vs-attributable-costs-rwe", "notes": "Determines the correct cost basis for intervention costs and offsets within the budget holder's perspective." }, { "relation_type": "see_also", "target_slug": "discounting-costs-effects-rwe", "notes": "Key contrast — unlike CEA, BIA uses a short horizon and does NOT discount; this concept frames why discounting is appropriate for CEA but not BIA." }, { "relation_type": "see_also", "target_slug": "generalizability-transportability-external-validity-rwe", "notes": "Eligible-population sizing and uptake realism depend on transporting epidemiology and treatment patterns to the budget holder's population." }, { "relation_type": "see_also", "target_slug": "probabilistic-sensitivity-analysis-hea-rwe", "notes": "BIA relies primarily on scenario and one-way sensitivity analyses; full PSA is optional rather than required by the guideline." }, { "relation_type": "see_also", "target_slug": "markov-transition-probabilities-rwe", "notes": "State-transition structures support dynamic BIAs when condition progression drives costs over the horizon." }, { "relation_type": "see_also", "target_slug": "discrete-event-simulation-rwe", "notes": "Discrete-event simulation supports BIAs where patient-level event timing and resource use must be modeled explicitly." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "hta", "journal", "pqa-cms" ] }, { "slug": "ispor-conjoint", "name": "ISPOR Conjoint Analysis / Discrete-Choice Experiment Good Research Practices", "short_definition": "ISPOR's good-research-practices and reporting checklist series for conjoint analysis and discrete-choice experiments (DCEs) in health, covering attribute development, experimental design, preference elicitation, and the statistical analysis of stated-preference data.", "long_description": "**What it is** — The **ISPOR Conjoint Analysis / Discrete-Choice Experiment (DCE) Good Research Practices** is a\nthree-report series issued by task forces of ISPOR (the Professional Society for Health Economics and Outcomes\nResearch). It is the field's de facto standard for designing, conducting, analyzing, and reporting *stated-preference*\nstudies in health. The series comprises: (1) **Bridges et al. 2011** — a 10-item conjoint-analysis-applications\n*checklist* spanning the research question, attribute and level identification, choice-task and experimental design,\npreference-elicitation format, instrument design, data-collection plan, statistical analysis, and reporting of results\nand limitations; (2) **Johnson et al. 2013** — good practices for *constructing experimental designs* (full vs partial\nprofiles, opt-out/status-quo alternatives, D-efficiency, fractional-factorial and Bayesian designs, blocking, and the\nnumber of choice tasks); and (3) **Hauber et al. 2016** — good practices for the *statistical analysis* of DCE data\n(random-utility theory, conditional/multinomial logit, mixed/random-parameters logit, latent-class models, hierarchical\nBayes, and tests of dominance, transitivity, and monotonicity). ISPOR maintains the series among its good-practices\nreports for outcomes research. It is a design-and-reporting good-practice standard — not a regulatory mandate.\n\n**When to use** — Apply this series whenever the deliverable is a quantitative measurement of how patients, caregivers,\nclinicians, or the public *trade off* attributes of a health intervention — a DCE, best-worst scaling (object/profile/multiprofile),\nranking, or rating-based conjoint task. Typical decision contexts: generating patient-preference information for an\n**FDA benefit-risk / Patient Preference Information** submission; supplying preference weights to a **multi-criteria\ndecision analysis (MCDA)** or benefit-risk framework in an **HTA/payer dossier**; quantifying acceptable risk or\nwillingness-to-pay/willingness-to-accept for a value argument; and peer-reviewed publication of a stated-preference\nstudy. Decision rule for which report governs which step: use **Bridges 2011** as the overarching checklist and for\nattribute/instrument development and reporting; reach for **Johnson 2013** when specifying the experimental design and\nchoice-set generation; and apply **Hauber 2016** when choosing and reporting the choice model. This series governs\n*stated*-preference work; it is the wrong instrument for *revealed*-preference analyses of real-world choices captured\nin claims or EHR data, which fall under observational-study reporting guidelines (STROBE/RECORD-PE) and causal-inference\ndesigns instead.\n\n**What it requires** — The substantive domains the checklist enforces, framed for a defensible preference study:\n(1) a clear *research question* and decision context that the elicited preferences must inform; (2) *attribute\nidentification* grounded in qualitative formative work (literature review, patient/clinician interviews, focus groups)\nso the attribute set is ecologically valid and not analyst-imposed; (3) *attribute-level selection* that is plausible,\nnon-dominated, and spans a clinically meaningful range; (4) *choice-task construction* — alternatives per task, full\nvs partial profiles, and an explicit decision on opt-out/status-quo options; (5) *experimental design* with a stated\nefficiency criterion (D-efficiency, fractional-factorial or Bayesian designs), blocking, and a justified number of\nchoice tasks (Johnson 2013); (6) *preference-elicitation format* (DCE vs BWS vs ranking/rating) matched to the\ncognitive demands of respondents; (7) *instrument design* with cognitive pretesting, framing checks, and pilot testing;\n(8) a *data-collection plan* specifying mode, sampling frame, and sample-size/power justification; (9) *statistical\nanalysis* using a random-utility-consistent model, explicit testing for *preference heterogeneity* (mixed logit,\nlatent-class), and diagnostic checks for dominance, monotonicity, and attribute non-attendance (Hauber 2016); and\n(10) transparent *reporting* of results, internal/external validity, and limitations.\n\n**When NOT to use — limitations and common misapplications** — This is a *reporting and design good-practice* series,\nnot a **risk-of-bias instrument** and not a **quality score**: a fully completed Bridges checklist does not certify that\na DCE is unbiased or that its preference weights are valid. Concrete failure modes: (a) treating stated preferences as\nif they were *causal* descriptions of real-world behavior — the stated-vs-revealed-preference gap means a DCE predicts\nhypothetical choices, not market or adherence behavior; (b) skipping qualitative attribute development and imposing an\nanalyst-chosen attribute set, yielding ecologically invalid results no design efficiency can rescue; (c) reporting only\npooled marginal utilities or a single conditional-logit model without testing for preference heterogeneity (mixed logit\nor latent class), masking clinically important subgroups; (d) deriving willingness-to-pay from a cost attribute that was\npoorly specified or non-linear, producing unstable monetary estimates; (e) \"checklist-as-theater\" — ticking items in an\nappendix while the underlying design (dominated alternatives, implausible levels, no pretesting) is weak; and (f) using\nthe *wrong guideline family* — applying STROBE, RECORD-PE, or a causal observational-design framework to a stated-preference\nstudy, or conversely using this DCE series to govern a revealed-preference claims/EHR analysis.\n\n**How it maps to this catalog** — The implementing concept in this repository is **preference-study**, which carries the\noperational detail for DCE/BWS design and analysis that this guideline requires. Attribute identification and instrument\ndevelopment map to the patient-reported-outcome and qualitative concepts: **pro-development** and **pro-validation** (formative\ndevelopment and psychometric/instrument testing), **qualitative-interview** and **qualitative-synthesis** (the formative\nwork that grounds the attribute set), and **pro-rwe** for downstream real-world preference evidence. The data-collection-plan\nand sample-size requirements map to **sample-size-power-precision-rwe**. Preference weights that feed value and benefit-risk\narguments connect to **hrqol** (utility-adjacent valuation) and to the health-economic concepts when DCE outputs populate\nan **MCDA** or benefit-risk model. Applied note: in an HTA/payer dossier, DCE-derived attribute importance weights are most\noften consumed by an MCDA or structured benefit-risk framework rather than by a claims/EHR fitness-for-use assessment — so\nthe chain of evidence runs from qualitative attribute development (pro-development, qualitative-interview) through an\nefficient experimental design and a heterogeneity-aware choice model (preference-study) to a transparent benefit-risk or\nvalue narrative. For an FDA Patient Preference Information submission, the same chain must additionally document the maximum\nacceptable risk or risk-tolerance estimate and its uncertainty, which depends directly on the experimental-design and\nstatistical-analysis good practices in Johnson 2013 and Hauber 2016.", "primary_category": "Guideline", "tags": [ "guideline", "good-research-practice", "conjoint-analysis", "discrete-choice-experiment", "stated-preference", "patient-preference", "ispor", "heor" ], "aliases": [ "ISPOR DCE", "ISPOR Conjoint Analysis Good Research Practices", "ISPOR Discrete-Choice Experiment Good Practices", "Bridges 2011 conjoint checklist", "ISPOR Conjoint Analysis Task Force reports" ], "applies_to_study_types": [ "preference_study" ], "data_sources": [ "primary" ], "citations": [ { "role": "introduce", "doi": "10.1016/j.jval.2010.11.013", "url": "https://doi.org/10.1016/j.jval.2010.11.013", "citation_text": "Bridges JFP, Hauber AB, Marshall D, et al. Conjoint analysis applications in health—a checklist: a report of the ISPOR Good Research Practices for Conjoint Analysis Task Force. Value in Health. 2011;14(4):403-413.", "year": 2011, "authors_short": "Bridges et al.", "notes": "Canonical statement paper and the 10-item conjoint-analysis checklist; the overarching good-practice reference for attribute development, instrument design, and reporting of stated-preference studies in health." }, { "role": "explain", "doi": "10.1016/j.jval.2012.08.2223", "url": "https://doi.org/10.1016/j.jval.2012.08.2223", "citation_text": "Reed Johnson F, Lancsar E, Marshall D, et al. Constructing experimental designs for discrete-choice experiments: report of the ISPOR Conjoint Analysis Experimental Design Good Research Practices Task Force. Value in Health. 2013;16(1):3-13.", "year": 2013, "authors_short": "Johnson et al.", "notes": "Experimental-design companion report — full vs partial profiles, opt-out/status-quo alternatives, D-efficiency, fractional-factorial and Bayesian designs, blocking, and number of choice tasks." }, { "role": "explain", "doi": "10.1016/j.jval.2016.04.004", "url": "https://doi.org/10.1016/j.jval.2016.04.004", "citation_text": "Hauber AB, González JM, Groothuis-Oudshoorn CGM, et al. Statistical methods for the analysis of discrete choice experiments: a report of the ISPOR Conjoint Analysis Good Research Practices Task Force. Value in Health. 2016;19(4):300-315.", "year": 2016, "authors_short": "Hauber et al.", "notes": "Statistical-analysis companion report — random-utility theory, conditional/mixed/latent-class logit, hierarchical Bayes, and tests for dominance, transitivity, monotonicity, and preference heterogeneity." }, { "role": "use", "url": "https://www.ispor.org/heor-resources/good-practices-for-outcomes-research", "citation_text": "ISPOR Good Practices for Outcomes Research — maintained task-force reports including the Conjoint Analysis / DCE series.", "year": 2016, "authors_short": "ISPOR", "notes": "Maintained landing page for the ISPOR good-practices reports; use to confirm the current version of each conjoint/DCE task-force report." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "preference-study", "notes": "Governs the design, analysis, and reporting expectations for stated-preference (conjoint/DCE/BWS) studies." }, { "relation_type": "used_with", "target_slug": "preference-study", "notes": "The primary implementing concept; carries the operational detail for DCE/BWS design, choice-model estimation, and heterogeneity analysis the guideline requires." }, { "relation_type": "used_with", "target_slug": "pro-development", "notes": "Attribute and level identification depends on qualitative formative development, the same discipline used to develop patient-reported-outcome instruments." }, { "relation_type": "used_with", "target_slug": "pro-validation", "notes": "Instrument design and pretesting parallel PRO psychometric validation; cognitive testing and pilot work apply to the choice instrument." }, { "relation_type": "see_also", "target_slug": "qualitative-interview", "notes": "Interviews and focus groups are the standard formative method for generating and bounding the attribute set before any experimental design." }, { "relation_type": "see_also", "target_slug": "sample-size-power-precision-rwe", "notes": "The data-collection plan must justify sample size and precision for the choice model, including subgroups when heterogeneity is expected." }, { "relation_type": "see_also", "target_slug": "hrqol", "notes": "DCE-derived preference weights are utility-adjacent and often feed value, MCDA, or benefit-risk arguments alongside HRQoL/utility evidence." }, { "relation_type": "see_also", "target_slug": "pro-rwe", "notes": "Connects stated-preference evidence to downstream real-world patient-experience and preference data." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "hta", "journal" ] }, { "slug": "ispor-indirect", "name": "ISPOR Good Practices for Indirect Treatment Comparisons & Network Meta-Analysis", "short_definition": "The ISPOR Indirect Treatment Comparisons Good Research Practices Task Force reports (Jansen 2011 Part 1, Hoaglin 2011 Part 2) plus the 2014 ISPOR-AMCP-NPC questionnaire — the reference standard for conducting, reporting, and appraising indirect treatment comparisons (ITC) and network meta-analysis (NMA) for health-care decision making.", "long_description": "**What it is** — The **ISPOR Indirect Treatment Comparisons (ITC) Good Research Practices Task Force**\nreports, issued by the International Society for Pharmacoeconomics and Outcomes Research (ISPOR), are the\nfield's reference standard for indirect and mixed treatment comparisons and network meta-analysis (NMA). The\nguidance comes in two complementary parts. **Part 1 (Jansen et al., 2011)** explains *how to interpret* ITC/NMA:\nthe assumptions (similarity, homogeneity, consistency), the difference between anchored and unanchored\ncomparisons, fixed- vs random-effects models, and how to read relative-effect and ranking outputs. **Part 2\n(Hoaglin et al., 2011)** is the *conducting and reporting* companion: a checklist of items a credible ITC/NMA\nmust document, from the systematic-review base and network diagram through statistical model, software, and\npresentation of results. The later **ISPOR-AMCP-NPC ITC/NMA Study Questionnaire (Jansen et al., 2014)** turns\nthe principles into a structured **relevance-and-credibility appraisal instrument** that payers and HTA bodies\nuse to decide whether a submitted ITC/NMA can be trusted. Together these are the closest thing the indirect-comparison\nliterature has to STROBE/PRISMA-grade governance, and NICE, CADTH, and other HTA agencies cite them directly.\n\n**When to use** — Reach for this guidance whenever a decision requires comparing treatments that have **not been\nstudied head-to-head**, but are linked through a connected network of randomized trials sharing common comparators.\nTypical contexts: an **HTA/payer dossier** (NICE, CADTH, ICER, G-BA) positioning a new drug against comparators it\nwas never trialled against; a **regulatory submission** where an anchored indirect comparison supplements direct\nevidence; or a **peer-reviewed evidence synthesis** going beyond pairwise meta-analysis. Decision rule for *which*\ntool applies: (1) if a **connected network of RCTs with a shared common comparator** exists, an **anchored ITC/NMA**\nunder ISPOR Good Practices is appropriate and preserves within-trial randomization. (2) If the network is\n**disconnected, or the comparison rests on single-arm/observational data with imbalanced effect modifiers**, the\nanchored ISPOR framework no longer holds and you must move to **population-adjusted methods** (MAIC, STC) or external-control\napproaches — and say so explicitly. (3) For appraising a *submitted* ITC/NMA rather than building one, apply the\n**2014 questionnaire** as the credibility checklist.\n\n**What it requires** — The substantive domains the guidance enforces are specific to network evidence, not generic\nstudy design:\n- **Systematic-review foundation** — the network must be assembled from a transparent, reproducible systematic review\n (PRISMA-grade search, eligibility, extraction); the ITC is only as good as the trials feeding it.\n- **Network geometry and connectedness** — a presented network diagram, evidence that the network is connected, and\n disclosure of how multi-arm trials and closed loops are handled.\n- **Transitivity / similarity of effect modifiers** — the central assumption: trials across the network must be\n sufficiently similar in distribution of effect modifiers (patient characteristics, dosing, outcome definitions,\n follow-up) for indirect comparison to be valid; this must be assessed, not asserted.\n- **Homogeneity** within each pairwise contrast and **consistency** between direct and indirect evidence on every\n closed loop — with formal inconsistency assessment (node-splitting, loop-specific or design-by-treatment models).\n- **Model specification** — choice of fixed- vs random-effects, the likelihood/link, priors (for Bayesian fits),\n handling of multi-arm correlation, convergence diagnostics, and the software used, all pre-specified and reported.\n- **Presentation of results** — relative effects for all pairs with credible/confidence intervals, and **ranking\n metrics** (SUCRA, rank probabilities, P-scores) reported with explicit caveats about their fragility.\n- **Credibility and relevance appraisal** (2014 questionnaire) — whether the analysis answers the decision-maker's\n PICO, and whether its assumptions are defensible enough to act on.\n\n**When NOT to use — limitations and common misapplications** — These reports are **good-practice / credibility\nguidance, not a mechanical quality score**: a completed Part 2 checklist or 2014 questionnaire documents *what was\ndone*, it does not certify the answer is unbiased. The dominant failure modes: (1) **Running an NMA over a network\nthat violates transitivity** — pooling trials whose populations differ systematically in effect modifiers produces a\nprecise but biased estimate; the checklist is satisfied while the inference is wrong. (2) **Ignoring inconsistency** —\nreporting a consistency model without ever testing direct-vs-indirect agreement on closed loops. (3) **Over-reading\nrankings** — presenting SUCRA/\"probability best\" as if it were a robust ordering when ranks are unstable and sensitive\nto imprecise nodes. (4) **Using the anchored ISPOR framework where it does not apply** — forcing an NMA through a\n*disconnected* network, or onto **single-arm/external-control evidence with imbalanced effect modifiers**, where\npopulation-adjusted methods (MAIC/STC) or formal external-control designs are required instead. (5) **Checklist-as-theatre**\n— appending a completed questionnaire to a dossier while the underlying systematic review is incomplete or the model\nis unspecified. (6) Mistaking this guidance for one that governs **observational RWE design** (confounding control,\ntime-zero, phenotyping) — those belong to RECORD-PE/HARPER/STaRT-RWE, not ISPOR ITC.\n\n**How it maps to this catalog** — The implementing concepts in this repository are the evidence-synthesis and\nexternal-comparison methods, not pharmacoepi design tools:\n- The **network geometry, transitivity, consistency, and ranking** requirements are implemented by\n **network-meta-analysis** — the core analytic method this guidance governs.\n- The **systematic-review base and pairwise foundations** are implemented by **meta-analysis-rct** (the usual\n substrate for an anchored ITC) and **meta-analysis-obs** when observational evidence enters the synthesis;\n **ipd-meta-analysis** implements the individual-patient-data variant that strengthens effect-modifier adjustment.\n- The **transitivity / similarity assumption** — that effect-modifier distributions are exchangeable across the\n network — is conceptually implemented by **generalizability-transportability-external-validity-rwe**, which makes\n the exchangeability logic explicit.\n- The **decision boundary** (when the anchored ISPOR framework fails and you must adjust for population differences\n or lean on external comparators) maps to **single-arm-external-control** and **rare-disease-external-controls-rwe**;\n the trial-protocol discipline behind a credible comparator maps to **target-trial-emulation**.\n\n**Applied note (RWE-anchored ITC).** ITC/NMA is fundamentally trial-data-centric, but RWE increasingly feeds these\nnetworks — e.g., a real-world external-control arm or an observational study contributing a node. When that happens,\nthe transitivity assumption becomes far harder to defend: claims/EHR/registry cohorts differ from RCT populations in\neffect modifiers, outcome ascertainment, and follow-up. Document the RWD source's fitness-for-use, characterize how\nits population differs from the trial nodes, down-weight or sensitivity-test that node, and consider population-adjusted\nmethods before letting observational evidence drive an indirect comparison submitted to an HTA body.", "primary_category": "Guideline", "tags": [ "guideline", "indirect-treatment-comparison", "network-meta-analysis", "evidence-synthesis", "hta", "ispor" ], "aliases": [ "ISPOR ITC", "ISPOR ITC Good Research Practices", "ISPOR Good Practices for Indirect Treatment Comparisons", "ISPOR NMA Good Practices", "ISPOR Task Force on Indirect Treatment Comparisons" ], "applies_to_study_types": [ "network_meta_analysis", "meta_analysis_rct", "meta_analysis_obs" ], "data_sources": [ "primary", "registry" ], "citations": [ { "role": "introduce", "doi": "10.1016/j.jval.2011.04.002", "url": "https://doi.org/10.1016/j.jval.2011.04.002", "citation_text": "Jansen JP, Fleurence R, Devine B, et al. Interpreting indirect treatment comparisons and network meta-analysis for health-care decision making: report of the ISPOR Task Force on Indirect Treatment Comparisons Good Research Practices: Part 1. Value in Health. 2011;14(4):417-428.", "year": 2011, "authors_short": "Jansen et al.", "notes": "Canonical ISPOR Good Research Practices statement on interpreting ITC/NMA — defines the similarity, homogeneity, and consistency assumptions and the anchored vs mixed-comparison framework. DOI verified against Crossref." }, { "role": "explain", "doi": "10.1016/j.jval.2011.01.011", "url": "https://doi.org/10.1016/j.jval.2011.01.011", "citation_text": "Hoaglin DC, Hawkins N, Jansen JP, et al. Conducting indirect-treatment-comparison and network-meta-analysis studies: report of the ISPOR Task Force on Indirect Treatment Comparisons Good Research Practices: Part 2. Value in Health. 2011;14(4):429-437.", "year": 2011, "authors_short": "Hoaglin et al.", "notes": "The conducting-and-reporting companion to Part 1 — the checklist of items a credible ITC/NMA must document (network diagram, model, software, results presentation). DOI verified against Crossref." }, { "role": "use", "doi": "10.1016/j.jval.2014.01.004", "url": "https://doi.org/10.1016/j.jval.2014.01.004", "citation_text": "Jansen JP, Trikalinos T, Cappelleri JC, et al. Indirect treatment comparison/network meta-analysis study questionnaire to assess relevance and credibility to inform health care decision making: an ISPOR-AMCP-NPC Good Practice Task Force report. Value in Health. 2014;17(2):157-173.", "year": 2014, "authors_short": "Jansen et al.", "notes": "Structured relevance-and-credibility appraisal instrument used by payers and HTA bodies to judge whether a submitted ITC/NMA is trustworthy. DOI verified against Crossref." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "network-meta-analysis", "notes": "Primary scope — ISPOR ITC Good Practices govern the conduct, reporting, and appraisal of network meta-analysis." }, { "relation_type": "applies_to", "target_slug": "meta-analysis-rct", "notes": "An anchored ITC is built on a systematic review and pairwise RCT meta-analysis; this guidance governs how those are extended to indirect comparisons." }, { "relation_type": "used_with", "target_slug": "network-meta-analysis", "notes": "network-meta-analysis is the catalog concept that implements the network geometry, transitivity, consistency, and ranking requirements this guidance enforces." }, { "relation_type": "used_with", "target_slug": "meta-analysis-rct", "notes": "Implements the systematic-review and pairwise-synthesis foundation on which an anchored ITC/NMA is constructed." }, { "relation_type": "see_also", "target_slug": "ipd-meta-analysis", "notes": "Individual-patient-data synthesis strengthens effect-modifier adjustment and the transitivity assumption central to ITC/NMA." }, { "relation_type": "see_also", "target_slug": "generalizability-transportability-external-validity-rwe", "notes": "Makes explicit the exchangeability/effect-modifier logic underlying the transitivity assumption required for valid indirect comparison." }, { "relation_type": "see_also", "target_slug": "single-arm-external-control", "notes": "When the network is disconnected or evidence is single-arm, the anchored ISPOR framework no longer applies and external-control / population-adjusted methods are required instead." }, { "relation_type": "see_also", "target_slug": "target-trial-emulation", "notes": "Shares the discipline of pre-specifying a credible comparator and PICO before synthesis." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "hta", "journal" ] }, { "slug": "ispor-modeling", "name": "ISPOR-SMDM Modeling Good Research Practices", "short_definition": "The seven-part ISPOR-SMDM Modeling Good Research Practices Task Force series — the consensus reference for building, parameterizing, validating, and reporting decision-analytic health-economic models (state-transition/Markov, microsimulation, dynamic transmission, discrete-event) used in cost-effectiveness, cost-utility, and budget-impact analyses.", "long_description": "**What it is** — The **ISPOR-SMDM Modeling Good Research Practices** series is a seven-part\nset of consensus reports issued jointly in 2012 by the International Society for Pharmacoeconomics\nand Outcomes Research (ISPOR) and the Society for Medical Decision Making (SMDM), co-published\nin *Value in Health* and *Medical Decision Making*. It is the field's reference standard for\nthe **conduct and reporting of decision-analytic (health-economic) models**. The series is\norganized by task force: TF-1 Overview (Caro), TF-2 Conceptualizing the model (Roberts),\nTF-3 State-transition/Markov models (Siebert), TF-4 Discrete-event simulation (Karnon), TF-5\nDynamic transmission models (Pitman), TF-6 Parameter estimation and uncertainty (Briggs), and\nTF-7 Model transparency and validation (Eddy). It is a *good-practices and modeling-reporting*\nframework — not an observational-study reporting checklist (STROBE/RECORD) and not a risk-of-bias\ntool. Its modern companions are CHEERS 2022 (reporting the economic evaluation that wraps the\nmodel) and the AdViSHE validation-reporting tool; ISPOR maintains the series under its Good\nPractices program.\n\n**When to use** — Use this series whenever a **decision-analytic model** generates the\ncomparative cost and health-outcome estimates in an HTA/payer dossier (NICE, ICER, CADTH, G-BA,\nHAS), a manufacturer global value dossier, or a peer-reviewed cost-effectiveness/cost-utility,\ncost-benefit, or budget-impact analysis. The choice of *which* report governs is driven by\nmodel architecture: cohort **state-transition/Markov** model with memoryless health states →\nTF-3; need for patient history, time-since-event, or many interacting attributes that explode a\nMarkov state space → **microsimulation/DES** under TF-4; infectious-disease or herd-immunity\nquestions where one person's state changes another's risk (non-linear force of infection) →\n**dynamic transmission** under TF-5. TF-2 (conceptualization) and TF-6 (parameter uncertainty)\nand TF-7 (validation/transparency) apply to **every** model regardless of type. Distinguish from\nsiblings: use **CHEERS 2022** to report the surrounding economic evaluation and **ISPOR budget-impact\ngood-practices** for affordability/budget-impact specifics; ISPOR-Modeling governs the engine, not\nthe wrapper.\n\n**What it requires** — The series enforces good practice across the model lifecycle. **Problem\nconceptualization (TF-2):** an explicit decision problem, PICO/scope, perspective, time horizon,\nand a model structure justified against the disease process — structure follows the problem, not\nsoftware convenience. **Structure and assumptions (TF-3/4/5):** correct cycle length and\nhalf-cycle correction for Markov models; justification of the Markov (memoryless) assumption\nversus the need for individual histories in microsimulation/DES; valid handling of the force of\ninfection and contact structure in transmission models. **Parameter estimation and uncertainty\n(TF-6):** every input traceable to a source with a defensible distribution; deterministic one-way\nand scenario analyses plus **probabilistic sensitivity analysis (PSA)** with appropriate\ndistributional choices and correlation; uncertainty characterized as parameter, structural,\nand methodological. **Transparency and validation (TF-7):** non-technical and technical\ndocumentation sufficient for independent reproduction, plus the validation hierarchy — **face**\nvalidity, **internal/verification** (the model does what was intended; debugging, extreme-value/null\ntests), **cross-validation** against other models, **external** validation against data not used to\nbuild it, and **predictive** validation. Where real-world data supply inputs (incidence,\ntransition probabilities, costs, utilities, treatment effects), those inputs must themselves be\nfit-for-purpose, with transparent estimands and confounding control — the model inherits the\ncredibility of its parameters.\n\n**When NOT to use — limitations and common misapplications** — (1) This is a framework for\n*models*, not for primary observational analyses. Following ISPOR-Modeling perfectly does **not**\nvalidate the upstream effectiveness estimate fed into the model; if a hazard ratio from claims is\nconfounded, a beautifully validated Markov model launders bias into a decision. (2) It is **not a\nrisk-of-bias instrument or a quality score** — there is no numeric grade; a model can satisfy\nevery reporting item and still rest on an indefensible structure or cherry-picked inputs.\n(3) **Validation-as-theater:** reporting \"the model was validated\" without specifying *which*\nvalidation (face/internal/external/predictive) and showing the comparison is a frequent failure\nthat HTA reviewers reject. (4) **Wrong architecture:** forcing a cohort Markov model onto an\ninfectious-disease question (ignoring herd effects) or onto a problem requiring patient memory\nviolates TF-3/4/5 and produces structurally biased results. (5) **PSA omitted or cosmetic:**\npoint estimates with no probabilistic uncertainty, or PSA with arbitrary distributions and no\nparameter correlation, fail TF-6. (6) Do not substitute this series for **CHEERS** when the\ndeliverable is the economic-evaluation manuscript — they are complementary, not interchangeable.\n\n**How it maps to this catalog** — The model engine and its uncertainty/validation requirements\nmap to: **markov-transition-probabilities-rwe** (TF-3 state-transition structure, cycle length,\ntransition-probability estimation from RWD), **discrete-event-simulation-rwe** (TF-4 DES/individual\nsimulation), **partitioned-survival-models-rwe** and **survival-extrapolation-hta-rwe**\n(oncology model structures and the long-horizon extrapolation TF-2/TF-7 demand be justified and\nvalidated), **health-economic-modeling-methods-rwe** (umbrella modeling concept),\n**probabilistic-sensitivity-analysis-hea-rwe** (TF-6 PSA), **discounting-costs-effects-rwe**,\n**qaly-utility-mapping-rwe**, and **icer-net-monetary-benefit-rwe** (outputs the model produces);\naffordability questions route to **budget-impact**. For the *parameters* the model consumes, the\nfitness and causal-credibility requirements map to **fit-for-purpose-data-assessment-rwe**,\n**estimands-ate-att-intercurrent-events-rwe**, **high-dimensional-propensity-score-hdps-rwe** and\n**active-comparator-new-user** (defensible treatment-effect inputs), **target-trial-emulation**\n(when the effectiveness input comes from an emulated trial), **diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe**\nand **claims-analysis** (where incidence, event rates, and resource-use parameters are estimated\nfrom claims/EHR), and **attrition-and-loss-to-follow-up-rwe** (so survival and transition inputs\nare not distorted by informative censoring). The wrapping economic-evaluation report is governed by\n**cost-effectiveness**, **cost-utility**, and CHEERS.\n\n**Applied note (claims/EHR/registry RWE).** When transition probabilities, time-to-event curves,\ncosts (PPPM/PPPY), or utility decrements are estimated from routinely collected data and piped into\na Markov or partitioned-survival model, treat each input as a small RWE study in its own right:\npre-specify the phenotype and time-zero, control confounding for any comparative-effectiveness input,\nand carry that input's sampling uncertainty (and, ideally, structural uncertainty over alternative\nalgorithms) into the TF-6 PSA — not just a fixed mean. Document the data source's fitness and known\npayer-specific quirks (e.g., Medicare FFS vs MA capture) so a reviewer can trace every model number\nback to an observable, defensible source.", "primary_category": "Guideline", "tags": [ "guideline", "health-economic-modeling", "decision-analytic-model", "cost-effectiveness", "markov", "microsimulation", "model-validation", "ispor", "hta" ], "aliases": [ "ISPOR Modeling", "ISPOR-SMDM Modeling Good Research Practices", "ISPOR-SMDM Modeling Task Force", "Modeling Good Research Practices", "ISPOR-SMDM modeling guidelines" ], "applies_to_study_types": [ "cost_effectiveness", "cost_utility", "cost_benefit" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "doi": "10.1016/j.jval.2012.06.012", "url": "https://doi.org/10.1016/j.jval.2012.06.012", "citation_text": "Caro JJ, Briggs AH, Siebert U, Kuntz KM; ISPOR-SMDM Modeling Good Research Practices Task Force. Modeling good research practices—overview: a report of the ISPOR-SMDM Modeling Good Research Practices Task Force-1. Value in Health. 2012;15(6):796-803.", "year": 2012, "authors_short": "Caro et al.", "notes": "Series overview (Task Force-1) defining the seven-part framework and the modeling good-practices it spans; the canonical anchor for the whole series." }, { "role": "explain", "doi": "10.1016/j.jval.2012.06.016", "url": "https://doi.org/10.1016/j.jval.2012.06.016", "citation_text": "Roberts M, Russell LB, Paltiel AD, et al. Conceptualizing a model: a report of the ISPOR-SMDM Modeling Good Research Practices Task Force-2. Value in Health. 2012;15(6):804-811.", "year": 2012, "authors_short": "Roberts et al.", "notes": "Task Force-2 — defining the decision problem, scope, perspective, time horizon, and matching model structure to the question (structure follows the problem)." }, { "role": "explain", "doi": "10.1016/j.jval.2012.06.014", "url": "https://doi.org/10.1016/j.jval.2012.06.014", "citation_text": "Siebert U, Alagoz O, Bayoumi AM, et al. State-transition modeling: a report of the ISPOR-SMDM Modeling Good Research Practices Task Force-3. Value in Health. 2012;15(6):812-820.", "year": 2012, "authors_short": "Siebert et al.", "notes": "Task Force-3 — state-transition (Markov) models: state definition, cycle length, half-cycle correction, and the memoryless assumption versus microsimulation." }, { "role": "use", "doi": "10.1016/j.jval.2012.04.012", "url": "https://doi.org/10.1016/j.jval.2012.04.012", "citation_text": "Eddy DM, Hollingworth W, Caro JJ, Tsevat J, McDonald KM, Wong JB; ISPOR-SMDM Modeling Good Research Practices Task Force. Model transparency and validation: a report of the ISPOR-SMDM Modeling Good Research Practices Task Force-7. Value in Health. 2012;15(6):843-850.", "year": 2012, "authors_short": "Eddy et al.", "notes": "Task Force-7 — the validation hierarchy (face, internal/verification, cross-, external, predictive) and the transparency documentation HTA reviewers expect; the practical checklist most often cited at submission." }, { "role": "use", "url": "https://www.ispor.org/heor-resources/good-practices/article/modeling-good-research-practices---overview", "citation_text": "ISPOR Good Practices — Modeling Methods (ISPOR-SMDM Modeling Good Research Practices Task Force reports), maintained landing page.", "year": 2012, "authors_short": "ISPOR", "notes": "Maintained ISPOR portal linking all seven task-force reports (TF-1 through TF-7); use to retrieve the DES (TF-4), dynamic-transmission (TF-5), and parameter-uncertainty (TF-6) reports not individually cited above." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "cost-effectiveness", "notes": "Governs the decision-analytic model that produces incremental cost and effect estimates in a cost-effectiveness analysis." }, { "relation_type": "applies_to", "target_slug": "cost-utility", "notes": "Governs the model engine underlying QALY-based cost-utility analyses for HTA submission." }, { "relation_type": "applies_to", "target_slug": "cost-benefit", "notes": "Applies to decision-analytic models monetizing health outcomes in cost-benefit analyses." }, { "relation_type": "used_with", "target_slug": "markov-transition-probabilities-rwe", "notes": "TF-3 state-transition structure; this concept implements transition-probability estimation and cycle-length/half-cycle decisions." }, { "relation_type": "used_with", "target_slug": "discrete-event-simulation-rwe", "notes": "TF-4 individual-level / discrete-event simulation when a memoryless Markov structure is inadequate." }, { "relation_type": "used_with", "target_slug": "partitioned-survival-models-rwe", "notes": "Common oncology model architecture whose structure and extrapolation must be justified and validated under TF-2/TF-7." }, { "relation_type": "used_with", "target_slug": "survival-extrapolation-hta-rwe", "notes": "Long-horizon extrapolation feeding the model is exactly the kind of structural assumption TF-6/TF-7 require be characterized and validated." }, { "relation_type": "used_with", "target_slug": "probabilistic-sensitivity-analysis-hea-rwe", "notes": "Implements the TF-6 requirement for probabilistic uncertainty characterization (PSA)." }, { "relation_type": "used_with", "target_slug": "health-economic-modeling-methods-rwe", "notes": "Umbrella catalog concept for the modeling methods this guideline governs." }, { "relation_type": "see_also", "target_slug": "icer-net-monetary-benefit-rwe", "notes": "ICER / net-monetary-benefit are the decision outputs the validated model produces." }, { "relation_type": "see_also", "target_slug": "discounting-costs-effects-rwe", "notes": "Discounting of future costs and effects is a methodological choice the model must apply and report." }, { "relation_type": "see_also", "target_slug": "qaly-utility-mapping-rwe", "notes": "Utility/QALY inputs feed cost-utility models and inherit the fitness/uncertainty requirements of TF-6." }, { "relation_type": "see_also", "target_slug": "budget-impact", "notes": "Affordability/budget-impact questions are governed by ISPOR budget-impact good practices, a complementary (not interchangeable) framework." }, { "relation_type": "see_also", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "RWD inputs (transitions, costs, utilities, event rates) must be fit-for-purpose before they parameterize the model." }, { "relation_type": "see_also", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Comparative-effectiveness inputs require a clearly defined estimand; the model inherits the credibility of that estimand." }, { "relation_type": "see_also", "target_slug": "target-trial-emulation", "notes": "When the effectiveness parameter comes from observational data, an emulated target trial yields a more defensible model input." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "hta", "journal", "fda", "ema" ] }, { "slug": "ispor-rwd-questionnaire", "name": "ISPOR Questionnaire for Relevance and Credibility of Observational/RWD Studies", "short_definition": "ISPOR Good Practice Task Force questionnaire for appraising the relevance and credibility of observational and real-world data studies for health-care decision making, with companion ISPOR-ISPE good-practice and database-reporting frameworks for RWD-specific application.", "long_description": "**What it is.** The \"ISPOR RWD Questionnaire\" in this catalog refers to the **ISPOR Good Practice Task Force\nquestionnaire for assessing the *relevance* and *credibility* of observational (real-world data) studies for\nhealth-care decision making** (Berger et al., 2014, an ISPOR-AMCP-NPC Good Practice Task Force Report). It is a\nstructured appraisal instrument — a set of yes/no/can't-answer questions grouped into a *relevance* domain (does the\nstudy answer the decision-maker's actual question: population, intervention, comparators, outcomes, setting,\ngeneralizability?) and a *credibility* domain (design, data, analysis, reporting, interpretation, conflicts of\ninterest). It was developed under the auspices of ISPOR (the professional society for health economics and outcomes\nresearch) and is maintained as part of ISPOR's Good Practices for Outcomes Research reports. Two later ISPOR-ISPE\njoint task-force reports extend the same appraisal logic specifically to routinely collected RWD: the *good-practices*\nrecommendations for RWD studies of treatment/comparative effectiveness (Berger et al., 2017) and the *reporting*\nframework for healthcare-database studies to improve reproducibility and validity assessment (Wang et al., 2017).\nTreat the 2014 questionnaire as the appraisal instrument and the 2017 pair as its RWD-specific operationalization —\nthey are complementary, not competing, documents.\n\n**When to use.** Reach for this questionnaire when you must *judge whether a completed (or proposed) observational/RWD\nstudy should inform a decision* — a payer or HTA coverage/formulary review weighing a manufacturer's RWE dossier; an\nevidence team triaging published database studies for a comparative-effectiveness or safety question; a journal or\nregulatory reviewer screening a non-interventional submission; or an internal \"go/no-go\" credibility check before a\nstudy is cited in a value story. Use the *relevance* arm first: a methodologically flawless study that answers the\nwrong PICOTS question is useless to the decision at hand, and the questionnaire forces that judgment before any\ncredibility scoring. Use it alongside, not instead of, design-stage and reporting-stage tools: pre-specify with HARPER\nor STaRT-RWE, report with STROBE/RECORD(-PE), and appraise relevance-plus-credibility with this ISPOR instrument. When\nthe object of appraisal is specifically a *healthcare-database* study (claims, EHR, registry, linked), pull in the\nWang 2017 reporting framework so that the credibility questions are answerable from the documentation the study should\nhave provided.\n\n**What it requires.** The instrument's domains map directly onto the operational decisions that make or break an RWD\nstudy, and a credible answer to each requires the artifacts a transparent study should already have produced:\n- **Relevance / PICOTS alignment** — the study population, exposure/intervention, comparator(s), outcomes, timing, and\n setting must match the decision question; generalizability/transportability to the target population is assessed\n explicitly, not assumed.\n- **Data fitness-for-use** — the data source(s) must be characterized (provenance, capture mechanism, linkage,\n lag/completeness, payer mix) and shown adequate for the exposures, outcomes, and covariates needed.\n- **Phenotype / algorithm validity** — exposure, outcome, and covariate definitions must be operationalized with code\n lists and, where claimed, validation metrics (PPV, sensitivity).\n- **Time-zero alignment** — index-date definition must avoid immortal time and align eligibility, treatment\n assignment, and start of follow-up.\n- **Estimand and intercurrent events** — the target estimand (population, treatment strategy, handling of switching,\n discontinuation, death) must be stated, not left implicit.\n- **Confounding control** — design (active comparator, new-user) and analysis (PS/hdPS, g-methods) must be adequate to\n the confounding structure, with balance diagnostics shown.\n- **Attrition and missing data** — loss to follow-up, censoring, and missingness must be reported and addressed.\n- **Sensitivity / quantitative bias analysis** — robustness to key assumptions (unmeasured confounding via E-value,\n negative controls, alternative specifications) must be demonstrated.\nCrucially, the questionnaire also requires *transparency about pre-specification and conflicts of interest* — whether\nthe protocol and analysis plan predated the analysis, and who funded and conducted the work.\n\n**When NOT to use — limitations and common misapplications.** This is a **relevance-and-credibility appraisal\ninstrument, not a risk-of-bias tool and not a numeric quality score.** Do not convert its yes/no answers into a summed\n\"quality score\" and rank studies by it — the developers explicitly designed it for structured judgment, not\nscoring, and summary scores obscure which specific threat invalidates a study. For formal, signalling-question\nrisk-of-bias assessment of a non-randomized study (e.g., within a systematic review), use **ROBINS-I**, which is the\nfit-for-purpose instrument; this questionnaire complements but does not replace it. Completing the questionnaire does\n**not make an observational study causal** — answering \"yes\" to the design questions documents that confounding was\n*addressed*, not that it was *eliminated*; residual and unmeasured confounding remain, which is why the sensitivity-\nanalysis domain matters. Do not use it as a *protocol template* (that is HARPER / StaRT-RWE) or as a *reporting\nchecklist* for the final manuscript (that is STROBE / RECORD-PE) — substituting one for another leaves real gaps that\nreviewers will find. Beware **appraisal-as-theater**: ticking boxes without inspecting the underlying code lists,\nvalidation studies, balance tables, and attrition diagrams produces a clean-looking checklist over an uncredible\nstudy. Finally, applying the generic 2014 questionnaire to a complex *healthcare-database* study without the Wang 2017\nreporting expectations will leave the data-fitness and reproducibility questions effectively unanswerable.\n\n**How it maps to this catalog.** Each appraisal domain has an implementing concept that supplies the operational\ndetail a credible answer requires:\n- **Relevance / PICOTS** → `picots-framework-rwe` (and `generalizability-transportability-external-validity-rwe` for\n the target-population judgment).\n- **Data fitness-for-use** → `fit-for-purpose-data-assessment-rwe`, with `claims-analysis` and\n `medicare-ffs-ma-commercial-claims-differences-rwe` for source-specific capture and payer-mix nuances.\n- **Phenotype / algorithm validity** → `diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe`.\n- **Design / confounding control** → `active-comparator-new-user` and `high-dimensional-propensity-score-hdps-rwe`,\n with `target-trial-emulation` as the pre-specification scaffold that makes the relevance and credibility questions\n answerable by construction.\n- **Estimand / intercurrent events** → `estimands-ate-att-intercurrent-events-rwe`.\n- **Attrition / missing data** → `attrition-and-loss-to-follow-up-rwe`.\n- **Sensitivity / quantitative bias analysis** → `e-value-sensitivity-analysis`.\n- **Reporting artifacts (attrition flow, balance diagnostics)** → `visualizations-pharmacoepidemiology-rwe`.\n**Applied note (claims/EHR/registry RWE).** A payer assessing a manufacturer's claims-based comparative-effectiveness\nstudy should run the relevance arm against `picots-framework-rwe`, then demand the credibility evidence concretely:\nthe data-fitness narrative (`fit-for-purpose-data-assessment-rwe`), the outcome code list and PPV\n(`diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe`), confirmation that the design is active-comparator/new-user\nwith a balance table (`active-comparator-new-user`, `high-dimensional-propensity-score-hdps-rwe`), the attrition\nfunnel (`attrition-and-loss-to-follow-up-rwe`), and an E-value or negative-control result\n(`e-value-sensitivity-analysis`). A \"yes\" with no supporting artifact is a \"no.\"", "primary_category": "Guideline", "tags": [ "guideline", "critical-appraisal", "relevance-and-credibility", "real-world-evidence", "hta", "ispor" ], "aliases": [ "ISPOR RWD Questionnaire", "ISPOR Relevance and Credibility Questionnaire", "ISPOR-AMCP-NPC Questionnaire", "Berger 2014 ISPOR questionnaire" ], "applies_to_study_types": [ "cohort_retrospective", "claims_analysis", "ehr_study", "linked_data", "multi_database" ], "data_sources": [ "claims", "ehr", "registry", "linked", "multi-database" ], "citations": [ { "role": "introduce", "doi": "10.1016/j.jval.2013.12.011", "url": "https://doi.org/10.1016/j.jval.2013.12.011", "citation_text": "Berger ML, Martin BC, Husereau D, et al. A questionnaire to assess the relevance and credibility of observational studies to inform health care decision making: an ISPOR-AMCP-NPC Good Practice Task Force report. Value in Health. 2014;17(2):143-156.", "year": 2014, "authors_short": "Berger et al.", "notes": "The canonical ISPOR questionnaire instrument. Two domains — relevance and credibility — assessed via structured yes/no/can't-answer questions; designed for judgment, not as a summed quality score." }, { "role": "explain", "doi": "10.1016/j.jval.2017.08.3018", "url": "https://doi.org/10.1016/j.jval.2017.08.3018", "citation_text": "Wang SV, Schneeweiss S, Berger ML, et al. Reporting to improve reproducibility and facilitate validity assessment for healthcare database studies V1.0. Value in Health. 2017;20(8):1009-1022.", "year": 2017, "authors_short": "Wang et al.", "notes": "ISPOR-ISPE reporting framework that operationalizes the credibility/data-fitness questions for healthcare-database (claims/EHR) studies, enabling reproducibility and validity assessment." }, { "role": "use", "doi": "10.1016/j.jval.2017.08.3019", "url": "https://doi.org/10.1016/j.jval.2017.08.3019", "citation_text": "Berger ML, Sox H, Willke RJ, et al. Good practices for real-world data studies of treatment and/or comparative effectiveness: recommendations from the joint ISPOR-ISPE Special Task Force on real-world evidence in health care decision making. Value in Health. 2017;20(8):1003-1008.", "year": 2017, "authors_short": "Berger et al.", "notes": "ISPOR-ISPE good-practice recommendations situating the questionnaire within the broader RWE decision-making lifecycle (pre-specification, registration, replication)." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "claims-analysis", "notes": "Primary appraisal target — claims-based comparative-effectiveness and safety studies submitted for HTA/payer or regulatory review." }, { "relation_type": "applies_to", "target_slug": "ehr-study", "notes": "Appraise EHR-based observational studies for relevance and credibility before using them in a decision." }, { "relation_type": "applies_to", "target_slug": "linked-data", "notes": "Linked claims-EHR-registry studies; data-fitness and reproducibility questions are answered with the Wang 2017 reporting framework." }, { "relation_type": "applies_to", "target_slug": "multi-database", "notes": "Multi-database studies; relevance/transportability and consistency across sources are explicit credibility considerations." }, { "relation_type": "see_also", "target_slug": "picots-framework-rwe", "notes": "Implements the relevance arm — the gating judgment of whether the study answers the decision-maker's question." }, { "relation_type": "see_also", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "Implements the data-fitness-for-use credibility domain." }, { "relation_type": "complements", "target_slug": "target-trial-emulation", "notes": "Pre-specifying the emulated trial makes both relevance and credibility questions answerable by construction; the questionnaire appraises the result." }, { "relation_type": "complements", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Supplies the phenotype/algorithm validity (code lists, PPV) that a credible answer to the outcome-definition questions requires." }, { "relation_type": "complements", "target_slug": "active-comparator-new-user", "notes": "Design-level confounding control the credibility arm assesses; pairs with high-dimensional-propensity-score-hdps-rwe." }, { "relation_type": "complements", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "Analytic confounding control and balance diagnostics underpinning credible confounding-control answers." }, { "relation_type": "complements", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Implements the estimand/intercurrent-events domain." }, { "relation_type": "complements", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Implements the attrition/missing-data credibility domain." }, { "relation_type": "complements", "target_slug": "e-value-sensitivity-analysis", "notes": "Implements the sensitivity / quantitative-bias-analysis domain (residual unmeasured confounding)." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "hta", "journal", "fda", "ema" ] }, { "slug": "ispor-rwe-tf", "name": "ISPOR-ISPE Good Practices for Real-World Data Studies (Berger et al. 2017)", "short_definition": "The joint ISPOR-ISPE Special Task Force good-practice recommendations for designing, conducting, and reporting real-world data studies of treatment effects and comparative effectiveness, paired with the V1.0 reproducibility reporting framework for healthcare-database studies. It governs a priori study registration, transparent design and implementation, and validity assessment of decision-grade observational evidence.", "long_description": "**What it is.** The *Good Practices for Real-World Data Studies of Treatment and/or Comparative\nEffectiveness* recommendations are the consensus output of the **joint ISPOR-ISPE Special Task Force\non Real-World Evidence in Health Care Decision Making** (Berger et al., 2017), published simultaneously\nin *Value in Health* and *Pharmacoepidemiology and Drug Safety*. ISPOR (the Professional Society for\nHealth Economics and Outcomes Research) and ISPE (the International Society for Pharmacoepidemiology)\njointly maintain the document. It is a **good-practice / design-and-conduct framework** — not a\nnumbered reporting checklist like STROBE/RECORD-PE and not a risk-of-bias instrument like ROBINS-I.\nIts central thesis is that real-world evidence can be decision-grade *if and only if* the study process\nis credible and reproducible, so the recommendations are organized around (1) a priori hypothesis\ntesting versus exploratory analysis, (2) public study **registration and pre-specified protocols**,\nand (3) transparent, auditable implementation. The Task Force deliberately split its work: this\ngood-practices statement covers *what makes an RWD study trustworthy*, while its companion, **Wang,\nSchneeweiss, Berger et al. (2017) \"Reporting to Improve Reproducibility and Facilitate Validity\nAssessment for Healthcare Database Studies V1.0,\"** provides the structured reporting template that\noperationalizes those principles for claims/EHR studies. Treat the two papers as one toolkit: the\nrecommendations set the standard, the V1.0 template enforces it.\n\n**When to use.** Apply this framework whenever a **non-interventional study of treatment effects or\ncomparative effectiveness** built on routinely collected data (claims, EHR, registries, linked, or\nmulti-database networks) is intended to inform a decision — regulatory (FDA/EMA), HTA/payer dossier,\nformulary/coverage, or a high-impact peer-reviewed manuscript. It is the right reference at the\n*planning* stage of any hypothesis-evaluating RWD study and again at the *governance* stage when you\nmust demonstrate that the analysis was pre-specified rather than data-dredged. Decision rules for\nchoosing THIS framework over siblings: use the **ISPOR-ISPE good practices** to set the overarching\ncredibility expectations (registration, pre-specification, fitness-for-purpose, sensitivity analysis)\nfor a *hypothesis-evaluating treatment-effect study*; switch to **HARPER / START-RWE** when you need a\nfill-in protocol/structured template for the design itself; use **STROBE + RECORD-PE** for the final\nmanuscript's reporting items and flow diagram; use the **ENCePP Guide/Checklist and EU PAS Register**\nwhen the study is an EU PASS or otherwise falls under GVP Module VIII; and consult the **FDA RWE\nFramework** and FDA's RWD/EHR-claims guidance for US regulatory submissions. The ISPOR-ISPE document\nsits above the reporting checklists: it tells you how to make the study believable; the checklists tell\nyou how to write it up.\n\n**What it requires.** The recommendations enforce a specific set of credibility-bearing practices,\nand the V1.0 reporting template makes each one concrete for real-world data:\n- **A priori declaration of intent.** State whether the study is hypothesis-evaluating (confirmatory)\n or exploratory, and post the **protocol and analysis plan to a public study register before looking at\n outcome data.** Post-hoc, register-after-the-fact analyses are explicitly disfavored for\n decision-grade claims.\n- **Design transparency.** Fully specify the study design (cohort, case-control, self-controlled),\n eligibility, exposure and comparator definitions, **time-zero/index-date alignment**, follow-up, and\n the causal contrast/estimand — at a level of detail that allows independent re-execution.\n- **Data fitness-for-use.** Document the data source(s), why they are fit for the question\n (relevance and reliability), capture/lags, linkage, and provenance, before relying on them.\n- **Operational definitions and phenotype/algorithm validation.** Provide complete, versioned code\n lists and the validation evidence (PPV/sensitivity) for exposure, outcome, and covariate algorithms.\n- **Confounding control and assumptions.** Pre-specify the confounding-adjustment strategy\n (e.g., propensity or high-dimensional propensity methods, active-comparator new-user design) and\n test its assumptions (balance, positivity).\n- **Attrition and missing data.** Report the cohort-attrition cascade transparently and handle\n loss-to-follow-up and missingness explicitly.\n- **Sensitivity and quantitative bias analysis.** Pre-specify sensitivity analyses for key design\n choices and address residual/unmeasured confounding (e.g., negative controls, E-value).\n- **Reproducibility.** Tie every analytic decision to documented code and a protocol version so an\n independent team could reproduce the result — the explicit goal of the V1.0 template.\n\n**When NOT to use — limitations and common misapplications.**\n- **It is not a reporting checklist or a quality score.** Do not treat the good-practices paper as a\n line-item box-ticking exercise the way you would STROBE or RECORD-PE; and the companion V1.0 template\n is a *reproducibility/reporting* structure, **not a risk-of-bias instrument** — it documents what was\n done, it does not grade internal validity. For risk-of-bias appraisal of a non-randomized study, use\n ROBINS-I (or ROBINS-E); for HTA suitability of an RWD source, use the ISPOR/ISPE suitability and RWD\n questionnaire tools.\n- **Following the framework does not make a study causal.** Registering a protocol and completing the\n V1.0 template documents transparency and reproducibility; it does not by itself remove confounding by\n indication, immortal-time bias, or selection bias. The design (active-comparator new-user, target-trial\n emulation) and the analysis carry the causal burden — the framework only makes those choices visible.\n- **Checklist-as-theater.** A registered protocol that is silently amended after seeing the data, or a\n completed template that points to unvalidated phenotypes, satisfies the letter and defeats the purpose.\n- **Wrong tool for the design context.** It does not cover RWE for **medical devices** (use FDA's\n device-specific RWE guidance), pragmatic/randomized real-world trials (use the FDA pragmatic-trials\n guidance, PRECIS-2, CONSORT-pragmatic), or EU-specific PASS procedural requirements (use ENCePP/EU PAS\n Register and GVP Module VIII). For *purely descriptive* RWE with no treatment-effect estimand, the\n treatment-effect/comparative-effectiveness emphasis here is heavier than needed.\n\n**How it maps to this catalog.** Each requirement is implemented by a concrete concept in this repo:\n- *A priori protocol/SAP and registration* → `study-protocol-or-sap-elements`, `estimand-analysis-traceability-rwe`,\n `pass-imposed` / `pass-voluntary` (registration context).\n- *Design transparency and the trial-like target* → `target-trial-emulation`, `picots-framework-rwe`,\n `time-zero-index-date-alignment-rwe`.\n- *Estimand and intercurrent events* → `estimands-ate-att-intercurrent-events-rwe`.\n- *Confounding control* → `active-comparator-new-user`, `high-dimensional-propensity-score-hdps-rwe`,\n `propensity-score-methods-psm-iptw`, `dags-backdoor-criterion-drug-studies`.\n- *Data fitness-for-use* → `fit-for-purpose-data-assessment-rwe`, `claims-analysis`,\n `medicare-ffs-ma-commercial-claims-differences-rwe`, `continuous-enrollment-observable-time-rwe`.\n- *Phenotype/algorithm validation* → `diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe`,\n `claims-outcome-algorithm-ppv-sensitivity-rwe`, `algorithm-validation`.\n- *Attrition and missing data* → `attrition-and-loss-to-follow-up-rwe`,\n `database-feasibility-attrition-funnel-rwe`, `missing-data-pattern-table-rwe`.\n- *Sensitivity / quantitative bias analysis* → `e-value-sensitivity-analysis`,\n `quantitative-bias-analysis-toolkit-rwe`, `negative-control-outcomes-rwe`,\n `empirical-calibration-negative-controls-rwe`.\n\n**Applied note (claims/EHR/registry RWE).** For a comparative-effectiveness claims study, the\nframework converts to a concrete sequence: register a pre-specified protocol with the full estimand and\nanalysis plan; document why the database is fit-for-purpose (`fit-for-purpose-data-assessment-rwe`,\nnoting Medicare FFS vs MA capture gaps per `medicare-ffs-ma-commercial-claims-differences-rwe`); build\nan active-comparator new-user cohort with a justified washout and time-zero\n(`active-comparator-new-user`, `time-zero-index-date-alignment-rwe`); attach validated code lists and PPV\nevidence to every phenotype (`diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe`); control\nconfounding with hdPS (`high-dimensional-propensity-score-hdps-rwe`); report the attrition funnel and\ncovariate balance; and pre-specify sensitivity and negative-control analyses\n(`e-value-sensitivity-analysis`, `negative-control-outcomes-rwe`). The deliverable that satisfies the\nTask Force is a registered protocol plus a completed V1.0 reporting template that lets an independent\nteam rerun the study from the documented code.", "primary_category": "Guideline", "tags": [ "guideline", "good-practice", "rwe", "comparative-effectiveness", "reproducibility", "study-registration" ], "aliases": [ "ISPOR-ISPE RWE Good Practices", "ISPOR-ISPE Joint Task Force on Real-World Evidence", "Berger 2017 Good Practices for Real-World Data Studies", "Wang 2017 Reproducibility V1.0", "ISPOR-RWE-TF" ], "applies_to_study_types": [ "cer_observational", "new_user", "active_comparator_new_user", "claims_analysis", "ehr_study", "target_trial_emulation" ], "data_sources": [ "claims", "ehr", "registry", "linked", "multi-database" ], "citations": [ { "role": "introduce", "doi": "10.1016/j.jval.2017.08.3019", "url": "https://doi.org/10.1016/j.jval.2017.08.3019", "citation_text": "Berger ML, Sox H, Willke RJ, Brixner DL, Eichler HG, Goettsch W, Madigan D, Makady A, Schneeweiss S, Tarricone R, Wang SV, Watkins J, Daniel Mullins C. Good Practices for Real-World Data Studies of Treatment and/or Comparative Effectiveness: Recommendations from the Joint ISPOR-ISPE Special Task Force on Real-World Evidence in Health Care Decision Making. Value in Health. 2017;20(8):1003-1008.", "year": 2017, "authors_short": "Berger et al.", "notes": "Canonical good-practices statement of the joint ISPOR-ISPE Special Task Force. Co-published in Pharmacoepidemiology and Drug Safety (2017;26(9):1033-1039, DOI 10.1002/pds.4297) with identical content." }, { "role": "explain", "doi": "10.1016/j.jval.2017.08.3018", "url": "https://doi.org/10.1016/j.jval.2017.08.3018", "citation_text": "Wang SV, Schneeweiss S, Berger ML, Brown J, de Vries F, Douglas I, Gagne JJ, Gini R, Klungel O, Mullins CD, Nguyen MD, Rassen JA, Smeeth L, Sturkenboom M, on behalf of the joint ISPE-ISPOR Special Task Force on Real World Evidence in Health Care Decision Making. Reporting to Improve Reproducibility and Facilitate Validity Assessment for Healthcare Database Studies V1.0. Value in Health. 2017;20(8):1009-1022.", "year": 2017, "authors_short": "Wang et al.", "notes": "Companion Task Force output. Provides the structured V1.0 reporting/reproducibility template that operationalizes the good practices for claims/EHR studies; precursor to HARPER/START-RWE. Co-published in Pharmacoepidemiology and Drug Safety (DOI 10.1002/pds.4295)." }, { "role": "use", "url": "https://www.ispor.org/heor-resources/good-practices/article/good-practices-for-real-world-data-studies-of-treatment-and-or-comparative-effectiveness-recommendations-from-the-joint-ispor-ispe-special-task-force-on-real-world-evidence-in-health-care-decision-making", "citation_text": "ISPOR Good Practices Reports — Joint ISPOR-ISPE Special Task Force on Real-World Evidence in Health Care Decision Making (maintained landing page and resources).", "year": 2017, "authors_short": "ISPOR / ISPE", "notes": "Society-maintained resource page; entry point to the good-practices report and related ISPOR/ISPE RWE tools (RWD suitability questionnaire, fit-for-purpose guidance)." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "cer-observational", "notes": "Sets the credibility, registration, and reproducibility expectations for observational comparative effectiveness research." }, { "relation_type": "applies_to", "target_slug": "active-comparator-new-user", "notes": "The good practices favor a trial-like design; ACNU is the standard analytic core for the treatment-effect studies this framework governs." }, { "relation_type": "applies_to", "target_slug": "target-trial-emulation", "notes": "Pre-specification of the hypothetical trial protocol is the design embodiment of the a priori, transparent-design recommendations." }, { "relation_type": "applies_to", "target_slug": "claims-analysis", "notes": "The V1.0 reporting template is built explicitly for healthcare-database (claims/EHR) studies." }, { "relation_type": "applies_to", "target_slug": "ehr-study", "notes": "Same reproducibility and validity-assessment expectations apply to EHR-based treatment-effect studies." }, { "relation_type": "requires", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "Documenting data relevance and reliability (fitness-for-use) is a prerequisite under the framework." }, { "relation_type": "requires", "target_slug": "study-protocol-or-sap-elements", "notes": "A priori protocol and statistical analysis plan, registered before outcome data look, is a core requirement." }, { "relation_type": "requires", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Versioned, validated phenotype/algorithm definitions are required for exposure, outcome, and covariates." }, { "relation_type": "complements", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "hdPS implements the pre-specified confounding-control expectation in high-dimensional claims data." }, { "relation_type": "complements", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "The causal-contrast/estimand specification the framework demands is implemented here." }, { "relation_type": "complements", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Transparent attrition reporting and handling of loss-to-follow-up satisfy the framework's attrition/missing-data expectations." }, { "relation_type": "complements", "target_slug": "e-value-sensitivity-analysis", "notes": "Quantitative bias analysis for residual/unmeasured confounding implements the sensitivity-analysis requirement." }, { "relation_type": "see_also", "target_slug": "harper", "notes": "HARPER provides the structured fill-in protocol template that succeeds the Task Force's V1.0 reporting framework." }, { "relation_type": "see_also", "target_slug": "record-pe", "notes": "Use RECORD-PE (with STROBE) for the final manuscript reporting items; this framework governs design and conduct." }, { "relation_type": "see_also", "target_slug": "fda-rwe-framework", "notes": "For US regulatory submissions, pair these good practices with FDA's RWE Framework and RWD guidances." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "hta", "journal" ] }, { "slug": "ispor-suitability", "name": "ISPOR SUITABILITY Checklist", "short_definition": "ISPOR good-practices checklist for assessing whether real-world data from electronic health records are trustworthy and fit-for-purpose to inform health technology assessment, structured around data delineation (characteristics, provenance, governance) and data fitness (reliability, relevance).", "long_description": "**What it is** — The **SUITABILITY checklist** (Fleurence et al., *Value in Health* 2024) is a good-practices report\nof an ISPOR Task Force that gives HTA bodies, payers, and evidence generators a structured way to judge whether\n**real-world data drawn from electronic health records (EHRs)** are trustworthy enough, and relevant enough, to inform\na health technology assessment. It is maintained by **ISPOR — The Professional Society for Health Economics and\nOutcomes Research** as part of its Good Practices series. The checklist is organized into two complementary arms.\n*Data delineation* establishes what the data are and whether they can be trusted, through three blocks: (1) **data\ncharacteristics** (structure, variables, coding systems, longitudinality, population captured); (2) **data provenance**\n(how the data were generated, transformed, curated, and versioned from the point of care to the analytic dataset);\nand (3) **data governance** (access, privacy, ethical and regulatory controls, and reproducibility). *Data fitness-for-\npurpose* then asks whether those data can actually answer the decision question, through (4) **reliability** (accuracy,\ncompleteness, plausibility, and consistency of the key variables) and (5) **relevance** (whether the population,\nexposures, outcomes, follow-up, and time window match the HTA question). SUITABILITY is a *data-assessment* instrument:\nit interrogates the substrate, not the analysis.\n\n**When to use** — Use SUITABILITY when an HTA submission, payer dossier, or value assessment rests on **EHR-derived\nreal-world evidence** and a reviewer or generator must defend (or scrutinize) the data source before trusting any\neffect estimate. It is built for the HTA/payer decision context — NICE-style technology appraisals, CADTH/INESSS\nreviews, joint EU HTA clinical assessments, and the cost-effectiveness or comparative-effectiveness analyses that feed\nthem — and for peer-reviewed reporting of EHR-based HTA evidence. **Decision rule for which tool applies:** reach for\nSUITABILITY specifically when the dossier's evidence comes from **EHR** data feeding an **HTA** decision. If the data\nare **claims-only** or **registry-only**, or the question is regulatory rather than HTA, a sibling instrument fits\nbetter — the **ISPOR RWD questionnaire** for general RWD source vetting, **ENCePP** / **GVP Module VIII** / **ISPE-\nSCOPE** for pharmacoepidemiologic and regulatory PASS work, and **NICE's RWE framework** as the HTA-body's own\ndata-suitability expectations. SUITABILITY does not replace those; it is the EHR-for-HTA-specialized lens.\n\n**What it requires** — Substantively, SUITABILITY forces documentation of: the data model and variable provenance\nend-to-end (raw EHR → ETL/curation → analytic dataset, with versioning); governance and reproducibility of the data\npipeline; **data-fitness-for-use** evidence — the reliability of the variables that carry the analysis (key\nexposures, outcomes, covariates) including completeness, plausibility, and internal consistency; and **relevance** of\nthe captured population, time window, and endpoints to the HTA question. Framed for real-world data, that means it\nexpects: explicit **phenotype / algorithm definitions** for diseases, exposures, and outcomes, with **validation**\nevidence (PPV, sensitivity) where the variable drives the result; clear handling of **observable time** and\n**attrition / loss to follow-up** in a visit-driven EHR; transparent treatment of **missing data**; and alignment of\nthe data window with **time-zero** and the intended **estimand**. It does *not* itself prescribe the confounding-\ncontrol or analytic strategy — but it asks whether the data can support the one chosen.\n\n**When NOT to use — limitations and common misapplications** — SUITABILITY is a **fit-for-purpose data-assessment\nchecklist, NOT a risk-of-bias instrument and NOT a study-quality score.** Passing it tells you the EHR data are\nusable; it tells you nothing about whether the *design* is sound or the estimate *causal* — a SUITABILITY-clean\ndataset analyzed with a prevalent-user, immortal-time-ridden design is still biased. Concrete failure modes: (1)\n**Wrong substrate** — applying it to a claims-only or registry-only study where the **ISPOR RWD questionnaire**,\n**ENCePP**, or **ISPE-SCOPE** is the correct tool; SUITABILITY's provenance/curation logic is tuned to the messy,\nunstructured, visit-driven nature of EHRs. (2) **Wrong decision context** — treating it as a regulatory (FDA/EMA)\nsubmission standard; it is HTA/payer turf, and FDA's RWD/EHR-claims guidance or EMA/GVP govern the regulatory lane.\n(3) **Checklist-as-theater** — ticking boxes without producing the underlying validation/provenance evidence, which\ndefeats the purpose. (4) **Substituting it for design rigor** — using a completed SUITABILITY assessment as if it\ncertified the analysis; it must be paired with sound design (target-trial emulation, active-comparator new-user) and\nbias analysis. (5) **Conflating reliability with relevance** — accurate data for the wrong population or time window\nstill fails fitness-for-purpose.\n\n**How it maps to this catalog** — SUITABILITY's two arms route directly to implementing concepts. The overarching\nfitness judgment is operationalized by **fit-for-purpose-data-assessment-rwe**. *Data delineation* →\n*characteristics/provenance/governance*: **ohdsi-cdm** (a common data model that makes characteristics and provenance\nauditable and reproducible) and **continuous-enrollment-observable-time-rwe** (defining observable person-time so\n\"what the data capture\" is explicit). *Data fitness → reliability*: variable accuracy is implemented by\n**ehr-phenotyping-algorithms-rwe** and **diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe** for cohort/condition\ndefinitions, **claims-outcome-algorithm-ppv-sensitivity-rwe** for the validation metrics (PPV, sensitivity) the\nchecklist expects, **attrition-and-loss-to-follow-up-rwe** for completeness over follow-up, and\n**missing-data-pattern-table-rwe** for missingness transparency. *Data fitness → relevance*: matching data to the\ndecision question is carried by **estimands-ate-att-intercurrent-events-rwe** (does the data support the target\nestimand and time-zero?) and, once relevance is established, the analytic engines that turn fit data into defensible\nestimates — **target-trial-emulation**, **active-comparator-new-user**, **high-dimensional-propensity-score-hdps-rwe**,\nand **empirical-calibration-negative-controls-rwe** for residual-confounding diagnostics. Use **claims-analysis** as\nthe contrast: when the source is claims rather than EHR, that concept (with the ISPOR RWD questionnaire) is where the\nsuitability assessment lives instead.\n\n**Applied note (EHR for HTA).** In a NICE-style appraisal built on a US/EU EHR network, SUITABILITY would have the\nsubmitter document the ETL from source EHR to analytic dataset with versions (provenance), show the governance and\nre-run path (governance), report PPV/sensitivity for the outcome phenotype against chart review (reliability), quantify\nloss to follow-up when patients leave the health system (reliability/relevance), and demonstrate that the captured\npopulation, line of therapy, and time horizon match the appraisal's decision problem (relevance) — *before* the\ncommittee weighs the comparative-effectiveness estimate at all.", "primary_category": "Guideline", "tags": [ "guideline", "data-suitability", "fit-for-purpose", "electronic-health-records", "real-world-data", "health-technology-assessment", "ispor", "good-practices" ], "aliases": [ "SUITABILITY", "SUITABILITY checklist", "ISPOR SUITABILITY", "ISPOR EHR data suitability checklist", "Assessing Real-World Data From EHRs for HTA" ], "applies_to_study_types": [ "ehr_study", "cer_observational", "cost_effectiveness" ], "data_sources": [ "ehr", "linked", "registry" ], "citations": [ { "role": "introduce", "doi": "10.1016/j.jval.2024.01.019", "url": "https://doi.org/10.1016/j.jval.2024.01.019", "citation_text": "Fleurence RL, Kent S, Adamson B, et al. Assessing real-world data from electronic health records for health technology assessment: the SUITABILITY checklist: a good practices report of an ISPOR task force. Value in Health. 2024;27(6):692-701.", "year": 2024, "authors_short": "Fleurence et al.", "notes": "Canonical ISPOR Task Force good-practices report defining the SUITABILITY checklist and its two arms (data delineation and data fitness-for-purpose)." }, { "role": "explain", "doi": "10.1016/j.jval.2024.04.020", "url": "https://doi.org/10.1016/j.jval.2024.04.020", "citation_text": "Hamilton Lopez M, et al. Perspectives on improving value assessment with the ISPOR SUITABILITY checklist. Value in Health. 2024;27(6).", "year": 2024, "authors_short": "Hamilton Lopez et al.", "notes": "Companion perspective discussing how HTA bodies and decision makers can operationalize the checklist in value assessment." }, { "role": "use", "url": "https://www.ispor.org/heor-resources/good-practices/article/assessing-real-world--data-from-electronic-health-records-for-health-technology-assessment--the-suitability-checklist", "citation_text": "ISPOR. Assessing Real-World Data From Electronic Health Records for Health Technology Assessment — The SUITABILITY Checklist. ISPOR Good Practices (maintained resource).", "year": 2024, "authors_short": "ISPOR", "notes": "Maintained ISPOR landing page with the checklist and supporting good-practices materials." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "ehr-study", "notes": "Primary use case — assessing EHR-derived data feeding an HTA decision." }, { "relation_type": "applies_to", "target_slug": "cer-observational", "notes": "Applies when observational comparative-effectiveness evidence for HTA is built on EHR data." }, { "relation_type": "applies_to", "target_slug": "cost-effectiveness", "notes": "Applies when EHR-derived inputs (effectiveness, utilization, costs) feed a cost-effectiveness model in a dossier." }, { "relation_type": "requires", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "The fitness-for-purpose verdict is operationalized by this concept (reliability + relevance for the decision question)." }, { "relation_type": "used_with", "target_slug": "ehr-phenotyping-algorithms-rwe", "notes": "Reliability of EHR-derived variables depends on transparent, validated phenotyping algorithms." }, { "relation_type": "used_with", "target_slug": "claims-outcome-algorithm-ppv-sensitivity-rwe", "notes": "Supplies the PPV/sensitivity validation evidence the reliability arm expects for key outcomes." }, { "relation_type": "used_with", "target_slug": "ohdsi-cdm", "notes": "A common data model makes data characteristics and provenance auditable and reproducible (delineation arm)." }, { "relation_type": "see_also", "target_slug": "continuous-enrollment-observable-time-rwe", "notes": "Observable person-time defines what the EHR data actually capture — a delineation/relevance input." }, { "relation_type": "see_also", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Loss to follow-up in visit-driven EHRs threatens completeness (reliability) and relevance." }, { "relation_type": "see_also", "target_slug": "target-trial-emulation", "notes": "SUITABILITY certifies the data, not the design; pair with target-trial emulation for a defensible estimate." }, { "relation_type": "alternative_to", "target_slug": "claims-analysis", "notes": "For claims-only sources, suitability assessment lives with claims-analysis and the ISPOR RWD questionnaire rather than the EHR-specialized SUITABILITY checklist." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "hta", "journal" ] }, { "slug": "jbi-case-control", "name": "JBI Critical Appraisal Checklist for Case-Control Studies", "short_definition": "A 10-item critical-appraisal (risk-of-bias screening) checklist from JBI for judging the methodological trustworthiness of case-control and nested case-control studies during evidence synthesis of etiology and risk.", "long_description": "**What it is** — The **JBI Critical Appraisal Checklist for Case-Control Studies** is a 10-item\ninstrument maintained by **JBI (formerly the Joanna Briggs Institute)** as part of its suite of\ndesign-specific critical-appraisal tools. It belongs to the JBI methodology for **systematic reviews\nof etiology and risk** (Chapter 7 of the *JBI Manual for Evidence Synthesis*), where it is used by two\nindependent reviewers to judge whether a case-control study's design, conduct, and analysis are\nmethodologically sound enough to trust and to include in a synthesis. Each item is rated **Yes / No /\nUnclear / Not applicable**, and the appraisal informs an explicit, reviewer-documented include/exclude\nand weight-of-evidence decision — it is *not* a numeric scale. The tool's published statement is Moola\net al. (2015); its currently maintained form lives in the open-access JBI Manual and on the JBI\ncritical-appraisal-tools site.\n\n**When to use** — Reach for this checklist when you are **appraising primary case-control or nested\ncase-control studies** inside a **systematic review, scoping/CAT, or HTA evidence-synthesis** workflow,\nmost often for a **peer-reviewed journal** or an **HTA/payer evidence assessment**. The decision rule\namong siblings is by *design and purpose*: use the **JBI Cohort** checklist for cohort designs, **JBI\nPrevalence** for prevalence/cross-sectional burden studies, **JBI Case Series / Case Reports** for\nuncontrolled descriptive designs — and this **Case-Control** checklist only when the included study\nsamples on *outcome status* (cases vs controls) and looks back at exposure. If your task is **reporting**\nyour own study rather than appraising others', JBI is the wrong family entirely: use **STROBE**\n(and **RECORD / RECORD-PE** for routinely-collected health data), or **HARPER** for the protocol/structure\nof a pharmacoepidemiologic study. If you need a **formal risk-of-bias instrument** for a comparative\nobservational effect estimate feeding a GRADE assessment, **ROBINS-I** is the more granular,\ndomain-based tool; JBI appraisal is lighter-weight and synthesis-oriented.\n\n**What it requires** — The 10 items map onto the bias domains a case-control study lives or dies by, and\neach has a real-world-data analogue:\n- **Comparable groups / appropriate matching** — cases and controls drawn from the same source population\n so controls represent the population that produced the cases; matching done and accounted for in analysis.\n In claims/EHR work this is the **risk-set / source-population** question — see *nested-case-control* and\n *case-control*.\n- **Same case/control identification criteria** — a transparent, validated **phenotype** applied identically\n to both groups (*diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe*,\n *claims-outcome-algorithm-ppv-sensitivity-rwe*, *algorithm-validation*).\n- **Exposure measured in a standard, valid, reliable way, identically for cases and controls** — guards\n against **differential/recall misclassification**; in RWD this is consistent\n **exposure-episode construction** and lookback definition (*exposure-episode-construction-rwe*,\n *washout-clean-lookback-period-rwe*, *misclassification-bias-correction-rwe*).\n- **Confounders identified and handled** — explicit confounder enumeration and a control strategy\n (matching, stratification, regression, or **propensity/disease-risk scores**); see\n *dags-backdoor-criterion-drug-studies* and *high-dimensional-propensity-score-hdps-rwe*.\n- **Outcomes/cases assessed in a standard way; exposure period long enough; appropriate statistical\n analysis** — covering case validity, biologically/clinically adequate **induction-latency windows**\n (*exposure-lag-induction-latency-window-rwe*), and analysis that respects matching and sparse-data\n behavior.\n\n**When NOT to use — limitations and common misapplications** — (1) **It is an appraisal/RoB-screening\ntool, not a quality score.** JBI explicitly warns against summing \"Yes\" answers into a total; a study with\n8/10 is not \"better\" than one with 6/10 if the two failed items are fatal (e.g., controls from a different\nsource population). Tallying scores is the single most common abuse. (2) **It is not ROBINS-I.** It does not\ndecompose bias by domain with signalling questions, so for a comparative-effect estimate destined for GRADE,\nit under-resolves confounding and selection bias — use ROBINS-I. (3) **It under-probes nested\ncase-control–specific issues.** Risk-set sampling, time-zero alignment, immortal-time, and incidence-density\nvs cumulative sampling are where nested designs fail, yet the JBI items don't interrogate them directly;\npair the appraisal with *time-zero-index-date-alignment-rwe* and *immortal-time-bias-handling*. (4) **Wrong\ntool for the job** — using JBI Case-Control to appraise a cohort or prevalence study, or using it as your\n*reporting* checklist (STROBE/RECORD-PE territory). (5) **Checklist-as-theater** — completing the form\nwithout independent dual review, consensus, and a documented effect on inclusion/synthesis adds no validity;\npassing the checklist does **not** make an observational association causal.\n\n**How it maps to this catalog** — Treat each JBI item as a requirement and follow the implementing concept:\ncase/control definition → *case-control*, *nested-case-control*, *diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe*,\n*claims-outcome-algorithm-ppv-sensitivity-rwe*, *algorithm-validation*; comparable source population /\nselection → *selection-bias-sensitivity-analysis-rwe*; exposure measurement → *exposure-episode-construction-rwe*,\n*washout-clean-lookback-period-rwe*, *exposure-lag-induction-latency-window-rwe*, *misclassification-bias-correction-rwe*;\nconfounder identification and control → *dags-backdoor-criterion-drug-studies*,\n*high-dimensional-propensity-score-hdps-rwe*, *baseline-characteristics-and-covariate-balance-rwe*;\nresidual-bias quantification → *e-value-sensitivity-analysis*, *quantitative-bias-analysis-toolkit-rwe*,\n*negative-control-outcomes-rwe*; design data feasibility → *fit-for-purpose-data-assessment-rwe*.\n\n**Applied note (claims/EHR/registry RWE).** A nested case-control study inside a claims or EHR cohort will\noften look strong on the JBI checklist while harboring the biases the checklist barely touches. Use the\nappraisal as a floor — confirm a *validated* phenotype (PPV/sensitivity), an exposure window measured\nidentically for cases and risk-set–sampled controls, and explicit confounding control — then go beyond JBI:\ndocument time-zero/risk-set sampling, run a negative-control-outcome or E-value sensitivity analysis, and\nstate the estimand. JBI tells you whether the study is appraisable and roughly trustworthy; it does not\ncertify it as fit for a regulatory or causal claim.", "primary_category": "Guideline", "tags": [ "guideline", "critical-appraisal", "risk-of-bias", "case-control", "evidence-synthesis", "jbi" ], "aliases": [ "JBI Critical Appraisal Checklist for Case-Control Studies", "JBI Case-Control Checklist", "Joanna Briggs Institute Case-Control Checklist", "JBI CA Case-Control" ], "applies_to_study_types": [ "case_control", "nested_case_control" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "doi": "10.1097/XEB.0000000000000064", "url": "https://doi.org/10.1097/XEB.0000000000000064", "citation_text": "Moola S, Munn Z, Sears K, et al. Conducting systematic reviews of association (etiology): the Joanna Briggs Institute's approach. International Journal of Evidence-Based Healthcare. 2015;13(3):163-169.", "year": 2015, "authors_short": "Moola et al.", "notes": "Published statement of the JBI etiology/risk synthesis approach that introduces the case-control critical-appraisal checklist and its appraisal logic." }, { "role": "explain", "doi": "10.46658/JBIMES-20-08", "url": "https://doi.org/10.46658/JBIMES-20-08", "citation_text": "Moola S, Munn Z, Tufanaru C, et al. Chapter 7: Systematic reviews of etiology and risk. In: Aromataris E, Munn Z (eds). JBI Manual for Evidence Synthesis. JBI; 2020.", "year": 2020, "authors_short": "Moola et al.", "notes": "Currently maintained methodology chapter housing the case-control checklist, item-by-item guidance, and the explicit warning against scoring." }, { "role": "use", "url": "https://jbi.global/critical-appraisal-tools", "citation_text": "JBI Critical Appraisal Tools — Checklist for Case Control Studies (maintained fillable checklist and guidance). JBI.", "year": 2020, "authors_short": "JBI", "notes": "Stable landing page for the downloadable checklist used in practice." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "case-control", "notes": "The primary appraisal target — sampling on outcome status with retrospective exposure ascertainment." }, { "relation_type": "applies_to", "target_slug": "nested-case-control", "notes": "Applies, but pair with risk-set/time-zero concepts the JBI items do not directly probe." }, { "relation_type": "see_also", "target_slug": "case-control", "notes": "Design concept the checklist appraises; defines source-population and control-selection logic." }, { "relation_type": "see_also", "target_slug": "nested-case-control", "notes": "Nested design where risk-set sampling, incidence-density vs cumulative sampling, and time-zero dominate validity beyond the JBI items." }, { "relation_type": "complements", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Implements the JBI requirement that cases/controls be identified by standard, validated criteria." }, { "relation_type": "complements", "target_slug": "dags-backdoor-criterion-drug-studies", "notes": "Operationalizes the \"confounders identified\" item via explicit causal-structure reasoning." }, { "relation_type": "complements", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "Implements the confounder-control item in high-dimensional claims/EHR data." }, { "relation_type": "complements", "target_slug": "misclassification-bias-correction-rwe", "notes": "Addresses the differential-exposure-measurement item that case-control designs are prone to." }, { "relation_type": "see_also", "target_slug": "e-value-sensitivity-analysis", "notes": "Quantifies robustness to the residual/unmeasured confounding the checklist can only flag." }, { "relation_type": "see_also", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "Establishes whether the RWD source can support a credible case-control appraisal at all." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "journal", "hta" ] }, { "slug": "jbi-case-series", "name": "JBI Critical Appraisal Checklist for Case Series", "short_definition": "A 10-item JBI critical-appraisal (risk-of-bias) tool for judging the internal validity and reporting completeness of case series, used primarily to grade study quality during evidence synthesis and journal peer review — not a reporting checklist and not a quality score.", "long_description": "**What it is.** The **JBI Critical Appraisal Checklist for Case Series** is one of the family of\ndesign-specific critical-appraisal tools maintained by the **Joanna Briggs Institute (JBI)** as part of the\nJBI Manual for Evidence Synthesis. A *case series* is a descriptive study that reports a group of patients\nwith the same exposure, condition, or intervention but **without a comparator group** — it sits at the lower\nend of the evidence hierarchy and cannot, by design, estimate a causal effect. The JBI checklist comprises\n**10 questions** answered Yes / No / Unclear / Not applicable, covering: (1) clear inclusion criteria; (2)\nvalid and reliable measurement of the condition for all participants; (3) valid identification methods for the\ncondition; (4) consecutive inclusion of participants; (5) complete inclusion of participants; (6) clear\nreporting of demographics; (7) clear reporting of clinical information; (8) reporting of outcomes or follow-up\nresults; (9) clear reporting of presenting site(s)/clinic(s) demographic information; and (10) appropriate\nstatistical analysis. It is an **appraisal (risk-of-bias-style) instrument** that supports a transparent\nreviewer judgement about whether a series' findings are credible enough to inform a synthesis — it is\nexplicitly **not** a numeric quality score and **not** a reporting guideline.\n\n**When to use.** Reach for the JBI Case Series checklist when you are **appraising the methodological quality\nof a no-comparator case series** as part of a systematic review, scoping review, or evidence map (the dominant\nuse case), or when a journal/editor asks for structured critical appraisal of a submitted series. It is the\ncorrect JBI instrument only when the design is a genuine **case series**: multiple patients, common\nexposure/condition, **no control or comparison arm**, and an aim that is descriptive (characterizing\npresentation, course, or short-term outcomes) rather than comparative. Decision rule for choosing among\nsiblings: a **single patient** → JBI Case Reports checklist or the CARE reporting guideline; **a series with\nno comparator** → this checklist; the moment an **internal or external comparison group** is introduced, the\ndesign becomes a cohort or case-control study and you must appraise it with the JBI Cohort / Case-Control\ntools (and report it under STROBE, or RECORD/RECORD-PE for routinely collected health data). In HTA/payer\ncontexts the checklist is useful for grading the certainty of descriptive real-world or single-arm safety\nevidence feeding an evidence dossier, where case series sometimes carry weight for rare diseases or novel\nexposures.\n\n**What it requires.** Mapped to real-world data practice, the 10 items enforce: **transparent eligibility**\n(explicit, reproducible inclusion criteria rather than a convenience sample); **valid and reliable\nascertainment of the condition** for *every* participant (a phenotype/algorithm question in claims/EHR work);\n**valid identification methods** (how the condition was confirmed — coded diagnosis, chart adjudication,\nlaboratory confirmation); **consecutive and complete inclusion** (the central defense against selection bias\nin a series — were all eligible patients captured, or only the memorable ones?); **complete demographic and\nclinical reporting**; **outcome/follow-up reporting** with attention to who was lost; and **appropriate\nstatistical handling** (descriptive statistics with uncertainty, not inferential claims a no-comparator design\ncannot support). For RWE the load-bearing items are consecutive/complete inclusion (selection bias), valid\ncondition measurement (phenotype validity), and follow-up completeness (attrition).\n\n**When NOT to use — limitations and common misapplications.**\n- **It is an appraisal tool, not a reporting checklist.** Do not use it to *structure* the write-up of a\n series; for reporting a single case use CARE, and for a series adapt CARE/STROBE elements (there is no\n dedicated EQUATOR statement for the case-series design).\n- **It is not a quality score.** Summing Yes answers into a number and thresholding (\"7/10 = high quality\")\n is a misuse JBI explicitly cautions against; the items inform a reasoned overall judgement.\n- **Wrong tool for the design.** A single patient is a *case report* (use the JBI Case Reports tool / CARE),\n not a case series. If a comparison group exists — even an external/historical control — the study is no\n longer a case series; appraise it as a cohort/case-control study (JBI Cohort/Case-Control + STROBE /\n RECORD-PE / HARPER for the analytic design).\n- **Appraisal does not make a series causal.** A flawless checklist score does not let a no-comparator series\n estimate a treatment effect, incidence rate ratio, or hazard ratio; without a comparator the design supports\n description and hypothesis generation only.\n- **Checklist-as-theater.** Pasting a completed grid into the appendix without using it to inform inclusion,\n sensitivity analyses, or certainty grading (e.g., GRADE) adds no value.\n\n**How it maps to this catalog.** Each JBI item points to a concept in this repository that implements the\nunderlying methodological requirement:\n- *Inclusion criteria / consecutive & complete inclusion (items 1, 4, 5)* → `descriptive-epidemiology-rwe`,\n `safety-signal-case-definition-rwe`, and `selection-bias-sensitivity-analysis-rwe` for diagnosing\n convenience-vs-consecutive sampling bias.\n- *Valid measurement and identification of the condition (items 2, 3)* →\n `diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe`, `outcome-algorithm-construction-rwe`, and\n `algorithm-validation` (PPV/sensitivity of the case definition), with chart adjudication handled in\n `endpoint-adjudication-chart-review-rwe`.\n- *Demographic and clinical reporting (items 6, 7, 9)* →\n `baseline-characteristics-and-covariate-balance-rwe`.\n- *Outcome and follow-up reporting (item 8)* → `attrition-and-loss-to-follow-up-rwe` and\n `missing-data-pattern-table-rwe`.\n- *Appropriate statistical analysis (item 10)* → `descriptive-epidemiology-rwe` and\n `incidence-rate-calculation-rwe` for honest descriptive summaries; if a comparator is added the analysis\n moves into `target-trial-emulation`, `active-comparator-new-user`, and\n `estimands-ate-att-intercurrent-events-rwe` — and the appraisal tool changes accordingly.\n\n**Applied note (claims/EHR/registry RWE).** A claims- or EHR-derived \"case series\" of patients dispensed a\nnew agent is only consecutive and complete if the data infrastructure captures *all* eligible patients with\ncontinuous observability — so items 4–5 become questions about `continuous-enrollment-observable-time-rwe` and\n`database-feasibility-attrition-funnel-rwe`. Item 2–3 (valid condition measurement) are answered by a\nvalidated phenotype rather than a clinician's note. Reviewers appraising registry case series should confirm\nwhether enrollment was consecutive (population-based registry) or selective (referral/voluntary registry),\nbecause a convenience registry fails item 4 even when each record is well documented.", "primary_category": "Guideline", "tags": [ "guideline", "critical-appraisal", "risk-of-bias", "quality-assessment", "case-series", "evidence-synthesis", "jbi" ], "aliases": [ "JBI Case Series", "JBI Critical Appraisal Checklist for Case Series", "JBI CAT — Case Series", "Joanna Briggs Institute Case Series Checklist" ], "applies_to_study_types": [ "case_series" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "doi": "10.11124/JBISRIR-D-19-00099", "url": "https://doi.org/10.11124/JBISRIR-D-19-00099", "citation_text": "Munn Z, Barker TH, Moola S, et al. Methodological quality of case series studies: an introduction to the JBI critical appraisal tool. JBI Database of Systematic Reviews and Implementation Reports. 2019;18(10):2127-2133.", "year": 2019, "authors_short": "Munn et al.", "notes": "Canonical methodological paper introducing the JBI case-series critical-appraisal tool, the rationale for each item, and the guidance that it is an appraisal instrument rather than a numeric quality score." }, { "role": "explain", "doi": "10.1136/bmjebm-2017-110853", "url": "https://doi.org/10.1136/bmjebm-2017-110853", "citation_text": "Murad MH, Sultan S, Haffar S, Bazerbachi F. Methodological quality and synthesis of case series and case reports. BMJ Evidence-Based Medicine. 2018;23(2):60-63.", "year": 2018, "authors_short": "Murad et al.", "notes": "Widely cited companion framework for appraising and synthesizing case series and reports; clarifies the limits of no-comparator descriptive evidence and complements the JBI item set." }, { "role": "use", "url": "https://jbi.global/critical-appraisal-tools", "citation_text": "Joanna Briggs Institute. JBI Critical Appraisal Tools — Checklist for Case Series. JBI, Adelaide.", "year": 2020, "authors_short": "Joanna Briggs Institute", "notes": "Maintained, downloadable checklist and the current JBI Manual for Evidence Synthesis guidance on applying the tool." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "case-series", "notes": "The instrument is intended for appraising the methodological quality of no-comparator case series." }, { "relation_type": "see_also", "target_slug": "case-report", "notes": "For a single patient use the JBI Case Reports tool / CARE guideline, not the case-series checklist — a common wrong-tool-for-design error." }, { "relation_type": "see_also", "target_slug": "cohort-retrospective", "notes": "If the study includes a comparison group it is a cohort (or case-control) study; appraise with the JBI Cohort/Case-Control tools and report under STROBE/RECORD, not this checklist." }, { "relation_type": "used_with", "target_slug": "selection-bias-sensitivity-analysis-rwe", "notes": "Items 4-5 (consecutive and complete inclusion) are the checklist's central defense against selection bias in convenience series." }, { "relation_type": "used_with", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Items 2-3 (valid, reliable measurement and identification of the condition) are answered in claims/EHR work by a validated phenotype/case-definition algorithm." }, { "relation_type": "used_with", "target_slug": "algorithm-validation", "notes": "Establishing PPV/sensitivity of the case definition underpins the validity items for routinely collected data." }, { "relation_type": "used_with", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Item 8 (outcomes / follow-up reporting) requires transparent accounting of who was lost to follow-up." }, { "relation_type": "used_with", "target_slug": "descriptive-epidemiology-rwe", "notes": "Item 10 (appropriate statistical analysis) for a no-comparator series means honest descriptive statistics, not inferential effect estimates." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "journal", "hta" ] }, { "slug": "jbi-cohort", "name": "JBI Critical Appraisal Tool for Cohort Studies", "short_definition": "A domain-based critical-appraisal / risk-of-bias instrument maintained by JBI for assessing cohort studies during evidence synthesis; the 2024 revision reframes the older 11-item checklist as structured risk-of-bias judgments rather than a summed quality score.", "long_description": "**What it is.** The **JBI Critical Appraisal Tool for Cohort Studies** is a structured instrument for judging the\ninternal validity (risk of bias) of cohort studies during systematic reviews and evidence synthesis. It is developed\nand maintained by **JBI** (formerly the Joanna Briggs Institute, University of Adelaide), the organization behind the\nJBI suite of design-specific appraisal tools (cohort, case-control, RCT, prevalence, case series, analytical\ncross-sectional, etc.) used across JBI- and Cochrane-adjacent reviews. The original tool was an 11-item checklist\nscored Yes / No / Unclear / Not applicable. The **2024 revision (Barker et al.)** is a substantive redesign: items were\nrewritten, mapped to explicit bias domains (group comparability and selection, exposure measurement, confounding\nidentification and management, outcome measurement, follow-up completeness and attrition, and appropriateness of the\nstatistical analysis), and — critically — the tool was reframed as a **per-domain risk-of-bias judgment** that converges\ntoward an overall judgment, NOT a numeric quality score. It is an appraisal tool, not a reporting checklist.\n\n**When to use.** Use the JBI Cohort tool when you are the *reviewer/appraiser* of a cohort study within an evidence\nsynthesis (systematic review, scoping-to-synthesis, rapid review, or evidence brief) that feeds a peer-reviewed\npublication or an **HTA / payer dossier**. It is the natural appraisal instrument when your review protocol already\nfollows JBI methodology for reviews of etiology/association or effectiveness. Decision rules versus siblings: appraise a\n**cohort** (including most non-interventional RWE cohorts built from claims/EHR/registry) with this tool; switch to the\n**JBI Case-Control** tool for case-control designs, the **JBI Analytical Cross-Sectional** tool for prevalence/\ncross-sectional designs, and the **JBI RCT** tool for randomized trials. For target-trial-emulation studies that mimic an\nRCT but are analyzed as cohorts, the cohort tool is appropriate, supplemented by emulation-specific scrutiny. ROBINS-I is\na reasonable alternative when the review demands a per-outcome, signed-bias framework tied to a target trial; many HTA\nbodies accept either, and some prefer ROBINS-I for intervention questions. JBI Cohort is for **appraising** a study, not\nfor **reporting** your own — that is the job of STROBE (general observational), RECORD / RECORD-PE (routinely collected\nhealth data and pharmacoepidemiology), or HARPER (HARmonized Protocol Template for pharmacoepi).\n\n**What it requires.** The tool forces a reviewer to interrogate the bias-relevant features of the study, framed here for\nreal-world data:\n- **Comparability and selection of groups** — were exposed and unexposed (or comparator) groups recruited from the same\n source population and comparable on baseline prognosis? In RWE this is the active-comparator / new-user question.\n- **Exposure measurement** — was exposure measured validly and the same way in both groups? For claims/EHR this is\n algorithm/phenotype validity (e.g., dispensing-based exposure, time-zero alignment).\n- **Confounding** — were the important confounders identified, *and* were strategies stated to deal with them\n (restriction, matching, propensity scores, regression)? The revision separates *identification* from *management*.\n- **Outcome measurement** — were outcomes measured validly, reliably, and blind to exposure where relevant? In RWE this\n maps to outcome-algorithm performance (PPV/sensitivity).\n- **Follow-up and attrition** — was follow-up long enough and complete, and was incomplete follow-up addressed?\n- **Statistical analysis** — was the analysis appropriate (including handling of time, confounding, and missing data)?\nThe reviewer reaches a domain-level risk-of-bias judgment (e.g., low / moderate / high / unclear) and an overall\njudgment supported by recorded rationale — not a tally of \"Yes\" answers.\n\n**When NOT to use — limitations and common misapplications.**\n- **It is not a reporting guideline.** Completing JBI Cohort tells you whether a study is *at risk of bias*, not whether\n it is *well reported*. Using it where STROBE / RECORD-PE / HARPER is required (e.g., a journal or regulator asking for\n reporting completeness) is a category error.\n- **It is not a quality score.** The single most common misapplication — endemic with the older 11-item version — is\n **summing the \"Yes\" answers into a numeric quality score** and dichotomizing studies at an arbitrary cut-point\n (e.g., \"≥7/11 = high quality\"). The 2024 revision was designed precisely to stop this: item counts are not\n interchangeable, a single fatally biased domain can invalidate a study with many \"Yes\" items, and meta-analytic\n weighting by such scores is methodologically indefensible. Report domain judgments, not a score.\n- **Appraisal is not causal inference.** A \"low risk of bias\" rating does not make an observational study causal; it\n means the *reported* design and analysis guard against known biases. Unmeasured confounding can remain invisible to any\n checklist.\n- **Checklist-as-theater.** Marking items without recording the specific evidence and rationale (the study text, the\n algorithm, the attrition figures) produces a defensible-looking artifact with no analytic content. HTA reviewers\n discount unjustified ratings.\n- **Wrong tool for the design.** Applying the cohort tool to a case-control or cross-sectional study (or vice versa)\n mis-frames the relevant biases.\n\n**How it maps to this catalog.** Each JBI domain is *implemented* by one or more concepts here:\n- Group comparability / selection → **active-comparator-new-user**, **target-trial-emulation**,\n **selection-bias-sensitivity-analysis-rwe**.\n- Exposure measurement / time-zero → **diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe**,\n **immortal-time-bias-handling**.\n- Confounding identification and management → **high-dimensional-propensity-score-hdps-rwe**,\n **propensity-score-methods-psm-iptw**, **unmeasured-confounding-probabilistic-bias-analysis-rwe**,\n **negative-control-outcomes-rwe**, **estimands-ate-att-intercurrent-events-rwe**.\n- Outcome measurement → **claims-outcome-algorithm-ppv-sensitivity-rwe**.\n- Follow-up / attrition → **attrition-and-loss-to-follow-up-rwe**.\n- Data substrate / overall feasibility → **claims-analysis**.\n\n**Applied note (claims/EHR/registry RWE).** When appraising a claims- or EHR-based cohort, do not accept a face-value\n\"exposure was measured\" or \"outcomes were ascertained.\" Demand the operational evidence behind each domain: the washout\nand continuous-enrollment definition behind incident-user status; the phenotype/algorithm and its validation (PPV,\nsensitivity) behind exposure and outcomes; the covariate-measurement window and confounding-control strategy; the\nattrition funnel and how disenrollment, death, and switching were handled; and the time-zero alignment that prevents\nimmortal time. The JBI domains give you the questions; the catalog concepts give you the standard against which a \"low\nrisk of bias\" judgment can actually be defended in an HTA dossier.", "primary_category": "Guideline", "tags": [ "guideline", "critical-appraisal", "risk-of-bias", "cohort", "evidence-synthesis", "jbi" ], "aliases": [ "JBI Cohort", "JBI Critical Appraisal Checklist for Cohort Studies", "JBI cohort checklist", "JBI risk-of-bias tool for cohort studies (2024 revision)", "Joanna Briggs Institute cohort appraisal tool" ], "applies_to_study_types": [ "cohort_prospective", "cohort_retrospective" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "doi": "10.11124/JBIES-24-00103", "url": "https://doi.org/10.11124/JBIES-24-00103", "citation_text": "Barker TH, Habibi N, Aromataris E, et al. The revised JBI critical appraisal tool for the assessment of risk of bias for cohort studies. JBI Evidence Synthesis. 2024;22(3):378-388.", "year": 2024, "authors_short": "Barker et al.", "notes": "Canonical statement of the current (2024) domain-based revision; reframes the tool as a risk-of-bias judgment rather than a summed quality score." }, { "role": "explain", "doi": "10.11124/JBIES-22-00125", "url": "https://doi.org/10.11124/JBIES-22-00125", "citation_text": "Barker TH, Stone JC, Sears K, et al. Revising the JBI quantitative critical appraisal tools to improve their applicability: an overview of methods and the development process. JBI Evidence Synthesis. 2022;21(3):478-493.", "year": 2022, "authors_short": "Barker et al.", "notes": "Describes the methodology and rationale for the redesign across the JBI quantitative appraisal suite, including the move from item-counting to domain-based bias assessment." }, { "role": "explain", "doi": "10.1097/XEB.0000000000000064", "url": "https://doi.org/10.1097/XEB.0000000000000064", "citation_text": "Moola S, Munn Z, Sears K, et al. Conducting systematic reviews of association (etiology): the Joanna Briggs Institute's approach. International Journal of Evidence-Based Healthcare. 2015;13(3):163-169.", "year": 2015, "authors_short": "Moola et al.", "notes": "Situates the cohort appraisal tool within JBI methodology for reviews of etiology/association, where appraising cohort evidence is central." }, { "role": "use", "url": "https://jbi.global/critical-appraisal-tools", "citation_text": "JBI Critical Appraisal Tools — maintained tool downloads and guidance, JBI.", "year": 2024, "authors_short": "JBI", "notes": "Official maintained source for the current downloadable tool and instructions." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "cohort-prospective", "notes": "Appraise prospective cohort studies (including RWE cohorts) for risk of bias during evidence synthesis." }, { "relation_type": "applies_to", "target_slug": "cohort-retrospective", "notes": "Appraise retrospective/database cohort studies (claims, EHR, registry) for risk of bias during evidence synthesis." }, { "relation_type": "see_also", "target_slug": "active-comparator-new-user", "notes": "Implements the group-comparability/selection domain — comparable groups from the same source population at a shared time zero." }, { "relation_type": "see_also", "target_slug": "target-trial-emulation", "notes": "Strengthens comparability and time-zero appraisal for emulated-trial cohorts analyzed under the cohort tool." }, { "relation_type": "see_also", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Supplies the exposure/condition-measurement evidence the exposure-measurement domain demands." }, { "relation_type": "see_also", "target_slug": "claims-outcome-algorithm-ppv-sensitivity-rwe", "notes": "Supplies outcome-algorithm performance (PPV/sensitivity) for the outcome-measurement domain." }, { "relation_type": "see_also", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "Implements the confounding-management domain in high-dimensional claims/EHR data." }, { "relation_type": "see_also", "target_slug": "propensity-score-methods-psm-iptw", "notes": "Standard confounding-control techniques underlying the confounding-management domain judgment." }, { "relation_type": "see_also", "target_slug": "unmeasured-confounding-probabilistic-bias-analysis-rwe", "notes": "Quantitative bias analysis for residual/unmeasured confounding that no checklist alone can detect." }, { "relation_type": "see_also", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Implements the follow-up-completeness/attrition domain." }, { "relation_type": "see_also", "target_slug": "immortal-time-bias-handling", "notes": "Time-zero alignment that the exposure-measurement and analysis domains must scrutinize." }, { "relation_type": "see_also", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Clarifies the estimand and intercurrent-event handling underlying the analysis-appropriateness domain." }, { "relation_type": "see_also", "target_slug": "selection-bias-sensitivity-analysis-rwe", "notes": "Tools for probing the selection/comparability domain." }, { "relation_type": "see_also", "target_slug": "claims-analysis", "notes": "General data-fitness considerations for the claims/EHR substrate underlying every domain in RWE cohorts." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "hta", "journal" ] }, { "slug": "jbi-cross-sectional", "name": "JBI Critical Appraisal Tool for Analytical Cross-Sectional Studies", "short_definition": "JBI's structured critical-appraisal / risk-of-bias instrument for analytical (association/etiology) cross-sectional studies, used to judge the methodological trustworthiness of an individual study during evidence synthesis — not a reporting checklist and not a numeric quality score.", "long_description": "**What it is** — The **JBI Critical Appraisal Tool for Analytical Cross-Sectional Studies** is a\nrisk-of-bias / critical-appraisal instrument maintained by **JBI (formerly the Joanna Briggs Institute)**\nas part of the JBI Manual for Evidence Synthesis. It is one of a *family* of design-specific JBI tools\n(cohort, case-control, case series, prevalence, qualitative, RCT, quasi-experimental). Its purpose is to\nlet two independent appraisers judge, item by item, whether the *internal validity* of a single analytical\ncross-sectional study can be trusted before that study is included in or weighted within a systematic review,\nmeta-analysis, or HTA evidence base. The original eight-item tool was operationalised by Moola and colleagues\n(the JBI etiology/association approach); JBI released a **revised tool in 2025/2026 (Barker et al.)** that\nreframes the items explicitly as **risk-of-bias domains** with response options Yes / No / Unclear / Not\napplicable and per-item guidance. It is a judgment instrument — it tells you how much to believe a study,\nnot how to write one up.\n\n**When to use** — Reach for this tool when you are **appraising** (not authoring) an *analytical*\ncross-sectional study — one that examines an association between an exposure and an outcome measured at the\nsame point in time (e.g., a prevalence-of-disease study that also estimates exposure-outcome odds ratios).\nThe typical decision contexts are: a **systematic review or meta-analysis** of observational evidence where\neach included study needs a documented risk-of-bias assessment; an **HTA/payer evidence synthesis** that\nmust grade the certainty of the included real-world studies (often feeding a GRADE assessment); and\n**peer-reviewed methods reporting** where reviewers expect a transparent appraisal table. Decision rules for\npicking the *right* member of the family: if the included study merely *describes* a prevalence or proportion\nwith no exposure-outcome contrast, use the **JBI Prevalence** tool instead; if subjects are followed over\ntime, use **JBI Cohort**; if cases and controls are sampled on outcome, use **JBI Case-Control**. And\ncritically: if your task is to *report* your own cross-sectional study, this is the wrong document — use the\nreporting guideline **STROBE (cross-sectional)** instead.\n\n**What it requires** — The tool enforces appraisal across eight substantive domains, each answered with\nevidence from the paper: (1) clearly defined **inclusion/eligibility criteria** for the sample; (2) detailed\ndescription of **study subjects and setting** (the source population and sampling frame); (3) **valid and\nreliable measurement of the exposure** — in RWD terms, the phenotype/algorithm used to define exposure; (4)\n**objective, standardised criteria for measuring the condition/outcome** — the outcome algorithm and its\nvalidation; (5) explicit **identification of confounding factors**; (6) explicit **strategies to deal with\nconfounding** (restriction, matching, adjustment, weighting); (7) **valid and reliable measurement of\noutcomes**; and (8) **appropriate statistical analysis**. For real-world data the load-bearing items are 3,\n4, 5, 6 and 7: misclassification of exposure or outcome from imperfect claims/EHR algorithms, and\nuncontrolled confounding, are exactly where cross-sectional RWD studies fail. The revised (Barker) tool\npushes appraisers to reason about the *direction and magnitude* of bias each domain introduces rather than\nticking a box.\n\n**When NOT to use — limitations and common misapplications** —\n- **It is a risk-of-bias instrument, not a reporting checklist.** Using it to structure how you *write* a\n cross-sectional study is a category error; STROBE-CSS is the reporting tool. Conversely, completing STROBE\n does not appraise a study — it only documents it.\n- **Do not convert the items into a numeric quality score and threshold** (e.g., \"include if ≥6/8\"). JBI\n explicitly advises against summing items into a cut-off; doing so masks which specific domain is fatally\n biased and gives spurious precision. Report each domain's judgment.\n- **Wrong tool in the family.** Routing a descriptive prevalence study, a cohort, or a case-control study\n through the analytical cross-sectional tool (or vice versa) is a frequent reviewer error.\n- **Passing the checklist does not make the study causal.** A cross-sectional design measures exposure and\n outcome simultaneously, so **temporality is unestablished** — reverse causation and prevalence-incidence\n (Neyman) bias survive a clean appraisal. A \"low risk of bias\" rating is a statement about execution, not\n about whether the design can support a causal claim.\n- **Appraisal-as-theater.** A single appraiser filling boxes without dual independent review, adjudication\n of disagreements, and a narrative on bias direction defeats the purpose.\n\n**How it maps to this catalog** — This guideline tells a reviewer *what to interrogate*; the following\nconcepts tell them *how each requirement is operationalised in RWD*:\n- Item 1–2 (eligibility, sampling frame) → `cross-sectional`, `descriptive-epidemiology-rwe`,\n `selection-bias-sensitivity-analysis-rwe`.\n- Item 3 (exposure measurement validity) → `diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe`,\n `claims-analysis`, `ehr-study`.\n- Items 4 & 7 (outcome/condition measurement validity) → `claims-outcome-algorithm-ppv-sensitivity-rwe`,\n `misclassification-bias-correction-rwe`.\n- Items 5–6 (identifying and handling confounding) → `dags-backdoor-criterion-drug-studies`,\n `quantitative-bias-analysis-toolkit-rwe`.\n- Item 8 (appropriate analysis of a prevalence/association estimate) → `prevalence-point-period-annual-rwe`.\n\n**Applied note (claims/EHR/registry RWE).** For a claims- or EHR-based analytical cross-sectional study, the\nappraisal must look behind every \"valid measurement\" claim: an exposure defined by a single diagnosis code\nhas low positive predictive value (weakening item 3); an outcome captured only when a patient happens to\nhave an encounter introduces differential ascertainment (item 4/7); and because exposure and outcome are read\nfrom the same cross-section of the data, you cannot tell which came first — so even a tool-clean study should\nbe downgraded for temporality and for unmeasured confounding when grading certainty for an HTA dossier.", "primary_category": "Guideline", "tags": [ "guideline", "critical-appraisal", "risk-of-bias", "jbi", "cross-sectional", "evidence-synthesis" ], "aliases": [ "JBI Analytical Cross-Sectional Checklist", "JBI Critical Appraisal Checklist for Analytical Cross-Sectional Studies", "JBI cross-sectional risk-of-bias tool", "Joanna Briggs Institute cross-sectional appraisal tool" ], "applies_to_study_types": [ "cross_sectional" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "doi": "10.1097/XEB.0000000000000064", "url": "https://doi.org/10.1097/XEB.0000000000000064", "citation_text": "Moola S, Munn Z, Sears K, et al. Conducting systematic reviews of association (etiology): The Joanna Briggs Institute's approach. International Journal of Evidence-Based Healthcare. 2015;13(3):163-169.", "year": 2015, "authors_short": "Moola et al.", "notes": "Canonical methodology paper articulating the JBI approach to appraising analytical cross-sectional (etiology/association) studies and the original eight-item critical-appraisal tool." }, { "role": "explain", "doi": "10.11124/JBIES-24-00523", "url": "https://doi.org/10.11124/JBIES-24-00523", "citation_text": "Barker TH, Hasanoff S, Aromataris E, et al. The revised JBI critical appraisal tool for the assessment of risk of bias for analytical cross-sectional studies. JBI Evidence Synthesis. 2025;24(3):401-408.", "year": 2025, "authors_short": "Barker et al.", "notes": "Revised tool that reframes the items as explicit risk-of-bias domains with Yes/No/Unclear/Not-applicable response options and per-item guidance on bias direction." }, { "role": "use", "doi": "10.46658/JBIMES-20-08", "url": "https://doi.org/10.46658/JBIMES-20-08", "citation_text": "Moola S, Munn Z, Tufanaru C, et al. Chapter 7: Systematic reviews of etiology and risk. In: Aromataris E, Munn Z (eds). JBI Manual for Evidence Synthesis. JBI; 2020.", "year": 2020, "authors_short": "Moola et al. (JBI Manual)", "notes": "Maintained authoritative chapter housing the appraisal checklist and its application within JBI evidence-synthesis workflow; the JBI critical-appraisal tools landing page is https://jbi.global/critical-appraisal-tools." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "cross-sectional", "notes": "Use to appraise the internal validity / risk of bias of an analytical cross-sectional study being considered for inclusion in a synthesis." }, { "relation_type": "see_also", "target_slug": "strobe-css", "notes": "STROBE (cross-sectional) is the REPORTING guideline for cross-sectional studies; JBI appraises, STROBE reports. Do not substitute one for the other." }, { "relation_type": "see_also", "target_slug": "jbi-prevalence", "notes": "Sister JBI tool for purely descriptive prevalence/incidence studies; route descriptive (non-association) studies there rather than through the analytical tool." }, { "relation_type": "used_with", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Implements the \"valid and reliable measurement of exposure\" item when exposure is defined from claims/EHR codes." }, { "relation_type": "used_with", "target_slug": "claims-outcome-algorithm-ppv-sensitivity-rwe", "notes": "Implements the \"objective/valid measurement of the condition and outcomes\" items via outcome-algorithm validation (PPV, sensitivity)." }, { "relation_type": "used_with", "target_slug": "dags-backdoor-criterion-drug-studies", "notes": "Implements the \"identification and handling of confounding\" items by making the confounding structure explicit." }, { "relation_type": "used_with", "target_slug": "quantitative-bias-analysis-toolkit-rwe", "notes": "Quantifies the magnitude/direction of residual bias the appraisal flags (misclassification, unmeasured confounding) rather than leaving it qualitative." }, { "relation_type": "see_also", "target_slug": "selection-bias-sensitivity-analysis-rwe", "notes": "Supports appraisal of the sampling-frame and inclusion-criteria items where selection into the cross-section may be informative." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "journal", "hta" ] }, { "slug": "jbi-prevalence", "name": "JBI Critical Appraisal Checklist for Prevalence Studies", "short_definition": "A JBI critical-appraisal (risk-of-bias) instrument for systematic reviews of prevalence and cumulative-incidence data, scoring whether a study's sampling frame, case ascertainment, coverage, and statistical reporting support a trustworthy prevalence estimate.", "long_description": "**What it is** — The **JBI Critical Appraisal Checklist for Prevalence Studies** is a nine-item critical-appraisal (risk-of-bias) instrument maintained by **JBI** (formerly the Joanna Briggs Institute, Adelaide) as part of its suite of design-specific appraisal tools and the *JBI Manual for Evidence Synthesis*. It was developed by Munn and colleagues specifically because generic observational-study appraisal tools (e.g., the Newcastle-Ottawa Scale, STROBE-derived checklists) do not interrogate the things that actually threaten a **prevalence** estimate: how the sample was drawn, whether it represents the target population, how the condition was measured, and whether the proportion and its precision were reported correctly. The nine items ask whether (1) the sample frame appropriately addressed the target population, (2) study participants were sampled appropriately, (3) the sample size was adequate, (4) study subjects and setting were described in detail, (5) data analysis covered the identified sample with sufficient coverage, (6) valid methods identified the condition, (7) the condition was measured in a standard, reliable way for all participants, (8) appropriate statistical analysis was used, and (9) the response rate was adequate (or low response rate was managed appropriately). It is a tool for *appraising included studies inside a systematic review of prevalence*, not a reporting checklist for authors and not a numeric quality score.\n\n**When to use** — Use this checklist when you are conducting (or reviewing) a **systematic review or meta-analysis of prevalence, point-prevalence, period-prevalence, or cumulative-incidence (risk) estimates** and need to assess the methodological quality / risk of bias of each included cross-sectional or descriptive study. Typical decision contexts: an HTA/payer epidemiology section that must justify the population size or eligible-patient counts behind a budget-impact or cost-effectiveness model; a burden-of-disease or unmet-need chapter in a value dossier; a peer-reviewed prevalence systematic review; or a regulatory background-epidemiology section supporting an orphan-designation or natural-history submission. Decision rules for choosing THIS tool over siblings: if your review question is \"**how common is the condition?**\" use JBI Prevalence; if it is \"**does exposure A cause/associate with outcome B?**\" use a tool for analytic observational studies (JBI for cohort/case-control, or ROBINS-I / Newcastle-Ottawa); if it is \"**how accurate is a diagnostic test?**\" use QUADAS-2 / JBI diagnostic-test-accuracy; if you are *authoring* and need a *reporting* checklist (not appraisal) reach for **STROBE** (and, for routinely-collected health data, **RECORD**), which are complementary to — not substitutes for — this appraisal tool.\n\n**What it requires** — Translated into real-world-data terms, the nine items enforce four substantive domains. (1) **Sampling and target-population fit** (items 1-3, 9): the source population and sampling frame must map to the population the estimate is meant to describe, with adequate size and an adequate, non-differential response/capture rate — in claims/EHR this becomes whether the enrolled or empaneled population is representative of the target and whether denominators are correctly defined. (2) **Setting and case definition transparency** (items 4, 6): the population, setting, and condition definition must be described in enough detail to reproduce — in RWD this is precisely **phenotype/algorithm specification** (code lists, diagnosis-window logic) and documentation of the data source. (3) **Coverage and measurement validity/reliability** (items 5, 7): the analysis must cover the identified sample with low attrition and the condition must be measured the same validated way in everyone — in RWD this is **algorithm validation** (PPV/sensitivity), continuous-enrollment requirements, and consistent ascertainment across subgroups. (4) **Correct statistical reporting** (item 8): the proportion must be reported with an appropriate confidence interval and, where relevant, stratification — not a point estimate alone. Note what the tool deliberately does **not** assess: it is not designed for confounding control, time-zero alignment, estimands/intercurrent events, or comparative effect estimation, because prevalence is a descriptive parameter, not a causal contrast.\n\n**When NOT to use — limitations and common misapplications** — This is a **risk-of-bias appraisal instrument, not a reporting guideline and not a validated quality score**: JBI itself discourages summing the nine items into a numeric cut-off, because an arbitrary \"7/9 = high quality\" threshold hides which domain failed and weights non-equivalent items equally. Do not use it to appraise **analytic/comparative** designs — applying a prevalence tool to a cohort study that estimates a hazard ratio leaves confounding, immortal time, and selection entirely unexamined; that is a wrong-tool error directly analogous to using STROBE where RECORD-PE is required. Do not treat completion of the checklist as evidence the estimate is **unbiased or generalizable** — a study can pass all nine items within a non-representative claims population and still produce a prevalence that does not transport to the decision population. Avoid **checklist-as-theater**: marking \"yes/no/unclear\" without extracting the underlying sampling frame, denominator, and algorithm validity adds no information. Finally, a single JBI per-study appraisal does not substitute for a **certainty-of-evidence** judgment across the body of prevalence estimates (e.g., GRADE-style downgrading for heterogeneity and indirectness).\n\n**How it maps to this catalog** — Several catalog concepts implement what individual JBI items demand for claims/EHR/registry reviews. Item 6 (valid identification of the condition) and item 4 (case-definition detail) are implemented by **diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe**, which gives the operational code-list and time-window logic, and its validity is what item 7 (standard, reliable measurement) actually scores. Item 5 (adequate coverage of the identified sample) and item 9 (response/capture rate) are implemented by **attrition-and-loss-to-follow-up-rwe**, which formalizes continuous-enrollment and loss-to-capture handling. Items 1-3 (sample frame and target-population fit) connect to **claims-analysis** for denominator construction and to data-source representativeness work (e.g., Medicare FFS vs Medicare Advantage vs commercial coverage differences) that determines whether the numerator/denominator describe the intended population. When a downstream review pivots from descriptive prevalence to a comparative or causal question, the relevant concepts shift to **active-comparator-new-user**, **high-dimensional-propensity-score-hdps-rwe**, **estimands-ate-att-intercurrent-events-rwe**, and **target-trial-emulation** — a signal that JBI Prevalence is no longer the right appraisal instrument. Applied note for RWD: when appraising a claims-based prevalence study, read item 1 as \"is the enrolled population the target population?\", item 5/9 as \"is the continuous-enrollment denominator complete and non-differential?\", and item 6/7 as \"is the phenotype validated (reported PPV/sensitivity) and applied identically to all enrollees?\" — these three questions carry most of the bias in real-world prevalence estimates.", "primary_category": "Guideline", "tags": [ "guideline", "critical-appraisal", "risk-of-bias", "prevalence", "systematic-review", "jbi", "descriptive-epidemiology" ], "aliases": [ "JBI Prevalence Checklist", "JBI Critical Appraisal Checklist for Studies Reporting Prevalence Data", "JBI prevalence critical appraisal tool", "Joanna Briggs Institute prevalence appraisal" ], "applies_to_study_types": [ "cross_sectional", "prevalence", "systematic_review", "meta_analysis_obs" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "doi": "10.15171/ijhpm.2014.71", "url": "https://doi.org/10.15171/ijhpm.2014.71", "citation_text": "Munn Z, Moola S, Riitano D, Lisy K. The development of a critical appraisal tool for use in systematic reviews addressing questions of prevalence. International Journal of Health Policy and Management. 2014;3(3):123-128.", "year": 2014, "authors_short": "Munn et al.", "notes": "Original development and item-by-item rationale for the JBI prevalence critical-appraisal checklist; defines why prevalence-specific appraisal differs from generic observational-study tools." }, { "role": "explain", "doi": "10.1097/XEB.0000000000000054", "url": "https://doi.org/10.1097/XEB.0000000000000054", "citation_text": "Munn Z, Moola S, Lisy K, Riitano D, Tufanaru C. Methodological guidance for systematic reviews of observational epidemiological studies reporting prevalence and cumulative incidence data. International Journal of Evidence-Based Healthcare. 2015;13(3):147-153.", "year": 2015, "authors_short": "Munn et al.", "notes": "Companion methods paper situating the checklist within the full JBI prevalence systematic-review workflow (question formulation, search, extraction, synthesis, and appraisal)." }, { "role": "use", "url": "https://jbi.global/critical-appraisal-tools", "citation_text": "JBI. Critical Appraisal Tools — Checklist for Prevalence Studies. JBI (Joanna Briggs Institute), Adelaide. Maintained tool and accompanying JBI Manual for Evidence Synthesis.", "year": 2024, "authors_short": "JBI", "notes": "Canonical, maintained source for the current checklist form and the JBI Manual chapter governing its application." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "cross-sectional", "notes": "Primary design appraised — cross-sectional studies reporting a prevalence proportion." }, { "relation_type": "applies_to", "target_slug": "prevalence-point-period-annual-rwe", "notes": "The checklist is purpose-built to appraise prevalence and cumulative-incidence estimates." }, { "relation_type": "applies_to", "target_slug": "systematic-review", "notes": "Used as the per-study critical-appraisal step inside a systematic review of prevalence." }, { "relation_type": "complements", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Implements items 4/6/7 (case definition, valid identification, standard measurement) for claims/EHR prevalence studies; its reported PPV/sensitivity is what those items score." }, { "relation_type": "complements", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Implements items 5/9 (coverage of the identified sample, adequate/non-differential response or capture rate) via continuous-enrollment and loss-to-capture handling." }, { "relation_type": "complements", "target_slug": "claims-analysis", "notes": "Provides denominator construction and source-population definition underlying items 1-3 (sample frame and target-population fit) for claims-based prevalence." }, { "relation_type": "see_also", "target_slug": "target-trial-emulation", "notes": "When the question shifts from descriptive prevalence to a comparative/causal estimand, JBI Prevalence no longer applies; move to target-trial emulation and analytic appraisal tools." }, { "relation_type": "see_also", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Marks the boundary of this tool — prevalence is a descriptive parameter, so estimands and intercurrent events are out of scope and signal a different question/tool." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "hta", "journal", "fda" ] }, { "slug": "jbi-qualitative", "name": "JBI Critical Appraisal Checklist for Qualitative Research", "short_definition": "A 10-item critical-appraisal instrument from the Joanna Briggs Institute used to assess the methodological congruity and trustworthiness of primary qualitative studies (interviews, focus groups, ethnography) for inclusion in qualitative evidence syntheses and patient-experience evidence.", "long_description": "**What it is** — The **JBI Critical Appraisal Checklist for Qualitative Research** is a 10-item methodological\nappraisal instrument maintained by the **Joanna Briggs Institute (JBI)** as part of the JBI Manual for Evidence\nSynthesis and the JBI suite of design-specific critical-appraisal tools. It is the qualitative counterpart to JBI's\nchecklists for RCTs, cohort, case-control, and prevalence studies. Its purpose is to judge whether a *primary\nqualitative study* was conducted and reported with internal congruity and methodological trustworthiness — i.e.,\nwhether the stated philosophical position, methodology, research question, data collection, analysis, and\ninterpretation hang together, and whether the researcher's influence and the participants' voices were handled\ndefensibly. It is a **critical-appraisal / risk-of-trustworthiness instrument**, not a reporting checklist and not a\nnumeric quality score; the canonical methodological articulation is Lockwood, Munn & Porritt (2015), with confidence\nin the *synthesized* findings graded separately via the ConQual approach (Munn et al., 2014).\n\n**When to use** — Use the JBI Qualitative checklist whenever you must appraise individual qualitative studies before\npooling or interpreting them: (1) the appraisal stage of a **JBI meta-aggregative qualitative systematic review** or\nany qualitative evidence synthesis; (2) the qualitative arm of a **mixed-methods systematic review or mixed-methods\nRWE study**; (3) appraising **patient-experience / concept-elicitation interviews** that underpin PRO instrument\ndevelopment or content-validity arguments; and (4) building the qualitative evidence base in an **HTA dossier** where\npatient and caregiver perspectives inform value, acceptability, or unmet-need claims. Decision rule for picking the\nright tool: appraise *primary qualitative studies* with **JBI Qualitative**; report a qualitative *study you are\nwriting* with **COREQ** (interviews/focus groups) or **SRQR**; report the *synthesis* with **ENTREQ / PRISMA**; and\ngrade confidence in the *pooled qualitative findings* with **GRADE-CERQual** or **JBI ConQual**. JBI Qualitative\nappraises inputs; it does not report your own study and does not grade your synthesis output.\n\n**What it requires** — The instrument enforces ten methodological-congruity judgements (each scored yes / no /\nunclear / not applicable, never summed into a single quality number): (1) congruity between the stated philosophical\nperspective and the research methodology; (2) congruity between the methodology and the research question or\nobjectives; (3) congruity between the methodology and the data-collection methods; (4) congruity between the\nmethodology and the representation and analysis of data; (5) congruity between the methodology and the\ninterpretation of results; (6) the researcher located culturally or theoretically (positionality/reflexivity); (7)\nthe influence of the researcher on the research, and vice versa, addressed; (8) participants and their voices\nadequately represented; (9) ethical approval by an appropriate body documented; and (10) conclusions that flow from\nthe analysis or interpretation of the data. In real-world evidence work the substantive analogues are *not* time-zero\nalignment or estimands but the qualitative-rigor parallels: an audit trail from raw transcript to theme (analytic\ntransparency), member checking or independent double-coding (the qualitative cousin of adjudication), saturation\njustification, and explicit reflexivity so the analyst's framing is visible rather than smuggled in.\n\n**When NOT to use — limitations and common misapplications** — (a) It is a **trustworthiness-appraisal tool, not a\nreporting checklist**: do not hand it to authors as a writing template, and do not substitute it for COREQ/SRQR when\nthe task is reporting your own qualitative study. (b) Do not **sum the ten items into a quality score or a cut-off**\nto include/exclude studies — JBI explicitly discourages numeric thresholds; meta-aggregation typically includes\nstudies regardless and uses appraisal to contextualize, not to gate. (c) Do not apply it to **quantitative or\nmixed-methods designs**: appraise quantitative arms with the matching JBI/quantitative-RWE tools and reserve this\nchecklist for the qualitative component only. (d) **Completing the checklist does not make findings generalizable or\ncausal** — it speaks to dependability and credibility, not to transferability claims, which require their own\nargument. (e) **Checklist-as-theater**: marking items \"yes\" without quoting the evidence in the study (the reflexivity\nstatement, the ethics approval number, the saturation rationale) defeats the purpose; appraisal must cite the text.\n(f) Confusing appraisal of *inputs* with confidence in *outputs* — that is ConQual/CERQual's job, downstream of this\nchecklist.\n\n**How it maps to this catalog** — The instrument operationalizes the appraisal step for the catalog's qualitative\ndesign concepts: **qualitative-interview** and **qualitative-ethnographic** are the primary designs it is built to\nappraise, and **qualitative-synthesis** is the meta-aggregative review where appraised studies are pooled (with\nConQual grading the pooled findings). It governs the qualitative component of **mixed-methods** RWE designs and is the\nrigor backbone for **pro-development** and **pro-rwe** content-validity / concept-elicitation interviews and for\n**preference-study** qualitative phases. Items 6–7 (reflexivity, researcher influence) and item 4 (analysis\nrepresentation) connect to **endpoint-adjudication-chart-review-rwe** as the qualitative analogue of independent,\ndocumented double-coding; item 8 (voice representation) and contextual interpretation connect to\n**sdoh-social-determinants-of-health** where lived-experience context shapes the evidence. Applied note for\nclaims/EHR/registry RWE: structured administrative data themselves are out of scope for this checklist, but qualitative\nsubstudies *nested in* RWE programs are squarely in scope — patient/clinician interviews that justify a phenotype's\nface validity, concept-elicitation work behind a PRO endpoint used in a registry, or caregiver-burden interviews in an\nHTA submission should each be appraised with JBI Qualitative before their findings are allowed to influence a\nregulatory or payer narrative (e.g., FDA Patient-Focused Drug Development / Voice-of-the-Patient evidence and EMA\npatient-experience inputs).", "primary_category": "Guideline", "tags": [ "guideline", "critical-appraisal", "qualitative", "risk-of-bias", "evidence-synthesis", "jbi", "meta-aggregation", "patient-experience" ], "aliases": [ "JBI Qualitative", "JBI Critical Appraisal Checklist for Qualitative Research", "JBI QARI checklist", "Joanna Briggs Institute qualitative appraisal tool" ], "applies_to_study_types": [ "qualitative_interview", "qualitative_ethnographic", "qualitative_synthesis" ], "data_sources": [ "primary" ], "citations": [ { "role": "introduce", "url": "https://jbi-global-wiki.refined.site/space/MANUAL/355862497", "citation_text": "JBI Manual for Evidence Synthesis, Chapter 2: Systematic reviews of qualitative evidence — JBI Critical Appraisal Checklist for Qualitative Research. Joanna Briggs Institute. https://jbi.global/critical-appraisal-tools", "year": 2024, "authors_short": "Joanna Briggs Institute", "notes": "Maintained source of record for the 10-item checklist and its scoring conventions; the checklist itself has no standalone journal DOI and is curated within the JBI Manual and JBI critical-appraisal tools library." }, { "role": "explain", "doi": "10.1097/XEB.0000000000000062", "url": "https://doi.org/10.1097/XEB.0000000000000062", "citation_text": "Lockwood C, Munn Z, Porritt K. Qualitative research synthesis: methodological guidance for systematic reviewers utilizing meta-aggregation. International Journal of Evidence-Based Healthcare. 2015;13(3):179-187.", "year": 2015, "authors_short": "Lockwood et al.", "notes": "Canonical methodological articulation of the JBI meta-aggregative approach that situates the qualitative critical-appraisal checklist within review conduct and explains congruity-based appraisal." }, { "role": "explain", "doi": "10.1186/1471-2288-14-108", "url": "https://doi.org/10.1186/1471-2288-14-108", "citation_text": "Munn Z, Porritt K, Lockwood C, Aromataris E, Pearson A. Establishing confidence in the output of qualitative research synthesis: the ConQual approach. BMC Medical Research Methodology. 2014;14:108.", "year": 2014, "authors_short": "Munn et al.", "notes": "Defines ConQual for grading confidence in pooled qualitative findings — the downstream step that appraisal of individual studies feeds, clarifying what JBI Qualitative does and does not do." }, { "role": "use", "url": "https://jbi.global/critical-appraisal-tools", "citation_text": "JBI Critical Appraisal Tools — qualitative checklist (current maintained version and user guide), Joanna Briggs Institute.", "year": 2024, "authors_short": "Joanna Briggs Institute", "notes": "Downloadable current checklist and accompanying guidance used in practice." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "qualitative-interview", "notes": "Primary design the checklist is built to appraise (interviews and focus groups)." }, { "relation_type": "applies_to", "target_slug": "qualitative-ethnographic", "notes": "Ethnographic and observational qualitative studies appraised on the same 10 congruity items." }, { "relation_type": "used_with", "target_slug": "qualitative-synthesis", "notes": "Appraisal step preceding JBI meta-aggregation; pooled findings are then graded with ConQual." }, { "relation_type": "see_also", "target_slug": "mixed-methods", "notes": "Use to appraise the qualitative component of mixed-methods RWE; appraise quantitative arms with the matching tools." }, { "relation_type": "see_also", "target_slug": "pro-development", "notes": "Rigor backbone for concept-elicitation and content-validity interviews underpinning PRO instruments." }, { "relation_type": "see_also", "target_slug": "pro-rwe", "notes": "Appraise patient-experience qualitative evidence informing PRO use in real-world studies." }, { "relation_type": "see_also", "target_slug": "preference-study", "notes": "Appraise the qualitative attribute-elicitation phase that precedes quantitative preference elicitation." }, { "relation_type": "see_also", "target_slug": "endpoint-adjudication-chart-review-rwe", "notes": "Items on analysis representation and researcher influence parallel documented, independent double-coding of qualitative data." }, { "relation_type": "see_also", "target_slug": "sdoh-social-determinants-of-health", "notes": "Voice-representation and contextual-interpretation items connect to lived-experience context in RWE." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "hta", "journal", "fda", "ema" ] }, { "slug": "jbi-sr", "name": "JBI Critical Appraisal Checklist for Systematic Reviews and Research Syntheses", "short_definition": "A JBI critical-appraisal (risk-of-bias) instrument used to assess the methodological rigor of a published systematic review or research synthesis before its results are trusted or incorporated into an umbrella review, evidence synthesis, or HTA dossier.", "long_description": "**What it is** — The **JBI Critical Appraisal Checklist for Systematic Reviews and Research Syntheses**\nis an 11-item critical-appraisal (risk-of-bias) tool maintained by **JBI** (formerly the Joanna Briggs\nInstitute, Adelaide) as part of the JBI suite of design-specific appraisal instruments and the JBI Manual\nfor Evidence Synthesis. It is the instrument a reviewer applies to *an already-published systematic\nreview* to judge whether that review was conducted with enough methodological rigor that its conclusions\ncan be believed and reused. It is the appraisal layer of JBI's umbrella-review / \"review of reviews\"\nmethodology: when you synthesize the findings of multiple systematic reviews, this checklist is how you\nformally rate the trustworthiness of each included review. It sits alongside sibling JBI checklists for\nRCTs, cohort, case-control, prevalence, qualitative, and economic-evaluation studies — each tuned to the\nbiases of its design.\n\n**When to use** — Use the JBI SR checklist when the *unit of appraisal is a systematic review itself*,\nnot a primary study. The canonical decision contexts are: (1) conducting an **umbrella review / overview\nof reviews**, where each candidate review must be appraised before its effect estimates are pooled or\nnarratively synthesized; (2) building the evidence-synthesis backbone of an **HTA or payer dossier**,\nwhere reviewers must defend why some published reviews were weighted and others discounted; (3)\n**peer-reviewed evidence synthesis** that cites prior reviews as inputs rather than re-extracting all\nprimary data; and (4) any **comparative-effectiveness landscape** where multiple overlapping reviews\nexist and the analyst must rank their credibility. Decision rule for which tool: appraise a *systematic\nreview* with JBI SR (or its sibling **AMSTAR 2**); appraise the *primary studies inside* a review with\nthe design-matched JBI checklist (RCT, cohort, case-control) or **ROBINS-I**; *report* your own review\nwith **PRISMA 2020** and *register the protocol* with **PRISMA-P** — those are reporting guidelines, not\nappraisal tools, and are not interchangeable with JBI SR.\n\n**What it requires** — The checklist enforces 11 methodological domains that map to the credible-synthesis\nquestions a regulator or HTA reviewer will ask: (1) an explicit and appropriate **review question**; (2)\n**inclusion criteria** appropriate to that question; (3) a **search strategy** adequate to find the\nrelevant evidence; (4) **adequate sources and resources** searched (databases, grey literature, languages);\n(5) **appropriate appraisal criteria** applied to the included studies; (6) **critical appraisal by two or\nmore reviewers independently**; (7) **methods to minimize error in data extraction**; (8) **appropriate\nmethods to combine studies** (meta-analysis assumptions, heterogeneity, model choice); (9) **assessment of\npublication bias**; (10) **policy/practice recommendations supported by the reported data**; and (11)\n**appropriate, specific directives for new research**. For real-world-data evidence synthesis these\ndomains acquire teeth: \"appropriate appraisal criteria\" must mean the included observational studies were\njudged on **design transparency**, **data fitness-for-use**, **phenotype/algorithm validation**,\n**time-zero alignment**, **estimand specification with intercurrent-event handling**, **confounding\ncontrol**, **attrition/missing-data accounting**, and **quantitative sensitivity/bias analysis** — not a\ngeneric \"low/high quality\" stamp. \"Appropriate methods to combine studies\" must confront the fact that\npooling effect estimates across non-randomized database studies with different time-zero, confounding, and\nexposure definitions can manufacture spurious precision.\n\n**When NOT to use — limitations and common misapplications** — (1) **It is a risk-of-bias instrument, not\na reporting checklist and not a numeric quality score.** Tallying \"yes\" answers into a sum and thresholding\nit (\"8/11 = high quality\") is a documented misuse — JBI explicitly discourages converting the items into a\ncutoff score; each domain must be judged and reported individually. (2) **It appraises the review, not the\nevidence base.** A methodologically flawless systematic review of biased observational studies is still a\nrigorous synthesis of weak evidence; passing JBI SR does not upgrade the certainty of the underlying RWE,\nwhich is the job of GRADE and of design-level appraisal of the primary studies. (3) **Wrong unit of\nappraisal:** using JBI SR to appraise a single cohort or RCT (use the design-matched JBI checklist or\nROBINS-I) or, conversely, using a primary-study tool to appraise a review. (4) **Checklist-as-theater:**\ncompleting the form to satisfy a journal or dossier template without the two-independent-reviewer process,\ndocumented disagreements, or an audit trail defeats the instrument. (5) **Substituting it for reporting\ncompliance:** a review can score well on JBI SR yet still need PRISMA 2020 for transparent reporting — the\ntwo are complementary, not redundant. (6) Completing the checklist **does not make an observational\nsynthesis causal**; estimand and confounding judgments still rest on the primary-study designs.\n\n**How it maps to this catalog** — When you apply the JBI SR appraisal criteria to a synthesis of\nreal-world-data studies, the substantive judgments are implemented by concepts in this repo. Item 5\n(\"appropriate appraisal criteria\") for included database studies is implemented by\n**target-trial-emulation** (the reference frame for judging whether each study aligned eligibility,\ntreatment assignment, and time zero), **active-comparator-new-user** (the design that controls confounding\nby indication and immortal-time bias you should look for), **high-dimensional-propensity-score-hdps-rwe**\n(the confounding-control adequacy you should demand), and **estimands-ate-att-intercurrent-events-rwe**\n(whether each study even specified what it was estimating). Whether the included studies validated their\noutcomes and exposures — a precondition for trusting any pooled estimate — is implemented by\n**diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe**. Item 7 (data-extraction error) and item 8\n(combining studies) for claims/EHR evidence depend on **attrition-and-loss-to-follow-up-rwe** (whether\ncohorts were comparably retained) and **claims-analysis** (whether the data substrate could support the\nquestion at all).\n\n**Applied note (claims/EHR/registry RWE).** In an HTA umbrella review of, say, comparative\ncardiovascular safety drawn from several claims- and EHR-based systematic reviews, JBI SR is the gate at\nthe *review* level, but its credibility hinges on the *primary-study* judgments above. A review that\npooled studies with incompatible time-zero definitions, unvalidated outcome phenotypes, and unaddressed\nMedicare Advantage claims attrition should fail item 5 and item 8 even if its search and extraction were\nimmaculate. Record each domain's judgment with the catalog concept that justified it, so the dossier shows\n*why* a given published review was up- or down-weighted — not merely that a checklist was filled in.", "primary_category": "Guideline", "tags": [ "guideline", "critical-appraisal", "risk-of-bias", "systematic-review", "evidence-synthesis", "umbrella-review", "jbi", "quality-assessment" ], "aliases": [ "JBI SR", "JBI Critical Appraisal Checklist for Systematic Reviews", "JBI checklist for systematic reviews and research syntheses", "Joanna Briggs Institute systematic review appraisal tool" ], "applies_to_study_types": [ "systematic_review", "meta_analysis_rct", "meta_analysis_obs", "umbrella" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "doi": "10.1097/XEB.0000000000000055", "url": "https://doi.org/10.1097/XEB.0000000000000055", "citation_text": "Aromataris E, Fernandez R, Godfrey CM, Holly C, Khalil H, Tungpunkom P. Summarizing systematic reviews: methodological development, conduct and reporting of an umbrella review approach. International Journal of Evidence-Based Healthcare. 2015;13(3):132-140.", "year": 2015, "authors_short": "Aromataris et al.", "notes": "Canonical JBI methodology paper establishing the umbrella-review approach and the role of critical appraisal of included systematic reviews; underpins the JBI SR checklist." }, { "role": "explain", "doi": "10.1097/XEB.0000000000000064", "url": "https://doi.org/10.1097/XEB.0000000000000064", "citation_text": "Moola S, Munn Z, Sears K, et al. Conducting systematic reviews of association (etiology): the Joanna Briggs Institute's approach. International Journal of Evidence-Based Healthcare. 2015;13(3):163-169.", "year": 2015, "authors_short": "Moola et al.", "notes": "Elaborates JBI synthesis methodology for etiology/association reviews — the observational-evidence context most relevant to appraising RWE syntheses with the JBI SR checklist." }, { "role": "use", "url": "https://jbi.global/critical-appraisal-tools", "citation_text": "JBI. Critical Appraisal Tools — Checklist for Systematic Reviews and Research Syntheses. JBI (Joanna Briggs Institute), Adelaide. Maintained instrument and JBI Manual for Evidence Synthesis.", "year": 2024, "authors_short": "JBI", "notes": "Authoritative source for the current 11-item checklist and accompanying guidance in the JBI Manual." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "systematic-review", "notes": "Primary use case — appraising the methodological rigor of a published systematic review." }, { "relation_type": "applies_to", "target_slug": "meta-analysis-rct", "notes": "Appraise systematic reviews that pool randomized trials before reusing their pooled estimates." }, { "relation_type": "applies_to", "target_slug": "meta-analysis-obs", "notes": "Appraise reviews pooling observational/RWE studies, where combine-studies and appraisal-criteria items carry the most weight." }, { "relation_type": "part_of", "target_slug": "umbrella-review", "notes": "The JBI SR checklist is the appraisal layer of the JBI umbrella-review / overview-of-reviews methodology." }, { "relation_type": "see_also", "target_slug": "target-trial-emulation", "notes": "Implements item 5 (appropriate appraisal criteria) for included RWE studies — the reference frame for judging eligibility, treatment-assignment, and time-zero alignment." }, { "relation_type": "see_also", "target_slug": "active-comparator-new-user", "notes": "The design whose presence/absence signals whether included studies controlled confounding by indication and immortal-time bias." }, { "relation_type": "see_also", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "The confounding-control adequacy to demand when appraising included observational studies." }, { "relation_type": "see_also", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Whether each included study specified its estimand and intercurrent-event handling — a precondition for credible pooling (item 8)." }, { "relation_type": "see_also", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Whether included claims/EHR studies validated outcome and exposure phenotypes before their results were trusted or combined." }, { "relation_type": "see_also", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Supports item 7/8 judgments on comparable cohort retention across the studies being synthesized." }, { "relation_type": "used_with", "target_slug": "claims-analysis", "notes": "Apply when the systematic review under appraisal synthesizes claims/EHR-based evidence." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "hta", "journal", "ema" ] }, { "slug": "mace-endpoint-specification-checklist-rwe", "name": "MACE Endpoint Specification Checklist", "short_definition": "A checklist for specifying major adverse cardiovascular event composites in RWE, including component list, death source, event-date rules, validation, and component reporting.", "long_description": "**What it is** - This guideline is the checklist layer for major adverse cardiovascular event\ncomposites in RWE. The companion MACE concept explains the endpoint family; this guideline states\nwhat must be specified when a study uses MACE, MACCE, or an expanded cardiovascular composite.\nIt treats \"MACE\" as an under-specified shorthand until the component list, event-date hierarchy,\ndeath source, validation evidence, and first-event/recurrent-event rules are written down.\n\n**When to use** - Use it for cardiovascular safety or effectiveness analyses in claims, EHR,\nregistry, linked mortality, pragmatic trial, or post-marketing safety studies whenever the\nendpoint is a composite of cardiovascular death, myocardial infarction, stroke, hospitalization\nfor unstable angina, heart failure, revascularization, or similar events. Use it before code-list\ndevelopment and table shells, because a 3-point MACE, 4-point MACE, 5-point MACE, and MACCE can\nanswer different clinical questions and produce different effect estimates.\n\n**What it requires / checklist domains** - Name the exact variant and list every component in\nthe protocol, SAP, code appendix, table shell, and manuscript. State whether death is\ncardiovascular death, all-cause death, unknown-cause death, a competing event, or a separate\nendpoint. Specify event-date hierarchy, same-day ties, transfer collapse, de-duplication windows,\nfirst-event versus recurrent-event handling, and whether a component closes follow-up for the\ncomposite. Provide ICD/CPT/HCPCS/death-source algorithms with diagnosis position and care-setting\nrestrictions. Report validation metrics or validation citations for each component, not only the\ncomposite. Tabulate component-specific event counts and component-specific estimates.\n\n**When NOT to use - limitations and common misapplications** - Do not use \"MACE\" without a\ncomponent list. Do not assume all components have the same clinical importance, ascertainment\nquality, or direction of treatment effect. Do not let a frequent, less severe, or poorly validated\ncomponent dominate the composite without component-specific reporting. Do not combine all-cause\ndeath with nonfatal CV events without explaining the estimand and competing-risk implications.\nDo not borrow trial endpoint definitions into claims without a computable phenotype and\nvalidation plan. A composite can improve power, but it can also obscure heterogeneity and\nmisclassification.\n\n**How it maps to this catalog** - This guideline cross-references\n`major-adverse-cardiovascular-events-mace-rwe` for the endpoint concept,\n`composite-endpoint-construction-rwe` for first-component and interpretation rules,\n`outcome-algorithm-construction-rwe` and `claims-outcome-algorithm-ppv-sensitivity-rwe` for\nvalidation, `mortality-source-hierarchy-rwe` for death source choice,\n`acute-event-deduplication-window-rwe` and `hospitalization-transfer-collapse-rwe` for event\nconstruction, and `competing-risks-cause-specific-fine-gray-rwe` when death and nonfatal events\nneed separate causal interpretation.", "primary_category": "Guideline", "tags": [ "guideline", "checklist", "mace", "cardiovascular-outcomes", "composite-endpoint" ], "aliases": [ "MACE checklist", "cardiovascular composite endpoint checklist", "MACCE checklist" ], "applies_to_study_types": [ "claims_analysis", "registry_study", "comparative_effectiveness", "safety_surveillance" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "doi": "10.1161/CIRCULATIONAHA.117.033502", "url": "https://doi.org/10.1161/CIRCULATIONAHA.117.033502", "citation_text": "Hicks KA, Mahaffey KW, Mehran R, et al. 2017 Cardiovascular and Stroke Endpoint Definitions for Clinical Trials. Circulation. 2018;137(9):961-972.", "year": 2018, "authors_short": "Hicks et al.", "notes": "Standardized cardiovascular and stroke endpoint definitions developed through the Standardized Data Collection for Cardiovascular Trials Initiative with FDA involvement." }, { "role": "explain", "doi": "10.1186/s12874-021-01440-5", "url": "https://doi.org/10.1186/s12874-021-01440-5", "citation_text": "Bosco E, Hsueh L, McConeghy KW, Gravenstein S, Saade E. Major adverse cardiovascular event definitions used in observational analysis of administrative databases: a systematic review. BMC Medical Research Methodology. 2021;21:241.", "year": 2021, "authors_short": "Bosco et al.", "notes": "MACE definitions in observational administrative-data studies." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "major-adverse-cardiovascular-events-mace-rwe", "notes": "Checklist for MACE endpoint construction." } ], "index_definitions": [], "checklist_items": [ "Name the MACE variant and list every component in the protocol, SAP, table shell, and manuscript.", "State whether death is cardiovascular death, all-cause death, unknown-cause death, a competing event, or a separate endpoint.", "Specify event-date hierarchy, first-event versus recurrent-event handling, same-day ties, transfers, and de-duplication windows.", "Provide ICD/CPT/HCPCS/death-source algorithms with diagnosis position and care-setting restrictions.", "Report validation metrics or validation citations for each component, not only the composite.", "Tabulate component-specific event counts and component-specific estimates alongside the composite.", "Run sensitivity analyses that remove low-specificity or high-frequency components if they dominate the composite." ], "regulatory_relevance": [ "fda", "ema", "hta" ] }, { "slug": "moose", "name": "MOOSE", "short_definition": "Meta-analysis Of Observational Studies in Epidemiology (MOOSE) — a reporting checklist for meta-analyses and systematic reviews of observational (non-randomized) studies, spanning background, search strategy, methods, results, discussion, and conclusions.", "long_description": "**What it is.** MOOSE (Meta-analysis Of Observational Studies in Epidemiology) is a reporting\nguideline for **meta-analyses and systematic reviews of observational studies** — cohort,\ncase-control, and cross-sectional evidence rather than randomized trials. It originated as a\nconsensus proposal from a 27-member expert workshop convened by the editors of *JAMA* and\npublished by Stroup and colleagues in 2000, and a structured reporting-checklist version was\nreaffirmed and re-issued by Brooke and colleagues in 2021. MOOSE is catalogued and maintained\nby the EQUATOR Network as the observational-evidence counterpart to PRISMA. Its 35 recommended\nitems are organized into six reporting domains: reporting of background, reporting of search\nstrategy, reporting of methods, reporting of results, reporting of discussion, and reporting\nof conclusions. MOOSE is a *reporting* instrument: it standardizes what a manuscript must\ndisclose so a reader can judge how the synthesis was conducted — it does not score study\nquality and does not adjudicate risk of bias.\n\n**When to use.** Apply MOOSE whenever the deliverable is a **synthesis of observational\nstudies** intended for a peer-reviewed journal, an HTA/payer evidence dossier, or a regulatory\n(FDA/EMA) submission that leans on pooled real-world or epidemiologic evidence. Typical\ntriggers: a pooled estimate of a drug-outcome association across pharmacoepidemiologic cohorts;\na meta-analysis of incidence, prevalence, or natural-history estimates from registries and\nclaims; a comparative-effectiveness synthesis where no head-to-head trial exists. Decision\nrule for which guideline applies: if the included studies are **observational**, MOOSE is the\ndesign-specific reporting standard, used **alongside** PRISMA 2020 (the general systematic-review\nscaffold for flow diagrams, abstract, and search reporting) — they are complementary, not\nsubstitutes. If the synthesis pools **randomized trials**, use PRISMA 2020 with a trials focus,\nnot MOOSE. If you are reporting a single **primary** observational study (one cohort, one\ncase-control analysis, one claims database study), MOOSE does not apply — use STROBE, or\nRECORD/RECORD-PE for routinely collected health data. For the protocol of an observational\nreview, register and report with PRISMA-P; MOOSE governs the completed synthesis.\n\n**What it requires.** MOOSE's six domains force explicit disclosure of the substantive choices\nthat make an observational synthesis interpretable. **Background:** the problem definition,\nhypothesis, study outcome(s), exposure/intervention, study designs eligible, and the target\npopulation. **Search strategy:** qualifications of searchers, databases and registries used,\nsearch terms and date limits, use of hand-searching and contact with authors, inclusion of\nnon-English and unpublished/grey literature, and the handling of publication bias — the\nreporting analogue of data-fitness-for-use, since the \"data\" of a meta-analysis are the\nretrievable primary studies. **Methods:** the rules for judging study conformance to the\nquestion, rationale for selection, documentation of how studies were assessed (and how\nheterogeneity in exposure/outcome **definitions and phenotypes** across primary studies was\nhandled), the pooling method, statistical model (fixed vs random effects), tests for\nheterogeneity, sensitivity and subgroup analyses, and assessment of confounding within and\nacross the included studies. **Results:** a study-flow diagram (accounting for attrition from\nrecords screened to studies pooled), tabulation of descriptive and effect estimates with\nappropriate measures of variability, and graphical summaries. **Discussion and conclusions:**\nquantitative assessment of bias (including publication bias), justification of exclusions, the\nvalidity and generalizability of the pooled estimate, and guidance for future research. For\nreal-world-data syntheses, MOOSE compliance specifically means documenting how disparate\nexposure windows, outcome algorithms, time-zero conventions, and confounding-control strategies\nacross the source studies were reconciled before pooling.\n\n**When NOT to use — limitations and common misapplications.** MOOSE is a **reporting checklist,\nnot a risk-of-bias instrument and not a quality score**: completing all 35 items certifies that\na manuscript *disclosed* its methods, not that those methods were sound. Pair it with a genuine\nrisk-of-bias tool — ROBINS-I for the included studies, or the Newcastle-Ottawa Scale — and do\nnot report a MOOSE \"score\" as if it graded validity. Specific failure modes: (a) using MOOSE to\nreport a **single primary observational study** when STROBE or RECORD-PE is the correct standard;\n(b) treating MOOSE completion as evidence that the pooled estimate is **causal** — observational\nsynthesis inherits the confounding of its inputs, and a transparently reported meta-analysis of\nbiased studies is still biased (\"garbage in, garbage out\"); (c) **checklist-as-theater** — filing\na completed checklist in an appendix while the manuscript body omits the heterogeneity, publication-bias,\nand sensitivity analyses the items demand; (d) applying MOOSE to a **meta-analysis of RCTs**,\nwhere PRISMA 2020 governs; (e) skipping MOOSE because the team already has PRISMA 2020 — PRISMA\nis the general scaffold and MOOSE adds the observational-specific reporting expectations, so both\nare completed.\n\n**How it maps to this catalog.** MOOSE's reporting domains map onto concrete methods concepts in\nthis repository that implement what each item demands. Search-strategy and study-eligibility\nreporting are operationalized by the synthesis concepts **meta-analysis-obs**, **systematic-review**,\nand **network-meta-analysis** (for indirect/mixed comparisons). The MOOSE requirement to reconcile\noutcome and exposure definitions across primary studies is implemented by\n**diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe** (phenotype/algorithm transparency and\nvalidation) and, for the underlying data, **claims-analysis** and **fit-for-purpose-data-assessment-rwe**.\nMOOSE's confounding-assessment items are implemented by **active-comparator-new-user** and\n**high-dimensional-propensity-score-hdps-rwe** (how the included studies controlled confounding,\nwhich a synthesis must summarize and contrast). The estimand and intercurrent-event reporting that\na defensible pooled effect requires is implemented by **estimands-ate-att-intercurrent-events-rwe**,\nand where the synthesis emulates a trial protocol, **target-trial-emulation**. MOOSE's study-flow\nand attrition reporting is implemented by **attrition-and-loss-to-follow-up-rwe**, and the PICOTS\nframing that anchors background and eligibility items is implemented by **picots-framework-rwe**.\nApplied note for claims/EHR/registry RWE: when meta-analyzing pharmacoepidemiologic studies built\non routinely collected data, the highest-yield MOOSE work is documenting cross-study heterogeneity\nin code lists, lookback/washout windows, time-zero conventions, and PS strategies — differences\nthat frequently explain between-study heterogeneity more than clinical effect modification, and\nthat a transparent MOOSE-compliant report must surface before any pooled estimate is trusted.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "meta-analysis", "observational", "evidence-synthesis", "rwe" ], "aliases": [ "MOOSE", "Meta-analysis Of Observational Studies in Epidemiology", "MOOSE checklist" ], "applies_to_study_types": [ "meta_analysis_obs", "systematic_review" ], "data_sources": [ "claims", "ehr", "registry", "linked", "multi-database" ], "citations": [ { "role": "introduce", "doi": "10.1001/jama.283.15.2008", "url": "https://doi.org/10.1001/jama.283.15.2008", "citation_text": "Stroup DF, Berlin JA, Morton SC, et al. Meta-analysis of observational studies in epidemiology: a proposal for reporting. Meta-analysis Of Observational Studies in Epidemiology (MOOSE) group. JAMA. 2000;283(15):2008-2012.", "year": 2000, "authors_short": "Stroup et al.", "notes": "Original consensus statement from the JAMA-convened workshop; defines the six reporting domains and the recommended checklist of items for meta-analyses of observational studies." }, { "role": "explain", "doi": "10.1001/jamasurg.2021.0522", "url": "https://doi.org/10.1001/jamasurg.2021.0522", "citation_text": "Brooke BS, Schwartz TA, Pawlik TM. MOOSE reporting guidelines for meta-analyses of observational studies. JAMA Surgery. 2021;156(8):787-788.", "year": 2021, "authors_short": "Brooke et al.", "notes": "Reaffirms MOOSE and presents the structured 35-item reporting checklist for contemporary use; clarifies that MOOSE is a reporting standard distinct from risk-of-bias appraisal." }, { "role": "use", "url": "https://www.equator-network.org/reporting-guidelines/meta-analysis-of-observational-studies-in-epidemiology-a-proposal-for-reporting-meta-analysis-of-observational-studies-in-epidemiology-moose-group/", "citation_text": "Meta-analysis Of Observational Studies in Epidemiology (MOOSE), EQUATOR Network — maintained guideline record and checklist resources.", "year": 2000, "authors_short": "EQUATOR Network", "notes": "Stable maintained landing page with the checklist and links for journal/regulatory appendices." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "meta-analysis-obs", "notes": "MOOSE is the design-specific reporting standard for meta-analyses of observational studies; this concept implements the pooling and heterogeneity methods it requires reported." }, { "relation_type": "applies_to", "target_slug": "systematic-review", "notes": "Governs the reporting of systematic reviews of observational evidence, used alongside PRISMA 2020 as the general scaffold." }, { "relation_type": "see_also", "target_slug": "network-meta-analysis", "notes": "When the observational synthesis uses indirect or mixed comparisons, MOOSE reporting expectations extend to the network structure and transitivity assumptions." }, { "relation_type": "used_with", "target_slug": "picots-framework-rwe", "notes": "PICOTS anchors MOOSE's background and eligibility items (population, intervention/exposure, comparator, outcomes, timing, setting/study designs)." }, { "relation_type": "used_with", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Implements the MOOSE requirement to document and reconcile outcome/exposure phenotype definitions across the primary studies before pooling." }, { "relation_type": "used_with", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "Supports MOOSE confounding-assessment items by characterizing how included studies controlled confounding in routinely collected data." }, { "relation_type": "used_with", "target_slug": "active-comparator-new-user", "notes": "A common design among included pharmacoepidemiologic studies; MOOSE requires its confounding-control and time-zero choices be summarized and contrasted across studies." }, { "relation_type": "used_with", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Implements the estimand and intercurrent-event clarity MOOSE expects so that a pooled effect has a coherent target of estimation." }, { "relation_type": "used_with", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Implements MOOSE's study-flow and attrition reporting from records screened through studies pooled." }, { "relation_type": "see_also", "target_slug": "target-trial-emulation", "notes": "When the synthesis emulates a hypothetical trial protocol, target-trial discipline sharpens the MOOSE methods and eligibility reporting." }, { "relation_type": "see_also", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "The synthesis analogue of data fitness — judging whether the retrievable primary studies and their underlying data are adequate for the question." }, { "relation_type": "see_also", "target_slug": "claims-analysis", "notes": "Apply when the pooled primary studies are claims- or EHR-based; cross-study differences in code lists and windows often drive heterogeneity MOOSE requires surfaced." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "hta", "journal" ] }, { "slug": "nice-reference-case", "name": "NICE Reference Case", "short_definition": "The standardized set of methodological requirements (NHS/PSS perspective, lifetime horizon, QALYs via EQ-5D, 3.5% discounting, probabilistic analysis) that NICE requires for the base-case cost-utility analysis in technology appraisals and highly specialised technologies evaluations, so that ICERs are comparable across decisions.", "long_description": "**What it is** — The **NICE Reference Case** is the fixed methodological specification that the National\nInstitute for Health and Care Excellence (England) requires for the **base-case economic evaluation** in its\ntechnology appraisals and highly specialised technologies (HST) evaluations. It is maintained by NICE and lives\nas Chapter 4 (\"Economic evaluation\") of *NICE health technology evaluations: the manual* (process guide **PMG36**,\npublished 31 January 2022, superseding the 2013 *Guide to the methods of technology appraisal*, PMG9). The reference\ncase is not a free-standing checklist or a reporting tool; it is a **normative methods standard** that pins down the\nanalytic choices most likely to drive a cost-effectiveness result so that the incremental cost-effectiveness ratio\n(ICER) submitted for one technology is comparable with the ICER for every other. Its core elements: the **perspective\non outcomes** is all direct health effects for patients and, where relevant, carers; the **perspective on costs** is\nthe **NHS and Personal Social Services (PSS)** (productivity and broader societal costs are excluded from the base\ncase); the **time horizon** is long enough to capture all material differences in cost and outcome (usually lifetime\nwhen survival is affected); health gain is measured in **quality-adjusted life years (QALYs)** with the **EQ-5D** as\nthe preferred utility instrument valued by a UK general-population tariff; **costs and effects are discounted at the\nsame annual rate, currently 3.5%**; comparators are those defined in the NICE **scope**; and uncertainty must be\ncharacterised through **probabilistic sensitivity analysis (PSA)**. NICE permits departures from the reference case,\nbut only when each departure is justified and the reference-case base case is reported alongside it.\n\n**When to use** — Use the reference case whenever you are building or appraising a **cost-utility / cost-effectiveness\nmodel intended for a NICE HTA decision** in England (technology appraisal or HST). It governs the **base case**: the\nmanufacturer/company submission, the Evidence Review Group / External Assessment Group critique, and the appraisal\ncommittee all read the model against it. Decision rules for which framework applies: (1) **PMG36 (2022)** is the\ncurrent manual and applies to evaluations started under the integrated process; **PMG9 (2013)** governed earlier\nappraisals and is the correct reference for historical or in-flight submissions begun under it — match the manual to\nthe appraisal's timeline, do not retrofit. (2) For **reporting** an economic evaluation in a journal or as a\nsubmission appendix, the reference case is *not* the reporting instrument — pair it with **CHEERS 2022** (the\nreporting checklist for health economic evaluations). (3) The reference case is **NICE-specific**: for other agencies\nuse their own reference case (e.g., ICER value framework in the US, CADTH in Canada, PBAC in Australia, the EU Joint\nClinical Assessment for relative effectiveness) — methods such as the discount rate and perspective differ by\njurisdiction and are not transferable.\n\n**What it requires** — The reference case enforces a small set of high-leverage methodological commitments, and for\nreal-world evidence (RWE) feeding the model each one creates concrete obligations: (a) **NHS/PSS cost perspective** —\nunit costs and resource use must be costed from an NHS payer view; RWE cost inputs drawn from non-UK or societal\ndatasets must be re-based to NHS/PSS, and the cost concept (all-cause vs disease-attributable, per-patient-per-month\nvs per-episode) must be stated explicitly. (b) **Lifetime horizon** — short-horizon RWE almost always requires\n**extrapolation** of survival and treatment effect beyond observed follow-up, with the extrapolation method and its\nuncertainty pre-specified. (c) **QALYs via EQ-5D** — utilities must map to a UK EQ-5D tariff; mapping algorithms from\ndisease-specific instruments must be justified. (d) **Comparators from the scope** — the effectiveness contrast must\nbe against the relevant comparator(s), which is exactly where comparative RWE design discipline matters: a defensible\neffectiveness estimate demands an **active-comparator, new-user** structure, **time-zero alignment**, **confounding\ncontrol**, and a clearly stated **estimand** with handling of intercurrent events. (e) **Single-arm / external-control\nevidence** (common in HST and oncology) must address comparability of the external control to the trial arm and the\nattendant selection and confounding biases. (f) **Discounting at 3.5%** for both costs and effects. (g) **PSA** to\npropagate parameter uncertainty into the ICER, plus scenario/deterministic analysis for structural and judgemental\nassumptions; bias in RWE inputs (measurement error in phenotypes, residual confounding) should be probed with\n**quantitative bias analysis** and **negative controls** rather than assumed away.\n\n**When NOT to use — limitations and common misapplications** — (1) **It is a methods standard, not a reporting\nchecklist.** Completing the reference case does not satisfy reporting expectations; CHEERS 2022 does that, and a clean\nreference-case base case can still be opaquely reported. (2) **A reference-case-compliant model is not automatically\ncredible** — compliance fixes *comparability of methods*, not the *validity of the inputs*; a model with perfect\nperspective, horizon, and discounting but a confounded RWE effectiveness estimate produces a precise, comparable,\nwrong ICER. (3) **Presenting departures without the base case.** Analysts sometimes submit only their preferred\n(non-reference-case) analysis; NICE requires the reference-case base case *and* the justified departure side by side.\n(4) **Wrong manual for the timeline** — applying PMG36 to an appraisal conducted under PMG9 (or vice versa) misstates\nthe applicable rules (e.g., the 2022 severity modifier replacing the earlier end-of-life modifier). (5) **Importing\nforeign-perspective or societal costs unadjusted** — US claims-derived or societal-perspective costs cannot be\ndropped into an NHS/PSS base case without re-basing. (6) **Treating PSA as optional** or as a cosmetic add-on rather\nthan the required characterisation of decision uncertainty. (7) **Conflating the reference case with the Budget Impact\nTest** — the BIT (the £20m-per-year commercial-arrangement trigger) is a *separate* NICE process about affordability,\nnot the cost-utility reference case; budget-impact analysis is governed by its own conventions, not by the reference\ncase.\n\n**How it maps to this catalog** — The reference case sets the *requirements*; these catalog concepts *implement* the\nRWE inputs that feed a reference-case model: comparative effectiveness from observational data is built with\n**active-comparator-new-user** and formalised as a **target-trial-emulation**, balanced with\n**propensity-score-methods-psm-iptw** and, when confounders are unmeasured, **high-dimensional-propensity-score-hdps-rwe**;\nthe causal target is specified with **estimands-ate-att-intercurrent-events-rwe**. Where the evidence is single-arm,\nsee **single-arm-external-control**, **rare-disease-external-controls-rwe**, and\n**external-adjustment-validation-substudy-bias-correction-rwe**. Endpoints and exposures from claims/EHR are defined\nwith **diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe** and validated via\n**claims-outcome-algorithm-ppv-sensitivity-rwe**; follow-up integrity is governed by\n**attrition-and-loss-to-follow-up-rwe**. Cost inputs for the NHS/PSS perspective draw on\n**healthcare-costs-pppm-pppy-pmpm**, **all-cause-vs-attributable-costs-rwe**, and **cost-outlier-handling-rwe**;\nthe 3.5% discounting requirement is implemented in **discounting-costs-effects-rwe**; whether the RWE population maps\nto the NHS decision population is assessed with **generalizability-transportability-external-validity-rwe**. Residual\nbias in RWE inputs is stress-tested with **quantitative-bias-analysis-toolkit-rwe**,\n**negative-control-outcomes-rwe**, and **empirical-calibration-negative-controls-rwe**. *Applied note (claims/EHR/registry RWE):*\nwhen a UK or non-UK claims/EHR source supplies the comparative effectiveness or cost input to a NICE submission, the\ncatalog requirement is to (i) state the estimand and emulate the relevant trial, (ii) re-base costs to NHS/PSS and\ndeclare the cost concept, (iii) extrapolate to a lifetime horizon with pre-specified survival models, and (iv) carry\nthe resulting parameter and structural uncertainty into the PSA — and to report all of this against CHEERS 2022, not\nmerely assert reference-case compliance.", "primary_category": "Guideline", "tags": [ "guideline", "hta", "reference-case", "cost-utility", "cost-effectiveness", "nice", "nhs-pss-perspective", "qaly" ], "aliases": [ "NICE Reference Case", "PMG36 reference case", "NICE health technology evaluations manual reference case", "NICE technology appraisal reference case", "NICE methods guide reference case", "NICE HST reference case" ], "applies_to_study_types": [ "cost_effectiveness", "cost_utility" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "url": "https://www.nice.org.uk/process/pmg36/chapter/economic-evaluation-2", "citation_text": "National Institute for Health and Care Excellence (NICE). NICE health technology evaluations: the manual (PMG36), Chapter 4: Economic evaluation — the reference case. Published 31 January 2022.", "year": 2022, "authors_short": "NICE", "notes": "Canonical agency statement of the current reference case (PMG36, 2022); no journal statement paper exists, so the stable NICE manual URL is the authoritative source. Specifies NHS/PSS perspective, QALYs/EQ-5D, 3.5% discount rate, lifetime horizon, scope comparators, and required PSA." }, { "role": "explain", "doi": "10.1016/j.jval.2023.05.015", "url": "https://doi.org/10.1016/j.jval.2023.05.015", "citation_text": "Angelis A, Lange A, Kanavos P, et al. The evolving nature of health technology assessment: a critical appraisal of NICE's new methods manual. Value in Health. 2023;26(10):1503-1509.", "year": 2023, "authors_short": "Angelis et al.", "notes": "Peer-reviewed critical appraisal of the 2022 NICE methods manual (PMG36), including the reference case and the severity modifier that replaced the end-of-life modifier — the strongest NICE-specific scholarly companion to the agency document." }, { "role": "use", "url": "https://www.nice.org.uk/process/pmg9/chapter/the-reference-case", "citation_text": "National Institute for Health and Care Excellence (NICE). Guide to the methods of technology appraisal 2013 (PMG9), Chapter 5: The reference case. Published 04 April 2013.", "year": 2013, "authors_short": "NICE", "notes": "Historical reference case (PMG9, 2013); the applicable framework for appraisals conducted before PMG36. Match the manual to the appraisal's timeline." }, { "role": "use", "doi": "10.1016/j.jval.2021.11.1351", "url": "https://doi.org/10.1016/j.jval.2021.11.1351", "citation_text": "Husereau D, Drummond M, Augustovski F, et al. Consolidated Health Economic Evaluation Reporting Standards 2022 (CHEERS 2022) statement: updated reporting guidance for health economic evaluations. Value in Health. 2022;25(1):3-9.", "year": 2022, "authors_short": "Husereau et al.", "notes": "Reporting companion. The reference case fixes the methods; CHEERS 2022 governs how the economic evaluation is reported. Use the two together, not interchangeably." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "cost-utility", "notes": "The reference case governs the base-case cost-utility (QALY-based) analysis in NICE appraisals." }, { "relation_type": "applies_to", "target_slug": "cost-effectiveness", "notes": "Applies to the cost-effectiveness analysis submitted for a NICE technology appraisal or HST evaluation." }, { "relation_type": "used_with", "target_slug": "discounting-costs-effects-rwe", "notes": "Implements the reference-case requirement to discount both costs and health effects at 3.5% annually." }, { "relation_type": "used_with", "target_slug": "target-trial-emulation", "notes": "Comparative effectiveness inputs from observational data should emulate the trial implied by the scope comparator before entering a reference-case model." }, { "relation_type": "used_with", "target_slug": "active-comparator-new-user", "notes": "The defensible design for the comparative effectiveness estimate that feeds the model's relative effect against the scope comparator." }, { "relation_type": "used_with", "target_slug": "propensity-score-methods-psm-iptw", "notes": "Confounding control for RWE effectiveness inputs supplied to the reference-case base case." }, { "relation_type": "used_with", "target_slug": "single-arm-external-control", "notes": "Common in HST/oncology submissions; external-control comparability and bias must be addressed before the relative effect enters the model." }, { "relation_type": "used_with", "target_slug": "healthcare-costs-pppm-pppy-pmpm", "notes": "Cost inputs must be costed and re-based to the NHS/PSS perspective required by the reference case." }, { "relation_type": "used_with", "target_slug": "generalizability-transportability-external-validity-rwe", "notes": "Assesses whether the RWE population maps to the NHS decision population the appraisal serves." }, { "relation_type": "used_with", "target_slug": "quantitative-bias-analysis-toolkit-rwe", "notes": "Stress-tests residual bias in RWE inputs so decision uncertainty in the ICER is honestly characterised." }, { "relation_type": "see_also", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "The causal estimand and intercurrent-event handling must be explicit for the effectiveness input." }, { "relation_type": "see_also", "target_slug": "cost-effectiveness", "notes": "General method entry; the reference case is NICE's jurisdiction-specific standardisation of it." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "hta", "journal" ] }, { "slug": "nice-rwe-framework", "name": "NICE Real-World Evidence Framework", "short_definition": "The National Institute for Health and Care Excellence (NICE) corporate guidance (ECD9, 23 June 2022) describing best practice for planning, conducting, and reporting real-world evidence used in NICE technology appraisals, guidelines, and other HTA decision contexts, including a structured data-suitability assessment (DataSAT) and expectations for transparent, decision-grade comparative-effect studies.", "long_description": "**What it is.** The **NICE Real-World Evidence (RWE) Framework** is corporate guidance issued by the\nNational Institute for Health and Care Excellence (reference **ECD9**, published 23 June 2022) that sets out\nhow NICE expects real-world data and evidence to be planned, generated, assessed, and reported when it is used\nto inform NICE decisions — technology appraisals, highly specialised technologies, medical-technology and\ndiagnostics guidance, and clinical guidelines. It is **not** a reporting checklist in the EQUATOR sense and\nnot a single statistical method; it is an HTA-agency *reference framework* that pulls together three things:\n(1) a description of where RWE can add value in NICE decision making (characterising disease and treatment\npatterns, generating comparative-effect estimates where trials are absent or insufficient, validating\nsurrogate-to-final-outcome links, and informing economic models); (2) a structured approach to **assessing\ndata suitability** — the **Data Suitability Assessment Tool (DataSAT)** covering data provenance, governance,\nrelevance, and quality (completeness, accuracy, validity, coverage); and (3) methodological expectations for\nthe **conduct of quantitative RWE studies of comparative effects**, emphasising pre-specification, study\nregistration, transparency, and explicit handling of confounding and bias. The framework is maintained by\nNICE (Science, Evidence and Analytics directorate) and is the UK HTA counterpart to the FDA RWE program\nguidances and the EMA/ENCePP and Heads of Medicines Agencies–EMA Big Data work; it is deliberately aligned\nin spirit with HARPER, STaRT-RWE, and the STROBE/RECORD-PE reporting tradition rather than replacing them.\n\n**When to use.** Use the NICE RWE Framework whenever real-world data will feed a **NICE (or, by extension, a\nUK/devolved-nation HTA) submission or appraisal** — for example, a single-arm trial benchmarked against an\n**external control** drawn from registries or electronic health records; a **comparative-effectiveness**\nanalysis substituting for a missing head-to-head trial; **disease epidemiology, treatment pathways, or\nresource-use and cost** inputs for the economic model; or **surrogate-endpoint** validation. Decision rules\nfor when *this* framework applies rather than a sibling: choose the NICE framework when the decision context\nis **HTA/reimbursement in the NICE remit** and you need agency-specific expectations on data suitability and\nacceptability of comparative-effect estimates. Use the **FDA RWE framework/guidances** when the destination\nis a US regulatory (label/effectiveness) decision, and **EMA/ENCePP** guidance (ENCePP Methods/Checklist,\nGVP, registry-based-studies guideline) when the destination is an EU regulatory submission or an imposed\n**PASS**. These are complementary, not interchangeable: a single study can be designed to satisfy more than\none, but the *acceptance criteria and emphasis differ* — NICE foregrounds transparency, data suitability, and\nfitness of the comparator/economic inputs for a UK decision problem (the PICO of the appraisal), whereas the\nregulatory frameworks foreground label-relevant effectiveness and pharmacovigilance.\n\n**What it requires.** Framed for routinely collected data, the framework's substantive expectations cluster as:\n(1) **Design transparency and pre-specification** — a registered protocol and statistical analysis plan with\nthe target estimand stated *before* analysis; NICE explicitly encourages study registration (e.g.,\nENCePP/EU PAS or equivalent) and reporting against an established guideline. (2) **Data fitness-for-use** —\nthe **DataSAT** dimensions: provenance and governance (how data were collected and curated, legal basis),\n**relevance** (does the dataset's population, exposures, outcomes, and follow-up map to the decision PICO?),\nand **quality** (completeness, accuracy, internal/external validity, and coverage). (3) **Phenotype and\nalgorithm validation** — operational definitions of exposures, outcomes, and covariates with validation\nevidence (e.g., PPV/sensitivity of code-based outcome and diagnosis algorithms). (4) **Time-zero alignment**\n— index-date definition that avoids immortal-time bias and aligns eligibility, treatment assignment, and the\nstart of follow-up (the target-trial discipline). (5) **Estimands and intercurrent events** — explicit target\npopulation, treatment strategies, and handling of switching, discontinuation, and competing events.\n(6) **Confounding control** — appropriate methods (active-comparator new-user design, propensity-score and\nhigh-dimensional PS approaches, g-methods where time-varying confounding is present) with covariate-balance\ndiagnostics. (7) **Attrition and missing data** — transparent attrition accounting and principled missing-data\nhandling. (8) **Sensitivity and quantitative bias analysis** — negative controls, E-values, and probabilistic\nbias analysis to probe residual confounding, misclassification, and selection. The framework asks that each of\nthese be *justified for the specific decision problem*, not merely performed.\n\n**When NOT to use — limitations and common misapplications.** (1) **It is not a reporting checklist and not a\nrisk-of-bias score.** Using the NICE framework where a *reporting* instrument is required (STROBE for\nobservational studies, RECORD/RECORD-PE for routinely collected health data, CHEERS for economic evaluations)\nis a category error — the framework points to those tools; it does not substitute for them, and there is no\n\"NICE framework score.\" (2) **Completing DataSAT does not make a study causal or unbiased.** A dataset can pass\ndata-suitability review and still yield a confounded comparative estimate; suitability is necessary, not\nsufficient. (3) **Framework-as-theater** — appending a DataSAT table to a dossier without the design and\nanalysis actually controlling confounding, or registering a protocol after analysis, defeats the purpose and\nis visible to NICE technical teams. (4) **Wrong agency/context** — do not present a NICE-framed dossier as if\nit discharges FDA or EMA expectations, or vice versa; the decision problem (UK PICO vs label population)\ndiffers. (5) **Out of remit** — it governs RWE used in *NICE* decisions; for pure regulatory effectiveness or\npharmacovigilance, defer to FDA/EMA/ENCePP. (6) **Not a license to skip the comparator question** — using a\nnon-comparable external control or a prevalent-user cohort because the data were \"suitable\" reintroduces the\nbiases the methodological chapter warns against.\n\n**How it maps to this catalog.** The framework's requirements are implemented by concrete concepts here:\ndata fitness-for-use → **fit-for-purpose-data-assessment-rwe** (with payer/source nuance in\n**medicare-ffs-ma-commercial-claims-differences-rwe** and the **claims-analysis** base concept); design\ntransparency and the decision PICO → **picots-framework-rwe** and **study-protocol-or-sap-elements**;\nthe design backbone for comparative effects → **target-trial-emulation** and **active-comparator-new-user**;\ntime-zero alignment → **time-zero-index-date-alignment-rwe**; phenotype/algorithm validation →\n**diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe** and **claims-outcome-algorithm-ppv-sensitivity-rwe**;\nestimands and intercurrent events → **estimands-ate-att-intercurrent-events-rwe** (with\n**competing-risks-cause-specific-fine-gray-rwe** for competing events); confounding control →\n**high-dimensional-propensity-score-hdps-rwe**, **propensity-score-methods-psm-iptw**, and\n**marginal-structural-models-g-methods** for time-varying confounding; attrition →\n**attrition-and-loss-to-follow-up-rwe** and **database-feasibility-attrition-funnel-rwe**; sensitivity and\nquantitative bias analysis → **e-value-sensitivity-analysis**, **negative-control-outcomes-rwe**, and\n**quantitative-bias-analysis-toolkit-rwe**. For the economic side that distinguishes NICE from regulators,\nthe framework's model inputs map to **survival-extrapolation-hta-rwe**, **cost-effectiveness**, and\n**healthcare-costs-pppm-pppy-pmpm**. **Applied note (claims/EHR/registry RWE):** for a UK appraisal using an\nexternal control to benchmark a single-arm oncology trial, the practical NICE-aligned chain is —\nrun DataSAT on the CPRD/registry source (relevance to the appraisal PICO, completeness of outcomes and\nmortality linkage), pre-register a **target-trial-emulation** protocol with an **active-comparator-new-user**\ncontrast where feasible, fix time zero to avoid immortal time, balance with **hdPS**, account for **attrition**\nwith a CONSORT-style funnel, and bound residual confounding with an **E-value** and **negative controls**\nbefore the estimate is allowed into the economic model.", "primary_category": "Guideline", "tags": [ "guideline", "hta", "nice", "real-world-evidence", "framework", "data-suitability", "comparative-effectiveness" ], "aliases": [ "NICE RWE Framework", "NICE real-world evidence framework", "NICE ECD9", "ECD9", "DataSAT", "Data Suitability Assessment Tool" ], "applies_to_study_types": [ "cer_observational", "claims_analysis", "ehr_study", "disease_registry", "single_arm_external_control", "cost_effectiveness" ], "data_sources": [ "claims", "ehr", "registry", "linked", "multi-database" ], "citations": [ { "role": "introduce", "doi": null, "url": "https://www.nice.org.uk/corporate/ecd9", "citation_text": "National Institute for Health and Care Excellence. NICE real-world evidence framework. Corporate document ECD9. Published 23 June 2022.", "year": 2022, "authors_short": "NICE", "notes": "Canonical statement of the framework. As NICE corporate guidance it has no journal DOI; this is the stable agency landing page (the full PDF and chapters, including \"Assessing data suitability\" / DataSAT and \"Methods for real-world studies of comparative effects\", are linked from it)." }, { "role": "explain", "doi": "10.57264/cer-2023-0135", "url": "https://doi.org/10.57264/cer-2023-0135", "citation_text": "Duffield S, et al. The real-world impact of the National Institute for Health and Care Excellence's real-world evidence framework. Journal of Comparative Effectiveness Research. 2023;12(12):e230135.", "year": 2023, "authors_short": "Duffield et al.", "notes": "NICE-authored evaluation of the framework's first-year use and impact on appraisals; clarifies how DataSAT and the comparative-effects expectations are applied in practice." }, { "role": "use", "doi": null, "url": "https://www.nice.org.uk/corporate/ecd9/chapter/assessing-data-suitability", "citation_text": "NICE. Assessing data suitability (Data Suitability Assessment Tool, DataSAT). Chapter of the NICE real-world evidence framework (ECD9).", "year": 2022, "authors_short": "NICE", "notes": "Operational chapter and DataSAT instrument for the data-fitness-for-use assessment the framework requires; use directly when documenting provenance, relevance, and quality." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "cer-observational", "notes": "Governs comparative-effectiveness observational studies whose results inform a NICE appraisal or guideline." }, { "relation_type": "applies_to", "target_slug": "single-arm-external-control", "notes": "Central use case — benchmarking a single-arm trial against a real-world external control, where DataSAT and comparator suitability are decisive." }, { "relation_type": "applies_to", "target_slug": "claims-analysis", "notes": "Applies when claims data feed a NICE decision; pair with DataSAT for source fitness and with confounding-control concepts for the comparative estimate." }, { "relation_type": "applies_to", "target_slug": "ehr-study", "notes": "Applies to EHR-based RWE (e.g., CPRD) used in NICE appraisals and guidelines." }, { "relation_type": "applies_to", "target_slug": "cost-effectiveness", "notes": "Governs real-world inputs to the NICE economic model — the HTA emphasis that distinguishes this framework from regulatory RWE guidances." }, { "relation_type": "requires", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "DataSAT implements the data-fitness-for-use requirement (provenance, relevance, quality) that the framework places before any analysis." }, { "relation_type": "requires", "target_slug": "picots-framework-rwe", "notes": "The framework demands that RWE be matched to the appraisal's decision problem; PICOTS makes that population/comparator/outcome alignment explicit." }, { "relation_type": "complements", "target_slug": "target-trial-emulation", "notes": "The framework's comparative-effects chapter is, in effect, target-trial discipline; pre-specify the hypothetical trial protocol before emulation." }, { "relation_type": "complements", "target_slug": "active-comparator-new-user", "notes": "Preferred design for the comparative-effect estimates the framework expects, controlling confounding by indication and prevalent-user bias." }, { "relation_type": "complements", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "Confounding-control approach the framework's methods expectations point toward when key confounders are unmeasured in routine data." }, { "relation_type": "complements", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Implements the framework's requirement to state the target estimand and handle intercurrent events before analysis." }, { "relation_type": "complements", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Implements the phenotype/algorithm validation the framework requires for exposures, outcomes, and covariates." }, { "relation_type": "complements", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Implements the transparent attrition accounting the framework expects." }, { "relation_type": "complements", "target_slug": "e-value-sensitivity-analysis", "notes": "Quantitative bias analysis to bound residual confounding, as the framework's sensitivity-analysis expectations require." }, { "relation_type": "complements", "target_slug": "survival-extrapolation-hta-rwe", "notes": "Real-world survival inputs for the NICE economic model, reflecting the framework's HTA-specific scope." }, { "relation_type": "see_also", "target_slug": "medicare-ffs-ma-commercial-claims-differences-rwe", "notes": "Source-specific completeness/coverage nuances that feed a DataSAT quality and relevance assessment." }, { "relation_type": "see_also", "target_slug": "regulatory-readiness-rwe", "notes": "Compare NICE's HTA expectations with FDA/EMA regulatory-readiness expectations; a study may need to satisfy both, with different acceptance criteria." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "hta", "fda", "ema" ] }, { "slug": "nih-qat", "name": "NIH Study Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies", "short_definition": "A 14-item critical-appraisal (risk-of-bias) instrument from the NHLBI for judging the internal validity of observational cohort and cross-sectional studies, scored item-by-item and synthesized into an overall good/fair/poor rating rather than a numeric quality score.", "long_description": "**What it is.** The **NIH Study Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies (NIH\nQAT)** is a 14-item critical-appraisal instrument developed and maintained by the **National Heart, Lung, and Blood\nInstitute (NHLBI)** of the U.S. National Institutes of Health, built originally to support NHLBI evidence-based\nclinical practice guideline panels and now one of the most widely used risk-of-bias tools in published observational\nevidence syntheses. It is a **quality / risk-of-bias appraisal tool**, not a reporting checklist: each item asks\nwhether a feature that protects internal validity was actually present (clear research question; defined and uniformly\nrecruited population; ≥50% participation; pre-specified eligibility; sample-size/power justification; exposure measured\n*before* outcome; sufficient timeframe to see an effect; examination of exposure as continuous or by levels; valid and\nreliable exposure measurement applied consistently; blinded/independent exposure assessors; adequate follow-up with\nloss <20%; valid and reliable outcome measurement; blinded outcome assessors; and **measurement and statistical\nadjustment of key confounders**). The reviewer answers Yes / No / Cannot Determine / Not Reported / Not Applicable per\nitem and then makes a **reasoned overall judgment of good, fair, or poor** internal validity. NIH QAT is part of a\nfamily of NHLBI tools (separate instruments exist for controlled intervention studies, systematic\nreviews/meta-analyses, case-control studies, and before-after [pre-post] studies with no control group).\n\n**When to use.** Use the NIH QAT to appraise the **risk of bias of individual observational cohort or cross-sectional\nstudies** inside a systematic review, scoping/evidence map, AHRQ-style evidence report, HTA evidence section, or the\nquality-appraisal step of a peer-reviewed manuscript — including real-world-data studies built on claims, EHR, or\nregistry sources. Decision rules for *which* instrument applies: (1) one or more cohorts or a cross-sectional sample →\n**this** tool; (2) case-control sampling on outcome status → the NIH **case-control** tool; (3) single-arm\nbefore/after with no concurrent control → the NIH **before-after** tool; (4) an RCT or controlled intervention study →\nthe NIH **controlled-intervention** tool, never this one. If the synthesis question is the **causal effect of an\nintervention** and you need a structured, signalling-question, domain-level judgment for decision-grade evidence,\nprefer **ROBINS-I**; reach for NIH QAT when you need a lighter, faster, transparent appraisal across many\nobservational studies and a defensible good/fair/poor verdict.\n\n**What it requires.** The 14 items map directly onto the bias structure of real-world evidence and force the appraiser\nto interrogate exactly the things that sink observational drug and device studies: a clear, answerable question;\n**defined and uniformly recruited source population** (the catalog's data-fitness and feasibility-funnel concerns);\n**temporality** — exposure ascertained before the outcome, which is the appraisal-side mirror of correct **time-zero /\nindex-date alignment** and protects against immortal-time and reverse-causation errors; **valid, reliable, consistently\napplied exposure and outcome measurement**, i.e., validated **phenotype/claims algorithms** with known PPV and\nsensitivity rather than unvalidated code lists; **sufficient follow-up time** and **attrition <20%** with handling of\nloss to follow-up; and, the single most consequential item, **identification, measurement, and statistical adjustment\nof key potential confounders**. A study can satisfy every reporting checklist and still be rated *poor* here if\nconfounding by indication is unaddressed or the outcome algorithm is unvalidated.\n\n**When NOT to use — limitations and common misapplications.**\n- **Do not sum the 14 items into a numeric quality score.** NHLBI explicitly designed the tool for a *reasoned*\n good/fair/poor judgment; tallying \"11/14\" treats all items as equally weighted (they are not — confounding and\n temporality dominate) and manufactures false precision. A study with one fatal flaw (no confounding adjustment) is\n *poor* regardless of how many other boxes are checked.\n- **It is not a reporting guideline.** It judges whether validity-protecting features were *present*, not whether the\n manuscript *reported* them transparently. Use STROBE (or RECORD/RECORD-PE for routinely-collected health data) for\n reporting; NIH QAT cannot substitute, and \"Not Reported\" answers conflate poor reporting with poor conduct.\n- **It does not make an observational study causal.** A *good* rating means low risk of bias relative to the question\n asked; it does not license a causal claim absent a sound design (target-trial logic, valid comparator, confounding\n control) and sensitivity/quantitative-bias analysis.\n- **Wrong instrument for the design.** Applying the cohort/cross-sectional tool to case-control or before-after data\n misfires (e.g., the participation-rate and temporality items behave differently under outcome-based sampling).\n- **Appraisal-as-theater.** Filling in the grid without engaging the study's actual analytic choices (comparator,\n estimand, algorithm validation) produces a defensible-looking but empty rating. Reviewers see through it.\n- **Causal-effect, decision-grade contexts** are better served by ROBINS-I; the NIH QAT's closest sibling is the\n **Newcastle-Ottawa Scale (NOS)**, and JBI critical-appraisal checklists are reasonable alternatives — choose\n deliberately, do not default to NIH QAT because it is familiar.\n\n**How it maps to this catalog.** The NIH QAT's items are appraisal-side questions; the catalog concepts are the\nconduct-side answers that let a study earn a *good* rating:\n- *Temporality / exposure-before-outcome (items 6-7)* → `time-zero-index-date-alignment-rwe`,\n `active-comparator-new-user`, `washout-clean-lookback-period-rwe`, `target-trial-emulation`.\n- *Valid, reliable, consistent exposure/outcome measurement (items 9, 11)* →\n `diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe`, `algorithm-validation`,\n `claims-outcome-algorithm-ppv-sensitivity-rwe`.\n- *Defined, uniformly recruited population (items 3-5)* → `fit-for-purpose-data-assessment-rwe`,\n `database-feasibility-attrition-funnel-rwe`, `continuous-enrollment-observable-time-rwe`.\n- *Follow-up and attrition <20% (items 8, 13)* → `attrition-and-loss-to-follow-up-rwe`.\n- *Key-confounder measurement and adjustment (item 14)* → `high-dimensional-propensity-score-hdps-rwe`,\n `propensity-score-methods-psm-iptw`, `dags-backdoor-criterion-drug-studies`,\n `estimands-ate-att-intercurrent-events-rwe`, `quantitative-bias-analysis-toolkit-rwe`,\n `e-value-sensitivity-analysis`.\n\n**Applied note (claims/EHR/registry RWE).** When the appraised study is a claims or EHR cohort, the two items that\nmost often decide good-vs-poor are measurement validity and confounding. For measurement, a *good* rating requires a\nvalidated outcome algorithm (reported PPV/sensitivity), not a bare ICD/NDC list — score item 11 as \"No\" or \"Cannot\nDetermine\" when validation is absent. For confounding, look for a defensible comparator (active comparator over\nnon-user), a high-dimensional propensity score on a pre-index lookback, and a negative-control/quantitative-bias\nanalysis; the absence of confounding adjustment is, by itself, sufficient grounds for an overall *poor* rating.", "primary_category": "Guideline", "tags": [ "guideline", "risk-of-bias", "critical-appraisal", "quality-assessment", "observational", "cohort", "cross-sectional" ], "aliases": [ "NIH QAT", "NHLBI Quality Assessment Tool", "NIH Study Quality Assessment Tool", "Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies", "NHLBI risk-of-bias tool" ], "applies_to_study_types": [ "cohort_prospective", "cohort_retrospective", "cross_sectional" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "url": "https://www.nhlbi.nih.gov/health-topics/study-quality-assessment-tools", "citation_text": "National Heart, Lung, and Blood Institute (NHLBI). Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies. Study Quality Assessment Tools. National Institutes of Health.", "year": 2014, "authors_short": "NHLBI / NIH", "notes": "Canonical source and maintained instrument; NHLBI provides the 14 items, the Yes/No/CD/NR/NA response set, and the instruction to render an overall good/fair/poor rating rather than a summed score. The tool has no separate journal statement paper." }, { "role": "explain", "doi": "10.1186/s40779-020-00238-8", "url": "https://doi.org/10.1186/s40779-020-00238-8", "citation_text": "Ma LL, Wang YY, Yang ZH, Huang D, Weng H, Zeng XT. Methodological quality (risk of bias) assessment tools for primary and secondary medical studies: what are they and which is better? Military Medical Research. 2020;7(1):7.", "year": 2020, "authors_short": "Ma et al.", "notes": "Comparative review situating the NIH QAT among risk-of-bias instruments by design type and clarifying its role as an internal-validity (not reporting) appraisal." }, { "role": "use", "doi": "10.1111/jep.12889", "url": "https://doi.org/10.1111/jep.12889", "citation_text": "Quigley JM, Thompson JC, Halfpenny NJ, Scott DA. Critical appraisal of nonrandomized studies—A review of recommended and commonly used tools. Journal of Evaluation in Clinical Practice. 2019;25(1):44-52.", "year": 2019, "authors_short": "Quigley et al.", "notes": "Evaluates the NIH QAT against domains expected of nonrandomized-study appraisal tools; underscores the limits of checklist-style scoring and the centrality of confounding and analysis items." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "cohort-prospective", "notes": "Use to appraise the internal validity of prospective cohort studies." }, { "relation_type": "applies_to", "target_slug": "cohort-retrospective", "notes": "Use to appraise the internal validity of retrospective (including claims/EHR) cohort studies." }, { "relation_type": "applies_to", "target_slug": "cross-sectional", "notes": "Use to appraise the internal validity of cross-sectional studies." }, { "relation_type": "see_also", "target_slug": "robins-i", "notes": "Prefer ROBINS-I for decision-grade causal-effect appraisal with structured signalling questions; NIH QAT is the lighter, faster alternative for high-volume observational appraisal." }, { "relation_type": "alternative_to", "target_slug": "nos", "notes": "The Newcastle-Ottawa Scale is the NIH QAT's closest sibling for cohort/case-control appraisal; both render a qualitative validity judgment rather than a validated numeric score." }, { "relation_type": "see_also", "target_slug": "jbi-cohort", "notes": "JBI critical-appraisal checklist for cohort studies is an alternative observational appraisal instrument." }, { "relation_type": "see_also", "target_slug": "jbi-cross-sectional", "notes": "JBI critical-appraisal checklist for analytical cross-sectional studies is an alternative appraisal instrument." }, { "relation_type": "used_with", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Item 11 (valid, reliable outcome measurement) is satisfied by validated phenotype/claims algorithms with reported operating characteristics." }, { "relation_type": "used_with", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "Item 14 (measure and adjust for key confounders) is operationalized in RWE by hdPS and other confounding control methods." }, { "relation_type": "used_with", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Items 8 and 13 (sufficient follow-up; loss to follow-up <20%) are appraised against attrition handling." }, { "relation_type": "used_with", "target_slug": "time-zero-index-date-alignment-rwe", "notes": "Item 6 (exposure measured before outcome) is the appraisal mirror of correct time-zero/index-date alignment." }, { "relation_type": "used_with", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "Defined-population and measurement items presuppose a data source assessed as fit for purpose." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "journal", "hta" ] }, { "slug": "nos", "name": "Newcastle-Ottawa Scale (NOS)", "short_definition": "A star-based critical-appraisal/risk-of-bias instrument for non-randomized observational studies (cohort and case-control) used within systematic reviews and meta-analyses, scoring three domains (selection, comparability, outcome/exposure ascertainment) out of a maximum of nine stars.", "long_description": "**What it is.** The **Newcastle-Ottawa Scale (NOS)** is a critical-appraisal / risk-of-bias instrument\nfor non-randomized observational studies, developed by Wells, Shea, O'Connell and colleagues at the\nOttawa Hospital Research Institute (OHRI) in collaboration with the University of Newcastle (Australia)\nand maintained by OHRI rather than by EQUATOR or Cochrane. It exists as **two separate forms** — one\nfor **cohort studies** and one for **case-control studies** — that share a common architecture but use\ndifferent items. Appraisal is done by awarding **stars** across **three domains**: **Selection** (up to\n4 stars — representativeness/definition of the exposed and non-exposed or case and control groups, and\nascertainment of exposure or outcome at baseline), **Comparability** (up to 2 stars — whether the study\ncontrolled for the most important confounder and for additional confounders, with the reviewer naming\nthe factors), and **Outcome** (cohort) or **Exposure** (case-control) ascertainment (up to 3 stars —\nblinding/record linkage, length and adequacy of follow-up, or non-response/ascertainment method). The\nmaximum score is **nine stars**. NOS is deliberately lightweight: a trained reviewer can complete a form\nin roughly 20-30 minutes, which is why it became the default appraisal tool for observational evidence\nin thousands of meta-analyses.\n\n**When to use.** Use NOS when you must appraise the methodological quality / risk of bias of individual\n**cohort or case-control studies** as part of a **systematic review or meta-analysis** of observational\nevidence — the standard context in a peer-reviewed evidence-synthesis manuscript, an HTA/payer evidence\ndossier that summarizes the observational literature, or a regulatory background section that grades\nexisting non-randomized studies. Decision rules for choosing NOS vs siblings: (1) pick the **cohort\nform** for prospective/retrospective cohorts and the **case-control form** for case-control designs —\ndo not mix them. (2) For appraising **non-randomized studies of *interventions*** (comparative drug or\nprocedure effects), the **Cochrane Handbook now recommends ROBINS-I** (Sterne 2016) over NOS; reserve\nNOS for etiologic/prognostic exposure-outcome questions or where a fast, transparent appraisal across\nmany studies is needed. (3) NOS appraises **study-level conduct**, not **reporting completeness** — if\nthe task is to check whether a routinely-collected-data study is fully reported, use **RECORD/RECORD-PE**\n(a STROBE extension), not NOS. (4) NOS is **not** an evidence-certainty grading system; certainty of the\npooled body of evidence is graded with **GRADE**, which sits downstream of per-study appraisal.\n\n**What it requires.** Completing a NOS form requires the appraiser to make and document explicit\njudgments about: (a) **selection** — whether the exposed/case group is representative and the\ncomparison group is drawn from the same source population, and whether exposure (case-control) or\noutcome (cohort) was secure and absent at baseline; (b) **comparability** — which confounders the\nstudy controlled for, with the reviewer **pre-specifying the most important confounder** (and a second\nfactor) that must be adjusted to earn the comparability stars; and (c) **outcome/exposure\nascertainment** — objective/record-linked vs self-reported ascertainment, blinding, and **adequacy and\nlength of follow-up** with accounting for those lost. For real-world-data studies this maps onto\nexposure and outcome **phenotype/algorithm validity** (was the claims/EHR case definition validated?),\n**data fitness-for-use** (is the source population representative and the comparison group from the same\ndata?), and **attrition / loss to follow-up** (is follow-up long enough and is dropout accounted for?).\n\n**When NOT to use — limitations and common misapplications.** NOS is widely used and widely criticized,\nand a defensible appraisal must acknowledge this. (1) **It is an appraisal tool, not a validated quality\n*score*.** Summing stars into a single number and applying a cut-off (the ubiquitous \"≥7 stars = high\nquality\") is **not validated**; the developers never established score thresholds, and meta-regression or\nsubgroup analysis on NOS totals can mislead. (2) **Poor inter-rater reliability and ambiguous wording.**\nStang (2010) showed the \"comparability\" star hinges entirely on which confounder the reviewer chooses to\nrequire, \"adequacy of follow-up\" is arbitrary, and several items are open to divergent interpretation;\nHartling (2013) empirically found **low reliability between independent reviewers**. Pre-specify, in your\nprotocol, the required confounders and the follow-up threshold before appraisal. (3) **It does not\ncapture the dominant biases in pharmacoepidemiology.** NOS has no item for **immortal time bias**,\n**time-zero / index-date misalignment**, **prevalent-user (depletion-of-susceptibles) bias**, or\n**confounding by indication** — a claims or EHR study can earn the full nine stars and still be biased to\nthe point of uselessness for a regulatory or HTA decision. (4) **Wrong tool for the question:** using NOS\nto appraise non-randomized *intervention* effects where ROBINS-I is expected, using it as a reporting\nchecklist (STROBE/RECORD-PE), or using it to grade body-of-evidence certainty (GRADE). (5)\n**Appraisal-as-theater** — filling in stars to satisfy a journal without letting the result change the\nsynthesis or the interpretation.\n\n**How it maps to this catalog.** NOS asks *whether* a study controlled bias; the catalog concepts\nspecify *how*. The **comparability** domain (control of confounding) is operationalized by\n`active-comparator-new-user` (design-stage confounding-by-indication control) and\n`high-dimensional-propensity-score-hdps-rwe` (analytic confounding control) — both address exactly the\nstar NOS reserves for \"controlled for the most important factor.\" The **selection** domain (secure,\ncomparable groups defined from the same source) is implemented through\n`diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe` (validated exposure/outcome definitions) and\n`claims-analysis` (source-population and data-fitness considerations). The **outcome/follow-up** domain\nmaps to `attrition-and-loss-to-follow-up-rwe`. The biases NOS structurally misses are precisely where\nthe catalog adds value: `target-trial-emulation` and the **time-zero** alignment it enforces close the\nimmortal-time and prevalent-user gaps, and `estimands-ate-att-intercurrent-events-rwe` makes explicit\nthe causal contrast and intercurrent-event handling that a star count cannot represent. **Applied note\nfor claims/EHR/registry RWE:** treat a high NOS score as necessary-but-insufficient. Before trusting a\nhighly-starred observational study in an evidence dossier, separately confirm that exposure/outcome\nphenotypes were validated, that time zero was aligned for both arms (no immortal time), that the design\nused an active comparator and new-user restriction where confounding by indication is plausible, and\nthat loss to follow-up was assessed as potentially informative — none of which NOS interrogates.", "primary_category": "Guideline", "tags": [ "guideline", "critical-appraisal", "risk-of-bias", "quality-assessment", "observational-studies", "systematic-review", "evidence-synthesis" ], "aliases": [ "NOS", "Newcastle-Ottawa Scale", "Newcastle-Ottawa Quality Assessment Scale", "NOQAS" ], "applies_to_study_types": [ "cohort_prospective", "cohort_retrospective", "case_control" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "url": "https://www.ohri.ca/programs/clinical_epidemiology/oxford.asp", "citation_text": "Wells GA, Shea B, O'Connell D, Peterson J, Welch V, Losos M, Tugwell P. The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses. Ottawa Hospital Research Institute. (cohort and case-control forms; maintained, undated.)", "year": 2000, "authors_short": "Wells et al.", "notes": "Canonical source. NOS has no journal statement paper; the OHRI page hosts the maintained cohort and case-control forms and the coding manual." }, { "role": "explain", "doi": "10.1007/s10654-010-9491-z", "url": "https://doi.org/10.1007/s10654-010-9491-z", "citation_text": "Stang A. Critical evaluation of the Newcastle-Ottawa Scale for the assessment of the quality of nonrandomized studies in meta-analyses. European Journal of Epidemiology. 2010;25(9):603-605.", "year": 2010, "authors_short": "Stang", "notes": "Definitive critique — ambiguous items, reviewer-dependent comparability star, arbitrary follow-up criterion; warns against treating star totals as a validated quality score." }, { "role": "explain", "doi": "10.1016/j.jclinepi.2013.03.003", "url": "https://doi.org/10.1016/j.jclinepi.2013.03.003", "citation_text": "Hartling L, Milne A, Hamm MP, et al. Testing the Newcastle Ottawa Scale showed low reliability between individual reviewers. Journal of Clinical Epidemiology. 2013;66(9):982-993.", "year": 2013, "authors_short": "Hartling et al.", "notes": "Empirical evidence of poor inter-rater reliability; motivates protocol pre-specification of required confounders and follow-up thresholds." }, { "role": "use", "doi": "10.1136/bmj.i4919", "url": "https://doi.org/10.1136/bmj.i4919", "citation_text": "Sterne JAC, Hernán MA, Reeves BC, et al. ROBINS-I: a tool for assessing risk of bias in non-randomised studies of interventions. BMJ. 2016;355:i4919.", "year": 2016, "authors_short": "Sterne et al.", "notes": "Cochrane-recommended successor for non-randomized studies of interventions; prefer over NOS when appraising comparative intervention effects." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "cohort-prospective", "notes": "Use the NOS cohort form to appraise prospective cohort studies in an observational evidence synthesis." }, { "relation_type": "applies_to", "target_slug": "cohort-retrospective", "notes": "Use the NOS cohort form for retrospective (e.g., claims/EHR) cohort studies." }, { "relation_type": "applies_to", "target_slug": "case-control", "notes": "Use the separate NOS case-control form; the exposure-ascertainment domain replaces the cohort outcome/follow-up domain." }, { "relation_type": "see_also", "target_slug": "target-trial-emulation", "notes": "NOS scores cannot detect immortal-time or time-zero misalignment; target-trial emulation closes exactly the biases NOS structurally omits." }, { "relation_type": "complements", "target_slug": "active-comparator-new-user", "notes": "The NOS comparability domain asks whether confounding was controlled; the active-comparator new-user design is how that control is achieved at the design stage for confounding by indication." }, { "relation_type": "complements", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "Implements the analytic confounding control that the NOS comparability stars credit." }, { "relation_type": "complements", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Secure, validated exposure/outcome definitions underpin the NOS selection and ascertainment domains in claims/EHR data." }, { "relation_type": "complements", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Operationalizes the adequacy-and-length-of-follow-up item of the NOS outcome domain." }, { "relation_type": "see_also", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "NOS records no estimand or intercurrent-event handling; specify these separately for a causal interpretation." }, { "relation_type": "see_also", "target_slug": "claims-analysis", "notes": "Source-population representativeness and data fitness-for-use that the NOS selection domain depends on for routinely-collected data." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "hta", "journal" ] }, { "slug": "ohdsi-cdm", "name": "OMOP Common Data Model and OHDSI Analytic Standards", "short_definition": "An open community data standard (OMOP CDM) plus a stack of standardized vocabularies, open-source analytic tooling (ATLAS, HADES), and methodological best practices maintained by the OHDSI collaborative for conducting reproducible, transparent observational research across federated databases.", "long_description": "**What it is.** The **Observational Medical Outcomes Partnership Common Data Model (OMOP CDM)** is an open,\nperson-centric relational schema that transforms heterogeneous observational health databases — administrative\nclaims, EHR, registries — into a single, consistent structure with standardized field definitions and standardized\nconcept vocabularies (a unified terminology layer mapping ICD, SNOMED, RxNorm, CPT/HCPCS, LOINC, NDC, and dozens\nmore to shared \"standard concepts\"). It is maintained by **OHDSI (Observational Health Data Sciences and\nInformatics)**, a non-profit, open-science collaborative, and is paired with an analytic-standards stack:\n**ATLAS** (a web tool for cohort/phenotype definition and study design), **HADES** (validated R packages for\npopulation-level effect estimation, patient-level prediction, and characterization), the **ACHILLES/Data Quality\nDashboard** tooling for data characterization and quality, and the **Book of OHDSI** as the canonical methods\nreference. Crucially, OMOP/OHDSI is *not* a reporting checklist (like STROBE, RECORD-PE, or CHEERS) and *not* a\nrisk-of-bias instrument (like ROBINS-I): it is a data-and-methods *standardization framework* whose value is\nreproducibility, executable study specifications, and network-scale evidence generation.\n\n**When to use.** Reach for OMOP/OHDSI when (a) a study must run identically across multiple databases or sites —\na **distributed/federated network study** where code travels to the data and only aggregate results return; (b)\nyou want an **executable, machine-readable study specification** (phenotype + analytic settings) that another team\ncan rerun and audit, supporting regulatory or HTA reproducibility expectations; (c) you are conducting\n**large-scale comparative effectiveness, safety surveillance, characterization, or patient-level prediction** and\nwant methods (propensity scores, negative-control calibration, large-scale diagnostics) baked into validated\ntooling; or (d) you are building reusable phenotype libraries that must map consistently across coding systems.\nDecision rule: choose OMOP/OHDSI when **standardization, reproducibility, and multi-database portability** are\nfirst-order requirements. For a single-database, bespoke analysis with idiosyncratic variables, the ETL cost of\nconverting to the CDM may not pay off, and a native-schema analysis (still governed by the relevant *reporting*\nguideline, e.g., RECORD-PE/HARPER) can be the simpler, defensible choice. OMOP/OHDSI is a complement to — never a\nsubstitute for — those reporting and design guidelines.\n\n**What it requires.** Operationally, conformance and good practice demand: (1) **ETL with documented provenance**\n— every source code mapped to a standard concept, with a maintained source-to-CDM specification; (2)\n**data-quality and fitness-for-use assessment** before analysis (Data Quality Dashboard checks for conformance,\ncompleteness, plausibility; ACHILLES characterization), because standardized structure does not guarantee fitness\nfor a given question; (3) **transparent, shareable phenotype/cohort definitions** with explicit entry/exit events,\ninclusion logic, and — critically — **phenotype validation** (the CDM standardizes *codes*, not clinical truth);\n(4) **explicit time-zero / index-event logic and observation-period anchoring** in the cohort definition; (5)\n**pre-specified estimands and confounding control** (new-user/active-comparator cohorts, large-scale propensity\nscores) executed through HADES; (6) **attrition reporting** via the standardized cohort-diagnostics/attrition\noutputs; and (7) **empirical bias control and diagnostics** — negative-control calibration of p-values and\nconfidence intervals, equipoise/covariate-balance checks, and pre-registration of the study package — which the\nOHDSI methodology elevates from optional sensitivity analysis to a default expectation.\n\n**When NOT to use — limitations and common misapplications.** OMOP/OHDSI standardizes *data and code*; it does\n**not** confer causal validity. Converting a database to the CDM and running ATLAS does not make a confounded\ncomparison unconfounded, nor a poorly validated phenotype accurate — the standardized concept set still inherits\nevery error of the source coding. Common failure modes: (i) treating a successful ETL as evidence of\nfitness-for-use and skipping data-quality/phenotype validation; (ii) **information loss during ETL** —\nsource-specific nuance (lab units, free-text, payer-specific benefit structure, Medicare Advantage vs FFS claims\ncompleteness) that is flattened or dropped when mapping to standard concepts, silently biasing downstream cohorts;\n(iii) over-trusting ATLAS \"rule-based\" phenotypes as validated when they have only face validity; (iv) assuming\nnetwork heterogeneity (different databases give different answers) is noise rather than a signal about\ngeneralizability or residual bias; (v) using OMOP/OHDSI as if it replaced a *reporting* guideline — a manuscript\nbuilt on OMOP still must satisfy STROBE/RECORD-PE, an HTA dossier still needs CHEERS for any economic component,\nand a PASS still follows ENCePP/HARPER. Mistaking the data standard for a quality or reporting standard is the\ncardinal error.\n\n**How it maps to this catalog.** OMOP/OHDSI is the standardized substrate; the concepts in this catalog supply the\nmethodological content it operationalizes. Fitness-for-use assessment → **fit-for-purpose-data-assessment-rwe** and\n**database-feasibility-attrition-funnel-rwe**. Phenotype/cohort definition and validation →\n**diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe** and **claims-outcome-algorithm-ppv-sensitivity-rwe**.\nDesign and time-zero alignment → **target-trial-emulation** and **active-comparator-new-user** (ATLAS cohorts are\nthe natural place to encode the new-user/active-comparator structure). Confounding control →\n**high-dimensional-propensity-score-hdps-rwe** (the OHDSI large-scale PS is the network-ready analogue). Estimand\nspecification → **estimands-ate-att-intercurrent-events-rwe** and **estimand-analysis-traceability-rwe**. Attrition\n→ **attrition-and-loss-to-follow-up-rwe**. Empirical bias control, a defining OHDSI practice →\n**empirical-calibration-negative-controls-rwe**, **negative-control-outcomes-rwe**, and\n**negative-control-exposures-rwe**. Data-source caveats that ETL must preserve →\n**medicare-ffs-ma-commercial-claims-differences-rwe** and **claims-analysis**.\n\n**Applied note (claims/EHR/registry).** In claims, the ETL must preserve enrollment/eligibility spans (the OMOP\n`observation_period`) faithfully, since \"absence of a code\" is only interpretable against continuous observability;\nMedicare Advantage person-time, which lacks fee-for-service claims, can masquerade as a clean washout if the ETL\ndoes not flag it. In EHR, drug *orders* vs *administrations* vs linked *dispensings* map to different OMOP domains\nand must be chosen deliberately for exposure definitions, and visit-driven capture means observation periods are\ninferred, not given. In registries, rich severity/staging often has no clean standard concept and is the first\ncasualty of mapping — validate that the CDM representation still answers the question before trusting network output.", "primary_category": "Guideline", "tags": [ "guideline", "methodological", "data-standard", "ohdsi", "omop-cdm", "reproducibility", "distributed-network-study", "standardized-vocabularies" ], "aliases": [ "OMOP CDM", "OMOP Common Data Model", "OHDSI", "OHDSI Analytic Standards", "Observational Medical Outcomes Partnership Common Data Model", "Book of OHDSI" ], "applies_to_study_types": [ "multi_database", "claims_analysis", "ehr_study", "cer_observational", "drug_utilization", "signal_detection" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "doi": "10.3233/978-1-61499-564-7-574", "url": "https://doi.org/10.3233/978-1-61499-564-7-574", "citation_text": "Hripcsak G, Duke JD, Shah NH, et al. Observational Health Data Sciences and Informatics (OHDSI): opportunities for observational researchers. Studies in Health Technology and Informatics. 2015;216:574-578.", "year": 2015, "authors_short": "Hripcsak et al.", "notes": "Foundational statement of the OHDSI mission, the OMOP CDM, standardized vocabularies, and the open analytic stack underpinning reproducible network-scale observational research." }, { "role": "explain", "doi": "10.1136/amiajnl-2011-000376", "url": "https://doi.org/10.1136/amiajnl-2011-000376", "citation_text": "Overhage JM, Ryan PB, Reich CG, Hartzema AG, Stang PE. Validation of a common data model for active safety surveillance research. Journal of the American Medical Informatics Association. 2012;19(1):54-60.", "year": 2012, "authors_short": "Overhage et al.", "notes": "Demonstrates that transforming disparate databases to the OMOP CDM preserves analytic content for safety surveillance — the empirical basis for the CDM as a standardization layer." }, { "role": "explain", "doi": "10.1162/99608f92.147cc28e", "url": "https://doi.org/10.1162/99608f92.147cc28e", "citation_text": "Schuemie MJ, Ryan PB, Pratt N, et al. How confident are we about observational findings in health care: a benchmark study. Harvard Data Science Review. 2020;2(1).", "year": 2020, "authors_short": "Schuemie et al.", "notes": "Articulates the OHDSI methodological discipline — large-scale empirical calibration with negative/positive controls and full study diagnostics — showing why CDM standardization alone is insufficient without bias control." }, { "role": "use", "url": "https://ohdsi.github.io/TheBookOfOhdsi/", "citation_text": "Observational Health Data Sciences and Informatics. The Book of OHDSI. OHDSI; 2021 (continuously maintained). Canonical reference for OMOP CDM, standardized vocabularies, ATLAS, and HADES methodology.", "year": 2021, "authors_short": "OHDSI Collaborative", "notes": "Maintained, open methodology reference and practical guide to building, validating, and executing OHDSI study packages across a database network." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "multi-database", "notes": "Core use case — distributed/federated network studies run identical code across OMOP-mapped databases." }, { "relation_type": "applies_to", "target_slug": "claims-analysis", "notes": "Claims ETL must preserve enrollment spans and source-coding nuance lost in concept mapping." }, { "relation_type": "applies_to", "target_slug": "ehr-study", "notes": "EHR exposure/outcome definitions depend on domain choice (order vs administration vs dispensing) in the CDM." }, { "relation_type": "applies_to", "target_slug": "cer-observational", "notes": "Comparative effectiveness via HADES population-level estimation with large-scale propensity scores." }, { "relation_type": "applies_to", "target_slug": "drug-utilization", "notes": "Standardized drug vocabularies (RxNorm) and exposure domains support portable utilization characterization." }, { "relation_type": "applies_to", "target_slug": "signal-detection", "notes": "Originating OMOP use case — standardized active safety surveillance across a database network." }, { "relation_type": "used_with", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "Standardized structure does not equal fitness-for-use; run data-quality and characterization before analysis." }, { "relation_type": "used_with", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "ATLAS encodes phenotype logic, but the CDM standardizes codes not clinical truth — validate phenotypes." }, { "relation_type": "used_with", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "The OHDSI large-scale propensity score is the network-ready analogue for confounding control." }, { "relation_type": "used_with", "target_slug": "empirical-calibration-negative-controls-rwe", "notes": "Negative-control calibration of effect estimates is a defining default of the OHDSI methodology." }, { "relation_type": "used_with", "target_slug": "active-comparator-new-user", "notes": "New-user/active-comparator structure is the standard design to encode in an ATLAS comparative cohort." }, { "relation_type": "see_also", "target_slug": "target-trial-emulation", "notes": "OMOP cohorts and HADES estimation provide a reproducible substrate for emulating a target trial." }, { "relation_type": "see_also", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Pre-specify the estimand and intercurrent-event handling before executing the OHDSI study package." }, { "relation_type": "see_also", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Use standardized cohort-diagnostics attrition outputs to report participant flow transparently." }, { "relation_type": "see_also", "target_slug": "medicare-ffs-ma-commercial-claims-differences-rwe", "notes": "ETL must flag Medicare Advantage person-time, which lacks FFS claims and corrupts observability assumptions." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "hta" ] }, { "slug": "pcori-standards", "name": "PCORI Methodology Standards", "short_definition": "Consensus methodology standards from the Patient-Centered Outcomes Research Institute (PCORI) that define minimum rigor for patient-centered comparative clinical effectiveness research, including observational designs using claims, EHR, and registry data.", "long_description": "**What it is.** The **PCORI Methodology Standards** are a set of consensus, peer-reviewed\nminimum-rigor standards for **patient-centered comparative clinical effectiveness research (CER)**,\ndeveloped and maintained by the **Methodology Committee of the Patient-Centered Outcomes Research\nInstitute (PCORI)** — a U.S. quasi-governmental funder created under the 2010 Affordable Care Act.\nThe current version (2024 update) comprises **more than fifty individual standards organized into\ncross-cutting and method-specific areas**: a set of cross-cutting standards (Formulating Research Questions; Patient\nCenteredness; Data Integrity and Rigorous Analyses; Preventing and Handling Missing Data;\nHeterogeneity of Treatment Effects; Usual Care as a Comparator) and design- or method-specific\nareas (Causal Inference Methods; Data Registries; Data Networks; Adaptive and Bayesian Trials;\nStudies of Medical Tests; Systematic Reviews; Research Designs Using Clusters; Complex\nInterventions; Qualitative Methods; Mixed Methods; Individual Participant-Level Data\nMeta-Analysis). The standards are mandatory for PCORI-funded work and are enforced through merit\nreview, contract terms, and the program-wide peer review of PCORI-funded results. Unlike a\nreporting checklist (STROBE, RECORD, CONSORT) or a risk-of-bias instrument (ROBINS-I), the PCORI\nStandards are **conduct standards**: they prescribe what the study must *do*, not merely what the\nmanuscript must *say*.\n\n**When to use.** Apply the PCORI Standards whenever you design, fund, conduct, or review\npatient-centered comparative effectiveness research — most directly for **PCORI-funded studies**,\nwhere compliance is contractual. Beyond that funding context they function as a widely cited\nreference standard for the *conduct* of comparative effectiveness and real-world evidence (RWE)\nstudies, including non-interventional cohort designs in claims/EHR/registry data, pragmatic trials,\nand evidence syntheses. Decision rule for which area governs: the **cross-cutting** areas\n(research questions, patient-centeredness, data integrity, missing data, heterogeneity, usual-care\ncomparator) apply to essentially every study; then layer the **design-specific** area that matches\nyour architecture — Causal Inference Methods for confounder-adjusted observational CER, Data\nRegistries / Data Networks for distributed or registry-based RWE, Systematic Reviews for syntheses,\nComplex Interventions for multi-component programs. The Standards are complementary to, not a\nsubstitute for, the relevant **reporting** guideline (use STROBE/RECORD-PE for the write-up,\nCONSORT for trials, PRISMA for reviews) and to FDA/EMA RWE guidance for regulatory submissions and\nHTA reference cases for payer dossiers.\n\n**What it requires.** For observational RWE, the binding requirements concentrate in a few areas\nand map cleanly onto modern pharmacoepidemiologic practice: (1) **Design transparency and a\npre-specified protocol/analysis plan** with a clearly framed, answerable research question\n(Formulating Research Questions); (2) **Data fitness-for-use** — justify that the data source can\nvalidly capture the exposure, outcome, covariates, and follow-up for the question, and document\ndata provenance and quality (Data Integrity; Data Registries; Data Networks); (3) **Valid\nascertainment and phenotype/algorithm validation** for exposures, outcomes, and covariates,\nincluding the operating characteristics of claims/EHR algorithms (Data Integrity; Studies of\nMedical Tests for diagnostic-accuracy work); (4) **Sound causal-inference design** — define the\ntarget estimand, align time zero, choose an appropriate comparator (with explicit standards on\n**Usual Care as a Comparator**), and control measured confounding while reasoning about unmeasured\nconfounding (Causal Inference Methods); (5) **Heterogeneity of treatment effects** — pre-specify\nsubgroups and the analytic approach rather than data-dredging (HTE area); (6) **Prevention and\nprincipled handling of missing data and attrition**, including documentation of the missingness\nmechanism and sensitivity analyses rather than naive complete-case defaults (Preventing and\nHandling Missing Data); and (7) **rigorous, pre-specified analysis with sensitivity / quantitative\nbias analysis** to probe robustness to the key threats (Data Integrity and Rigorous Analyses).\nPatient-centeredness — engaging patients in question formulation and outcome selection — is a\ncross-cutting requirement distinctive to PCORI relative to most epidemiologic standards.\n\n**When NOT to use — limitations and common misapplications.** (1) **The Standards are conduct\nstandards, not a reporting checklist** — satisfying them does not replace STROBE/RECORD-PE for the\nmanuscript, and they are **not a numeric quality score or a risk-of-bias instrument**; do not\nconvert them into a tick-box \"PCORI score\" or use them in place of ROBINS-I when grading study\nbias. (2) **Compliance does not make an observational study causal.** Meeting the Causal Inference\nMethods standards (propensity adjustment, time-zero alignment) constrains *design* bias but cannot\ncertify that ignorability holds; residual and unmeasured confounding remain, which is exactly why\nthe Standards require negative controls / quantitative bias analysis. (3) **Standards-as-theater** —\nasserting that a protocol \"follows PCORI Standards\" without demonstrable phenotype validation, a\npre-registered analysis plan, or executed sensitivity analyses is non-compliance dressed as\ncompliance. (4) **Wrong area for the design** — applying the Systematic Reviews standards to a\nprimary cohort study, or skipping the Data Networks standards for a distributed/multi-site RWE\nstudy, leaves the governing requirements unmet. (5) **Scope limits** — the Standards are U.S.-,\nCER-, and patient-centeredness-oriented; they are not a regulatory framework (use FDA/EMA RWE\nguidance) and not an HTA reference case (use the national HTA body's methods guide) for those\ndecision contexts.\n\n**How it maps to this catalog.** Several PCORI requirement areas are *implemented* by concrete\nconcepts in this repository: the **Causal Inference Methods** and comparator standards are\noperationalized by `target-trial-emulation` (estimand-anchored design, explicit time zero),\n`active-comparator-new-user` (comparator choice and immortal-time avoidance), and\n`high-dimensional-propensity-score-hdps-rwe` and `propensity-score-methods-psm-iptw` (measured\nconfounding control). The **estimand / heterogeneity** requirements map to\n`estimands-ate-att-intercurrent-events-rwe` and `estimand-analysis-traceability-rwe`. **Data\nintegrity, fitness-for-use, and phenotype validation** are implemented by\n`diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe`,\n`claims-outcome-algorithm-ppv-sensitivity-rwe`, `procedure-identification-and-measurement-in-claims-ehr`,\n`claims-analysis`, and `medicare-ffs-ma-commercial-claims-differences-rwe` (data-source-specific\nvalidity). **Missing data and attrition** map to `attrition-and-loss-to-follow-up-rwe`,\n`database-feasibility-attrition-funnel-rwe`, and `missing-data-pattern-table-rwe`. The required\n**sensitivity / quantitative bias analysis** is implemented by `e-value-sensitivity-analysis`,\n`empirical-calibration-negative-controls-rwe`, `negative-control-outcomes-rwe`, and\n`selection-bias-sensitivity-analysis-rwe`.\n\n**Applied note (claims/EHR/registry RWE).** For a PCORI-funded comparative-effectiveness cohort in\nadministrative claims, demonstrable compliance means: a registered protocol with a target-trial\nframing and named estimand; documented enrollment/data-quality criteria establishing fitness-for-use\n(e.g., continuous medical + pharmacy enrollment so absence of a fill is observed rather than\nmissing — see `medicare-ffs-ma-commercial-claims-differences-rwe`); validated exposure and outcome\nalgorithms with reported PPV/sensitivity (`claims-outcome-algorithm-ppv-sensitivity-rwe`); a\npre-specified attrition funnel and missing-data plan; a propensity-based analysis with balance\ndiagnostics; and a sensitivity battery (negative-control outcomes, E-value) that probes the\nunmeasured-confounding assumption the design cannot eliminate.", "primary_category": "Guideline", "tags": [ "guideline", "methodology-standards", "comparative-effectiveness-research", "patient-centered", "pcori", "causal-inference", "real-world-evidence" ], "aliases": [ "PCORI Standards", "PCORI Methodology Standards", "Patient-Centered Outcomes Research Institute Methodology Standards", "PCORI methodology standards and report" ], "applies_to_study_types": [ "cer_observational", "pragmatic_trial", "new_user", "active_comparator_new_user", "cohort_prospective", "cohort_retrospective" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "doi": "10.1001/jama.2012.466", "url": "https://doi.org/10.1001/jama.2012.466", "citation_text": "Methodology Committee of the Patient-Centered Outcomes Research Institute (PCORI). Methodological standards and patient-centeredness in comparative effectiveness research: the PCORI perspective. JAMA. 2012;307(15):1636-1640.", "year": 2012, "authors_short": "PCORI Methodology Committee", "notes": "Peer-reviewed statement establishing the PCORI Methodology Standards framework and its patient-centeredness orientation; the canonical reference for the standards." }, { "role": "explain", "doi": "10.1056/NEJMp1207437", "url": "https://doi.org/10.1056/NEJMp1207437", "citation_text": "Gabriel SE, Normand SLT. Getting the methods right — the foundation of patient-centered outcomes research. New England Journal of Medicine. 2012;367(9):787-790.", "year": 2012, "authors_short": "Gabriel & Normand", "notes": "Foundational perspective from the Methodology Committee chair and vice-chair on why methodological rigor anchors patient-centered outcomes research." }, { "role": "explain", "doi": "10.1007/s11606-020-06093-6", "url": "https://doi.org/10.1007/s11606-020-06093-6", "citation_text": "Esmail L, Gerson N, Garrison L, et al. Improving comparative effectiveness research of complex health interventions: standards from the Patient-Centered Outcomes Research Institute (PCORI). Journal of General Internal Medicine. 2020;35(3):875-881.", "year": 2020, "authors_short": "Esmail et al.", "notes": "Worked elaboration of a design-specific standards area (Complex Interventions), illustrating how a PCORI standards area is articulated and applied." }, { "role": "use", "url": "https://www.pcori.org/research-related-projects/about-our-research/research-methodology/pcori-methodology-standards", "citation_text": "PCORI Methodology Standards (current version) and the PCORI Methodology Report. Patient-Centered Outcomes Research Institute — maintained standards and supporting documentation.", "year": 2024, "authors_short": "PCORI", "notes": "Authoritative, version-controlled source for the full set of standards (2024 update, including Usual Care as a Comparator)." } ], "relations": [ { "relation_type": "used_with", "target_slug": "target-trial-emulation", "notes": "Implements the Causal Inference Methods standards — estimand-anchored design with explicit time zero for observational CER." }, { "relation_type": "used_with", "target_slug": "active-comparator-new-user", "notes": "Operationalizes the comparator and design-validity standards (comparator choice, immortal-time avoidance, time-zero alignment)." }, { "relation_type": "used_with", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "Implements the Causal Inference Methods requirement to control measured confounding in high-dimensional claims/EHR data." }, { "relation_type": "used_with", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Implements the standards on defining the target estimand and pre-specifying treatment-effect contrasts and intercurrent-event handling." }, { "relation_type": "used_with", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Implements the Data Integrity / fitness-for-use standards through validated exposure, outcome, and covariate phenotypes." }, { "relation_type": "used_with", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Implements the Preventing and Handling Missing Data standards via documented attrition and informative-censoring assessment." }, { "relation_type": "used_with", "target_slug": "empirical-calibration-negative-controls-rwe", "notes": "Implements the required sensitivity / quantitative bias analysis probing residual and unmeasured confounding." }, { "relation_type": "used_with", "target_slug": "claims-analysis", "notes": "Apply when the comparative-effectiveness study is conducted in claims/EHR/registry real-world data." }, { "relation_type": "see_also", "target_slug": "e-value-sensitivity-analysis", "notes": "A concrete sensitivity-analysis tool satisfying the standards' requirement to quantify robustness to unmeasured confounding." }, { "relation_type": "see_also", "target_slug": "medicare-ffs-ma-commercial-claims-differences-rwe", "notes": "Data-source-specific fitness-for-use considerations underpinning the Data Integrity standards in U.S. claims." }, { "relation_type": "applies_to", "target_slug": "cer-observational", "notes": "Primary use case — observational comparative effectiveness research." }, { "relation_type": "applies_to", "target_slug": "pragmatic-trial", "notes": "Pragmatic CER trials fall under PCORI conduct standards (including adaptive/Bayesian and cluster-design areas where applicable)." }, { "relation_type": "applies_to", "target_slug": "active-comparator-new-user", "notes": "Recommended design framing for confounding-by-indication control under the Causal Inference Methods standards." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "hta", "journal" ] }, { "slug": "precis-2", "name": "PRECIS-2 (Pragmatic-Explanatory Continuum Indicator Summary 2)", "short_definition": "A trial-design and appraisal tool that scores nine design domains on a 5-point pragmatic-explanatory continuum, making explicit how closely a trial's conditions match the routine practice setting in which its results are meant to apply.", "long_description": "**What it is.** **PRECIS-2 (Pragmatic-Explanatory Continuum Indicator Summary 2)** is a structured\ndesign-and-appraisal tool that characterizes how *pragmatic* (oriented to usual-care decision-making\nacross diverse settings) or how *explanatory* (oriented to maximizing the chance of detecting a\nmechanistic effect under ideal conditions) a trial is, across **nine domains**: (1) **Eligibility** —\nhow closely participants resemble those who would receive the intervention in usual care; (2)\n**Recruitment** — how participants are identified and enrolled (routine appointments vs. dedicated\nrecruitment drives); (3) **Setting** — how closely the care setting matches usual practice; (4)\n**Organisation** — the expertise and resources delivering the intervention; (5) **Flexibility:\ndelivery** — how much latitude practitioners have in *how* the intervention is delivered; (6)\n**Flexibility: adherence** — how much latitude participants have, and what is done to enforce\nadherence; (7) **Follow-up** — how closely follow-up intensity matches usual care; (8) **Primary\noutcome** — how directly relevant the outcome is to participants; (9) **Primary analysis** — the\nextent to which all data are included (e.g., intention-to-treat) versus restricted (per-protocol).\nEach domain is rated 1 (very explanatory) to 5 (very pragmatic) and plotted on a nine-spoke wheel,\nforcing designers to justify each score against the trial's purpose. It is **not** a reporting\nchecklist and **not** a risk-of-bias instrument. PRECIS-2 was developed by Loudon, Treweek, Zwarenstein,\nand colleagues (Loudon et al., BMJ 2015), revising the original PRECIS (Thorpe et al., 2009); guidance,\nthe scoring wheel, and worked exemplars are maintained at the open www.precis-2.org toolkit.\n\n**When to use.** Use PRECIS-2 **prospectively at the design stage** to align each design choice with\nthe decision the trial is meant to inform: a regulatory efficacy trial supporting marketing\nauthorization sits toward the explanatory end, whereas a comparative-effectiveness or\nimplementation trial intended to inform HTA, payer, and routine-practice decisions should be\ndeliberately pragmatic on the domains that matter. It is equally useful **retrospectively** to\nappraise where a published trial sits and to judge whether its conditions support transporting the\nresult to a real-world target population. Decision rule for picking the right tool: PRECIS-2 answers\n*\"how applicable to usual care is this trial's design?\"* For *reporting* a pragmatic trial, use the\n**CONSORT extension for pragmatic trials** instead; for randomized-trial *risk of bias*, use **RoB 2**;\nfor *non-randomized/observational* RWE bias, use **ROBINS-I**. PRECIS-2 complements but does not\nreplace any of these. In RWE work it is most relevant to pragmatic randomized trials, hybrid\neffectiveness-implementation trials, and registry-based randomized trials (RRCTs) that randomize\nwithin a claims/EHR/registry infrastructure.\n\n**What it requires.** PRECIS-2 requires the design team to (a) **fix the intended applicability\nquestion first** — name the population, setting, and decision-maker the results must serve, because a\ndomain score is only meaningful relative to that purpose; (b) **score all nine domains independently**,\nwith a written rationale per domain, ideally by several raters to expose disagreement; (c) **plot the\nwheel** and inspect for unintended explanatory pull (e.g., a pragmatic eligibility criterion undercut\nby an explanatory, resource-intensive follow-up schedule that no real clinic would run); and (d)\n**iterate the protocol** so the design profile matches intent. In a registry/EHR-embedded trial the\ndomains map onto operational RWD decisions: *Eligibility* and *Recruitment* depend on the phenotype\nalgorithm and the database feasibility funnel; *Setting* and *Organisation* depend on the data\npartners; *Follow-up* depends on routinely captured encounters rather than study visits; *Primary\noutcome* depends on an outcome algorithm with known PPV/sensitivity; and *Primary analysis* depends on\nthe estimand (intention-to-treat vs. per-protocol) and how intercurrent events are handled.\n\n**When NOT to use — limitations and common misapplications.** (1) **It is not a quality or\nvalidity score.** A high pragmatic score is not \"better\" — pragmatism is a design *target*, not a\nvirtue; a high overall rating does not mean the trial is well conducted or low-risk-of-bias. Treating\nthe wheel as a quality grade is the most common abuse. (2) **It does not assess bias.** A pragmatic\ntrial can still be confounded, unblinded inappropriately, or underpowered; do not substitute PRECIS-2\nfor RoB 2 or ROBINS-I. (3) **It is not a reporting checklist** — completing the wheel does not satisfy\nCONSORT or its pragmatic extension. (4) **It applies to trials, not to purely observational\ndesigns.** PRECIS-2 presumes a randomized comparison; do not apply it to a single-arm or\nnon-randomized cohort study as if the domain scores conferred causal credibility. (5) **Scores are\njudgment-based and rater-dependent**; a single unblinded rater rationalizing a funded protocol\nproduces theater. (6) **It does not, by itself, establish transportability** — a pragmatic profile\nmakes generalization more plausible but does not replace formal effect-transport or external-validity\nanalysis to a named target population.\n\n**How it maps to this catalog.** When PRECIS-2 is used to design or appraise a registry/EHR-embedded\npragmatic or hybrid trial, the catalog concepts below implement the operational requirements behind\neach domain:\n- **Eligibility / Recruitment** → implemented by **diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe**\n (who the RWD identifies as eligible) and **database-feasibility-attrition-funnel-rwe** (how many\n survive each criterion); **active-comparator-new-user** disciplines the comparison when the pragmatic\n trial randomizes between two active treatments.\n- **Follow-up / attrition** → **attrition-and-loss-to-follow-up-rwe** quantifies how pragmatic,\n routine-care follow-up trades completeness for realism.\n- **Primary outcome** → **claims-outcome-algorithm-ppv-sensitivity-rwe** establishes whether the\n routinely captured outcome is measured well enough to support the pragmatic intent.\n- **Primary analysis** → **estimands-ate-att-intercurrent-events-rwe** and\n **estimand-analysis-traceability-rwe** formalize the intention-to-treat vs. per-protocol choice the\n Primary Analysis domain scores, and how intercurrent events (switching, discontinuation) are\n handled.\n- **Design backbone / transportability** → **target-trial-emulation** provides the protocol\n scaffold a registry trial mirrors, and **medicare-ffs-ma-commercial-claims-differences-rwe** flags\n when the data source's population limits the Setting/Eligibility pragmatism that can honestly be\n claimed; residual confounding in pragmatic comparisons can be probed with\n **empirical-calibration-negative-controls-rwe**.\n\n*Applied note (registry-based RCT in claims/EHR).* For a pragmatic RRCT embedded in linked\nclaims-EHR data, score Eligibility/Recruitment against the actual phenotype-derived screening funnel\n(not the protocol's aspirational criteria), score Follow-up against routinely captured encounters\nrather than study visits, and score Primary Outcome against the measured PPV/sensitivity of the\noutcome algorithm. A trial can advertise pragmatic eligibility yet collapse to an explanatory profile\nonce an outcome algorithm requires confirmatory testing only done at academic centers — the wheel\nmakes that contradiction visible before the protocol is locked.", "primary_category": "Guideline", "tags": [ "guideline", "trial-design", "pragmatic-trial", "applicability", "explanatory-pragmatic-continuum", "registry-based-rct" ], "aliases": [ "PRECIS-2", "PRECIS-2 tool", "Pragmatic-Explanatory Continuum Indicator Summary 2", "PRECIS", "pragmatic-explanatory continuum" ], "applies_to_study_types": [ "pragmatic_trial", "registry_trial" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "doi": "10.1136/bmj.h2147", "url": "https://doi.org/10.1136/bmj.h2147", "citation_text": "Loudon K, Treweek S, Sullivan F, Donnan P, Thorpe KE, Zwarenstein M. The PRECIS-2 tool: designing trials that are fit for purpose. BMJ. 2015;350:h2147.", "year": 2015, "authors_short": "Loudon et al.", "notes": "Canonical statement paper defining the nine PRECIS-2 domains, the 5-point scoring scale, and the pragmatic-explanatory wheel." }, { "role": "explain", "doi": "10.1016/j.jclinepi.2008.12.011", "url": "https://doi.org/10.1016/j.jclinepi.2008.12.011", "citation_text": "Thorpe KE, Zwarenstein M, Oxman AD, et al. A pragmatic-explanatory continuum indicator summary (PRECIS): a tool to help trial designers. Journal of Clinical Epidemiology. 2009;62(5):464-475.", "year": 2009, "authors_short": "Thorpe et al.", "notes": "Original PRECIS instrument that PRECIS-2 revised; explains the conceptual origin of the continuum and the design-purpose framing." }, { "role": "use", "url": "https://www.precis-2.org/", "citation_text": "PRECIS-2 toolkit — scoring wheel, domain guidance, worked exemplars, and rater resources (Loudon, Treweek, Zwarenstein, maintained).", "year": 2015, "authors_short": "PRECIS-2 toolkit", "notes": "Maintained open toolkit with the interactive wheel and applied scoring guidance." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "pragmatic-trial", "notes": "Primary use case — scoring and designing pragmatic randomized trials along the nine domains." }, { "relation_type": "applies_to", "target_slug": "registry-trial", "notes": "Applicable to registry-based and EHR-embedded randomized trials, where domain scores map onto RWD operational choices (phenotype, routine follow-up, outcome algorithm)." }, { "relation_type": "see_also", "target_slug": "target-trial-emulation", "notes": "Target-trial emulation supplies the protocol scaffold a registry RCT mirrors; PRECIS-2 then scores how pragmatic that design is." }, { "relation_type": "see_also", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "The Primary Analysis domain (ITT vs per-protocol, intercurrent-event handling) is formalized by the estimands framework." }, { "relation_type": "see_also", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Eligibility and Recruitment domains in RWD trials are operationalized by the phenotype algorithm that identifies eligible patients." }, { "relation_type": "see_also", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "The Follow-up domain trades realism against completeness; attrition handling quantifies the cost of pragmatic routine-care follow-up." }, { "relation_type": "see_also", "target_slug": "claims-outcome-algorithm-ppv-sensitivity-rwe", "notes": "The Primary Outcome domain depends on whether a routinely captured outcome is measured with adequate PPV and sensitivity." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "hta", "journal" ] }, { "slug": "prisma-2020", "name": "PRISMA 2020", "short_definition": "The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 statement — an EQUATOR-network reporting guideline (27-item checklist plus a four-phase flow diagram) that specifies the minimum set of items authors must report so a systematic review or meta-analysis can be appraised, reproduced, and trusted. It is a reporting standard, not a risk-of-bias tool, a quality score, or a method for conducting a review.", "long_description": "**What it is.** PRISMA 2020 (Preferred Reporting Items for Systematic Reviews and\nMeta-Analyses) is the flagship reporting guideline for systematic reviews, published\nby Page and colleagues in *BMJ* in 2021 as a major update to the original 2009\nstatement. It is maintained under the EQUATOR Network and curated by the PRISMA\nsteering group at prisma-statement.org. The statement is a **27-item checklist**\n(organised across Title, Abstract, Introduction, Methods, Results, Discussion, and\n\"Other information\" — funding, registration, data/code availability) plus an updated\n**four-phase flow diagram** (identification, screening, eligibility, inclusion) that\nseparately tracks records from databases/registers and from other sources, and\ndocuments records removed before screening. It defines what a review *report* must\ncontain — search strategy, eligibility criteria, study selection, data collection,\nrisk-of-bias assessment of included studies, synthesis methods (including any\nmeta-analysis and certainty-of-evidence rating such as GRADE), and results — so that\na reader can judge the review's conduct and reproduce its decisions. PRISMA 2020 is\nendorsed by hundreds of journals and is the expected reporting standard for evidence\nsynthesis submitted to peer review, HTA bodies, and regulators.\n\n**When to use.** Use PRISMA 2020 whenever you report a **systematic review** —\nincluding a meta-analysis of randomized trials or of observational/non-interventional\nstudies — for a peer-reviewed journal, an HTA/payer dossier (NICE, ICER, CADTH,\nG-BA), an FDA or EMA submission that relies on a structured evidence synthesis, or\nany deliverable where the review's conclusions must withstand external scrutiny.\nDecision rules for choosing the right instrument: (1) For the **protocol** of the\nreview (registered on PROSPERO or filed as a SAP), use the sibling **PRISMA-P**, not\nPRISMA 2020 — PRISMA 2020 governs the final report, PRISMA-P governs the plan. (2)\nFor a **scoping review**, use the **PRISMA-ScR** extension; for searching, use\n**PRISMA-S**; for abstracts, **PRISMA for Abstracts**; for living, network\nmeta-analysis, individual-participant-data, or harms-focused reviews, use the\ncorresponding PRISMA extension. (3) PRISMA governs the *synthesis*; it does not\ngovern the *primary studies* feeding it — a review of database (claims/EHR/registry)\nstudies is reported with PRISMA, while each included real-world study should itself\nhave followed STROBE/RECORD/RECORD-PE/HARPER.\n\n**What it requires.** The checklist enforces transparency across the review\nlifecycle, much of which carries directly into real-world-evidence synthesis:\npre-specified, reproducible **eligibility criteria** (Item 5) framed in PICO/PICOTS\nterms; a **full, reproducible search strategy** for every database with dates (Items\n6–7); explicit **selection and data-collection** processes with number of reviewers\nand automation tools (Items 8–9); pre-specified **outcomes and effect measures**\n(Items 10, 12); **risk-of-bias assessment of each included study** using a named\ntool (Item 11) — e.g., RoB 2 for trials, ROBINS-I for non-randomized intervention\nstudies — and a **certainty-of-evidence** rating (Item 15); a transparent\n**synthesis** description, including how studies were grouped, the meta-analytic\nmodel, heterogeneity handling, and **sensitivity analyses** (Items 13a–13f); the\n**flow diagram** documenting every record's fate (Item 16); and \"Other information\"\nitems demanding **protocol registration**, **funding/conflicts**, and **data,\ncode, and analytic-material availability** (Items 24–27). For an RWE meta-analysis\nthese map onto the same fitness-for-use, phenotype-transparency, and\nestimand-alignment concerns the included studies should have addressed — PRISMA\nmakes the synthesis author disclose them rather than letting heterogeneity in design,\ndata source, or time-zero definition disappear into a pooled estimate.\n\n**When NOT to use — limitations and common misapplications.** PRISMA 2020 is a\n**reporting checklist, not a risk-of-bias instrument and not a quality score**.\nCompleting all 27 items certifies that the report is *complete and transparent*; it\nsays nothing about whether the review was *well conducted* or whether the included\nevidence is *valid* — for that you need a critical-appraisal tool such as **AMSTAR 2**\nor **ROBIS** applied to the review, and **RoB 2 / ROBINS-I** applied to the primary\nstudies. Do not compute a \"PRISMA score\" by counting ticked boxes; the items are not\nweighted and adherence is not a quality metric. PRISMA does **not** make a synthesis\nof observational data causal — pooling confounded estimates yields a more precise\nconfounded estimate, and PRISMA only asks you to *report* that you assessed bias, not\nto remove it. Using PRISMA 2020 for a **review protocol** (PRISMA-P territory), a\n**scoping review** (PRISMA-ScR), or a **single primary study** (STROBE/RECORD-PE/\nHARPER) is the wrong-instrument error. Finally, beware *checklist-as-theater*:\nciting PRISMA and supplying a flow diagram while omitting the search strategy, the\nper-study risk-of-bias judgments, or the deviations from protocol defeats the\npurpose — reviewers increasingly request the completed item-level checklist with page\nreferences, not a blanket statement of adherence.\n\n**How it maps to this catalog.** When the included evidence is real-world data, the\nPRISMA-reported synthesis inherits the design discipline of its component studies,\nwhich this catalog implements concept-by-concept. Eligibility, comparator, and\ntime-zero alignment of the underlying studies are governed by\n**active-comparator-new-user** and **target-trial-emulation** — a meta-analysis\nshould not pool prevalent-user and new-user designs without flagging it.\nFitness-for-use and source-data adequacy of the included databases sit with\n**claims-analysis**. Outcome and exposure definitions feeding the synthesis are\nimplemented by **diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe** (and its\nPPV-validation logic, which PRISMA Item 11/13 forces you to surface). The\ncausal-contrast and intercurrent-event framing that determines whether pooled\neffects are even commensurable is **estimands-ate-att-intercurrent-events-rwe**.\nConfounding control in the included studies — whose adequacy the risk-of-bias item\nasks you to judge — is implemented by **high-dimensional-propensity-score-hdps-rwe**.\nAttrition and missing-data reporting (the flow diagram's RWE analogue) is\n**attrition-and-loss-to-follow-up-rwe**. *Applied note (claims/EHR/registry):* a\nmeta-analysis of database studies of the same drug comparison often mixes\nfee-for-service and Medicare-Advantage populations, divergent phenotype algorithms,\nand different washout/time-zero rules; PRISMA 2020 does not adjudicate these but it\nforces them into the open via the eligibility, risk-of-bias, and synthesis items, so\nheterogeneity is interrogated rather than averaged away.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "systematic-review", "meta-analysis", "equator", "evidence-synthesis" ], "aliases": [ "PRISMA 2020", "PRISMA-2020", "Preferred Reporting Items for Systematic Reviews and Meta-Analyses", "PRISMA statement" ], "applies_to_study_types": [ "systematic_review", "meta_analysis_rct", "meta_analysis_obs" ], "data_sources": [ "claims", "ehr", "registry", "linked", "primary" ], "citations": [ { "role": "introduce", "doi": "10.1136/bmj.n71", "url": "https://doi.org/10.1136/bmj.n71", "citation_text": "Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71.", "year": 2021, "authors_short": "Page et al.", "notes": "Canonical PRISMA 2020 statement — the 27-item checklist and the updated four-phase flow diagram that supersede the 2009 statement." }, { "role": "explain", "doi": "10.1136/bmj.n160", "url": "https://doi.org/10.1136/bmj.n160", "citation_text": "Page MJ, Moher D, Bossuyt PM, et al. PRISMA 2020 explanation and elaboration: updated guidance and exemplars for reporting systematic reviews. BMJ. 2021;372:n160.", "year": 2021, "authors_short": "Page et al.", "notes": "Item-by-item rationale, examples of complete reporting, and the methodological basis for each checklist element." }, { "role": "use", "url": "https://www.prisma-statement.org/", "citation_text": "PRISMA statement website (EQUATOR Network) — maintained checklists, flow-diagram generator, extensions (PRISMA-P, PRISMA-ScR, PRISMA-S, etc.), and translations.", "year": 2021, "authors_short": "PRISMA / EQUATOR Network", "notes": "Authoritative source for the fillable checklist, flow-diagram tools, and the full family of design-specific extensions." } ], "relations": [ { "relation_type": "see_also", "target_slug": "prisma-p", "notes": "PRISMA-P governs the review *protocol* and registration; PRISMA 2020 governs the final *report*. Use PRISMA-P first, then PRISMA 2020 at write-up." }, { "relation_type": "used_with", "target_slug": "claims-analysis", "notes": "When the included evidence is claims/EHR/registry data, PRISMA's eligibility and risk-of-bias items require disclosing data fitness-for-use, which this concept implements." }, { "relation_type": "used_with", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Outcome/exposure phenotype definitions and their PPV validation are what PRISMA Items 11 and 13 force a synthesis author to surface for each included RWE study." }, { "relation_type": "see_also", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Pooled effects are only commensurable if the included studies share an estimand; PRISMA's synthesis and risk-of-bias items make estimand heterogeneity visible." }, { "relation_type": "see_also", "target_slug": "active-comparator-new-user", "notes": "A defensible RWE meta-analysis should not pool prevalent-user and new-user designs without flagging it; this concept defines the design whose presence/absence PRISMA asks you to report." }, { "relation_type": "see_also", "target_slug": "target-trial-emulation", "notes": "Target-trial framing of the included observational studies underpins whether a PRISMA-reported synthesis of RWE supports a causal interpretation." }, { "relation_type": "complements", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "Adequacy of confounding control in included studies is exactly what PRISMA's risk-of-bias item (RoB 2 / ROBINS-I) asks you to judge; hdPS is a common implementation in the primary studies." }, { "relation_type": "complements", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "The PRISMA flow diagram's real-world analogue — documenting cohort attrition and loss to follow-up in the underlying database studies." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "hta", "journal" ] }, { "slug": "prisma-cosmin", "name": "PRISMA-COSMIN for OMIs", "short_definition": "A reporting guideline (PRISMA extension) specifying what must be reported in systematic reviews of outcome measurement instruments, including patient-reported outcome measures, across 54 sub-items.", "long_description": "**What it is** — **PRISMA-COSMIN for Outcome Measurement Instruments (OMIs) 2024** is a reporting\nguideline: a 34-item / 54-sub-item checklist that prescribes the *minimum content* a systematic review\nof measurement instruments must report. It is a content-specialized extension of the PRISMA 2020\nstatement, jointly developed and maintained by the COSMIN (COnsensus-based Standards for the selection\nof health Measurement INstruments) initiative and registered with the EQUATOR Network. \"OMIs\" includes\npatient-reported outcome measures (PROMs), clinician-reported, observer-reported, and performance-based\noutcome measures. The checklist organizes items around the parts of an OMI review that have no analogue\nin a treatment-effect review: which *construct* and *target population* the instrument measures, which\n*measurement properties* (content validity, structural validity, internal consistency, reliability,\nmeasurement error, construct validity, cross-cultural validity, responsiveness) were synthesized, how\nthe quality of each property's evidence was rated, and how a *recommendation for use* was derived. It is\na reporting standard only — it tells you what to write down, not how to judge an instrument.\n\n**When to use** — Use PRISMA-COSMIN when the unit of synthesis is an *instrument*, not a treatment\neffect: a systematic review whose question is \"which OMI best measures construct X in population Y, and\nis the evidence for its measurement properties adequate?\" This is the standard for the reporting section\nof any OMI/PROM systematic review submitted to a peer-reviewed journal, and for the instrument-selection\nevidence package that supports an HTA dossier, a core-outcome-set justification, or a regulatory\nfit-for-purpose argument (FDA PRO Guidance, EMA reflection papers) for a PRO endpoint. Decision rule for\npicking the right guideline: if you are *reviewing measurement instruments*, use PRISMA-COSMIN; if you\nare reviewing *clinical or treatment effects*, use base **PRISMA 2020** (or PRISMA-DTA for diagnostic\naccuracy); if you are *conducting* the OMI review's methodology and risk-of-bias assessment, that is the\nCOSMIN methodology (Prinsen 2018) and the COSMIN Risk of Bias checklist (Mokkink 2017), which\nPRISMA-COSMIN reports *on* but does not replace; if you are *designing or reporting a single primary PROM\nvalidation study*, use the COSMIN study-design checklist, not this reporting guideline.\n\n**What it requires** — The checklist enforces, among its domains: (1) an explicit eligibility frame\ndefining the construct, population, instrument type, and the measurement properties of interest;\n(2) transparent, reproducible search strategy across databases plus instrument-specific filters;\n(3) a study-selection and data-extraction process keyed to *measurement properties* rather than effect\nestimates; (4) reporting of the *content validity* assessment (often the most consequential and least\nreported property); (5) reporting of the COSMIN Risk of Bias rating per property per study; (6) the\nevidence-synthesis method per property and the GRADE-style rating of the *quality of evidence*; (7) the\nderivation of the categorized recommendation (A = recommended, B = potential, C = not recommended) for\neach instrument; and (8) reporting of feasibility, interpretability, and any conflicts of interest or\ndeveloper involvement. In RWE terms, the analogues of design transparency and fitness-for-use are the\nconstruct-to-population mapping and the content-validity chain; phenotype/algorithm validation maps onto\nmeasurement-property validation; and the \"estimand\" question becomes whether the instrument actually\nmeasures the intended construct in the intended use context.\n\n**When NOT to use — limitations and common misapplications** — PRISMA-COSMIN is a *reporting* checklist,\nnot a risk-of-bias instrument and not a quality score; completing all 54 sub-items documents\ntransparency but does not establish that any instrument is valid — that judgment comes from the COSMIN\nmethodology and Risk of Bias checklist. The single most common misapplication is conflating the three\nCOSMIN artifacts: PRISMA-COSMIN (what to report), the COSMIN methodology / Prinsen 2018 (how to conduct\nand synthesize the review), and the COSMIN Risk of Bias checklist / Mokkink 2017 (how to appraise each\nprimary study). Other failure modes: using base PRISMA 2020 for an OMI review (it has no fields for\nmeasurement properties, content validity, or per-instrument recommendations); using PRISMA-COSMIN to\nreport a review of *treatment effects that happen to use a PRO endpoint* (that is a PRISMA 2020 review);\napplying it to a single primary validation study (use the COSMIN study-design checklist); and\nchecklist-as-theater — pasting page numbers against items without the underlying content-validity and\nevidence-grading work actually being done. It also does not adjudicate which instrument to adopt; it\ngoverns the reporting of the evidence that informs that choice.\n\n**How it maps to this catalog** — The implementing concepts in this repository are the\nmeasurement/validation family, not the causal-design family. Instrument selection and the construct it\nmeasures connect to **pro-rwe** (using PROs as endpoints in real-world data) and **pro-development**\n(how an OMI is built). The measurement-property evidence that PRISMA-COSMIN requires you to report is\ngenerated and appraised in **pro-validation**, with the broader validation logic shared by\n**algorithm-validation** and **surrogate-endpoint-validation-rwe**. The eligibility-and-extraction\ndiscipline parallels **study-protocol-or-sap-elements**. Reading guide: report the construct/population\nframe via *pro-rwe* and *pro-development*; report the measurement-property and content-validity evidence\nvia *pro-validation*; situate instrument validity within the same evidentiary stance as\n*algorithm-validation* and *surrogate-endpoint-validation-rwe*.\n\n**Applied note (claims/EHR/registry RWE).** When a real-world study needs a PRO endpoint — e.g.,\nselecting PROMIS, EQ-5D, or a disease-specific PROM to be collected in a registry or linked to claims/EHR\n— a PRISMA-COSMIN–compliant systematic review is the defensible mechanism for choosing a\nfit-for-purpose instrument *before* the protocol locks the endpoint. It documents that the chosen OMI has\nadequate content validity and measurement properties in the actual target population (e.g., older,\nmultimorbid Medicare enrollees, not a clinical-trial sample), which is exactly the fitness-for-use\nargument an HTA body or regulator will probe when a PRO drives the value or label claim.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "prisma-extension", "cosmin", "outcome-measurement-instruments", "patient-reported-outcomes", "systematic-review" ], "aliases": [ "PRISMA-COSMIN", "PRISMA-COSMIN for OMIs", "PRISMA-COSMIN for OMIs 2024", "PRISMA extension for systematic reviews of outcome measurement instruments" ], "applies_to_study_types": [ "pro_development", "pro_validation", "systematic_review" ], "data_sources": [ "registry", "ehr", "linked" ], "citations": [ { "role": "introduce", "doi": "10.1007/s11136-024-03634-y", "url": "https://doi.org/10.1007/s11136-024-03634-y", "citation_text": "Elsman EBM, Mokkink LB, Terwee CB, et al. Guideline for reporting systematic reviews of outcome measurement instruments (OMIs): PRISMA-COSMIN for OMIs 2024. Quality of Life Research. 2024;33(8):2029-2046.", "year": 2024, "authors_short": "Elsman et al.", "notes": "Canonical PRISMA-COSMIN statement paper; defines the 34-item / 54-sub-item reporting checklist as a PRISMA 2020 extension." }, { "role": "explain", "doi": "10.1007/s11136-018-1798-3", "url": "https://doi.org/10.1007/s11136-018-1798-3", "citation_text": "Prinsen CAC, Mokkink LB, Bouter LM, et al. COSMIN guideline for systematic reviews of patient-reported outcome measures. Quality of Life Research. 2018;27(5):1147-1157.", "year": 2018, "authors_short": "Prinsen et al.", "notes": "Underlying COSMIN methodology for conducting and synthesizing OMI reviews; PRISMA-COSMIN reports on this process but does not replace it." }, { "role": "use", "doi": "10.1007/s11136-017-1765-4", "url": "https://doi.org/10.1007/s11136-017-1765-4", "citation_text": "Mokkink LB, de Vet HCW, Prinsen CAC, et al. COSMIN Risk of Bias checklist for systematic reviews of Patient-Reported Outcome Measures. Quality of Life Research. 2018;27(5):1171-1179 (online 2017).", "year": 2017, "authors_short": "Mokkink et al.", "notes": "The companion risk-of-bias appraisal tool for primary measurement-property studies; distinct from PRISMA-COSMIN, which is reporting-only." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "systematic-review", "notes": "The guideline governs reporting of systematic reviews whose unit of synthesis is an outcome measurement instrument." }, { "relation_type": "applies_to", "target_slug": "pro-validation", "notes": "Reviews synthesize the measurement-property (validation) evidence for each candidate instrument." }, { "relation_type": "applies_to", "target_slug": "pro-development", "notes": "Construct definition and instrument provenance from development inform eligibility and content-validity reporting." }, { "relation_type": "see_also", "target_slug": "pro-rwe", "notes": "PRISMA-COSMIN reviews are how a fit-for-purpose PRO endpoint is selected before a real-world study locks its outcome." }, { "relation_type": "see_also", "target_slug": "pro-validation", "notes": "Implements the measurement-property and content-validity evidence that the checklist requires to be reported per instrument." }, { "relation_type": "see_also", "target_slug": "pro-development", "notes": "Source of the construct-to-population mapping that anchors a PRISMA-COSMIN eligibility frame." }, { "relation_type": "see_also", "target_slug": "surrogate-endpoint-validation-rwe", "notes": "Shares the validation stance — establishing that a measured quantity faithfully reflects the intended construct or clinical benefit." }, { "relation_type": "see_also", "target_slug": "algorithm-validation", "notes": "Parallel logic — validating that a measurement (instrument or algorithm) captures the intended target in the intended population." }, { "relation_type": "see_also", "target_slug": "study-protocol-or-sap-elements", "notes": "The pre-specified eligibility, search, and extraction discipline mirrors protocol/SAP pre-specification for primary studies." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "hta", "journal" ] }, { "slug": "prisma-dta", "name": "PRISMA-DTA", "short_definition": "The PRISMA reporting extension for systematic reviews and meta-analyses of diagnostic test accuracy (DTA) studies — a 27-item checklist (with 2-item abstract extension) that specifies what must be reported about the review question, index test(s), reference standard, eligibility, search, risk-of-bias appraisal, and accuracy synthesis.", "long_description": "**What it is.** PRISMA-DTA (Preferred Reporting Items for a Systematic Review and Meta-analysis of\nDiagnostic Test Accuracy Studies) is the official PRISMA extension for *systematic reviews of\ndiagnostic test accuracy*, published as the McInnes et al. statement (JAMA, 2018) with a companion\nexplanation-and-elaboration paper (Salameh et al., BMJ, 2020). It is maintained under the EQUATOR\nNetwork and the PRISMA family alongside the parent PRISMA 2020 statement. The checklist contains 27\nitems adapted from PRISMA to the structure of accuracy reviews, plus a separate PRISMA-DTA for\nAbstracts (PRISMA-DTA-A) extension of 11 items for abstract reporting. Its purpose is *reporting\ntransparency*, not study conduct: it tells authors, peer reviewers, editors, and HTA assessors what\nfacts a DTA review must disclose so that a reader can judge how sensitivity, specificity, and other\naccuracy estimates were derived and how trustworthy they are. The defining feature versus generic\nPRISMA is the two-test structure of every accuracy question: an **index test** compared against a\n**reference standard** (\"gold standard\"), with paired estimates (sensitivity/specificity, or PPV/NPV\nat a defined prevalence), threshold effects, and the hierarchical/bivariate meta-analytic models\nthose data require.\n\n**When to use.** Use PRISMA-DTA whenever the report is a *systematic review or meta-analysis whose\nprimary objective is the accuracy of one or more tests* — diagnostic, screening, staging, prognostic\nclassification, or triage — against a reference standard. This is the correct extension for journal\nsubmission of a DTA review, for the evidence-synthesis section of an HTA/payer dossier on a\ndiagnostic or companion-diagnostic technology (e.g., NICE Diagnostics Assessment Programme), and for\nthe diagnostic-evidence component of an FDA or EMA submission that relies on a synthesis of accuracy\nstudies. Decision rules for picking PRISMA-DTA over a sibling guideline: (1) if the review estimates\n**test accuracy** (sensitivity/specificity, ROC/AUC, likelihood ratios) → PRISMA-DTA; (2) if the\nreview estimates **intervention or exposure effects** (risk ratios, hazard ratios) → parent PRISMA\n2020, with RECORD/RECORD-PE conventions for the underlying routinely-collected-data studies; (3) if\nyou are writing the **protocol** for the review rather than the completed review → PRISMA-P; (4) for\nthe *primary* DTA study being included, the relevant reporting guideline is STARD 2015, not\nPRISMA-DTA, and the within-study risk of bias is appraised with QUADAS-2. PRISMA-DTA governs the\nreview layer; STARD governs the primary-study layer; QUADAS-2 is the appraisal instrument the review\nmust apply and report.\n\n**What it requires.** The checklist enforces complete reporting of the substantive domains that make\nan accuracy synthesis interpretable: a structured title and abstract identifying it as a DTA review;\nan explicit review question framed by population, index test(s), comparator test(s) where relevant,\nreference standard(s), and target condition; pre-registration (e.g., PROSPERO) and any protocol\ndeviations; eligibility criteria including the *definition of the target condition and the reference\nstandard* used to confirm it; a reproducible search strategy; data items capturing test thresholds,\nreference-standard definition, and 2×2 counts (TP/FP/FN/TN); **risk-of-bias and applicability\nassessment using QUADAS-2** across the patient-selection, index-test, reference-standard, and\nflow-and-timing domains; the synthesis methods (bivariate/HSROC hierarchical models, handling of\nthreshold variability and partial/differential verification); results presented as paired accuracy\nestimates with confidence/credible regions and SROC plots; investigation of heterogeneity and\npublication/selective-reporting bias; and a discussion of applicability to the intended clinical\npathway and prevalence setting. For real-world-data DTA reviews — where the \"index test\" is often a\n*claims- or EHR-derived phenotype or computable algorithm* benchmarked against chart review or a\nvalidated reference — these same items map onto data-fitness-for-use, algorithm/phenotype\ndefinition and validation, verification completeness (the analogue of attrition), and sensitivity\nanalyses around the case definition.\n\n**When NOT to use — limitations and common misapplications.** PRISMA-DTA is a *reporting* checklist:\nit certifies that information was disclosed, not that the review is unbiased or high quality.\nCompleting all 27 items does not make a flawed synthesis valid, does not substitute for QUADAS-2\nappraisal of the included studies, and is not a quality score — there is no total to tally, and\nticking items must never be treated as a methodological pass/fail. Specific failure modes: (1)\napplying PRISMA-DTA to a review of *intervention effects* or an etiologic question — use parent\nPRISMA 2020 instead; (2) using generic PRISMA for an accuracy review, which omits the index/reference\nstructure, threshold reporting, and hierarchical-model items reviewers will demand; (3) confusing it\nwith STARD (a *primary-study* DTA guideline) or with QUADAS-2 (the *risk-of-bias instrument*); (4)\n\"checklist-as-theater,\" where authors cite PRISMA-DTA and append a checklist but report no 2×2 data,\nno reference-standard definition, and no risk-of-bias assessment, leaving the accuracy estimates\nun-auditable; (5) in RWD validation reviews, reporting algorithm PPV alone without the verification\nscheme and spectrum, which makes pooled accuracy uninterpretable. The checklist also does not address\n*clinical-utility* or decision-impact questions, which need different frameworks.\n\n**How it maps to this catalog.** PRISMA-DTA is the reporting wrapper around DTA-relevant concepts in\nthis repo. The accuracy-estimation core is implemented by **diagnostic-accuracy** (sensitivity,\nspecificity, ROC/AUC, bivariate/HSROC synthesis) and, for routinely-collected-data benchmarking, by\n**claims-outcome-algorithm-ppv-sensitivity-rwe** and\n**diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe** (defining and validating the computable\ncase definition that plays the role of the \"index test\"). The reference-standard and target-condition\ndefinition items connect to **procedure-identification-and-measurement-in-claims-ehr**. Verification\ncompleteness and differential loss of patients between index and reference standard map to\n**attrition-and-loss-to-follow-up-rwe** and **database-feasibility-attrition-funnel-rwe**. Estimand\nclarity for what the accuracy contrast actually targets connects to\n**estimands-ate-att-intercurrent-events-rwe** and **estimand-analysis-traceability-rwe**; the\nevidence-synthesis machinery aligns with **systematic-review**, **meta-analysis-obs**, and\n**network-meta-analysis** (for comparative/network DTA). Cross-database accuracy differences are\nsurfaced by **medicare-ffs-ma-commercial-claims-differences-rwe**. Applied note for claims/EHR/registry\nRWE: when a DTA \"review\" is in fact a synthesis of *algorithm-validation* studies, report each study's\nreference standard (chart review vs adjudication), the verification fraction and how unverified\nrecords were handled, the operating threshold (e.g., 1 inpatient OR 2 outpatient codes within a time\nwindow), and prevalence-dependent measures (PPV/NPV) separately from prevalence-independent ones\n(sensitivity/specificity) — exactly the items PRISMA-DTA forces into the open.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "prisma", "diagnostic-test-accuracy", "equator", "systematic-review", "meta-analysis" ], "aliases": [ "PRISMA-DTA", "PRISMA for Diagnostic Test Accuracy", "Preferred Reporting Items for a Systematic Review and Meta-analysis of Diagnostic Test Accuracy Studies", "PRISMA-DTA-A" ], "applies_to_study_types": [ "diagnostic_accuracy", "systematic_review", "meta_analysis_obs" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "doi": "10.1001/jama.2017.19163", "url": "https://doi.org/10.1001/jama.2017.19163", "citation_text": "McInnes MDF, Moher D, Thombs BD, et al. Preferred Reporting Items for a Systematic Review and Meta-analysis of Diagnostic Test Accuracy Studies: The PRISMA-DTA Statement. JAMA. 2018;319(4):388-396.", "year": 2018, "authors_short": "McInnes et al.", "notes": "Canonical PRISMA-DTA statement and 27-item checklist (with the PRISMA-DTA for Abstracts extension)." }, { "role": "explain", "doi": "10.1136/bmj.m2632", "url": "https://doi.org/10.1136/bmj.m2632", "citation_text": "Salameh JP, Bossuyt PM, McGrath TA, et al. Preferred reporting items for systematic review and meta-analysis of diagnostic test accuracy studies (PRISMA-DTA): explanation, elaboration, and checklist. BMJ. 2020;370:m2632.", "year": 2020, "authors_short": "Salameh et al.", "notes": "Item-by-item explanation and elaboration; the authoritative guide to what each PRISMA-DTA item requires and why." }, { "role": "explain", "doi": "10.7326/0003-4819-155-8-201110180-00009", "url": "https://doi.org/10.7326/0003-4819-155-8-201110180-00009", "citation_text": "Whiting PF, Rutjes AWS, Westwood ME, et al. QUADAS-2: A Revised Tool for the Quality Assessment of Diagnostic Accuracy Studies. Annals of Internal Medicine. 2011;155(8):529-536.", "year": 2011, "authors_short": "Whiting et al.", "notes": "The risk-of-bias and applicability instrument PRISMA-DTA requires reviewers to apply and report; distinct from the reporting checklist itself." }, { "role": "use", "doi": "10.1136/bmj.n71", "url": "https://doi.org/10.1136/bmj.n71", "citation_text": "Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71.", "year": 2021, "authors_short": "Page et al.", "notes": "Parent PRISMA statement; use this (not PRISMA-DTA) for reviews of intervention or exposure effects." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "diagnostic-accuracy", "notes": "Primary scope — reporting standard for systematic reviews and meta-analyses of test accuracy." }, { "relation_type": "applies_to", "target_slug": "systematic-review", "notes": "Governs the review layer; pair with PRISMA-P for the protocol and STARD for included primary studies." }, { "relation_type": "used_with", "target_slug": "diagnostic-accuracy", "notes": "Implements the accuracy estimands (sensitivity/specificity, ROC/AUC) and bivariate/HSROC synthesis that PRISMA-DTA requires reporting." }, { "relation_type": "used_with", "target_slug": "claims-outcome-algorithm-ppv-sensitivity-rwe", "notes": "For RWD DTA reviews, the algorithm validation metrics (PPV, sensitivity) are the per-study accuracy data PRISMA-DTA structures." }, { "relation_type": "see_also", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "The computable phenotype/algorithm plays the role of the index test benchmarked against a reference standard." }, { "relation_type": "see_also", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Verification completeness and differential verification map onto attrition/flow-and-timing domains of QUADAS-2." }, { "relation_type": "see_also", "target_slug": "systematic-review", "notes": "PRISMA-DTA is the DTA-specific reporting layer on top of general systematic-review conduct." }, { "relation_type": "see_also", "target_slug": "meta-analysis-obs", "notes": "Hierarchical accuracy synthesis is an observational meta-analysis; the synthesis-reporting items apply." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "hta", "journal" ] }, { "slug": "prisma-equity", "name": "PRISMA-Equity 2012 Extension", "short_definition": "PRISMA reporting extension for systematic reviews and meta-analyses whose a priori question is the distribution of intervention effects across socially stratifying (equity) factors, requiring authors to report how equity was defined, searched for, analyzed, and interpreted.", "long_description": "**What it is.** PRISMA-Equity 2012 is the official PRISMA extension for **equity-focused systematic reviews** — reviews\nwhose primary question is not only \"does the intervention work?\" but \"for whom, and does it widen or narrow health\ninequities?\" It was developed by the Campbell and Cochrane Equity Methods Group (Welch, Petticrew, Tugwell and colleagues)\nthrough a Bellagio consensus process and is hosted by the EQUATOR Network alongside the PRISMA family. It supplements the\nPRISMA checklist with equity-specific reporting items and is organized around the **PROGRESS-Plus** framework of\nsocially stratifying factors: Place of residence, Race/ethnicity/culture/language, Occupation, Gender/sex, Religion,\nEducation, Socioeconomic status, Social capital — plus context-relevant additions such as age, disability, sexual\norientation, and other circumstances tied to discrimination or vulnerability. PRISMA-Equity is a **reporting guideline**:\nit tells authors what to disclose so a reader (or HTA committee) can judge whether the review's equity claims are\ntrustworthy. It is not a search strategy, a risk-of-bias tool, or a quality score.\n\n**When to use.** Use PRISMA-Equity when health equity is an **a priori focus** of a systematic review or meta-analysis —\nthe protocol asks whether effects differ across PROGRESS-Plus strata, or whether the intervention is differentially\neffective/accessible in disadvantaged populations. Typical decision contexts: a peer-reviewed equity review or\nhealth-disparities synthesis; an HTA/payer evidence package where the appraisal body (e.g., NICE health-inequalities\nguidance, CADTH equity considerations, ICER's contextual considerations) asks how the evidence base bears on\ndisadvantaged groups; a guideline-development synthesis with an explicit equity remit. **Decision rule for which PRISMA\ndocument applies:** use base **PRISMA 2020** for any systematic review; layer **PRISMA-Equity** on top only when equity\nis a stated review objective. For a *protocol* of such a review, register with **PRISMA-P** and flag the equity aim\nthere. PRISMA-Equity sits alongside — not instead of — PRISMA 2020; it adds equity items to, and does not replace, the\nbase checklist.\n\n**What it requires.** PRISMA-Equity enforces equity-specific transparency at each review stage: (1) **frame the equity\nquestion** — name the PROGRESS-Plus factors of interest and the populations expected to be disadvantaged, with the\nlogic model linking the intervention to the equity outcome; (2) **eligibility and search** — report whether study\nselection and search terms were designed to capture disadvantaged populations and equity-relevant settings (e.g.,\nlow-/middle-income settings, grey literature); (3) **data items** — pre-specify which PROGRESS-Plus stratifiers were\nextracted and how subgroup/effect-modification analyses were planned; (4) **synthesis** — report effects *within and\nacross* equity strata, gradients (not just present/absent), and whether differential effects were tested rather than\nassumed; (5) **interpretation** — discuss how findings affect the equity gap and the applicability of evidence to\ndisadvantaged groups, including the limits imposed by under-representation. Because the *included* studies are\nfrequently real-world observational designs drawn from claims, EHR, or registry data, PRISMA-Equity reporting forces a\nreviewer to confront whether equity variables were even measurable in the source data — race/ethnicity, language,\nincome, and place are commonly absent, incomplete, or differentially missing in administrative data, and the review\nmust say so rather than silently dropping strata.\n\n**When NOT to use — limitations and common misapplications.**\n- **It is a reporting checklist, not a risk-of-bias or quality instrument.** A complete PRISMA-Equity checklist tells\n you the review *disclosed* its equity methods; it does not certify that those methods were sound. Critical appraisal\n of the included studies still requires ROBINS-I / RoB 2 (and equity-specific appraisal of differential bias), and\n confidence in the synthesis still requires GRADE. Treating a filled checklist as evidence of quality is the\n checklist-as-theater failure mode.\n- **Completing it does not make differential effects causal.** Reporting a subgroup gradient across PROGRESS-Plus\n strata is descriptive transparency; whether the gradient reflects effect modification rather than confounding or\n differential measurement is a separate analytic question.\n- **Wrong scope.** Do not bolt PRISMA-Equity onto a review that was never designed around equity, merely to appear\n thorough — the extension presumes equity is an a priori objective, and retrofitting it produces post-hoc subgroup\n fishing dressed up as equity reporting. Conversely, do not use base PRISMA-only reporting when equity *was* the\n review's purpose.\n- **Version mismatch.** PRISMA-Equity 2012 was written against PRISMA 2009; when used today it must be paired with the\n current **PRISMA 2020** base checklist (item numbering differs), not the obsolete 2009 wording.\n- **Not a substitute for equity reporting in primary studies.** For a single observational RWE study, equity reporting\n belongs in STROBE/RECORD plus subgroup pre-specification, not in a systematic-review extension.\n\n**How it maps to this catalog.** PRISMA-Equity governs *reporting* of an equity-focused synthesis; the substantive RWE\nconcepts that implement each requirement live elsewhere in this catalog. The PROGRESS-Plus stratifiers are operationally\nthe social determinants captured (or not) in real-world data — see `sdoh-social-determinants-of-health` for how Place,\nrace/ethnicity, and SES are measured and where they are missing in claims/EHR. The \"for whom does the evidence apply\"\nquestion is generalizability/applicability — see `generalizability-transportability-external-validity-rwe` and\n`special-populations-rwe-methods`. Differential inclusion of disadvantaged groups across primary studies is a selection\nproblem — see `selection-bias-sensitivity-analysis-rwe`. Because equity variables are systematically incomplete in\nadministrative data, the honest reporting of which strata could be analyzed depends on `missing-data-pattern-table-rwe`.\nThe review's population and equity strata are specified through the P (and the equity lens on every element) of\n`picots-framework-rwe`. The reviews themselves are `systematic-review` and `meta-analysis-obs`. **Applied note for\nclaims/EHR/registry RWE:** when the included evidence is administrative-data RWE, PRISMA-Equity's value is forcing the\nreviewer to state plainly that race/ethnicity may be imputed, that income and education are usually proxied by area-level\nmeasures, and that disadvantaged populations are often under-enrolled or differentially censored — so an apparent \"no\ndifference across strata\" may reflect measurement and representation gaps rather than equitable effects.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "health-equity", "progress-plus", "systematic-review", "prisma-extension" ], "aliases": [ "PRISMA-E", "PRISMA-Equity", "PRISMA Equity Extension", "PRISMA-Equity 2012" ], "applies_to_study_types": [ "systematic_review", "meta_analysis_obs" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "doi": "10.1371/journal.pmed.1001333", "url": "https://doi.org/10.1371/journal.pmed.1001333", "citation_text": "Welch V, Petticrew M, Tugwell P, et al. PRISMA-Equity 2012 extension: reporting guidelines for systematic reviews with a focus on health equity. PLoS Medicine. 2012;9(10):e1001333.", "year": 2012, "authors_short": "Welch et al.", "notes": "Canonical statement paper introducing the equity-focused PRISMA extension and its reporting items." }, { "role": "explain", "doi": "10.1016/j.jclinepi.2013.08.005", "url": "https://doi.org/10.1016/j.jclinepi.2013.08.005", "citation_text": "O'Neill J, Tabish H, Welch V, et al. Applying an equity lens to interventions: using PROGRESS ensures consideration of socially stratifying factors to illuminate inequities in health. Journal of Clinical Epidemiology. 2014;67(1):56-64.", "year": 2014, "authors_short": "O'Neill et al.", "notes": "Defines and operationalizes the PROGRESS-Plus framework of socially stratifying factors that PRISMA-Equity uses." }, { "role": "explain", "doi": "10.1136/bmj.n71", "url": "https://doi.org/10.1136/bmj.n71", "citation_text": "Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71.", "year": 2021, "authors_short": "Page et al.", "notes": "Current base PRISMA checklist; PRISMA-Equity (written against PRISMA 2009) must be layered on PRISMA 2020 today." }, { "role": "use", "url": "https://www.equator-network.org/reporting-guidelines/prisma-equity/", "citation_text": "PRISMA-Equity 2012 Extension — EQUATOR Network reporting-guideline page (maintained checklist and resources).", "year": 2012, "authors_short": "EQUATOR Network", "notes": "Stable maintained landing page with the checklist and supporting materials." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "systematic-review", "notes": "Use when the systematic review's a priori question concerns the distribution of effects across equity strata." }, { "relation_type": "applies_to", "target_slug": "meta-analysis-obs", "notes": "Use when synthesizing observational evidence with an explicit health-equity objective." }, { "relation_type": "used_with", "target_slug": "sdoh-social-determinants-of-health", "notes": "PROGRESS-Plus stratifiers are the social determinants captured (or missing) in real-world data; this concept implements their measurement." }, { "relation_type": "used_with", "target_slug": "picots-framework-rwe", "notes": "The equity lens is applied across PICOTS, especially Population, where the disadvantaged strata are specified." }, { "relation_type": "see_also", "target_slug": "generalizability-transportability-external-validity-rwe", "notes": "Equity reporting is fundamentally about applicability: to whom does the synthesized evidence transport?" }, { "relation_type": "see_also", "target_slug": "special-populations-rwe-methods", "notes": "Disadvantaged and under-represented populations central to equity reviews are addressed methodologically here." }, { "relation_type": "see_also", "target_slug": "selection-bias-sensitivity-analysis-rwe", "notes": "Differential inclusion or censoring of disadvantaged groups across primary studies is a selection-bias problem." }, { "relation_type": "see_also", "target_slug": "missing-data-pattern-table-rwe", "notes": "Equity variables (race/ethnicity, income, education) are systematically incomplete in administrative data; honest stratum-level reporting depends on documenting that missingness." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "journal", "hta" ] }, { "slug": "prisma-ipd", "name": "PRISMA-IPD", "short_definition": "Reporting guideline (a PRISMA 2020 extension) for systematic reviews and meta-analyses that collect, check, and re-analyze the individual participant data (raw line-level records) from each included study, rather than only published aggregate results.", "long_description": "**What it is.** PRISMA-IPD (Preferred Reporting Items for a Systematic Review and Meta-Analysis\nof Individual Participant Data) is the reporting guideline for the specific class of evidence\nsynthesis in which the reviewers obtain the *individual participant data* — the raw, line-level\nrecords for each randomized or enrolled subject — from the original investigators, harmonize and\nre-check those data, and analyze them centrally, instead of extracting summary effect estimates\nfrom publications. Published by Stewart and colleagues for the IPD Meta-analysis Methods Group as\nthe **PRISMA-IPD Statement** (JAMA, 2015), it extends the parent PRISMA checklist with items unique\nto the IPD process: documenting which studies provided data and which did not, how the supplied data\nwere integrity-checked, how variables were harmonized across datasets, and how the meta-analysis\nmodeled within- and between-study structure. It is maintained within the EQUATOR Network /\nPRISMA family alongside the parent PRISMA 2020 statement and the protocol extension PRISMA-P; the\nchecklist and flow diagram are hosted at prisma-statement.org. PRISMA-IPD is a *reporting* standard —\nit governs the transparency and completeness of what authors write, not the methodological quality\nof the review itself.\n\n**When to use.** Use PRISMA-IPD whenever the synthesis is built on individual participant data —\nwhether the underlying studies are randomized trials, observational cohorts, or a mix — and the\ndeliverable is a peer-reviewed manuscript, an HTA/payer evidence dossier, a regulatory (FDA/EMA)\nsubmission relying on pooled patient-level evidence, or a registered protocol. Decision rule for\nchoosing within the PRISMA family: if you are pooling *published aggregate results*, report under\nparent **PRISMA 2020**; if you are writing the *protocol* for any systematic review, use **PRISMA-P**;\nif you are obtaining and re-analyzing *patient-level records*, PRISMA-IPD is the correct extension and\nPRISMA 2020 alone is insufficient because it has no items for data solicitation, integrity checking,\nor harmonization. For network or diagnostic-accuracy IPD syntheses, layer the relevant topic\nextension (PRISMA-NMA, PRISMA-DTA) on top of the IPD items. In RWE, PRISMA-IPD increasingly applies\nto federated or pooled analyses of claims, EHR, and registry cohorts where partner sites contribute\npatient-level or common-data-model (e.g., OMOP) extracts to a central analysis.\n\n**What it requires.** Beyond the standard PRISMA reporting domains (eligibility criteria, search\nstrategy, selection process, synthesis methods, results, certainty of evidence), PRISMA-IPD enforces\nIPD-specific items that map directly onto RWD fitness-for-use concerns: (1) **which data were sought\nand obtained** — an explicit accounting of studies/sites approached, those that provided IPD, those\nthat did not, and the participants represented versus missing, so the reader can judge availability\nbias; (2) **data integrity and harmonization** — how supplied datasets were checked (range/logic\nchecks, randomization integrity, baseline balance, follow-up completeness) and how heterogeneous\nvariable definitions, coding systems, and outcome ascertainment were reconciled into a common\nanalysis-ready structure; (3) **the analytic model** — whether a one-stage (pooled with study as a\nfactor/random effect) or two-stage (study-specific estimates then meta-analyzed) approach was used,\nhow clustering by study was preserved, and how heterogeneity and subgroup/interaction effects were\nmodeled; (4) **risk of bias at the study and outcome level**, assessed using the line-level data; and\n(5) **handling of missing data and participants** across contributing datasets. For real-world data\nthese items force exactly the disclosures regulators and HTA bodies demand: transparent design,\ndata-source fitness-for-use, phenotype/outcome-algorithm definitions reconciled across sites,\ntime-zero/cohort-entry alignment, the target estimand and its handling of intercurrent events, and\nmissing-data/attrition accounting.\n\n**When NOT to use — limitations and common misapplications.** PRISMA-IPD is a reporting checklist, so\nthe most damaging misuses confuse reporting completeness with scientific validity. (a) It is **not a\nrisk-of-bias instrument and not a quality score** — completing every PRISMA-IPD item does not certify\nthe review is unbiased; risk of bias still requires a dedicated tool (e.g., RoB 2 for trials,\nROBINS-I for non-randomized studies, ROBIS for the review). (b) **A high checklist score does not make\npooled observational data causal** — central re-analysis of patient-level claims/EHR records inherits\nconfounding, selection, and misclassification that no reporting standard removes; the estimand and\nconfounding strategy must be argued on their own merits. (c) **Wrong member of the family**: using\nparent PRISMA 2020 (or, worse, a non-IPD checklist) for an IPD synthesis hides the availability,\nintegrity, and harmonization decisions that determine whether the pooled result is trustworthy; using\nPRISMA-IPD for an aggregate-data review imports items the study never engaged. (d) **Checklist-as-theater**:\nappending a filled grid at submission without the methods text actually describing data solicitation,\nintegrity checks, and the one-/two-stage choice satisfies a journal field but informs no reader. (e) It\ndoes not replace protocol registration (use PRISMA-P), nor does it adjudicate certainty of evidence\n(use GRADE).\n\n**How it maps to this catalog.** PRISMA-IPD's reporting items are *implemented* by concrete RWE methods\nin this repository: the data-availability/fitness items map to **fit-for-purpose-data-assessment-rwe**,\n**database-feasibility-attrition-funnel-rwe**, and **multi-database** federated analysis; harmonized\noutcome/exposure definitions map to **diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe** and\n**outcome-algorithm-construction-rwe** (with OMOP common-data-model patterns via **omop-cdm-method-patterns-rwe**);\ncohort-entry/time-zero alignment across contributing studies maps to **time-zero-index-date-alignment-rwe**;\nthe estimand and intercurrent-event reporting maps to **estimands-ate-att-intercurrent-events-rwe**; the\npooled-cohort confounding-control strategy that the reanalysis must still defend maps to\n**active-comparator-new-user**, **high-dimensional-propensity-score-hdps-rwe**, and\n**target-trial-emulation**; missing-data and participant-flow items map to\n**attrition-and-loss-to-follow-up-rwe** and **missing-data-pattern-table-rwe**; and the one-stage vs\ntwo-stage modeling choice and study-level pooling map to **ipd-meta-analysis**, **mixed-effects-models-longitudinal-rwe**,\nand **meta-analysis-rct** / **meta-analysis-obs**. **Applied note for claims/EHR/registry RWE:** in a\nfederated pooled analysis where partner sites map data to OMOP and return patient-level (or privacy-preserving\nsite-level) extracts, PRISMA-IPD is the natural reporting backbone — report the sites solicited versus\ncontributing (availability bias), the integrity/logic checks run on each site's extract, how phenotype\nalgorithms and time-zero were harmonized across data sources of differing completeness (e.g., Medicare\nFFS vs Medicare Advantage vs commercial claims), and whether effects were pooled one-stage or two-stage —\nwhile remembering the checklist documents these choices but does not validate them.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "prisma", "individual-participant-data", "evidence-synthesis", "meta-analysis" ], "aliases": [ "PRISMA-IPD", "PRISMA for Individual Participant Data", "PRISMA IPD Statement", "Preferred Reporting Items for a Systematic Review and Meta-analysis of Individual Participant Data" ], "applies_to_study_types": [ "ipd_meta_analysis", "meta_analysis_rct", "meta_analysis_obs", "systematic_review" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "doi": "10.1001/jama.2015.3656", "url": "https://doi.org/10.1001/jama.2015.3656", "citation_text": "Stewart LA, Clarke M, Rovers M, et al. Preferred Reporting Items for a Systematic Review and Meta-analysis of Individual Participant Data: the PRISMA-IPD Statement. JAMA. 2015;313(16):1657-1665.", "year": 2015, "authors_short": "Stewart et al.", "notes": "Canonical statement paper defining the PRISMA-IPD checklist and flow diagram for the IPD Meta-analysis Methods Group; the authoritative reference for the extension." }, { "role": "explain", "doi": "10.1136/bmj.n71", "url": "https://doi.org/10.1136/bmj.n71", "citation_text": "Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71.", "year": 2021, "authors_short": "Page et al.", "notes": "Parent PRISMA 2020 statement; PRISMA-IPD is read as an extension of this base checklist, which supplies the core search/selection/synthesis reporting items." }, { "role": "use", "url": "https://www.prisma-statement.org/extensions", "citation_text": "PRISMA Statement website — PRISMA extensions, including PRISMA-IPD checklist, flow diagram, and explanatory resources (EQUATOR Network).", "year": 2015, "authors_short": "PRISMA / EQUATOR Network", "notes": "Maintained checklist, flow diagram, and downloadable resources for applying the extension." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "ipd-meta-analysis", "notes": "The primary study type PRISMA-IPD governs — synthesis built on harmonized patient-level data." }, { "relation_type": "applies_to", "target_slug": "meta-analysis-rct", "notes": "Applies when the contributing IPD come from randomized trials." }, { "relation_type": "applies_to", "target_slug": "meta-analysis-obs", "notes": "Applies when the contributing IPD come from observational/RWD cohorts." }, { "relation_type": "see_also", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "Implements the data-availability and fitness-for-use items — which datasets/sites were sought, obtained, and adequate for the question." }, { "relation_type": "see_also", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Implements the harmonization of outcome/exposure definitions across contributing datasets." }, { "relation_type": "see_also", "target_slug": "time-zero-index-date-alignment-rwe", "notes": "Implements consistent cohort-entry/time-zero alignment across pooled studies." }, { "relation_type": "see_also", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Implements the target-estimand and intercurrent-event reporting the checklist demands." }, { "relation_type": "see_also", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Implements the participant-flow, missing-participant, and follow-up-completeness items." }, { "relation_type": "used_with", "target_slug": "multi-database", "notes": "Federated/pooled multi-database RWE analyses are the natural RWD home for PRISMA-IPD reporting." }, { "relation_type": "used_with", "target_slug": "claims-analysis", "notes": "Apply when the pooled patient-level data are drawn from claims (or EHR/registry) sources." }, { "relation_type": "used_with", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "Confounding control that a pooled observational re-analysis must still specify and defend." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "hta", "journal" ] }, { "slug": "prisma-nma", "name": "PRISMA-NMA (PRISMA Extension for Network Meta-Analysis)", "short_definition": "A reporting checklist that extends PRISMA to systematic reviews incorporating network meta-analysis, adding items for network geometry, the transitivity assumption, direct-versus-indirect consistency, and the cautious presentation of treatment-ranking statistics.", "long_description": "**What it is.** The **PRISMA Extension for Network Meta-Analysis (PRISMA-NMA)** is a reporting guideline\nfor systematic reviews that synthesize a connected network of interventions using direct and indirect\ncomparisons. Published by Hutton, Salanti, and colleagues in *Annals of Internal Medicine* (2015), it adds\nnetwork-specific items on top of the base PRISMA framework: a description of **network geometry** (the\nnetwork plot, nodes, edges, and weighting), an explicit account of the **transitivity / similarity\nassumption**, the methods used to assess **consistency** (agreement of direct and indirect evidence), and\nthe **presentation of ranking statistics** (SUCRA, P-scores, mean ranks). It is maintained as a PRISMA\nextension under the EQUATOR Network and is intended to sit alongside, not replace, the parent PRISMA 2020\nstatement. Like all PRISMA products it is a *transparency* instrument — a list of what a competent report\nmust disclose — not a methods manual, a risk-of-bias tool, or a quality score.\n\n**When to use.** Use PRISMA-NMA to report any systematic review whose synthesis includes **at least one\nindirect comparison** — i.e., any network meta-analysis or mixed-treatment comparison, whether the\nincluded evidence is from randomized trials or, increasingly, from observational comparative-effectiveness\nstudies. It applies to journal manuscripts, HTA/payer dossiers that submit an indirect treatment\ncomparison to value the relative effect of a new therapy against multiple competitors, and regulatory or\nscientific-advice packages that lean on network estimates. Decision rule for choosing the right PRISMA\nproduct: if the review pools **only head-to-head (pairwise) evidence**, the base **PRISMA 2020** statement\nsuffices; if it forms a **connected network with indirect comparisons**, use **PRISMA-NMA**; if you are\nregistering or publishing the **protocol**, use **PRISMA-P** (and pre-register in PROSPERO); a **scoping\nreview** uses PRISMA-ScR and a **diagnostic-test-accuracy** review uses PRISMA-DTA. PRISMA-NMA governs the\n*reporting* of the synthesis; the methodological appraisal of *confidence in the network estimates* is a\nseparate, companion step (e.g., CINeMA or GRADE for NMA).\n\n**What it requires.** Beyond the standard PRISMA items (eligibility, search, selection, data extraction,\nrisk-of-bias assessment of the included studies, summary of findings), PRISMA-NMA enforces a set of\nnetwork-specific disclosures:\n- **Network geometry** — a network plot with nodes and edges, node/edge weighting, the number of studies\n and participants per comparison, and any **disconnected subgraphs**. A network reported as connected\n when it is not is a reporting failure.\n- **Node definition (lumping vs splitting)** — how interventions were grouped into nodes (dose, regimen,\n formulation, drug class) and the sensitivity of conclusions to that grouping.\n- **Transitivity / similarity** — an explicit statement of the assumption that potential **effect\n modifiers are distributed comparably** across the comparisons being indirectly linked, with the\n clinical and methodological reasoning (and data) used to defend it.\n- **Consistency / incoherence** — the statistical approach used to check agreement of direct and indirect\n estimates (node-splitting / back-calculation, loop-specific tests, or the design-by-treatment\n interaction model) and what it showed.\n- **Effect estimates and uncertainty for every contrast** in the network, including comparisons informed\n only indirectly, with between-study heterogeneity (τ²) and, for Bayesian fits, priors and convergence.\n- **Ranking statistics, reported with restraint** — SUCRA, P-scores, or mean ranks presented *with their\n uncertainty* and explicitly framed as relative summaries, **not** as effect sizes or definitive\n orderings.\nFor an observational (\"real-world\") network the same RWE discipline that governs the individual studies\ncarries up to the synthesis: design transparency and data-fitness-for-use of each contributing database,\nphenotype/algorithm validation, **time-zero alignment**, and the **estimand** (and intercurrent-event\nhandling) of each node must be comparable before the nodes are networked.\n\n**When NOT to use — limitations and common misapplications.**\n- **It is not a methods or quality instrument.** Completing the PRISMA-NMA checklist documents *that* you\n reported the network; it does not certify that transitivity holds, that the model is correct, or that\n the evidence is trustworthy. Confidence in the estimates is assessed by **CINeMA / GRADE for NMA**, and\n the included studies' internal validity by **RoB 2 / ROBINS-I** — PRISMA-NMA does none of these and must\n not be scored as if it did.\n- **Wrong PRISMA product.** Reporting a network meta-analysis against the base **PRISMA 2020** checklist\n silently omits the network-specific items (geometry, transitivity, consistency, ranking) that are the\n entire reason an indirect comparison needs special reporting. Conversely, applying PRISMA-NMA to a\n purely pairwise review is over-scoping.\n- **Over-interpreting rankings** is the single most common misuse: declaring the \"best\" treatment from a\n SUCRA value when credible intervals overlap heavily, when the network is sparse, or when transitivity is\n doubtful. PRISMA-NMA exists in part to *discipline* this by forcing rankings to be shown with their\n uncertainty.\n- **Ignoring transitivity** — pooling across populations whose effect modifiers (severity, line of\n therapy, era, care setting) differ, then reporting a tidy network as if the indirect link were valid.\n- **Lumping clinically distinct interventions** into one node merely to force the network to connect, or\n reporting a **disconnected network** as connected.\n- **Checklist-as-theater** — a complete checklist appended to a report that does not actually contain a\n network plot, a transitivity argument, or consistency diagnostics.\n\n**How it maps to this catalog.** PRISMA-NMA is the reporting layer over the synthesis concept\n**network-meta-analysis** (and **meta-analysis-obs** when the evidence is observational); the upstream\ncomparative-effectiveness logic lives in **comparative-effectiveness-research-cer-methods** and\n**ispor-indirect** (the ISPOR good-practices for indirect/mixed-treatment comparisons). The question is\nscoped with **picots-framework-rwe**. The transitivity assumption is, at root, a\n**generalizability-transportability-external-validity-rwe** problem applied across studies: the indirect\nlink is only valid if effect-modifier distributions are exchangeable across the linked comparisons. For a\nreal-world-data network, node comparability is best secured by building each node from a\n**target-trial-emulation** (or at minimum an **active-comparator-new-user** cohort) with aligned\neligibility and **time-zero**, and by pre-specifying each node's **estimands-ate-att-intercurrent-events-rwe**\nso the contrasts being networked are the same estimand. Database-level fitness and contrast construction\nfor claims-derived nodes draw on **claims-analysis**. Confidence in the resulting network estimates is\ngraded with **grade** (extended for NMA). PRISMA-NMA tells the reader *which of these were done and how*;\nthe concepts above tell them *how to do each one*.\n\n**Applied note (claims / EHR / registry RWE).** A payer-facing indirect comparison that networks\ndrug-class contrasts assembled from several claims and EHR databases is the hardest transitivity case:\nchanneling, calendar time, formulary and care-setting differences vary by data source and act as effect\nmodifiers across nodes. PRISMA-NMA's transitivity and consistency items become the place where the analyst\nmust show — not assert — that the observational nodes are comparable: harmonized eligibility and time-zero\n(target-trial framing), validated outcome phenotypes with reported PPV/sensitivity, the same estimand per\nnode, and a documented sensitivity analysis on node definition. Reporting the network plot with per-edge\nstudy/participant counts and the τ² for each contrast lets a reviewer judge whether the indirect evidence\nis thin or robust before any ranking is read.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "network-meta-analysis", "indirect-comparison", "evidence-synthesis", "transitivity", "prisma" ], "aliases": [ "PRISMA-NMA", "PRISMA Extension for Network Meta-Analysis", "PRISMA for network meta-analysis", "Hutton 2015 PRISMA-NMA statement" ], "applies_to_study_types": [ "network_meta_analysis", "meta_analysis_obs" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "doi": "10.7326/M14-2385", "url": "https://doi.org/10.7326/M14-2385", "citation_text": "Hutton B, Salanti G, Caldwell DM, et al. The PRISMA extension statement for reporting of systematic reviews incorporating network meta-analyses of health care interventions: checklist and explanations. Annals of Internal Medicine. 2015;162(11):777-784.", "year": 2015, "authors_short": "Hutton et al.", "notes": "Canonical statement paper. Defines the 32-item PRISMA-NMA checklist (PRISMA items plus NMA-specific items on network geometry, transitivity/similarity, consistency, and ranking) with elaboration and worked examples." }, { "role": "explain", "doi": "10.1016/j.jclinepi.2025.111985", "url": "https://doi.org/10.1016/j.jclinepi.2025.111985", "citation_text": "Veroniki AA, Tricco AC, Rangira D, et al. Updating the PRISMA reporting guideline for network meta-analysis: a scoping review. Journal of Clinical Epidemiology. 2025;188:111985.", "year": 2025, "authors_short": "Veroniki et al.", "notes": "Scoping review motivating and informing the update to PRISMA-NMA; useful context on reporting gaps (especially transitivity justification and ranking-statistic over-interpretation) since 2015." }, { "role": "explain", "doi": "10.1136/bmj.n71", "url": "https://doi.org/10.1136/bmj.n71", "citation_text": "Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71.", "year": 2021, "authors_short": "Page et al.", "notes": "Parent PRISMA 2020 statement. PRISMA-NMA layers network-specific items onto this base; cite both when reporting an NMA." }, { "role": "use", "url": "https://www.prisma-statement.org/extensions", "citation_text": "PRISMA Extension for Network Meta-Analysis (PRISMA-NMA). EQUATOR Network / PRISMA Statement — maintained checklist, flow diagram, and explanatory resources.", "year": 2015, "authors_short": "EQUATOR Network", "notes": "Stable home for the current checklist and accompanying materials." } ], "relations": [ { "relation_type": "see_also", "target_slug": "network-meta-analysis", "notes": "PRISMA-NMA is the reporting standard for this synthesis method; the concept supplies the statistical machinery (network model, geometry, consistency) the checklist asks you to disclose." }, { "relation_type": "see_also", "target_slug": "prisma-2020", "notes": "Parent guideline. PRISMA-NMA adds network-specific items; report a network meta-analysis against both." }, { "relation_type": "see_also", "target_slug": "meta-analysis-obs", "notes": "Networks of observational evidence inherit confounding and the transitivity concern; the same fitness-for-use and estimand discipline must hold at each node before networking." }, { "relation_type": "see_also", "target_slug": "ispor-indirect", "notes": "ISPOR good-research-practices for indirect/mixed-treatment comparisons — the methodological companion that PRISMA-NMA expects you to have followed and to report." }, { "relation_type": "see_also", "target_slug": "comparative-effectiveness-research-cer-methods", "notes": "NMA is a CER tool for ranking multiple interventions when head-to-head trials are missing; CER framing scopes the question PRISMA-NMA then governs the reporting of." }, { "relation_type": "see_also", "target_slug": "grade", "notes": "Confidence in network estimates is graded with GRADE-for-NMA (CINeMA); this is the appraisal step PRISMA-NMA does not itself perform." }, { "relation_type": "used_with", "target_slug": "picots-framework-rwe", "notes": "Scope the population, interventions, comparators, outcomes, timing, and setting of the network with PICOTS before defining nodes and edges." }, { "relation_type": "used_with", "target_slug": "generalizability-transportability-external-validity-rwe", "notes": "Transitivity is exchangeability of effect modifiers across the linked comparisons — the same logic as transportability, applied across studies rather than from study to target." }, { "relation_type": "used_with", "target_slug": "target-trial-emulation", "notes": "For an observational network, build each node from a target-trial emulation with aligned eligibility and time-zero so node-level estimands are comparable before networking." }, { "relation_type": "used_with", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Pre-specify the estimand and intercurrent-event handling per node so the contrasts being networked estimate the same quantity." }, { "relation_type": "used_with", "target_slug": "active-comparator-new-user", "notes": "A defensible per-node contrast in claims/EHR data is typically an active-comparator, new-user cohort; aligned designs strengthen the transitivity argument across nodes." }, { "relation_type": "used_with", "target_slug": "claims-analysis", "notes": "Apply when nodes are built from claims- or EHR-derived comparative effects; database fitness and contrast construction feed the network." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "hta", "journal" ] }, { "slug": "prisma-p", "name": "PRISMA-P (PRISMA for Protocols)", "short_definition": "Reporting guideline that specifies the minimum content a systematic review or meta-analysis PROTOCOL should contain before the review is conducted; the protocol-stage companion to PRISMA 2020, maintained within the EQUATOR Network.", "long_description": "**What it is** — **PRISMA-P (Preferred Reporting Items for Systematic Review and Meta-Analysis Protocols)**\nis a 17-item reporting checklist that defines the minimum information a *protocol* for a systematic review or\nmeta-analysis should report, so that the review's methods are pre-specified, transparent, and auditable before\nany study is screened or any effect is pooled. It was published as the 2015 statement (Moher, Shamseer et al.)\nwith a companion explanation-and-elaboration paper (Shamseer et al.), and is hosted and maintained as a PRISMA\nextension within the **EQUATOR Network**. PRISMA-P is the *protocol-stage* member of the PRISMA family: it\ngoverns what you commit to in advance, whereas **PRISMA 2020** governs how you report the *completed* review.\nIts purpose is to make the planned eligibility criteria, search strategy, risk-of-bias plan, outcomes, and\nsynthesis approach explicit up front, so that deviations are visible and selective reporting and post-hoc\ndata-dredging are constrained.\n\n**When to use** — Apply PRISMA-P whenever you are *writing or registering the protocol* for a systematic review\nor meta-analysis, including reviews of randomized trials, reviews of observational/real-world-evidence (RWE)\nstudies, and mixed evidence. It is the appropriate checklist for a PROSPERO registration record, a protocol\nmanuscript (e.g., Systematic Reviews, BMJ Open), and the evidence-synthesis protocol underpinning an HTA/payer\ndossier or a regulatory evidence package that relies on a systematic review. Decision rule for choosing the\nright family member: use **PRISMA-P** for the *plan*; use **PRISMA 2020** for the *finished review report*; if\nthe review is a scoping review use **PRISMA-ScR**; if it is a network meta-analysis use the **PRISMA-NMA**\nextension for the report. PRISMA-P governs the protocol regardless of which downstream PRISMA extension reports\nthe results. It does **not** govern primary study protocols — a single observational cohort/case-control RWE\nstudy uses HARPER, the ENCePP checklist, or SPIRIT (for trials), not PRISMA-P.\n\n**What it requires** — The 17 items (organized as administrative information, introduction, and methods) compel\npre-specification of the elements that otherwise drift during a review: title/registration; protocol amendments\nand version control; rationale and objectives framed as a structured **PICOTS** question (population,\ninterventions/exposures, comparators, outcomes, timing, setting/study designs); **explicit eligibility\ncriteria** including study designs admitted; a reproducible **information-sources and search strategy** (named\ndatabases, draft search string for at least one database, planned date ranges); a **study-selection and\ndata-extraction** process (screening, duplication, calibration); a pre-specified list of **outcomes and other\nvariables**; a planned **risk-of-bias / quality-assessment** approach for the included studies (the tool to be\nused and the level at which it is applied); and the planned **data synthesis** (effect measures, handling of\nheterogeneity, quantitative pooling vs narrative, planned subgroup/sensitivity analyses, and assessment of\nmeta-bias such as publication bias and selective outcome reporting). For an RWE-heavy review these generic items\ncarry specific weight: the eligibility criteria must state which real-world data designs and data sources are\nadmissible; the risk-of-bias plan must name an instrument suited to non-randomized studies (e.g., ROBINS-I) and,\nwhere outcomes are algorithm-defined, the protocol should pre-specify how it will appraise phenotype/outcome\nalgorithm validity, time-zero alignment, and confounding control across the included studies; and the synthesis\nsection must pre-commit to whether heterogeneous observational estimands are poolable at all or only\nqualitatively synthesized.\n\n**When NOT to use — limitations and common misapplications** — PRISMA-P is a *reporting* checklist for a\n*protocol*; it is not a quality score, not a risk-of-bias instrument, and not a guarantee of a valid review.\nConcrete failure modes: (1) **Wrong family member** — using PRISMA-P to report the *completed* review (use\nPRISMA 2020) or, conversely, citing PRISMA 2020 for a protocol. (2) **Wrong checklist entirely** — applying\nPRISMA-P to a *primary* observational/RWE study protocol; a single claims/EHR cohort study should follow HARPER,\nthe ENCePP checklist, or STROBE/RECORD-PE for reporting, never PRISMA-P. (3) **Mistaking it for an appraisal\ntool** — PRISMA-P says nothing about whether the *included studies* are sound; the included studies are\nappraised with ROBINS-I/ROB 2 and the review's own conduct with AMSTAR 2. (4) **PROSPERO ≠ compliance** —\nregistering a record is not the same as completing the checklist; do both, and keep the registration synchronized\nwith protocol amendments. (5) **Checklist-as-theater** — ticking 17 boxes while leaving eligibility criteria,\nthe search string, the risk-of-bias plan, or the synthesis method vague defeats the purpose; the value is the\npre-specification, not the page count. (6) **Completing the checklist does not make a pooled observational\nestimate causal or unconfounded** — a fully PRISMA-P-compliant protocol can still plan a meta-analysis of\nbiased studies; transparency of the plan is necessary, not sufficient.\n\n**How it maps to this catalog** — In this repo, PRISMA-P's requirements are implemented by concepts a reviewer\ncan pre-specify against:\n- The review object and synthesis machinery: **systematic-review**, **meta-analysis-obs**, **meta-analysis-rct**\n (and, for the report stage, **network-meta-analysis**).\n- The structured question / eligibility spine (PRISMA-P items 8–10): **picots-framework-rwe** operationalizes\n the population/intervention/comparator/outcome/timing/setting frame the protocol must declare.\n- The protocol-content discipline itself: **study-protocol-or-sap-elements** supplies the pre-specification and\n amendment-control habits PRISMA-P's administrative and methods items demand.\n- Appraisal of the *included* RWE studies (planned in the risk-of-bias item, not implemented by PRISMA-P\n itself): when the review pools real-world studies, the protocol should commit to appraising them against\n **algorithm-validation** (outcome/phenotype algorithm validity), **generalizability-transportability-external-validity-rwe**,\n and the design-validity concepts those primary studies rely on — **target-trial-emulation**,\n **active-comparator-new-user**, **estimands-ate-att-intercurrent-events-rwe**,\n **diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe**, **attrition-and-loss-to-follow-up-rwe**, and\n **high-dimensional-propensity-score-hdps-rwe**. These are *not* PRISMA-P items for the protocol's own conduct;\n they are the lens the protocol pre-specifies for grading the evidence it will synthesize.\n\n**Applied note (claims/EHR/registry RWE).** A systematic review pooling claims- and EHR-based comparative\nstudies should, at the protocol stage, state which data designs are admissible (e.g., active-comparator new-user\ncohorts vs prevalent-user designs), pre-specify a non-randomized risk-of-bias instrument and how time-zero,\nconfounding, and algorithm-defined outcomes will be judged across heterogeneous data sources, and decide *in\nadvance* whether estimates from different databases and estimands are quantitatively poolable or only narratively\nsynthesized — the questions PRISMA-P items on eligibility, risk of bias, and synthesis exist to force.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "systematic-review", "protocol", "prisma", "evidence-synthesis", "equator" ], "aliases": [ "PRISMA-P", "PRISMA for Protocols", "Preferred Reporting Items for Systematic Review and Meta-Analysis Protocols", "PRISMA-P 2015" ], "applies_to_study_types": [ "systematic_review", "meta_analysis_rct", "meta_analysis_obs" ], "data_sources": [], "citations": [ { "role": "introduce", "doi": "10.1186/2046-4053-4-1", "url": "https://doi.org/10.1186/2046-4053-4-1", "citation_text": "Moher D, Shamseer L, Clarke M, et al. Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015 statement. Systematic Reviews. 2015;4(1):1.", "year": 2015, "authors_short": "Moher et al.", "notes": "Canonical PRISMA-P 2015 statement defining the 17-item protocol reporting checklist." }, { "role": "explain", "doi": "10.1136/bmj.g7647", "url": "https://doi.org/10.1136/bmj.g7647", "citation_text": "Shamseer L, Moher D, Clarke M, et al. Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015: elaboration and explanation. BMJ. 2015;349:g7647.", "year": 2015, "authors_short": "Shamseer et al.", "notes": "Item-by-item explanation and elaboration with examples of good protocol reporting." }, { "role": "use", "url": "https://www.equator-network.org/reporting-guidelines/prisma-protocols/", "citation_text": "PRISMA-P (PRISMA for Protocols) 2015 statement. EQUATOR Network reporting-guidelines library (maintained checklist, downloadable templates, and PRISMA extension links).", "year": 2015, "authors_short": "EQUATOR Network", "notes": "Canonical maintained landing page with the checklist in usable formats and links to related PRISMA extensions." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "systematic-review", "notes": "PRISMA-P governs the protocol that precedes a systematic review." }, { "relation_type": "applies_to", "target_slug": "meta-analysis-rct", "notes": "Use at the protocol stage for a meta-analysis of randomized trials." }, { "relation_type": "applies_to", "target_slug": "meta-analysis-obs", "notes": "Use at the protocol stage for a meta-analysis of observational/RWE studies; pre-specify a non-randomized risk-of-bias plan and whether estimates are poolable." }, { "relation_type": "see_also", "target_slug": "systematic-review", "notes": "PRISMA-P is the protocol-stage checklist; the completed review is reported with PRISMA 2020." }, { "relation_type": "used_with", "target_slug": "picots-framework-rwe", "notes": "PICOTS operationalizes PRISMA-P items 8-10 (objectives, eligibility criteria, study designs admitted)." }, { "relation_type": "used_with", "target_slug": "study-protocol-or-sap-elements", "notes": "Supplies the pre-specification and amendment-control discipline PRISMA-P's administrative/methods items require." }, { "relation_type": "see_also", "target_slug": "meta-analysis-obs", "notes": "Synthesis target for RWE reviews; the protocol must pre-commit to quantitative pooling vs narrative synthesis." }, { "relation_type": "see_also", "target_slug": "algorithm-validation", "notes": "When the review pools algorithm-defined RWE outcomes, the risk-of-bias plan should pre-specify how outcome/phenotype algorithm validity is appraised across included studies." }, { "relation_type": "see_also", "target_slug": "generalizability-transportability-external-validity-rwe", "notes": "The protocol should state how external validity of included real-world studies will be judged before synthesis." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "journal", "hta", "fda", "ema" ] }, { "slug": "prisma-s", "name": "PRISMA-S", "short_definition": "A PRISMA 2020 content extension specifying how to transparently and reproducibly report the literature search component of a systematic review — information sources, full search strategies, filters, deduplication, and search peer review — so that any reader can evaluate and rerun the search.", "long_description": "**What it is.** **PRISMA-S (PRISMA Extension for Reporting Literature Searches)** is a 16-item reporting checklist that\nextends the parent **PRISMA 2020 statement** to cover one thing in depth: the *search* underlying a systematic review.\nIt was developed by an international working group of information specialists and methodologists (Rethlefsen et al.,\n2021) under the **EQUATOR Network** umbrella and is maintained alongside the other PRISMA products at the PRISMA\nStatement website. Its sole purpose is **search transparency and reproducibility** — making the methods used to find\nstudies explicit enough that a knowledgeable reader could appraise the search's adequacy and, in principle, reproduce\nit. PRISMA-S is a *content* extension, not a standalone guideline: it expands the search-related PRISMA 2020 items into\ngranular, operational reporting requirements. It is deliberately narrow. It says nothing about study design, risk of\nbias, effect estimation, or synthesis methods.\n\n**When to use.** Use PRISMA-S whenever a manuscript, protocol, or dossier reports a systematic search of the literature\n— systematic reviews, meta-analyses (of RCTs or observational studies), network meta-analyses, scoping reviews, rapid\nreviews, and the search sections of HTA submissions and guideline evidence reviews. PRISMA-S is applied *in addition to*\nthe relevant PRISMA reporting set, not instead of it. Decision rule: report against **PRISMA 2020** for the review as a\nwhole, then layer **PRISMA-S** over the search to satisfy the search-reporting items at full granularity; if the review\nis a scoping review use **PRISMA-ScR** for the body and still layer PRISMA-S over its search, and likewise for\n**PRISMA-IPD** or **PRISMA-NMA**. PRISMA-S governs *reporting* at write-up and peer-review time; for *conducting and\npeer-reviewing* the search itself the methodological partner is **PRESS** (Peer Review of Electronic Search Strategies).\n\n**What it requires.** PRISMA-S enforces complete, rerunnable documentation across the search lifecycle. Its item groups\nrequire: (1) **Information sources** — every database (with platform/vendor, e.g., MEDLINE via Ovid vs PubMed), trial\nregistry, grey-literature source, web search engine, and the dates each was searched, plus any limits or coverage\ncaveats. (2) **Search strategies** — the *full, line-by-line* search string for at least one database reproduced\nverbatim (not a paraphrase), and ideally for every source, including all controlled-vocabulary terms, free-text terms,\nfield tags, Boolean operators, truncation, and the use of any published or methodological **search filters** (with\ntheir source). (3) **Supplementary search methods** — citation chasing (forward and backward), contacting authors or\nmanufacturers, examining reference lists, and any hand-searching. (4) **Search management and processing** — the\nsoftware or platform used to run/export searches, the **deduplication** method and tool, and total records before and\nafter dedup. (5) **Search peer review** — whether the search strategy was peer reviewed (e.g., using PRESS) and by whom.\nThe unifying standard is reproducibility: a reader should be able to reconstruct the search from the report alone.\n\n**When NOT to use — limitations and common misapplications.** PRISMA-S is a **reporting** checklist, not a risk-of-bias\ninstrument, not a quality score, and not a measure of search *adequacy* — checking every item only confirms that the\nsearch was *described* transparently, not that it was sensitive, unbiased, or appropriately scoped (that judgment needs\nPRESS and methodological appraisal). Common failures: (a) **paraphrasing instead of pasting** the actual search string,\nwhich defeats reproducibility — PRISMA-S requires the verbatim strategy, usually as a supplementary appendix; (b)\nreporting \"PubMed, Embase, Cochrane\" with no platform, no dates, and no strings (item-level non-compliance dressed up as\ncompliance); (c) **checklist-as-theater** — ticking boxes and citing PRISMA-S while the search section remains\nirreproducible; (d) using PRISMA-S where a *different* PRISMA product governs the review type, or omitting the parent\nPRISMA 2020 set and treating PRISMA-S as if it covered the whole review (it does not — it covers only the search); (e)\ntreating PRISMA-S compliance as evidence the review's *conclusions* are sound. PRISMA-S is also not a search-conduct\nguide: it tells you what to report, not how to design a sensitive strategy.\n\n**How it maps to this catalog.** PRISMA-S is an evidence-synthesis reporting tool, so it complements the synthesis\ndesigns in this repo rather than any causal-inference concept. It applies to **systematic-review**, **meta-analysis-rct**,\n**meta-analysis-obs**, **network-meta-analysis**, and **scoping-review**: each of these implements a search that PRISMA-S\ngoverns the reporting of, and **ipd-meta-analysis** when an IPD synthesis runs its own search. Pair it with\n**picots-framework-rwe**, which defines the Population/Intervention/Comparator/Outcome/Timing/Setting scope that the\nsearch strategy must operationalize and that PRISMA-S then asks you to report faithfully. There is intentionally **no**\ncross-reference to causal-design concepts (target-trial-emulation, propensity scores, phenotype algorithms, time-zero,\nestimands, attrition) — those govern how primary RWE *studies* are designed and have nothing to do with how a review's\n*literature search* is reported; bolting them on would be a category error.\n\n**Applied note for RWE evidence synthesis.** When a systematic review or meta-analysis synthesizes real-world evidence\n— for example, an HTA submission summarizing observational comparative-effectiveness or burden-of-disease studies — the\nindividual RWE studies are designed and appraised with the catalog's RWE concepts, but the *search that found them* is\nstill reported under PRISMA-S. That means documenting database platforms (and that registries such as ClinicalTrials.gov\nor EU PAS were searched for RWE/PASS studies), the verbatim strategies including any observational/RWE study-design\nfilters, grey-literature and conference-abstract sources (often decisive for unpublished RWE), the dedup workflow, and\nwhether the strategy was PRESS-reviewed. Keep this distinct from how the underlying RWE studies handled confounding,\ntime-zero, or attrition — PRISMA-S has no opinion on those, and conflating the two is the most common scoping error.", "primary_category": "Guideline", "tags": [ "guideline", "reporting-guideline", "systematic-review", "meta-analysis", "literature-search", "prisma-extension", "equator", "evidence-synthesis" ], "aliases": [ "PRISMA-S", "PRISMA Extension for Reporting Literature Searches", "PRISMA Search Extension", "PRISMA literature search extension" ], "applies_to_study_types": [ "systematic_review", "meta_analysis_rct", "meta_analysis_obs", "network_meta_analysis", "scoping_review" ], "data_sources": [ "bibliographic_databases", "trial_registries", "grey_literature" ], "citations": [ { "role": "introduce", "doi": "10.1186/s13643-020-01542-z", "url": "https://doi.org/10.1186/s13643-020-01542-z", "citation_text": "Rethlefsen ML, Kirtley S, Waffenschmidt S, et al. PRISMA-S: an extension to the PRISMA Statement for Reporting Literature Searches in Systematic Reviews. Systematic Reviews. 2021;10(1):39.", "year": 2021, "authors_short": "Rethlefsen et al.", "notes": "The canonical PRISMA-S statement and 16-item checklist; defines each search-reporting item and its rationale." }, { "role": "explain", "doi": "10.1136/bmj.n71", "url": "https://doi.org/10.1136/bmj.n71", "citation_text": "Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71.", "year": 2021, "authors_short": "Page et al.", "notes": "The parent statement PRISMA-S extends; PRISMA-S is applied together with PRISMA 2020, not instead of it." }, { "role": "explain", "doi": "10.1136/bmj.n160", "url": "https://doi.org/10.1136/bmj.n160", "citation_text": "Page MJ, Moher D, Bossuyt PM, et al. PRISMA 2020 explanation and elaboration: updated guidance and exemplars for reporting systematic reviews. BMJ. 2021;372:n160.", "year": 2021, "authors_short": "Page et al.", "notes": "Item-by-item elaboration and worked exemplars for the parent PRISMA 2020 reporting items, including the search." }, { "role": "see_also", "doi": "10.1016/j.jclinepi.2016.01.021", "url": "https://doi.org/10.1016/j.jclinepi.2016.01.021", "citation_text": "McGowan J, Sampson M, Salzwedel DM, Cogo E, Foerster V, Lefebvre C. PRESS Peer Review of Electronic Search Strategies: 2015 Guideline Statement. Journal of Clinical Epidemiology. 2016;75:40-46.", "year": 2016, "authors_short": "McGowan et al.", "notes": "The methodological partner for peer-reviewing the search strategy itself; PRISMA-S asks you to report whether a PRESS-style peer review was performed." }, { "role": "use", "url": "https://www.prisma-statement.org/prisma-2020", "citation_text": "PRISMA Statement website (EQUATOR Network) — maintained PRISMA 2020 and PRISMA-S checklists, flow-diagram templates, and extension index.", "year": 2021, "authors_short": "PRISMA / EQUATOR Network", "notes": "Authoritative, maintained source for the current checklist files and the full family of PRISMA extensions." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "systematic-review", "notes": "Governs how the literature search of a systematic review is reported; applied on top of PRISMA 2020." }, { "relation_type": "applies_to", "target_slug": "meta-analysis-rct", "notes": "The search that identifies trials for a meta-analysis of RCTs must be reported to PRISMA-S granularity." }, { "relation_type": "applies_to", "target_slug": "meta-analysis-obs", "notes": "Applies to the search underlying meta-analyses of observational/RWE studies, including grey-literature sources." }, { "relation_type": "applies_to", "target_slug": "network-meta-analysis", "notes": "NMAs run a systematic search; PRISMA-S governs its reporting (PRISMA-NMA governs the synthesis-specific items)." }, { "relation_type": "applies_to", "target_slug": "scoping-review", "notes": "Layer PRISMA-S over the search of a scoping review even though PRISMA-ScR governs the body of the report." }, { "relation_type": "see_also", "target_slug": "ipd-meta-analysis", "notes": "When an IPD meta-analysis conducts its own literature search, that search is reported under PRISMA-S." }, { "relation_type": "used_with", "target_slug": "picots-framework-rwe", "notes": "PICOTS defines the scope the search must operationalize; PRISMA-S then requires that the resulting strategy be reported verbatim and reproducibly." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "hta", "journal" ] }, { "slug": "prisma-scr", "name": "PRISMA-ScR (PRISMA Extension for Scoping Reviews)", "short_definition": "A 20-item reporting checklist (with two optional items) that extends the PRISMA statement to scoping reviews, standardizing how authors disclose the objectives, eligibility, search, charting, and synthesis of an exploratory evidence map. It governs reporting transparency, not study conduct or risk of bias.", "long_description": "**What it is.** PRISMA-ScR is the **Preferred Reporting Items for Systematic reviews and Meta-Analyses extension for Scoping Reviews**, published by Tricco and colleagues in 2018 (Annals of Internal Medicine). It is an official EQUATOR Network reporting guideline, developed by a Joanna Briggs Institute (JBI)/PRISMA working group through a Delphi process and maintained alongside the wider PRISMA family at prisma-statement.org. The checklist comprises **20 essential items plus 2 optional items** organized into the familiar PRISMA sections: title, abstract, introduction (rationale, objectives), methods (protocol/registration, eligibility, information sources, search, selection of sources, data charting, data items, critical appraisal *if done*, synthesis of results), results (selection, characteristics, results of individual sources, synthesis), discussion (summary, limitations, conclusions), and funding. Crucially, PRISMA-ScR is a **reporting** standard: it specifies what must appear in the manuscript so that a reader can judge what was done and reproduce it. It is not a methodology, a critical-appraisal/risk-of-bias instrument, or a quality score.\n\n**When to use.** Use PRISMA-ScR when the deliverable is a **scoping review** — an exploratory synthesis whose aim is to map the breadth and nature of evidence, clarify concepts, identify knowledge gaps, or examine how research on a topic has been conducted, rather than to answer a focused effect-estimate question. Decision rules versus siblings: (1) If you are reporting a **systematic review or meta-analysis of effects**, use **PRISMA 2020**, not PRISMA-ScR. (2) If you are reporting the **protocol** of a systematic review, use **PRISMA-P**; PRISMA-ScR has no separate protocol extension, so scoping-review protocols are typically written against JBI scoping-review methodology and registered (e.g., on OSF) with PRISMA-P used adaptively. (3) Scoping reviews do **not** require a pooled estimate, a PICO-driven question, or a formal risk-of-bias assessment, although appraisal *may* be reported (item 12 is conditional). PRISMA-ScR applies across decision contexts where a scoping review is the chosen evidence product: peer-reviewed journals, HTA/payer landscape and gap analyses that precede a full systematic review, and regulatory or methods-mapping work (e.g., scoping the published RWE evidence base for a therapeutic area before designing a pharmacoepidemiologic study). It does not directly govern primary FDA/EMA non-interventional study submissions — those are reported under STROBE/RECORD-PE/HARPER and ISPOR/ISPE good-practice guidance.\n\n**What it requires.** The checklist enforces transparent disclosure of the review's machinery. **Design transparency:** a stated rationale and explicit objectives framed (commonly) by population/concept/context; pre-specification via a protocol and its registration/availability; explicit eligibility criteria with justification. **Search reproducibility:** all information sources with dates of coverage, and a **full electronic search strategy for at least one database, presented verbatim** so it can be re-run. **Selection and charting:** the process for screening and selecting sources of evidence, the **data-charting** method (the scoping-review analogue of extraction — what was charted, by whom, and how charting was piloted and calibrated), and a clear account of how results were summarized (typically narrative/tabular/graphical mapping rather than statistical pooling). **Flow accounting:** a PRISMA-style flow diagram tracing records identified, screened, excluded with reasons, and included — the discipline that makes the evidence map auditable. **Conditional appraisal:** if critical appraisal of sources was performed, the method and results must be reported (and authors should state explicitly when it was *not* done). For scoping reviews built on real-world data sources or RWE literature, the same charting rigor should capture data-fitness descriptors of the included studies — design, data source (claims/EHR/registry), phenotype/algorithm definitions, time-zero handling, and stated estimands — so the map is informative for downstream method selection.\n\n**When NOT to use — limitations and common misapplications.** (1) **It is a reporting checklist, not a risk-of-bias tool and not a quality score.** Ticking all 22 items certifies disclosure, not validity; a fully PRISMA-ScR-compliant scoping review can still rest on a biased or non-reproducible search. Do not sum the items into a \"quality score.\" (2) **Completing the checklist does not make the underlying evidence causal or decision-grade.** A scoping review maps what exists; it does not synthesize effect sizes, grade certainty (no GRADE), or support a comparative-effectiveness claim — use a systematic review/meta-analysis for that. (3) **Wrong-extension errors:** using PRISMA-ScR to report a synthesis that actually estimates effects (should be PRISMA 2020), or forcing a focused effect question into the scoping format to avoid meta-analytic and risk-of-bias obligations. (4) **Checklist-as-theater:** pasting a completed checklist into an appendix while the search strategy is not reproducible, the flow numbers do not reconcile, or charting was undocumented. (5) **Mis-scoping:** labeling a review \"scoping\" to sidestep PROSPERO registration or risk-of-bias assessment when the question is genuinely about effects. (6) PRISMA-ScR governs *reporting of the review*; it imposes no standards on the *primary RWE studies* it includes — those are appraised against their own reporting and good-practice guidance.\n\n**How it maps to this catalog.** PRISMA-ScR sits at the evidence-synthesis layer; its requirements point downward to the primary-study concepts whose presence (or absence) a scoping review of RWE should chart and that any included study must satisfy. Map item-by-item: the **eligibility / data-charting** items should capture each included study's design and data-fitness — implemented here by `claims-analysis` (administrative-claims data structure and fitness-for-use) and `medicare-ffs-ma-commercial-claims-differences-rwe` (data-source heterogeneity to record when charting); the **objectives/concept** framing for comparative RWE maps to `active-comparator-new-user` and `target-trial-emulation` as the design templates a mature evidence map should distinguish; **phenotype/algorithm transparency** that the charting form should extract is implemented by `diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe` and `claims-outcome-algorithm-ppv-sensitivity-rwe`; **estimand clarity** of included studies maps to `estimands-ate-att-intercurrent-events-rwe` and `estimand-analysis-traceability-rwe`; **attrition/flow accounting** parallels `attrition-and-loss-to-follow-up-rwe` and `database-feasibility-attrition-funnel-rwe`; and **confounding-control sophistication** worth charting maps to `high-dimensional-propensity-score-hdps-rwe`. Applied note for claims/EHR/registry RWE: when scoping a therapeutic area's real-world evidence, build the charting form to extract, per study, the data source and provenance, the operational phenotype/outcome algorithm with validation metrics (PPV/sensitivity), the time-zero/new-user rule, the comparator strategy, the declared estimand, and the attrition funnel — this turns a PRISMA-ScR-compliant map into a directly actionable input for designing the confirmatory study, rather than a bibliography.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "scoping-review", "evidence-synthesis", "equator", "prisma" ], "aliases": [ "PRISMA-ScR", "PRISMA Extension for Scoping Reviews", "PRISMA for Scoping Reviews", "PRISMA ScR" ], "applies_to_study_types": [ "scoping_review" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "doi": "10.7326/M18-0850", "url": "https://doi.org/10.7326/M18-0850", "citation_text": "Tricco AC, Lillie E, Zarin W, et al. PRISMA Extension for Scoping Reviews (PRISMA-ScR): Checklist and Explanation. Annals of Internal Medicine. 2018;169(7):467-473.", "year": 2018, "authors_short": "Tricco et al.", "notes": "Canonical statement paper defining the 20+2 item PRISMA-ScR checklist with item-by-item explanation; developed by a JBI/PRISMA Delphi working group." }, { "role": "explain", "doi": "10.11124/JBIES-20-00167", "url": "https://doi.org/10.11124/JBIES-20-00167", "citation_text": "Peters MDJ, Marnie C, Tricco AC, et al. Updated methodological guidance for the conduct of scoping reviews. JBI Evidence Synthesis. 2020;18(10):2119-2126.", "year": 2020, "authors_short": "Peters et al.", "notes": "Companion JBI methodology that operationalizes scoping-review conduct (objective/concept/context framing, charting) underlying what PRISMA-ScR asks authors to report." }, { "role": "explain", "doi": "10.1136/bmj.n71", "url": "https://doi.org/10.1136/bmj.n71", "citation_text": "Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71.", "year": 2021, "authors_short": "Page et al.", "notes": "Parent PRISMA statement; cited to fix the decision boundary — effect-estimate syntheses use PRISMA 2020, exploratory evidence maps use PRISMA-ScR." }, { "role": "use", "url": "https://www.prisma-statement.org/extensions", "citation_text": "PRISMA Extension for Scoping Reviews (PRISMA-ScR), EQUATOR Network / PRISMA — maintained checklist, flow diagram, and author resources.", "year": 2018, "authors_short": "PRISMA / EQUATOR Network", "notes": "Stable maintained source for the current checklist and flow-diagram templates." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "scoping-review", "notes": "PRISMA-ScR is the designated reporting standard for scoping reviews." }, { "relation_type": "see_also", "target_slug": "target-trial-emulation", "notes": "A design template a mature RWE scoping review should distinguish and chart among included comparative studies." }, { "relation_type": "see_also", "target_slug": "active-comparator-new-user", "notes": "Core comparative design whose presence/absence the charting form should capture when mapping pharmacoepidemiologic evidence." }, { "relation_type": "see_also", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Phenotype/algorithm transparency that the data-charting items should extract from each included RWE study." }, { "relation_type": "see_also", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Estimand clarity of included studies is a key field for an actionable charting form." }, { "relation_type": "see_also", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "The PRISMA flow/attrition discipline parallels how primary RWE studies must account for cohort attrition." }, { "relation_type": "used_with", "target_slug": "claims-analysis", "notes": "Apply when scoping a body of claims/EHR-based real-world evidence; chart each study's data source and fitness-for-use." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "hta", "journal" ] }, { "slug": "probast", "name": "PROBAST (Prediction model Risk Of Bias ASsessment Tool)", "short_definition": "A structured risk-of-bias instrument for appraising prediction model studies (development, validation, or both). Four signalling-question domains — participants, predictors, outcome, and analysis — each rated low/high/unclear risk of bias, plus an overall applicability judgment. Pairs with TRIPOD (reporting) but is an appraisal tool, not a checklist authors fill in.", "long_description": "**What it is** — **PROBAST (Prediction model Risk Of Bias ASsessment Tool)** is a domain-structured,\nsignalling-question instrument for assessing the **risk of bias and applicability** of studies that\ndevelop, validate, or update a multivariable clinical prediction model (diagnostic or prognostic). It\nwas published by Wolff et al. in 2019 in *Annals of Internal Medicine* alongside a companion\nExplanation & Elaboration paper (Moons et al. 2019, same issue). PROBAST is the appraisal companion\nto the **TRIPOD** reporting guideline: TRIPOD tells authors what to *report*; PROBAST tells reviewers\nhow to *judge* whether a prediction model study is at low or high risk of bias. Both are maintained by\nthe **EQUATOR Network**. PROBAST has four **domains**: *D1 Participants* (source data and case-mix\nadequate?), *D2 Predictors* (well-defined, not using post-index information?), *D3 Outcome* (valid\ndefinition, appropriately assessed, blinded to predictors?), and *D4 Analysis* (sample size, missing\ndata handling, predictor selection, overfitting control, calibration reported?). Within each domain,\nsignalling questions (answered low/high/unclear) drive an overall domain risk-of-bias rating, and the\nfour domain ratings combine into an **overall risk-of-bias judgment**. A separate **applicability**\nassessment — across the same three phases as ROBINS-I — captures whether the study's population,\npredictors, or outcome definition are appropriate for the review question even if the study is\notherwise at low risk of bias.\n\n**When to use** — Apply PROBAST whenever you are *appraising* a prediction model study rather than\nreporting one: in a **systematic review or meta-analysis of prediction models** (the primary design\ncontext for which PROBAST was developed); in a **guideline panel or HTA subgroup** weighting model-\nbased evidence; in a **regulatory or payer submission** where a risk-stratification or enrichment\nmodel must be credibly evaluated; or in an **internal evidence-quality gate** before a model is cited\nor deployed. Decision rule: if the study's deliverable is a multivariable prediction model and the\nquestion is \"is this model at risk of bias?\", use **PROBAST**; if the question is \"is this model\n*reported* completely?\", use **TRIPOD**. If the study is an observational comparative-effectiveness\nstudy (not a prediction model), use **ROBINS-I** or **GRACE** instead. PROBAST is relevant whether\nthe model was built on claims, EHR, registry, or prospectively collected data; it adapts because its\nsignalling questions probe the methodological fundamentals rather than a specific data type.\n\n**What it requires (checklist domains)** — PROBAST enforces appraisal across four domains with\n~20 signalling questions in total. *Domain 1 — Participants*: was the data source and sampling frame\nappropriate to the prediction task? Were inclusion/exclusion criteria pre-specified and applied\nconsistently? Is there risk of selection bias (e.g., case-control designs applied to inherently\ncohort-structured questions)? *Domain 2 — Predictors*: are all predictors well-defined and measured\nat the correct time relative to the prediction horizon? Is there risk that predictor measurement was\ninfluenced by knowledge of outcome (look-ahead bias)? Were candidate predictors not selected\npost-hoc based on univariable screening in ways that inflate the final model? *Domain 3 — Outcome*:\nis the outcome defined clearly and measured validly? Was outcome assessment blinded to predictor\nvalues, or is there differential measurement? Was follow-up time adequate and consistent with the\nprediction horizon? *Domain 4 — Analysis*: was sample size adequate (events-per-variable or\nevents-per-parameter)? Was missing data handled appropriately (not complete-case only)? Were\nstatistical methods for model building (including overfitting control via regularisation, bootstrap,\nor cross-validation) appropriate? Was **calibration** assessed in addition to discrimination? Was\nmodel performance assessed in an independent validation sample rather than just in the training data?\nEach domain is rated **low / high / unclear** risk of bias; overall bias is rated high if *any* domain\nis high (a conservative rule that reflects how one flawed domain can invalidate a model's usefulness).\n\n**When NOT to use — limitations and common misapplications** — PROBAST is an *appraisal* tool, not a\nreporting checklist; do not hand it to authors as a writing guide (use TRIPOD for that). It is also\n**not a numeric quality score**: tallying signalling-question \"lows\" into a sum and using that as a\nmeta-analytic weight is a methodological error — PROBAST supports structured domain-level judgments,\nnot arithmetic. Common failure modes: (1) **Wrong study type** — applying PROBAST to a comparative-\neffectiveness observational study (use ROBINS-I or GRACE) or a single diagnostic test (use QUADAS-2)\nrather than a multivariable prediction model. (2) **Confusing PROBAST and TRIPOD** — a study can\nbe fully TRIPOD-compliant (reported everything) and still be at high PROBAST risk of bias (e.g., if\nthe predictor selection was data-driven and overfit). (3) **Ignoring the calibration item** — Domain 4\nrequires that calibration was assessed; reviewers often accept an AUC/C-statistic alone and rate D4\nlow-risk inappropriately. (4) **Over-applying the \"unclear = low\" shortcut** — unclear risk should\nnot default to low risk when the information was not reported; absent calibration reporting or missing\nsample-size justification should drive uncertain or high ratings. (5) **Applicability vs bias\nconfusion** — a model may be at low risk of bias but highly inapplicable to the target population\n(different healthcare system, different case-mix, older predictor definitions); PROBAST's applicability\ndomain is meant to capture this and should not be collapsed into the bias assessment. (6) **Checklist\ntheater** — completing PROBAST forms without engaging the underlying methodological questions produces\nmisleading reassurance; a high Domain 4 rating because calibration was never assessed is a\nsubstantive finding, not a technicality.\n\n**How it maps to this catalog** — PROBAST's four domains map directly to the methodological concepts\nthat *implement* what PROBAST only judges. Domain 1 (Participants / data adequacy) is implemented by\n**fit-for-purpose-data-assessment-rwe** and **claims-analysis** (data-source fitness and case-mix\nadequacy). Domain 2 (Predictors — no look-ahead, pre-specified) is implemented by\n**time-zero-index-date-alignment-rwe** (ensuring predictors are measured before the prediction\nhorizon) and **diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe** (how predictor variables are\ndefined in administrative data). Domain 3 (Outcome validity) is implemented by\n**claims-outcome-algorithm-ppv-sensitivity-rwe** and **ehr-phenotyping-algorithms-rwe** (validated\noutcome algorithms with PPV/sensitivity). Domain 4 (Analysis — sample size, overfitting, calibration)\nis implemented by **sample-size-power-precision-rwe** (events-per-variable justification),\n**multiple-imputation-longitudinal-rwe** (appropriate missing-data handling), and\n**prediction-model-validation-recalibration-rwe** (internal/external validation, calibration,\nrecalibration). The discrimination and calibration reporting that PROBAST requires is visualised\nthrough **visualizations-pharmacoepidemiology-rwe** (calibration plots, ROC/AUC, decision curves).\nThe reporting obligation PROBAST assumes is fulfilled with **TRIPOD** (the companion guideline in this\ncatalog). Whenever a systematic review of prediction models is appraised with PROBAST, the evidence\nsynthesis itself should follow **PRISMA-2020** or **PRISMA-DTA** and the overall certainty judgment\nmay incorporate **GRADE** principles.", "primary_category": "Guideline", "tags": [ "guideline", "risk-of-bias", "critical-appraisal", "prediction-model", "prognostic-model", "diagnostic-model", "validation", "calibration", "equator" ], "aliases": [ "PROBAST", "Prediction model Risk Of Bias ASsessment Tool", "prediction model risk of bias" ], "applies_to_study_types": [ "algorithm_validation", "cohort_prospective", "cohort_retrospective", "diagnostic_accuracy" ], "data_sources": [ "claims", "ehr", "registry", "linked", "primary" ], "citations": [ { "role": "introduce", "doi": "10.7326/M18-1376", "url": "https://doi.org/10.7326/M18-1376", "citation_text": "Wolff RF, Moons KGM, Riley RD, et al. PROBAST: A Tool to Assess the Risk of Bias and Applicability of Prediction Model Studies. Annals of Internal Medicine. 2019;170(1):51-58.", "year": 2019, "authors_short": "Wolff et al.", "notes": "Statement paper introducing the PROBAST four-domain risk-of-bias and applicability instrument for prediction model studies; published concurrently with the Explanation & Elaboration paper." }, { "role": "explain", "doi": "10.7326/M18-1377", "url": "https://doi.org/10.7326/M18-1377", "citation_text": "Moons KGM, Wolff RF, Riley RD, et al. PROBAST: A Tool to Assess Risk of Bias and Applicability of Prediction Model Studies: Explanation and Elaboration. Annals of Internal Medicine. 2019;170(1):W1-W33.", "year": 2019, "authors_short": "Moons et al.", "notes": "Item-by-item explanation with worked examples of low and high risk-of-bias ratings; the authoritative companion to the statement paper for applying PROBAST in practice." }, { "role": "use", "url": "https://www.probast.org/wp-content/uploads/2020/02/PROBAST_20190515.pdf", "citation_text": "PROBAST — EQUATOR Network maintained entry (tool, companion papers, and related TRIPOD resources for prediction model appraisal).", "year": 2019, "authors_short": "EQUATOR Network", "notes": "Maintained PROBAST tool PDF with signaling questions, applicability domains, and companion-paper links." } ], "relations": [ { "relation_type": "see_also", "target_slug": "tripod", "notes": "TRIPOD is the companion reporting guideline — TRIPOD governs what is reported; PROBAST judges whether the underlying methods are at risk of bias." }, { "relation_type": "used_with", "target_slug": "prediction-model-validation-recalibration-rwe", "notes": "Implements Domain 4 analysis requirements — internal/external validation, calibration, and recalibration are the analytic actions PROBAST Domain 4 appraises." }, { "relation_type": "used_with", "target_slug": "claims-outcome-algorithm-ppv-sensitivity-rwe", "notes": "Implements Domain 3 outcome-validity requirement; a validated algorithm with PPV/sensitivity is what a low-risk Domain 3 rating should require." }, { "relation_type": "used_with", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "Implements Domain 1 data-source adequacy; fitness-for-purpose assessment answers the core PROBAST D1 questions." }, { "relation_type": "used_with", "target_slug": "time-zero-index-date-alignment-rwe", "notes": "Implements Domain 2 predictor-timing requirement — predictors must be measured at or before the prediction time point; misalignment is a D2 look-ahead-bias risk." }, { "relation_type": "see_also", "target_slug": "sample-size-power-precision-rwe", "notes": "Events-per-variable and events-per-parameter calculations are the Domain 4 sample-size signalling question." }, { "relation_type": "see_also", "target_slug": "multiple-imputation-longitudinal-rwe", "notes": "Appropriate missing-data handling is a Domain 4 signalling question; complete-case analysis without justification drives a high D4 rating." }, { "relation_type": "see_also", "target_slug": "visualizations-pharmacoepidemiology-rwe", "notes": "Calibration plots and discrimination curves implement the Domain 4 performance-reporting items PROBAST requires." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "journal", "fda", "ema", "hta" ] }, { "slug": "prospero", "name": "PROSPERO", "short_definition": "International Prospective Register of Systematic Reviews — a free, open registry (maintained by the Centre for Reviews and Dissemination, University of York) where authors prospectively record the protocol of a systematic review of human health-related research before data extraction, to deter outcome switching, duplication, and undisclosed protocol deviation.", "long_description": "**What it is.** PROSPERO is the **International Prospective Register of Systematic Reviews**,\nan open-access registry created in 2011 and maintained by the **Centre for Reviews and\nDissemination (CRD), University of York**, with early support from the UK National Institute\nfor Health Research. It is a *prospective protocol registry*, not a reporting checklist and\nnot a critical-appraisal instrument. Authors deposit the key methodological intentions of a\nplanned systematic review — the question, eligibility criteria, search strategy, outcomes,\nrisk-of-bias plan, and synthesis plan — *before* they begin formal screening and data\nextraction, and receive a permanent registration number and time-stamped public record.\nIts purpose is to make the review's a-priori plan discoverable and citable so that the\npublished review can be checked against what was promised, reducing unplanned outcome\nswitching, selective reporting, and unintended duplication of effort. Scope is restricted\nto systematic reviews with a **health-related outcome** (interventions, diagnosis,\nprognosis, etiology, screening) where the protocol has not yet progressed past the data\nextraction stage; reviews already completed cannot be registered. PROSPERO is part of the\nbroader transparency ecosystem alongside the EQUATOR reporting guidelines (PRISMA family),\nbut registration and reporting are distinct acts.\n\n**When to use.** Register on PROSPERO **before screening/extraction** whenever you are\nconducting a systematic review or meta-analysis of human health-related studies — including\nreviews of randomized trials, reviews of non-randomized/observational (real-world) studies,\nnetwork meta-analyses, and diagnostic-test-accuracy reviews. Registration is expected by\nmost peer-reviewed journals, is a PRISMA 2020 reporting item (the published review must state\nthe registry and ID), and strengthens HTA/payer evidence dossiers and FDA/EMA submissions\nthat lean on a systematic review of the comparative evidence. Decision rule for *which*\nregistry: if the deliverable is a **systematic review/meta-analysis of human studies**, use\nPROSPERO. If the deliverable is a **primary non-interventional/pharmacoepidemiologic study**\n(cohort, case-control, target-trial emulation, or a regulatory post-authorization safety\nstudy/PASS on claims, EHR, or registry data), PROSPERO is the wrong registry — register the\nstudy in the **EU PAS Register / HMA-EMA Catalogue of RWD studies** (or\nClinicalTrials.gov where appropriate). If the deliverable is an **interventional clinical\ntrial**, use ClinicalTrials.gov / EudraCT-CTIS. PROSPERO registers the *review of* such\nstudies, never the primary studies themselves.\n\n**What it requires.** PROSPERO is a structured registration record, not a checklist scored\nfor completeness; the substantive fields it enforces are the pre-specified methodological\ncommitments of the planned review: (1) **review question and PICO/eligibility** — population,\ninterventions/exposures, comparators, outcomes, and study designs to be included; (2) the\n**information sources and search strategy** (databases, dates, planned search terms); (3) the\n**primary and secondary outcomes**, stated a priori so that later outcome switching is\ndetectable; (4) the planned **risk-of-bias / quality-assessment tool** (e.g., RoB 2,\nROBINS-I, AMSTAR-2 for umbrella reviews) and how certainty will be graded (e.g., GRADE);\n(5) the **data-extraction and synthesis plan** (qualitative synthesis, meta-analysis model,\nplanned subgroup/sensitivity analyses, handling of heterogeneity); (6) administrative items —\nanticipated/actual start and completion dates, the review team and roles, funding sources,\nand conflicts of interest. For reviews of real-world data, the strongest records also\npre-specify how the *underlying* studies' data-fitness and design features (phenotype/\nalgorithm validity, time-zero alignment, confounding control, attrition) will be appraised\nand synthesized — but these are appraisal criteria applied to the included studies, not\nfields PROSPERO imposes on a primary analysis.\n\n**When NOT to use — limitations and common misapplications.**\n- **It is not a reporting guideline.** A PROSPERO record does not substitute for PRISMA-P\n (protocol reporting) or PRISMA 2020 (final-review reporting); a complete registration with\n a poorly reported manuscript still fails peer review.\n- **It is not a risk-of-bias or quality instrument.** Registering a review says nothing about\n its methodological quality — that is the job of AMSTAR-2 (appraising the SR) and\n ROBINS-I / RoB 2 (appraising the included studies). Treating the PROSPERO ID as a quality\n signal is a category error.\n- **Registration is not adherence.** Discrepancies between the registered plan and the\n published review (outcome switching, added/dropped analyses) are well documented; the value\n of registration is realized only when reviewers and editors *check the record against the\n paper*. Registration-as-theater — a record filed and ignored — provides no protection.\n- **Wrong registry for primary RWE.** Using PROSPERO to register a primary\n pharmacoepidemiology study, target-trial emulation, or a PASS is a scope error; those\n belong in the EU PAS Register / HMA-EMA RWD Catalogue. PROSPERO will not accept primary\n studies, only reviews of them.\n- **Too late to register.** A review that has already passed the data-extraction stage is\n not eligible for prospective registration, and back-filling a record after results are\n known defeats the purpose.\n\n**How it maps to this catalog.** PROSPERO governs the *systematic-review* layer, so its\nnatural neighbors are the SR/MA guideline family, not the primary-RWE design concepts.\nThe protocol-reporting requirements it presupposes are implemented by **prisma-p**; the\nfinal-review reporting (which must cite the PROSPERO ID) by **prisma-2020**; network\nmeta-analyses by **prisma-nma** and diagnostic-accuracy reviews by **prisma-dta**; the\nsearch-reporting field by **prisma-s**. The risk-of-bias / certainty plan that PROSPERO asks\nyou to pre-specify is operationalized by **robins-i** (non-randomized studies of\ninterventions), **amstar-2** (critical appraisal of the systematic review itself), and\n**grade** (certainty of evidence). When the *included* studies are real-world database\nanalyses, the appraisal you pre-register draws on the primary-data concepts — e.g.,\n**claims-analysis** for the data substrate, and the design quality of any\n**target-trial-emulation** included in the review. Applied note for claims/EHR/registry\nRWE: a systematic review or meta-analysis of observational drug-safety or comparative-\neffectiveness studies built on claims/EHR data should register in PROSPERO and, in the\nregistered protocol, commit to ROBINS-I for the included studies and GRADE for certainty —\nbut the primary database studies being synthesized are themselves registered in the EU PAS\nRegister, not PROSPERO.", "primary_category": "Guideline", "tags": [ "guideline", "protocol-registration", "systematic-review", "meta-analysis", "transparency", "equator-adjacent" ], "aliases": [ "PROSPERO", "International Prospective Register of Systematic Reviews", "PROSPERO registry", "CRD PROSPERO" ], "applies_to_study_types": [ "systematic_review", "meta_analysis_rct", "meta_analysis_obs", "network_meta_analysis" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "doi": "10.1186/2046-4053-1-2", "url": "https://doi.org/10.1186/2046-4053-1-2", "citation_text": "Booth A, Clarke M, Dooley G, Ghersi D, Moher D, Petticrew M, Stewart L. The nuts and bolts of PROSPERO: an international prospective register of systematic reviews. Systematic Reviews. 2012;1:2.", "year": 2012, "authors_short": "Booth et al.", "notes": "Canonical statement paper describing PROSPERO's rationale, scope (prospective registration of health-related systematic-review protocols), and the registration fields." }, { "role": "explain", "doi": "10.1016/j.jclinepi.2018.01.003", "url": "https://doi.org/10.1016/j.jclinepi.2018.01.003", "citation_text": "Sideri S, Papageorgiou SN, Eliades T. Registration in the international prospective register of systematic reviews (PROSPERO) of systematic review protocols was associated with increased review quality. Journal of Clinical Epidemiology. 2018;100:103-110.", "year": 2018, "authors_short": "Sideri et al.", "notes": "Empirical evidence that prospective registration is associated with higher review quality, and documentation of registered-vs-published discrepancies (registration is not adherence)." }, { "role": "use", "url": "https://www.crd.york.ac.uk/prospero/", "citation_text": "PROSPERO — International Prospective Register of Systematic Reviews. Centre for Reviews and Dissemination, University of York. Maintained registration platform and guidance.", "year": 2011, "authors_short": "CRD, University of York", "notes": "Authoritative live registry and submission guidance; eligibility, fields, and record management." } ], "relations": [ { "relation_type": "see_also", "target_slug": "prisma-p", "notes": "PRISMA-P governs how the systematic-review *protocol* is reported; PROSPERO is where that protocol is prospectively registered. Use both — register on PROSPERO, report per PRISMA-P." }, { "relation_type": "see_also", "target_slug": "prisma-2020", "notes": "The published review must state the registry and PROSPERO ID (a PRISMA 2020 item); reviewers check the manuscript against the registered record." }, { "relation_type": "see_also", "target_slug": "prisma-nma", "notes": "For network meta-analyses registered in PROSPERO, NMA-specific reporting follows PRISMA-NMA." }, { "relation_type": "see_also", "target_slug": "prisma-dta", "notes": "Diagnostic-test-accuracy reviews are registrable in PROSPERO; report per PRISMA-DTA." }, { "relation_type": "used_with", "target_slug": "robins-i", "notes": "The pre-specified risk-of-bias plan for reviews including non-randomized/observational studies is typically ROBINS-I; PROSPERO asks authors to commit to the tool a priori." }, { "relation_type": "used_with", "target_slug": "amstar-2", "notes": "AMSTAR-2 critically appraises the systematic review itself; registration on PROSPERO does not certify quality and is no substitute for AMSTAR-2 appraisal." }, { "relation_type": "used_with", "target_slug": "grade", "notes": "PROSPERO records should pre-specify how certainty of evidence will be graded; GRADE is the standard framework." }, { "relation_type": "see_also", "target_slug": "target-trial-emulation", "notes": "When a review synthesizes target-trial-emulation or other primary RWE studies, those primary studies register in the EU PAS Register, not PROSPERO; PROSPERO registers the review of them." }, { "relation_type": "used_with", "target_slug": "claims-analysis", "notes": "Applies when a PROSPERO-registered review synthesizes claims-/EHR-based observational studies; the protocol should pre-specify appraisal of the underlying data fitness-for-use." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "hta", "journal" ] }, { "slug": "record-nlp", "name": "RECORD-NLP (anticipated RECORD extension for NLP-derived data)", "short_definition": "An anticipated, not-yet-published extension of the RECORD reporting guideline intended to govern reporting of observational studies on routinely collected health data whose exposures, covariates, or outcomes are derived by natural language processing of clinical free text. No finalized checklist exists as of mid-2026; the defensible reporting stack today is RECORD/RECORD-PE plus algorithm-validation and NLP-specific practice reporting.", "long_description": "**What it is** — **RECORD-NLP** denotes an *anticipated* extension of the **RECORD**\nstatement (REporting of studies Conducted using Observational Routinely-collected health\nData; Benchimol et al., 2015), aimed at observational/real-world-evidence (RWE) studies in\nwhich exposures, covariates, eligibility criteria, or outcomes are derived from **natural\nlanguage processing (NLP) of clinical free text** (notes, discharge summaries, pathology and\nradiology reports). It would sit in the EQUATOR Network's STROBE→RECORD family, parallel to\nthe published pharmacoepidemiology extension **RECORD-PE** (Langan et al., 2018). **Critical\nstatus caveat: as of mid-2026 there is no finalized, published RECORD-NLP checklist.** No\nconsensus statement, no item list, and no EQUATOR-hosted maintained checklist exists under\nthat name. The closest *active*, EQUATOR-registered effort for clinical-text NLP reporting is\nthe **CINEX** guideline (Clinical Information EXtraction; development protocol Reichenpfader\net al., 2025), and the most usable *interim* practice recommendations for NLP-assisted\nobservational research are the Fu et al. (2023) scoping-review recommendations. This catalog\nentry is therefore a **forward pointer and a reporting-stack recommendation**, not a citation\nof an existing instrument: it tells you what to report *now* when NLP feeds an RWE study, and\nwhat guideline to watch for.\n\n**When to use** — Treat this entry as the lookup you reach for when a study using routinely\ncollected health data depends on NLP to define any analytic variable — a phenotype, an\noutcome, a comorbidity covariate, a severity measure, or an eligibility flag — and you must\nreport it for a peer-reviewed journal, an HTA/payer dossier, a regulatory (FDA/EMA, ENCePP)\nsubmission, or a registered protocol/PASS. Because the named checklist does not yet exist, the\ndecision rule is a *stacking* rule rather than a single-instrument choice: (1) report the\nobservational study itself against **RECORD** (or **RECORD-PE** if it is pharmacoepidemiologic);\n(2) report the NLP/algorithm-derived variables against the catalog's algorithm-validation and\nphenotyping concepts; (3) add the **Fu (2023)** NLP-specific practice items; and (4) if a\nmachine-learning or large-language model produces the variable, add a model-reporting standard\n(TRIPOD+AI or MI-CLAIM) for the development and validation of that model. Use RECORD-NLP-as-a-pointer\nprecisely when a reader or reviewer asks \"which guideline covers the NLP part?\" — the honest\nanswer is \"none is finalized; here is the stack that covers it.\"\n\n**What it requires** — *Anticipated domains, by analogy to RECORD/RECORD-PE plus the documented\nNLP-specific gaps; these are not items from a published RECORD-NLP checklist and should be read\nas the substance such an extension would have to cover, not as an enforceable list.* Beyond the\ngeneric RECORD items (data-source provenance, linkage, population/codes used, cleaning and\nflow), an NLP-aware reporting standard would compel: the **corpus and text source** (which\ndocument types, sections, time windows, and what fraction of the cohort had usable text — text\navailability is itself a selection mechanism); the **NLP method** (rule-based vs statistical vs\ntransformer/LLM; model name, version, training data, and whether it was developed in-sample or\nported from another institution); the **annotation and reference standard** (how gold-standard\nlabels were created, annotator agreement, the chart-review sample); **validation metrics with\nconfidence intervals** (PPV, sensitivity, specificity, F1) measured *in the target population\nand time period*, not borrowed from the model's origin study; **error analysis and\nmisclassification** characterization, including whether errors are differential by arm,\nsubgroup, site, or calendar time; **drift and portability** (performance decay over time and\nacross institutions/EHR versions); and how NLP-derived measurement error was propagated into\nthe analysis (quantitative bias analysis or sensitivity analyses). These layer onto the standard\nRWE requirements an extension would still inherit: time-zero alignment, estimands and\nintercurrent events, confounding control, and attrition/missing-data accounting.\n\n**When NOT to use — limitations and common misapplications** — (1) **Do not cite \"RECORD-NLP\"\nas a completed checklist** in any FDA/EMA/HTA/journal submission or protocol today; it does not\nexist as a finalized instrument, and claiming compliance with it is misrepresentation. Cite\nRECORD/RECORD-PE plus the NLP-specific stack instead, and, if you want a *named* NLP reporting\neffort, watch and reference CINEX. (2) A reporting checklist — actual or anticipated — is **not a\nrisk-of-bias instrument and not a quality score**; completing items does not make an NLP-derived\nobservational estimate valid, unconfounded, or causal. Bias in the *included* NLP study is judged\nwith ROBINS-I and with the algorithm-validation evidence, not by a reporting tally. (3) **Wrong\nguideline for the design**: a primary RWE study without NLP needs RECORD/RECORD-PE, not an NLP\nextension; a systematic review of NLP-RWE studies needs PRISMA-P/PRISMA 2020; a *trial* uses\nCONSORT/SPIRIT. (4) **Reporting-the-model is not reporting-the-study**: TRIPOD+AI or MI-CLAIM\ndescribe how the NLP model was built and validated, but they do not cover the epidemiologic design\nthat consumes its output — you need both. (5) **Checklist-as-theater**: stating \"NLP was used to\nidentify outcomes\" with no corpus description, no in-sample validation metrics, and no error\nanalysis is the exact gap these efforts exist to close; the value is the operational detail, not\nthe sentence.\n\n**How it maps to this catalog** — Because the named checklist is not yet available, the substance\nof an NLP-aware RECORD extension is implemented *today* by concepts in this repo, which together\nform the recommended reporting stack:\n- The NLP/algorithm-derived variables — the core of what RECORD-NLP would add — map to\n **ehr-phenotyping-algorithms-rwe** (NLP and structured phenotype construction),\n **outcome-algorithm-construction-rwe** (building outcome definitions),\n **algorithm-validation** (the chart-review/reference-standard validation an NLP variable demands),\n **claims-outcome-algorithm-ppv-sensitivity-rwe** (PPV/sensitivity reporting and how\n misclassification biases estimates), and **diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe**\n (the code/window logic an NLP signal is often blended with or compared against).\n- The inherited RWE-design requirements an extension would still enforce map to\n **target-trial-emulation** (design transparency and time-zero), **active-comparator-new-user**\n (confounding control and incident-user structure), **estimands-ate-att-intercurrent-events-rwe**\n (estimand and intercurrent-event specification),\n **high-dimensional-propensity-score-hdps-rwe** (confounding adjustment, including\n NLP-augmented covariates), and **attrition-and-loss-to-follow-up-rwe** (flow and missing data).\n- The data-substrate context maps to **claims-analysis** (and EHR linkage), where text\n availability and document capture govern whether an NLP variable can even be measured.\n\n**Applied note (claims/EHR/registry RWE).** NLP variables live almost entirely in **EHR** and\nlinked claims–EHR substrates: claims carry codes but no free text, so an NLP-derived phenotype\nis only measurable for the subset with available notes, and that subset is differentially\ncaptured (sicker, more-engaged, single-institution patients). A defensible report therefore states\nwhat fraction of the analytic cohort had usable text, validates the NLP variable against a\nchart-reviewed gold standard **in that cohort and calendar window** (PPV and sensitivity with CIs),\nchecks whether misclassification is differential across arms, and propagates the measurement error\nthrough a quantitative bias analysis — exactly the items a finalized RECORD-NLP would be expected to\nrequire, and exactly the items the algorithm-validation concepts in this catalog already operationalize.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "record", "nlp", "routinely-collected-health-data", "phenotyping", "equator", "in-development" ], "aliases": [ "RECORD-NLP", "RECORD extension for natural language processing", "RECORD NLP extension", "NLP reporting extension for routinely collected health data" ], "applies_to_study_types": [ "ehr_study", "claims_analysis", "linked_data" ], "data_sources": [ "ehr", "linked", "claims", "registry" ], "citations": [ { "role": "introduce", "url": "https://www.equator-network.org/library/reporting-guidelines-under-development/", "citation_text": "EQUATOR Network. Reporting guidelines under development (registry of in-progress reporting guidelines, including extensions for routinely collected health data and NLP/clinical information-extraction studies). Accessed 2026.", "year": 2026, "authors_short": "EQUATOR Network", "notes": "Honest stable reference for a not-yet-published guideline; no finalized RECORD-NLP checklist or statement paper exists as of mid-2026, so the EQUATOR under-development registry is the canonical pointer." }, { "role": "explain", "doi": "10.1371/journal.pmed.1001885", "url": "https://doi.org/10.1371/journal.pmed.1001885", "citation_text": "Benchimol EI, Smeeth L, Guttmann A, et al. The REporting of studies Conducted using Observational Routinely-collected health Data (RECORD) Statement. PLOS Medicine. 2015;12(10):e1001885.", "year": 2015, "authors_short": "Benchimol et al.", "notes": "Parent RECORD statement; the STROBE extension whose items an NLP extension would inherit and build upon." }, { "role": "explain", "doi": "10.1136/bmj.k3532", "url": "https://doi.org/10.1136/bmj.k3532", "citation_text": "Langan SM, Schmidt SAJ, Wing K, et al. The reporting of studies conducted using observational routinely collected health data statement for pharmacoepidemiology (RECORD-PE). BMJ. 2018;363:k3532.", "year": 2018, "authors_short": "Langan et al.", "notes": "Published RECORD extension for pharmacoepidemiology; the precedent and template for how a RECORD-NLP extension would be structured and adopted." }, { "role": "explain", "doi": "10.2196/76776", "url": "https://doi.org/10.2196/76776", "citation_text": "Reichenpfader D, Denecke K, et al. Improving the Reporting Quality of Studies on Information Extraction From Clinical Texts: Protocol for the Development of a Consensus-Based Reporting Guideline (CINEX). JMIR Research Protocols. 2025;14:e76776.", "year": 2025, "authors_short": "Reichenpfader et al.", "notes": "Closest active, EQUATOR-registered reporting-guideline development effort for clinical-text NLP/information-extraction studies; the most likely concrete form NLP-specific reporting standards will take." }, { "role": "use", "doi": "10.1111/cts.13463", "url": "https://doi.org/10.1111/cts.13463", "citation_text": "Fu S, Wen A, Schaeferle GM, et al. Recommended practices and ethical considerations for natural language processing-assisted observational research: A scoping review. Clinical and Translational Science. 2023;16(3):398-411.", "year": 2023, "authors_short": "Fu et al.", "notes": "Interim, usable practice recommendations for NLP-assisted observational research (corpus, validation, error analysis, generalizability); the reporting items to apply today while a named extension is in development." } ], "relations": [ { "relation_type": "is_variant_of", "target_slug": "record", "notes": "RECORD-NLP would be an extension of the parent RECORD statement for routinely collected health data, adding items specific to NLP-derived variables." }, { "relation_type": "see_also", "target_slug": "record-pe", "notes": "The published RECORD pharmacoepidemiology extension is the structural template; many NLP-RWE studies are pharmacoepidemiologic and should also follow RECORD-PE." }, { "relation_type": "used_with", "target_slug": "ehr-phenotyping-algorithms-rwe", "notes": "Implements the core NLP/phenotype-construction reporting an NLP extension would require." }, { "relation_type": "used_with", "target_slug": "algorithm-validation", "notes": "Supplies the chart-review/reference-standard validation evidence (PPV, sensitivity) that an NLP-derived variable must report; the substantive heart of NLP reporting." }, { "relation_type": "used_with", "target_slug": "outcome-algorithm-construction-rwe", "notes": "Governs how an NLP-derived outcome definition is built and documented." }, { "relation_type": "used_with", "target_slug": "claims-outcome-algorithm-ppv-sensitivity-rwe", "notes": "PPV/sensitivity reporting and the propagation of NLP misclassification into effect estimates." }, { "relation_type": "see_also", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Code/window phenotype logic that NLP signals are commonly compared against or combined with." }, { "relation_type": "see_also", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "The estimand and intercurrent-event specification an NLP-RWE study still owes regardless of how variables are derived." }, { "relation_type": "see_also", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Cohort flow and missing-data accounting, including text-availability as a selection mechanism." }, { "relation_type": "see_also", "target_slug": "claims-analysis", "notes": "Data-substrate context; claims lack free text, so NLP variables require EHR or linked data and a measurable text-available subset." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "journal", "fda", "ema", "hta" ] }, { "slug": "record-pe", "name": "RECORD-PE (RECORD for Pharmacoepidemiology)", "short_definition": "Reporting guideline that extends RECORD (and, through it, STROBE) with pharmacoepidemiology-specific items for transparent reporting of treatment-effect and drug-utilization studies conducted using routinely collected health data such as claims, EHR, and registries.", "long_description": "**What it is** — **RECORD-PE (REporting of studies Conducted using Observational Routinely-collected health\nData for PharmacoEpidemiology)** is a reporting checklist that layers pharmacoepidemiology-specific items onto\nthe parent **RECORD** statement, which in turn extends **STROBE**, the base reporting guideline for observational\nepidemiology. The lineage matters operationally: STROBE (von Elm et al., 2007) defines the 22-item backbone for\ncohort, case-control, and cross-sectional studies; RECORD (Benchimol et al., 2015) adds items for studies that\nreuse routinely collected data (databases, linkage, code lists, data cleaning, population definition); and\nRECORD-PE (Langan et al., BMJ 2018) adds 15 PE-specific items and sub-items that address how drug exposure,\ncomparators, follow-up, and confounding are defined and reported. It is a *reporting* tool — it governs what a\ncompleted manuscript must transparently disclose, not how to design or appraise the study. It is published in\nthe BMJ and maintained as a RECORD/STROBE extension within the **EQUATOR Network**, developed with the\nInternational Society for Pharmacoepidemiology (**ISPE**).\n\n**When to use** — Apply RECORD-PE when **reporting a completed pharmacoepidemiology study that uses routinely\ncollected data** — comparative drug safety/effectiveness cohorts (e.g., active-comparator new-user designs),\ndrug-utilization studies, and database analyses in claims, EHR, registries, or linked sources — for a\npeer-reviewed journal, an HTA/payer evidence dossier, or a regulatory (FDA RWE, EMA) submission package. Use it\nalongside the **STROBE/RECORD** flow diagram and tables, since RECORD-PE does not replace its parents — a\ncompliant manuscript satisfies STROBE, RECORD, *and* the RECORD-PE additions together. Decision rules for\nchoosing the right family member: a **randomized trial** is reported with **CONSORT**, not RECORD-PE; a\n**non-pharmacoepidemiologic** observational study using routine data (e.g., a health-services or surveillance\nstudy with no drug-exposure contrast) is reported with **RECORD** (or plain STROBE) without the PE layer; a\n**systematic review/meta-analysis** is reported with PRISMA and its protocol with PRISMA-P; and **protocol-stage**\npre-specification of a single PE study belongs to **HARPER** (Wang et al., 2022) or the **ENCePP Checklist**, not\nto RECORD-PE, which is a *reporting* (post-hoc) instrument.\n\n**What it requires** — RECORD-PE forces transparent reporting of exactly the design choices that, left vague,\nlet bias hide in a routine-data study. Its substantive domains include: (1) **Data source and fitness-for-use** —\nnaming the database(s), the population they capture, linkage, the time period, and known limitations of the data\nfor the question. (2) **Exposure definition** — how drug exposure was operationalized from dispensing/prescribing\nrecords (code lists, days-supply, grace periods, stockpiling, exposure windows) and whether definitions were\nvalidated. (3) **Comparator and design** — the comparator group and rationale (active comparator vs non-user),\nnew-user vs prevalent-user status, and how the design controls confounding by indication. (4) **Time-zero / index\ndate alignment** — how follow-up start was defined for all groups so that immortal time and post-baseline\nadjustment are avoided. (5) **Outcome and covariate phenotypes** — algorithm/code-list definitions and any\nvalidation (PPV/sensitivity). (6) **Confounding control** — measured confounders, covariate assessment windows,\nand the analytic method (e.g., propensity scores, high-dimensional PS). (7) **Estimand and analysis** — the causal\ncontrast, treatment strategies, intercurrent-event handling, censoring rules, competing risks. (8) **Attrition,\nmissing data, and sensitivity / quantitative bias analysis** — cohort-derivation/attrition reporting, handling of\nmissingness, and pre-specified sensitivity and negative-control analyses. A recurring RECORD-PE expectation is the\n**public availability of code lists and algorithms**, so that exposure, outcome, and covariate definitions are\nreproducible.\n\n**When NOT to use — limitations and common misapplications** — RECORD-PE is a *reporting checklist*, and most\nfailures come from treating it as something it is not. (1) **It is not a risk-of-bias instrument** — completing\nRECORD-PE tells a reader what you did, not whether it was valid; appraise non-randomized studies with **ROBINS-I**,\nnot RECORD-PE. (2) **It is not a quality score** — there is no total, no threshold, and no \"RECORD-PE score\";\nmanufacturing one misrepresents the guideline. (3) **Checklist-as-theater** — a fully checked manuscript whose\nexposure window, time-zero rule, or confounding strategy is still described in one vague sentence has missed the\npoint; the deliverable is transparent, reproducible detail (and public code lists), not a completed table. (4)\n**A complete checklist does not make an observational study causal or unconfounded** — transparency is necessary,\nnot sufficient; residual confounding, selection bias, and unfit data survive a fully compliant report. (5) **Wrong\nfamily member / wrong extension** — using plain **STROBE** where the PE-specific items are required, using\n**RECORD-PE** to report a randomized trial (use CONSORT) or a non-PE routine-data study (use RECORD), or using a\n*reporting* checklist where a *protocol* template (HARPER, ENCePP) is what the regulator or HTA body actually asked\nfor, are all common and visible errors.\n\n**How it maps to this catalog** — RECORD-PE is the reporting layer; the concepts in this repo are what actually\n*implement* each of its requirements, and a reviewer should read them as the substantive backing for each checklist\nitem:\n- **Design and time-zero** (comparator, new-user status, immortal-time avoidance) → **active-comparator-new-user**\n and, for the trial-protocol framing of the whole study, **target-trial-emulation**.\n- **Exposure/outcome/covariate phenotypes and their validation** → **diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe**\n and **algorithm-validation** (the code lists and PPV/sensitivity evidence RECORD-PE asks you to disclose).\n- **Confounding control** → **high-dimensional-propensity-score-hdps-rwe** (and the PS balancing referenced within\n active-comparator-new-user).\n- **Estimand, treatment strategies, intercurrent events** → **estimands-ate-att-intercurrent-events-rwe**.\n- **Structured question / eligibility spine** → **picots-framework-rwe**.\n- **Data fitness-for-use and source choice** → **fit-for-purpose-data-assessment-rwe**,\n **medicare-ffs-ma-commercial-claims-differences-rwe**, and the general **claims-analysis** patterns.\n- **Attrition, missing data, and sensitivity/quantitative bias analysis** → **attrition-and-loss-to-follow-up-rwe**\n and **e-value-sensitivity-analysis**; external validity of the reported result → **generalizability-transportability-external-validity-rwe**.\n\n**Applied note (claims/EHR/registry RWE).** For a Medicare/commercial claims comparative cohort, RECORD-PE\ncompliance means the manuscript states the database and enrollment requirements (and, e.g., that Medicare\nAdvantage-only person-time was excluded because fee-for-service claims are unavailable), publishes the NDC/diagnosis\ncode lists used for exposure and outcome phenotypes with their validation metrics, makes the index-date/time-zero\nrule explicit so immortal time is auditable, reports the attrition funnel from source population to analytic cohort,\nand presents the confounding strategy (e.g., high-dimensional PS) with balance diagnostics and pre-specified\nsensitivity analyses (washout length, grace period, negative-control outcome). The reporting is the visible surface;\nthe catalog concepts above are where the methods themselves live.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "pharmacoepidemiology", "routinely-collected-data", "rwe", "equator", "strobe-extension" ], "aliases": [ "RECORD-PE", "RECORD for Pharmacoepidemiology", "REporting of studies Conducted using Observational Routinely-collected health Data for PharmacoEpidemiology" ], "applies_to_study_types": [ "new_user", "active_comparator_new_user", "claims_analysis", "ehr_study", "drug_utilization" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "doi": "10.1136/bmj.k3532", "url": "https://doi.org/10.1136/bmj.k3532", "citation_text": "Langan SM, Schmidt SAJ, Wing K, et al. The reporting of studies conducted using observational routinely collected health data statement for pharmacoepidemiology (RECORD-PE). BMJ. 2018;363:k3532.", "year": 2018, "authors_short": "Langan et al.", "notes": "Canonical RECORD-PE statement; defines the pharmacoepidemiology-specific reporting items layered onto RECORD." }, { "role": "explain", "doi": "10.1371/journal.pmed.1001885", "url": "https://doi.org/10.1371/journal.pmed.1001885", "citation_text": "Benchimol EI, Smeeth L, Guttmann A, et al. The REporting of studies Conducted using Observational Routinely-collected health Data (RECORD) Statement. PLOS Medicine. 2015;12(10):e1001885.", "year": 2015, "authors_short": "Benchimol et al.", "notes": "Parent RECORD statement that RECORD-PE extends; defines the routine-data reporting items (code lists, linkage, population/data cleaning)." }, { "role": "explain", "doi": "10.1371/journal.pmed.0040296", "url": "https://doi.org/10.1371/journal.pmed.0040296", "citation_text": "von Elm E, Altman DG, Egger M, et al. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement: guidelines for reporting observational studies. PLOS Medicine. 2007;4(10):e296.", "year": 2007, "authors_short": "von Elm et al.", "notes": "Grandparent STROBE statement; the 22-item observational-reporting backbone on which RECORD and RECORD-PE build." }, { "role": "explain", "doi": "10.1002/pds.5507", "url": "https://doi.org/10.1002/pds.5507", "citation_text": "Wang SV, Pottegård A, Crown W, et al. HARmonized Protocol Template to Enhance Reproducibility of hypothesis evaluating real-world evidence studies on treatment effects (HARPER). Pharmacoepidemiology and Drug Safety. 2023;32(1):44-55.", "year": 2023, "authors_short": "Wang et al.", "notes": "Protocol-stage companion; clarifies the protocol-vs-reporting boundary - HARPER pre-specifies a PE study, RECORD-PE reports the completed one." }, { "role": "use", "url": "https://www.equator-network.org/reporting-guidelines/record-pe/", "citation_text": "RECORD-PE reporting guideline. EQUATOR Network reporting-guidelines library (maintained checklist and downloadable items).", "year": 2018, "authors_short": "EQUATOR Network", "notes": "Maintained landing page with the RECORD-PE checklist in usable form and links to RECORD/STROBE." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "new-user-design", "notes": "RECORD-PE governs transparent reporting of new-user pharmacoepidemiology studies in routine data." }, { "relation_type": "applies_to", "target_slug": "active-comparator-new-user", "notes": "The reporting standard for active-comparator new-user comparative effectiveness/safety studies." }, { "relation_type": "applies_to", "target_slug": "claims-analysis", "notes": "Applies to claims-based PE analyses; requires disclosure of data fitness, code lists, and time-zero rules." }, { "relation_type": "applies_to", "target_slug": "ehr-study", "notes": "Applies to EHR-based PE studies, including phenotype/algorithm definitions and validation." }, { "relation_type": "applies_to", "target_slug": "drug-utilization", "notes": "Applies to drug-utilization studies using routinely collected prescribing/dispensing data." }, { "relation_type": "complements", "target_slug": "active-comparator-new-user", "notes": "Implements RECORD-PE's comparator/new-user and time-zero reporting items; the design behind the contrast." }, { "relation_type": "complements", "target_slug": "target-trial-emulation", "notes": "Frames the whole PE study as an emulated trial, sharpening the estimand and design items RECORD-PE asks to report." }, { "relation_type": "complements", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Implements the exposure/outcome/covariate phenotype definitions whose code lists RECORD-PE requires." }, { "relation_type": "complements", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "Implements the confounding-control reporting item (PS construction, covariate windows, balance diagnostics)." }, { "relation_type": "complements", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Implements the estimand / treatment-strategy / intercurrent-event reporting RECORD-PE demands." }, { "relation_type": "complements", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Implements the cohort-derivation/attrition and follow-up reporting items." }, { "relation_type": "see_also", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "Backs the data-source fitness-for-use disclosures RECORD-PE expects." }, { "relation_type": "see_also", "target_slug": "claims-analysis", "notes": "General claims patterns underpinning the data-source and operational-definition reporting." }, { "relation_type": "see_also", "target_slug": "picots-framework-rwe", "notes": "Structures the eligibility/question spine the report must state." }, { "relation_type": "see_also", "target_slug": "e-value-sensitivity-analysis", "notes": "Supports the sensitivity / quantitative-bias-analysis reporting item." }, { "relation_type": "see_also", "target_slug": "algorithm-validation", "notes": "Supplies the phenotype/algorithm validation metrics (PPV, sensitivity) RECORD-PE asks authors to report." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "hta", "journal" ] }, { "slug": "record", "name": "RECORD", "short_definition": "The REporting of studies Conducted using Observational Routinely-collected health Data (RECORD) Statement: an EQUATOR-registered reporting extension of STROBE that adds 13 routinely-collected-data items (data sources, codes and algorithms, validation, linkage, population selection, flow diagram) for studies using EHR, administrative claims, and disease registries.", "long_description": "**What it is** — RECORD (REporting of studies Conducted using Observational Routinely-collected health Data) is a\nreporting guideline published in 2015 (Benchimol et al., *PLOS Medicine*) that extends the STROBE statement for\nobservational research using data not collected for research purposes: electronic health records, administrative\nclaims, disease and product registries, and linkages among them. It is **not a new checklist** but an *extension* —\nauthors must satisfy the 22 STROBE items **and** the RECORD additions. The RECORD items target the failure points\nspecific to routinely-collected data: explicit naming of the data source(s) in the title/abstract (RECORD 1.1–1.3),\nthe population-selection process and the codes/algorithms used to define populations, exposures, outcomes, and\nconfounders, with their *validation* (RECORD 6.1–6.3, 7.1), data linkage and the linkage quality assessment\n(RECORD 12.1–12.3), a population flow diagram, and statements on data cleaning, access, and availability of code\nlists. RECORD is maintained by its author group and hosted in the EQUATOR Network library\n(record-statement.org). The principal published extension is **RECORD-PE** (Langan et al., 2018, *BMJ*) for\npharmacoepidemiology studies of medication effects. A reporting extension for NLP-derived variables (provisionally\n\"RECORD-NLP\") has been proposed but **is not yet a finalized published checklist** as of 2026 — treat it as\nanticipated guidance, not an established standard.\n\n**When to use** — Reach for RECORD whenever a non-interventional study analyses *routinely-collected* health data and\nthe output is destined for a peer-reviewed journal, an HTA/payer dossier, a regulatory submission (FDA RWE program,\nEMA), or a registered protocol. It is the reporting backbone for claims, EHR, registry, linked, and multi-database\nstudies. Decision rules for which member of the family applies:\n- If the study evaluates the **effect of a medication, vaccine, or other intervention** (comparative safety or\n effectiveness, drug utilisation), use **RECORD-PE** — it adds pharmacoepidemiology items (exposure windows,\n new-user/active-comparator design, time-zero, washout) that plain RECORD does not enforce.\n- If key variables (phenotypes, outcomes) are extracted from **clinical free text via natural-language processing**,\n document the text source, NLP model/method, and validation of the derived phenotype (the proposed RECORD-NLP\n extension is not yet finalized — until it is, report these elements under RECORD/RECORD-PE plus the relevant\n NLP-validation literature).\n- If the data were collected *for* research (prospective cohort, primary-data registry trial, surveys), plain\n **STROBE** (or the relevant STROBE extension) suffices — RECORD's data-provenance items add little.\n- RECORD is a *reporting* tool. For design pre-specification of an RWD study use a protocol template (HARPER,\n STaRT-RWE, ENCePP Checklist); for HTA reference-case alignment use NICE/CADTH frameworks. RECORD governs *how you\n report what you did*, not how you design it.\n\n**What it requires** — The substantive item clusters RECORD enforces, framed for real-world data:\n- **Data-source transparency** (RECORD 1.1, 6.1–6.2): name every database, the dates and geographic/health-system\n coverage, and the population-selection cascade from source to analytic cohort.\n- **Data fitness-for-use** (RECORD 6.3, 13): completeness, representativeness, and the cleaning/validation steps —\n the reader must be able to judge whether the data can answer the question.\n- **Phenotype / algorithm transparency and validation** (RECORD 6.1, 7.1): a *complete* list of codes and algorithms\n used to define populations, exposures, outcomes, and confounders, with references to validation (PPV, sensitivity)\n where available. Code lists should be made available.\n- **Population flow and attrition** (RECORD 13.1): a flow diagram from the source population through each eligibility\n and exclusion step to the analytic sample.\n- **Data linkage and its quality** (RECORD 12.1–12.3): linkage methods, the proportion linked, and the impact of\n incomplete linkage on the study population.\n- **Access and reproducibility** (RECORD 22.1): how others could access the data/code and any approvals required.\nRECORD does *not* itself prescribe estimands, confounding control, or sensitivity analysis — but because it requires\nyou to *report* exposure/outcome definitions, time windows, and analytic decisions transparently, applying it well\nforces the underlying RWE methods (time-zero alignment, active-comparator design, hdPS, quantitative bias analysis)\nto be specified and defended.\n\n**When NOT to use — limitations and common misapplications**\n- **RECORD is a reporting checklist, not a risk-of-bias instrument and not a quality score.** Ticking every item\n certifies *completeness of reporting*, not *validity*. A perfectly RECORD-compliant paper can still be hopelessly\n confounded. For critical appraisal use ROBINS-I, the Newcastle-Ottawa Scale, or ISPE/ISPOR good-practice\n recommendations — not RECORD.\n- **Completing the checklist does not make an observational study causal.** Transparent reporting of an\n immortal-time-biased or prevalent-user design is still a biased design, fully reported.\n- **Using STROBE alone for routinely-collected data.** STROBE omits the data-provenance, code-list, validation, and\n linkage items that are the whole point of RECORD; a claims/EHR study reported only to STROBE will be sent back by\n informed reviewers.\n- **Using plain RECORD where RECORD-PE is required.** A comparative drug-safety study reported to RECORD but not\n RECORD-PE will under-report exposure definition, new-user/active-comparator design, time-zero, and washout — the\n items that determine whether the comparison is interpretable.\n- **Wrong extension for the design** (e.g., RECORD where an NLP-derived phenotype demands RECORD-NLP).\n- **Checklist-as-theater.** Appending a completed RECORD table to a manuscript when the protocol never pre-specified\n the code lists, validation, or analytic decisions is retrofitting, not transparency. Pre-specify in the protocol\n (HARPER/STaRT-RWE), then report against RECORD.\n\n**How it maps to this catalog** — Each RECORD requirement is implemented by one or more concepts in this repo:\n- *Data fitness (RECORD 6.1–6.3, 13)* → `fit-for-purpose-data-assessment-rwe` and `database-feasibility-attrition-funnel-rwe`;\n payer-specific completeness via `medicare-ffs-ma-commercial-claims-differences-rwe` and `claims-analysis`.\n- *Population/exposure/outcome codes and algorithms with validation (RECORD 6.1, 7.1)* →\n `diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe`, `claims-outcome-algorithm-ppv-sensitivity-rwe`,\n `ehr-phenotyping-algorithms-rwe`, and `algorithm-validation`.\n- *Time-zero / index-date alignment and washout (reported under design/exposure)* →\n `time-zero-index-date-alignment-rwe`, `washout-clean-lookback-period-rwe`, `active-comparator-new-user`.\n- *Estimands and intercurrent events* → `estimands-ate-att-intercurrent-events-rwe`; design emulation via\n `target-trial-emulation`.\n- *Confounding control to report* → `high-dimensional-propensity-score-hdps-rwe`, `propensity-score-methods-psm-iptw`.\n- *Population flow / attrition (RECORD 13.1)* → `attrition-and-loss-to-follow-up-rwe` and\n `continuous-enrollment-observable-time-rwe`.\n- *Data linkage and its quality (RECORD 12.1–12.3)* → `linked-data` and, for the mother-infant case, `mother-infant-linkage-rwe`.\n- *Sensitivity / quantitative bias analysis to report* → `e-value-sensitivity-analysis` and\n `quantitative-bias-analysis-toolkit-rwe`.\n- *Reporting visuals* → `visualizations-pharmacoepidemiology-rwe` for the population flow diagram and balance/diagnostic plots.\n\n**Applied note (claims/EHR/registry RWE).** In a Medicare + commercial claims comparative-safety study, RECORD\ncompliance means: name each database and its dates and benefit type in the abstract; report the exact NDC/ICD/CPT\ncode lists and the phenotype rules (e.g., 1 inpatient or 2 outpatient diagnoses within a window) for exposures,\noutcomes, and confounders, with PPV/sensitivity references; show the population flow from enrollees through\ncontinuous-enrollment and washout filters to the analytic cohort; and, because it is a *drug* study, report against\n**RECORD-PE** so that the new-user/active-comparator structure, time-zero, and exposure windows are explicit. Where\nMedicare Advantage encounter completeness or fee-for-service claim capture affects the cohort, state it under data\nfitness — that disclosure is exactly what RECORD 6.3 exists to elicit.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "rwe", "routinely-collected-data", "strobe-extension", "equator" ], "aliases": [ "RECORD", "RECORD Statement", "REporting of studies Conducted using Observational Routinely-collected health Data", "STROBE-RECORD" ], "applies_to_study_types": [ "claims_analysis", "ehr_study", "linked_data", "multi_database", "cohort_retrospective", "disease_registry" ], "data_sources": [ "claims", "ehr", "registry", "linked", "multi-database" ], "citations": [ { "role": "introduce", "doi": "10.1371/journal.pmed.1001885", "url": "https://doi.org/10.1371/journal.pmed.1001885", "citation_text": "Benchimol EI, Smeeth L, Guttmann A, et al. The REporting of studies Conducted using Observational Routinely-collected health Data (RECORD) Statement. PLOS Medicine. 2015;12(10):e1001885.", "year": 2015, "authors_short": "Benchimol et al.", "notes": "The canonical RECORD statement and its 13 STROBE-extension items for routinely-collected health data; the explanation-and-elaboration document accompanies it." }, { "role": "explain", "doi": "10.1371/journal.pmed.0040296", "url": "https://doi.org/10.1371/journal.pmed.0040296", "citation_text": "von Elm E, Altman DG, Egger M, et al. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement: guidelines for reporting observational studies. PLoS Medicine. 2007;4(10):e296.", "year": 2007, "authors_short": "von Elm et al.", "notes": "The parent guideline. RECORD is an extension of STROBE; authors must satisfy STROBE items 1-22 in addition to the RECORD items." }, { "role": "explain", "doi": "10.1136/bmj.k3532", "url": "https://doi.org/10.1136/bmj.k3532", "citation_text": "Langan SM, Schmidt SAJ, Wing K, et al. The reporting of studies conducted using observational routinely collected health data statement for pharmacoepidemiology (RECORD-PE). BMJ. 2018;363:k3532.", "year": 2018, "authors_short": "Langan et al.", "notes": "The pharmacoepidemiology extension. Required instead of plain RECORD for studies of medication or vaccine effects; adds exposure-window, design, and time-zero items." }, { "role": "use", "url": "https://www.record-statement.org/", "citation_text": "RECORD Statement, EQUATOR Network — maintained checklists, extensions (RECORD-PE, RECORD-NLP), and guidance.", "year": 2015, "authors_short": "RECORD Working Committee", "notes": "Maintained source for the current checklist, the explanation document, and downloadable item tables." } ], "relations": [ { "relation_type": "see_also", "target_slug": "record-pe", "notes": "Pharmacoepidemiology extension; required instead of plain RECORD for medication/vaccine effect studies." }, { "relation_type": "see_also", "target_slug": "record-nlp", "notes": "Extension for studies whose variables are derived from clinical free text via NLP." }, { "relation_type": "see_also", "target_slug": "strobe", "notes": "Parent guideline; RECORD adds 13 items to the 22 STROBE items and cannot be applied without them." }, { "relation_type": "see_also", "target_slug": "harper", "notes": "Use HARPER to pre-specify the RWD study protocol; report the executed study against RECORD." }, { "relation_type": "see_also", "target_slug": "start-rwe", "notes": "STaRT-RWE structured template complements RECORD by pre-specifying exposure, outcome, and design tables that RECORD then requires to be reported." }, { "relation_type": "see_also", "target_slug": "encepp-checklist", "notes": "ENCePP methodological checklist supports protocol-stage design decisions that RECORD reporting later documents." }, { "relation_type": "see_also", "target_slug": "fda-rwe-noninterventional", "notes": "FDA guidance on non-interventional studies; RECORD/RECORD-PE supply the reporting transparency such submissions expect." }, { "relation_type": "requires", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "Implements RECORD 6.3/13 — completeness, representativeness, and fitness-for-use of the data source." }, { "relation_type": "requires", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Implements RECORD 6.1/7.1 — the code lists and phenotype rules defining populations, exposures, and outcomes." }, { "relation_type": "requires", "target_slug": "algorithm-validation", "notes": "Implements the RECORD 7.1 validation expectation (PPV, sensitivity) for codes and algorithms." }, { "relation_type": "requires", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Implements RECORD 13.1 — the population flow diagram from source to analytic cohort." }, { "relation_type": "used_with", "target_slug": "claims-outcome-algorithm-ppv-sensitivity-rwe", "notes": "Provides the outcome-algorithm validation metrics RECORD 7.1 asks authors to report." }, { "relation_type": "used_with", "target_slug": "linked-data", "notes": "Implements RECORD 12.1-12.3 — linkage method, linkage rate, and impact of incomplete linkage." }, { "relation_type": "used_with", "target_slug": "active-comparator-new-user", "notes": "Design whose exposure definition and time-zero RECORD-PE requires to be reported explicitly." }, { "relation_type": "used_with", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "Confounding-control method whose covariate construction RECORD requires to be transparently reported." }, { "relation_type": "used_with", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "The estimand and intercurrent-event handling RECORD-PE expects to be stated." }, { "relation_type": "used_with", "target_slug": "visualizations-pharmacoepidemiology-rwe", "notes": "Supplies the population flow diagram and balance/diagnostic figures RECORD reporting calls for." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "hta", "journal" ] }, { "slug": "robins-e", "name": "ROBINS-E (Risk Of Bias In Non-randomized Studies - of Exposures)", "short_definition": "A structured, signalling-question risk-of-bias instrument for appraising the effect estimate from a non-randomized follow-up (cohort) study of an exposure, adapting the ROBINS-I architecture to exposure questions (environmental, occupational, nutritional) rather than interventions; maintained by the ROBINS-E Development Group.", "long_description": "**What it is** — **ROBINS-E (Risk Of Bias In Non-randomized Studies - of Exposures)** is a\ndomain-based, signalling-question critical-appraisal tool that assesses the risk of bias in the\n*effect estimate* reported by a non-randomized **follow-up (cohort) study of an exposure**. It is\nthe exposures sibling of ROBINS-I (interventions): it inherits ROBINS-I's \"emulated target trial\"\nlogic — judge the observational study against the hypothetical randomized experiment it is trying\nto approximate — but rewrites the domains and signalling questions for the realities of exposure\nresearch, where the \"exposure\" is an environmental, occupational, dietary, or other non-prescribed\nagent that no one assigned. It is maintained by the **ROBINS-E Development Group** (an international\ncollaboration including the ROBINS-I authors and environmental-health methodologists, hosted at\nriskofbias.info) and was developed under the program to adapt **GRADE for environmental health**;\nthe launch tool was described by Higgins, Morgan, Rooney, Taylor, Thayer and colleagues in\n*Environment International* (2024). ROBINS-E is a **risk-of-bias instrument**, not a reporting\nchecklist and not a quality score: its output is a per-domain and overall judgment (Low risk / Some\nconcerns / High risk / Very high risk) for a *specified result*, structured around a clearly stated\n**PECO** (Population, Exposure, Comparator, Outcome) question.\n\n**When to use** — Apply ROBINS-E when you are appraising, in a systematic review or evidence\nsynthesis, a **non-randomized cohort/follow-up study estimating the causal effect of an exposure on\na health outcome**, and you need a transparent, reproducible bias assessment to feed a GRADE\ncertainty-of-evidence rating. Its decision context is environmental and occupational health risk\nassessment (e.g., EPA/IRIS, NTP, IARC monographs, EFSA opinions), nutritional epidemiology reviews,\nand Cochrane-style reviews of exposure questions, plus the peer-reviewed reviews that underpin them.\nThe governing decision rule for choosing the right sibling instrument: if the \"exposure\" is a\n**therapeutic intervention** (a drug, device, procedure, or program someone decided to give),\nappraise it with **ROBINS-I**, not ROBINS-E. If it is a non-assigned exposure (air pollution, PFAS,\nsilica dust, a dietary pattern, a behavior), ROBINS-E is the tool. ROBINS-E targets **follow-up\ndesigns**; it is not built for case-control or cross-sectional designs, which fall outside its\ncurrent scope and require other appraisal approaches.\n\n**What it requires** — ROBINS-E first fixes a **PECO and a specific numerical result** to be\nappraised (you assess a *result*, not a *study*), then works through **seven bias domains**, each\ndriven by signalling questions answered Yes / Probably yes / Probably no / No / No information: (1)\n**bias due to confounding** — were the important confounders of the exposure-outcome relationship\nidentified and adequately controlled, given that exposure was not randomized; (2) **bias in\nmeasurement of the exposure** — was exposure assessed validly and reliably, and was assessment\ndifferential with respect to the outcome (recall, exposure-misclassification, and the validity of\nthe exposure metric); (3) **bias in selection of participants into the study** — selection related\njointly to exposure and outcome, including selection at or after the start of follow-up; (4) **bias\ndue to post-exposure interventions** — actions taken after exposure that differ by exposure level\nand affect the outcome; (5) **bias due to missing data** — missingness in exposure, outcome, or\nconfounders and whether it could distort the estimate; (6) **bias in measurement of the outcome** —\noutcome ascertainment validity and whether it was differential by exposure; and (7) **bias in\nselection of the reported result** — selective reporting from multiple measurements, analyses, or\nsubgroups. Each domain rolls up to a domain-level judgment, and the domains combine (worst-domain-\ndominant logic, with a distinct \"Very high\" tier) into an overall risk-of-bias rating for that\nresult. The tool also asks the assessor to record the predicted **direction** of each bias, which\nfeeds the downstream GRADE judgment.\n\n**When NOT to use — limitations and common misapplications** — (1) **Wrong sibling instrument\n(the dominant error).** Using ROBINS-E to appraise a *drug/intervention* cohort — e.g., a claims-\nbased active-comparator new-user study of two antidiabetics — is a category error; that is the\n**ROBINS-I** lane. Conversely, forcing ROBINS-I onto an environmental exposure study misframes the\nconfounding and exposure-measurement domains. (2) **Wrong design.** ROBINS-E is for follow-up\nstudies; applying it to case-control or cross-sectional studies stretches it past its validated\nscope. (3) **A risk-of-bias tool is not a reporting checklist** — a study can be beautifully\nreported (STROBE/RECORD-compliant) and still be High risk in ROBINS-E, and vice versa; do not\nsubstitute one for the other. (4) **It is not a numeric quality score.** ROBINS-E deliberately\navoids summing items into a scale; converting domain ratings into points and averaging them\ndiscards the worst-domain logic the tool is built on. (5) **Result-level, not study-level.** A\nsingle paper can yield Low risk for one outcome and High risk for another; assessing \"the study\"\nrather than a defined PECO result is a misuse. (6) **It does not manufacture causality** —\nrating a study Low risk does not make an observational association causal; it only certifies that\n*internal* bias is judged low for that estimate. (7) **Known critiques** (Bero et al., 2018,\nraised during development) flag that early versions were difficult to apply consistently and risked\nover-penalizing or under-penalizing confounding in observational exposure science; assessor\ntraining, pilot calibration, and dual independent assessment with reconciliation are necessary to\nget reproducible ratings. (8) **Checklist-as-theater** — answering signalling questions without\nthe underlying methodological judgment (e.g., waving through \"confounding adequately controlled\"\nwithout scrutinizing the confounder set) defeats the instrument.\n\n**How it maps to this catalog** — ROBINS-E's seven domains are appraisal *lenses*; in this repo the\nunderlying methods a study must execute well to earn a Low-risk rating are implemented by concepts\nthe assessor can check against, domain by domain:\n- **Confounding (Domain 1):** the study should demonstrate principled confounder selection and\n residual-confounding accounting — **unmeasured-confounding-probabilistic-bias-analysis-rwe**,\n **e-value-sensitivity-analysis**, **negative-control-outcomes-rwe** /\n **negative-control-exposures-rwe**, and (where a propensity approach is used)\n **propensity-score-methods-psm-iptw**; the estimand being targeted should be explicit via\n **estimands-ate-att-intercurrent-events-rwe**.\n- **Exposure measurement (Domain 2):** validity of the exposure metric / phenotype maps to\n **algorithm-validation** (and **claims-outcome-algorithm-ppv-sensitivity-rwe** when an\n administrative-data exposure proxy is used).\n- **Participant selection (Domain 3):** **selection-bias-sensitivity-analysis-rwe**,\n **time-zero-index-date-alignment-rwe** (selection/immortal-time at the start of follow-up), and\n **immortal-time-bias-handling**.\n- **Missing data (Domain 5):** **missing-data-pattern-table-rwe** and\n **attrition-and-loss-to-follow-up-rwe**.\n- **Outcome measurement (Domain 6):** **algorithm-validation** again, for outcome-ascertainment\n validity and whether it is differential by exposure.\n- **Overall / synthesis:** the magnitude and direction of residual bias the tool asks you to record\n are quantified with the **quantitative-bias-analysis-toolkit-rwe**, and external validity of the\n appraised estimate (a GRADE indirectness concern downstream) with\n **generalizability-transportability-external-validity-rwe**. Note these are ROBINS-I-lane\n pharmacoepi designs only by analogy; **active-comparator-new-user**,\n **high-dimensional-propensity-score-hdps-rwe**, and **claims-analysis** belong to the\n intervention sibling and are *not* the natural exemplars for ROBINS-E.\n\n**Applied note (when an administrative-data cohort meets ROBINS-E).** ROBINS-E's home substrate is\nenvironmental/occupational/nutritional cohorts (e.g., a PFAS-serum cohort and kidney cancer, an\nambient-PM2.5 cohort and cardiovascular mortality, a dietary-pattern cohort and incident diabetes),\nwhere exposure measurement and confounding are the dominant biases. Claims/EHR data *can* host a\nROBINS-E-appropriate study only when the cohort studies a **non-intervention exposure** captured in\nthose data — for example, an occupational or environmental exposure recorded in linked\nrecords — in which case the exposure-measurement domain hinges on how well the administrative proxy\nvalidates against true exposure (**algorithm-validation**), and the confounding domain hinges on\nthe completeness of the recorded confounder set. The moment the \"exposure\" is a prescribed therapy,\nswitch to ROBINS-I.", "primary_category": "Guideline", "tags": [ "guideline", "risk-of-bias", "critical-appraisal", "non-randomized-studies", "exposures", "environmental-epidemiology", "grade", "evidence-synthesis" ], "aliases": [ "ROBINS-E", "Risk Of Bias In Non-randomized Studies - of Exposures", "Risk of Bias in Non-randomised Studies of Exposures", "ROBINS-E tool" ], "applies_to_study_types": [ "cohort_prospective", "cohort_retrospective" ], "data_sources": [ "registry", "primary", "linked" ], "citations": [ { "role": "introduce", "doi": "10.1016/j.envint.2024.108602", "url": "https://doi.org/10.1016/j.envint.2024.108602", "citation_text": "Higgins JPT, Morgan RL, Rooney AA, Taylor KW, Thayer KA, Silva RA, et al. A tool to assess risk of bias in non-randomized follow-up studies of exposure effects (ROBINS-E). Environment International. 2024;186:108602.", "year": 2024, "authors_short": "Higgins et al.", "notes": "Canonical launch description of the ROBINS-E tool, its seven bias domains, signalling questions, and PECO-anchored result-level assessment logic." }, { "role": "explain", "doi": "10.1016/j.envint.2018.07.015", "url": "https://doi.org/10.1016/j.envint.2018.07.015", "citation_text": "Morgan RL, Whaley P, Thayer KA, Schunemann HJ. Identifying the PECO: a framework for formulating good questions to explore the association of environmental and other exposures with health outcomes. Environment International. 2018;121(Pt 1):1027-1031.", "year": 2018, "authors_short": "Morgan et al.", "notes": "Defines the PECO question framework around which a ROBINS-E assessment is structured; the result being appraised must be tied to an explicit PECO." }, { "role": "explain", "doi": "10.1186/s13643-018-0915-2", "url": "https://doi.org/10.1186/s13643-018-0915-2", "citation_text": "Bero L, Chartres N, Diong J, Fabbri A, Ghersi D, Lam J, et al. The risk of bias in observational studies of exposures (ROBINS-E) tool: concerns arising from application to observational studies of exposures. Systematic Reviews. 2018;7(1):242.", "year": 2018, "authors_short": "Bero et al.", "notes": "Candid critique raised during development - inter-rater consistency and handling of confounding in exposure science - motivating assessor training, piloting, and dual independent assessment." }, { "role": "use", "url": "https://www.riskofbias.info/welcome/robins-e-tool", "citation_text": "ROBINS-E Development Group. ROBINS-E: Risk Of Bias In Non-randomized Studies - of Exposures. riskofbias.info (maintained tool, guidance documents, and detailed templates).", "year": 2024, "authors_short": "ROBINS-E Development Group", "notes": "Maintained landing page with the current tool version, full guidance, worked examples, and domain templates for routine use." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "cohort-prospective", "notes": "Appraise the bias in an exposure-effect estimate from a prospective non-randomized follow-up study." }, { "relation_type": "applies_to", "target_slug": "cohort-retrospective", "notes": "Appraise the bias in an exposure-effect estimate from a retrospective non-randomized follow-up study." }, { "relation_type": "see_also", "target_slug": "target-trial-emulation", "notes": "ROBINS-E imports ROBINS-I's emulated-target-trial logic - judge the cohort against the randomized exposure experiment it approximates - as the conceptual backbone of its confounding and selection domains." }, { "relation_type": "see_also", "target_slug": "unmeasured-confounding-probabilistic-bias-analysis-rwe", "notes": "Domain 1 (confounding) - quantify residual confounding the qualitative judgment flags." }, { "relation_type": "see_also", "target_slug": "e-value-sensitivity-analysis", "notes": "Domain 1 (confounding) - express how strong an unmeasured confounder would need to be to overturn the estimate." }, { "relation_type": "see_also", "target_slug": "negative-control-outcomes-rwe", "notes": "Domain 1 (confounding) - negative controls help detect residual/uncontrolled confounding that the confounding domain must weigh." }, { "relation_type": "see_also", "target_slug": "algorithm-validation", "notes": "Domains 2 and 6 (exposure and outcome measurement) - validity of the exposure metric and outcome ascertainment, including whether misclassification is differential." }, { "relation_type": "see_also", "target_slug": "selection-bias-sensitivity-analysis-rwe", "notes": "Domain 3 (selection of participants) - selection related jointly to exposure and outcome, including selection after the start of follow-up." }, { "relation_type": "see_also", "target_slug": "missing-data-pattern-table-rwe", "notes": "Domain 5 (missing data) - characterize missingness in exposure, outcome, and confounders." }, { "relation_type": "see_also", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Domain 5 (missing data) - differential loss to follow-up in a cohort exposure study." }, { "relation_type": "see_also", "target_slug": "quantitative-bias-analysis-toolkit-rwe", "notes": "Quantifies the magnitude and direction of bias ROBINS-E records qualitatively per domain." }, { "relation_type": "see_also", "target_slug": "generalizability-transportability-external-validity-rwe", "notes": "External validity of the appraised estimate - a downstream GRADE indirectness concern, distinct from ROBINS-E internal risk of bias." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "journal", "hta" ] }, { "slug": "robins-i", "name": "ROBINS-I (Risk Of Bias In Non-randomised Studies of Interventions)", "short_definition": "A domain-based risk-of-bias instrument for non-randomized studies of interventions that judges each study against a hypothetical target trial across seven bias domains and returns a qualitative overall judgment (Low / Moderate / Serious / Critical / No information) — not a reporting checklist and not a numeric quality score.", "long_description": "**What it is** — **ROBINS-I (Risk Of Bias In Non-randomised Studies of Interventions)** is a\nstructured, domain-based tool for the *critical appraisal* of an individual non-randomized study\nestimating the effect of an intervention. It was introduced by Sterne, Hernán, Higgins, Reeves and\ncolleagues (BMJ, 2016) and is developed and maintained by the ROBINS-I development group and the\nCochrane Bias Methods Group, distributed at riskofbias.info — it is **not** an EQUATOR reporting\nguideline. Its defining feature, which separates it from every reporting checklist, is the\n**target-trial framing**: the assessor first specifies the hypothetical pragmatic randomized trial\nwhose effect the observational study is trying to estimate, then judges bias as *deviation of the\nactual study from that target trial*. Appraisal proceeds through signaling questions in seven bias\ndomains — (1) confounding, (2) selection of participants into the study, (3) classification of\ninterventions, (4) deviations from intended interventions, (5) missing data, (6) measurement of\noutcomes, and (7) selection of the reported result — yielding a judgment per domain and an overall\njudgment on a five-level *qualitative* scale: Low, Moderate, Serious, Critical, or No information.\nIt is a risk-of-bias instrument, in the same family as RoB 2 (for randomized trials) and ROBINS-E\n(for exposure/etiologic studies), and is the appraisal counterpart to — not a substitute for —\nreporting checklists (STROBE, RECORD-PE), protocol templates (HARPER, ENCePP), and systematic-review\nappraisal tools (AMSTAR 2).\n\n**When to use** — Use ROBINS-I to appraise the internal validity of a *primary non-randomized study\nof an intervention* — typically a comparative cohort study of two treatment strategies in claims,\nEHR, registry, or linked data, including target-trial emulations and active-comparator new-user\ncohorts. It is the standard within-study risk-of-bias tool for the included non-randomized studies of\na Cochrane or other systematic review, and it feeds directly into **GRADE** certainty rating for a\nbody of non-randomized evidence (Schünemann et al., 2019). It is increasingly expected in HTA/payer\nevidence assessments and in regulatory contexts (FDA/EMA) where the credibility of an observational\ncomparative-effectiveness or safety estimate must be defended. Decision rule for choosing the right\nsibling: appraise a **randomized trial with RoB 2**, not ROBINS-I; appraise an **exposure/etiologic\nstudy with no defined intervention** (e.g., effect of an environmental or behavioral exposure on\nharm) with **ROBINS-E**, not ROBINS-I; appraise the **systematic review itself with AMSTAR 2**; and\n*report* the primary study with STROBE/RECORD-PE. ROBINS-I is for an interventional contrast you can\ncast as a target trial.\n\n**What it requires** — Substantively, ROBINS-I forces the appraiser to make explicit the things that\ndetermine whether an observational estimate is credible. The **target-trial step** demands a stated\nPICO/estimand (eligibility, the intervention strategies compared, the comparator, the outcome, and\nthe effect of interest — assignment vs adherence). The **confounding domain** requires a pre-specified\nlist of important confounders and time-varying confounders affected by prior treatment, and an\nappraisal of whether the analysis controlled them appropriately (e.g., propensity-score or\nhigh-dimensional adjustment, g-methods) — this is the domain that most often drives a Serious or\nCritical rating in real-world data. The **selection domain** interrogates time-zero alignment and\nimmortal-time bias (was selection and start of follow-up tied to the intervention decision?). The\n**classification-of-interventions domain** asks whether exposure status was defined and ascertained\nwithout knowledge of the outcome — central where exposure is an algorithm over dispensing or\nadministration records. The **deviations domain** addresses whether the analysis targets the intended\nassignment-vs-adherence estimand and handles switching/discontinuation (intercurrent events). The\n**missing-data domain** addresses attrition, loss to follow-up, and informative censoring. The\n**outcome-measurement domain** asks whether the outcome phenotype is valid and ascertained blind to\nexposure. The **reported-result domain** addresses selective reporting against a pre-specified\nanalysis plan. For claims/EHR/registry RWE these map onto data-fitness-for-use, phenotype/algorithm\nvalidation (PPV/sensitivity), and sensitivity/quantitative bias analysis.\n\n**When NOT to use — limitations and common misapplications** — ROBINS-I is a *risk-of-bias* tool, not\na reporting checklist, not a study-quality scale, and not a guarantee of validity. Concrete failure\nmodes: (1) **Wrong tool for the design** — applying ROBINS-I to a randomized trial (use RoB 2) or to a\nnon-interventional exposure/etiologic study with no intervention contrast (use ROBINS-E). (2)\n**Summing or scoring domains** — ROBINS-I yields qualitative per-domain and overall judgments; adding\nthem into a points total or a \"quality percentage\" is a misuse the developers explicitly reject, and a\nsingle Critical domain caps the overall rating at Critical regardless of the others. (3) **Skipping the\ntarget-trial step** — answering signaling questions without first specifying the hypothetical trial and\nthe estimand produces incoherent, non-reproducible ratings; the target trial *is* the instrument. (4)\n**Confusing appraisal with reporting** — completing ROBINS-I does not make a study well-reported (that\nis STROBE/RECORD-PE) and, crucially, **completing ROBINS-I does not make an observational estimate\ncausal or unconfounded** — a Low overall rating reflects the *appraiser's* judgment given measured\nconfounders, not proof of exchangeability. (5) **Confusing it with AMSTAR 2** — ROBINS-I appraises the\n*included primary studies*; AMSTAR 2 appraises the *review*. (6) **Checklist-as-theater** — recording\ndomain judgments without documenting the rationale and the confounder list defeats the auditability the\ntool exists to create. Note that **ROBINS-I V2** is in development, refining the signaling questions and\ndomain structure; cite the version actually applied.\n\n**How it maps to this catalog** — ROBINS-I tells the reviewer *what* must be controlled; the catalog\nconcepts implement *how*. Domain-by-domain crosswalk to implementing concepts in this repo:\n- **Target-trial / estimand framing** (the prerequisite step) → **target-trial-emulation** and\n **estimands-ate-att-intercurrent-events-rwe** specify the hypothetical trial, the estimand, and the\n handling of intercurrent events the appraisal judges against.\n- **Confounding domain** → **high-dimensional-propensity-score-hdps-rwe** and\n **active-comparator-new-user** implement the confounder-control and channeling-mitigation that move a\n study away from a Serious/Critical confounding rating.\n- **Selection / time-zero domain** → **active-comparator-new-user** (time-zero at initiation, no\n immortal time) and **target-trial-emulation** (aligned eligibility, assignment, and start of\n follow-up).\n- **Classification-of-interventions & outcome-measurement domains** →\n **diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe** supplies the validated exposure/outcome\n algorithms (1 IP / 2 OP rules, time windows, PPV) the appraisal requires.\n- **Missing-data domain** → **attrition-and-loss-to-follow-up-rwe** implements attrition accounting and\n informative-censoring handling.\n- **Data fitness underlying every domain** → **fit-for-purpose-data-assessment-rwe** and\n **claims-analysis** establish whether the source data can support the operational definitions at all.\n\n**Applied note (claims/EHR/registry RWE).** When appraising a claims- or EHR-based comparative study,\nresolve the target trial first (eligibility, the two treatment strategies, time zero, the estimand),\nthen attack confounding and selection: confirm an active-comparator new-user design with time-zero at\ninitiation, a high-dimensional propensity score over the lookback window, and that \"no prior fill\"\nreflects observed continuous enrollment rather than Medicare Advantage missingness. For the\nclassification and outcome domains, demand the phenotype's validation metrics (PPV/sensitivity) and\nascertainment blind to exposure. For missing data, require a CONSORT-style attrition flow and treatment\nof loss to follow-up as potentially informative, with quantitative bias analysis (e.g., E-value,\nnegative controls) supporting the reported-result domain. A clean ROBINS-I appraisal documents the\nconfounder list and the rationale for each domain judgment, not just the five-level labels.", "primary_category": "Guideline", "tags": [ "guideline", "risk-of-bias", "critical-appraisal", "non-randomized-studies", "comparative-effectiveness", "cochrane", "grade", "rwe" ], "aliases": [ "ROBINS-I", "Risk Of Bias In Non-randomised Studies of Interventions", "ROBINS-I V1" ], "applies_to_study_types": [ "cer_observational", "new_user", "active_comparator_new_user", "cohort_prospective", "cohort_retrospective", "target_trial_emulation" ], "data_sources": [ "claims", "ehr", "registry", "linked", "multi-database" ], "citations": [ { "role": "introduce", "doi": "10.1136/bmj.i4919", "url": "https://doi.org/10.1136/bmj.i4919", "citation_text": "Sterne JAC, Hernán MA, Reeves BC, et al. ROBINS-I: a tool for assessing risk of bias in non-randomised studies of interventions. BMJ. 2016;355:i4919.", "year": 2016, "authors_short": "Sterne et al.", "notes": "Canonical statement paper introducing ROBINS-I, the seven bias domains, the target-trial framing, and the five-level overall judgment scale." }, { "role": "explain", "doi": "10.1016/j.jclinepi.2018.01.012", "url": "https://doi.org/10.1016/j.jclinepi.2018.01.012", "citation_text": "Schünemann HJ, Cuello C, Akl EA, et al. GRADE guidelines: 18. How ROBINS-I and other tools to assess risk of bias in nonrandomized studies should be used to rate the certainty of a body of evidence. Journal of Clinical Epidemiology. 2019;111:105-114.", "year": 2019, "authors_short": "Schünemann et al.", "notes": "Defines how ROBINS-I domain and overall judgments feed GRADE certainty rating for a body of non-randomized evidence; clarifies that ROBINS-I outputs are qualitative judgments, not scores." }, { "role": "explain", "doi": "10.1136/bmj.l4898", "url": "https://doi.org/10.1136/bmj.l4898", "citation_text": "Sterne JAC, Savović J, Page MJ, et al. RoB 2: a revised tool for assessing risk of bias in randomised trials. BMJ. 2019;366:l4898.", "year": 2019, "authors_short": "Sterne et al.", "notes": "The randomized-trial sibling instrument; anchors the decision rule that RoB 2 (not ROBINS-I) is used for RCTs." }, { "role": "use", "url": "https://www.riskofbias.info/welcome/home", "citation_text": "ROBINS-I tool — current version, signaling questions, guidance, and worked examples. riskofbias.info (Cochrane Bias Methods Group / ROBINS-I development team).", "year": 2024, "authors_short": "ROBINS-I development team", "notes": "Maintained landing page for the instrument, including downloadable templates, the V2 development status, and the ROBINS-E exposure sibling; cite the version actually applied." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "cer-observational", "notes": "ROBINS-I is the standard within-study risk-of-bias tool for comparative-effectiveness observational studies of an intervention." }, { "relation_type": "applies_to", "target_slug": "active-comparator-new-user", "notes": "The active-comparator new-user cohort is the prototypical non-randomized study of an intervention that ROBINS-I appraises against a target trial." }, { "relation_type": "applies_to", "target_slug": "target-trial-emulation", "notes": "ROBINS-I's first step is to specify the hypothetical target trial; emulations make that explicit and are appraised on how faithfully they realize it." }, { "relation_type": "used_with", "target_slug": "target-trial-emulation", "notes": "Implements the target-trial / estimand framing that is the prerequisite step of a ROBINS-I appraisal." }, { "relation_type": "used_with", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "Implements the confounder control whose adequacy the ROBINS-I confounding domain judges." }, { "relation_type": "used_with", "target_slug": "active-comparator-new-user", "notes": "Implements time-zero alignment and channeling control that move a study away from Serious/Critical confounding and selection ratings." }, { "relation_type": "used_with", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Implements the estimand and intercurrent-event handling judged by the deviations-from-intended- interventions domain." }, { "relation_type": "used_with", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Supplies validated exposure/outcome algorithms for the classification-of-interventions and outcome-measurement domains." }, { "relation_type": "used_with", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Implements attrition accounting and informative-censoring handling judged by the missing-data domain." }, { "relation_type": "see_also", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "Establishes whether the source data can support the operational definitions every ROBINS-I domain depends on." }, { "relation_type": "see_also", "target_slug": "claims-analysis", "notes": "Operational claims-data foundations underlying exposure, outcome, and confounder ascertainment in the appraised study." }, { "relation_type": "see_also", "target_slug": "picots-framework-rwe", "notes": "PICOTS operationalizes the eligibility/intervention/comparator/outcome/timing/setting that the target-trial step must declare." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "hta", "journal" ] }, { "slug": "sentinel-methods", "name": "FDA Sentinel System Methods", "short_definition": "The FDA Sentinel System is a national active surveillance system for monitoring the safety of marketed medical products using routinely collected electronic healthcare data. Its methods infrastructure includes the Common Data Model (Sentinel CDM), standardised analytic tools (CIDA, SCDM queries, propensity-score toolkits), and published methods guidance for distributed database querying, study design, and bias control. Sentinel methods are among the most operationally mature and regulatorily authoritative frameworks for pharmacoepidemiological evidence in the United States.", "long_description": "**What it is** — The **FDA Sentinel System** is a congressionally mandated active post-market\nsurveillance programme administered by the US Food and Drug Administration under the FDA Amendments\nAct (FDAAA) of 2007. It queries a distributed network of administrative claims and electronic health\nrecord data from US insurers, health systems, and data partners — currently covering more than\n500 million patient-years of longitudinal data — using a **Common Data Model (Sentinel CDM)** that\nstandardises data structure across partners without centralising data. The Sentinel System's\n**methods infrastructure** — the subject of this catalog entry — consists of: (1) the **Sentinel\nCommon Data Model (CDM)** (defining standardised tables, coding conventions, and derived variables);\n(2) the **Modular Programs** (reusable SAS/R analytic tools distributed as open-source code through\nthe Sentinel Innovation Center for querying the CDM for cohort identification, covariate extraction,\nand outcome ascertainment); (3) the **CIDA (Cohort Identification and Data Analysis) framework**\n(a pre-specified, protocol-driven workflow for active surveillance queries that controls for false\ndiscovery in repeated looks); (4) the **propensity-score toolkits** (including high-dimensional\npropensity score [hdPS] modules implemented for distributed execution); and (5) published **methods\npapers and guidance** that document design decisions, analytic choices, and validation studies. Platt\net al. (2018, *NEJM*) provided the first major public description of how the Sentinel System\nfunctions as a routine safety-assessment tool. The system is operated through a partnership between\nFDA and the **Sentinel Innovation Center** (hosted at Harvard Pilgrim Health Care Institute / Harvard\nMedical School).\n\n**When to use** — Reference Sentinel methods when: (1) **Designing a pharmacoepidemiology or\ndrug-safety study in US administrative data** — the Sentinel CDM conventions for enrollment,\ndiagnosis, procedure, drug, and laboratory data are the de facto US standard for distributed claims\nanalysis; aligning a study's data infrastructure with Sentinel CDM conventions maximises\nreproducibility and comparability to FDA's own analyses. (2) **Adapting or citing Sentinel modular\nprograms** — the open-source Sentinel programs for cohort identification (including active-comparator\nnew-user designs), covariate extraction (including hdPS), and outcome ascertainment represent\npeer-reviewed, FDA-validated implementations that can be adapted for non-Sentinel databases. (3)\n**Preparing a drug-safety RWE submission to FDA** — FDA's reviewers are familiar with Sentinel\nmethods conventions; a study that aligns with CIDA framework design and hdPS implementation\nprovides a more credible regulatory-science evidence package. (4) **Active post-market surveillance\nprotocol design** — for safety signals requiring repeated interim analysis, the CIDA framework's\nsequential testing and group-sequential methods (including maxSPRT, CUSUM, and likelihood ratio tests)\nprovide validated designs for continuous surveillance. (5) **Validation of phenotype algorithms in\nclaims data** — the Sentinel Data Core and published Sentinel validation studies are a reference\nfor PPV/sensitivity benchmarks for common outcome algorithms (e.g., MI, stroke, fracture, GI bleed).\nDecision rule: Sentinel methods are US claims-centric; for EHR-based surveillance or European data,\nequivalent frameworks include the EU-ADR, OHDSI/OMOP, and ENCePP approaches.\n\n**What it requires (checklist domains)** — Using Sentinel methods rigorously requires adherence to\nits principal analytic conventions: *Data model alignment*: data must conform to or be mapped to the\n**Sentinel CDM** table and field definitions (enrollment spans, medical claims, pharmacy claims,\nlaboratory, vital status) before any Sentinel modular program can be applied. *Continuous enrollment\nrequirement*: Sentinel analyses define **observable time** using continuous enrollment windows\n(typically ≥183 days pre-index for baseline covariate capture); gaps in enrollment are handled by\ncensoring. *Active-comparator new-user design*: the Sentinel modular programs implement ACNU design\nby default — both the treatment and comparator must be new users (no prior use in the baseline\nwindow) with a concurrent active comparator; prevalent-user analyses require explicit justification\nand are methodologically disfavoured. *Propensity score / hdPS implementation*: the Sentinel hdPS\ntoolkit operationalises the Schneeweiss et al. algorithm at distributed scale — empirically\nidentified covariates from prior healthcare utilisation supplemented by pre-specified clinical\ncovariates; propensity-score trimming and assessment of balance (standardised mean differences) are\nrequired outputs. *Outcome ascertainment*: outcomes must be defined by validated algorithms with\nreported PPV/sensitivity from Sentinel validation studies or equivalent; single-code, unvalidated\noutcome definitions are not acceptable in Sentinel queries. *Sequential testing / CIDA*:\nfor active surveillance (repeated looks), the CIDA framework specifies the maximum information\nfraction, the sequential test (typically MaxSPRT for rare events), and the signalling threshold;\na non-pre-specified sequential analysis or one that ignores inflation from repeated looks is\nmethodologically inadequate. *Distributed execution*: analyses run locally at each data partner\nagainst the CDM; only aggregate statistics (not patient-level data) are returned to the coordinating\ncentre, preserving privacy.\n\n**When NOT to use — limitations and common misapplications** — (1) **Assuming Sentinel coverage\nis population-representative** — Sentinel data are primarily employer-sponsored and Medicare\nAdvantage claims; Medicaid, Medicare fee-for-service, uninsured, and VA populations are covered\nby separate or supplemental systems; generalisability to all US patients requires acknowledgment of\nthese coverage gaps. (2) **Applying Sentinel CDM conventions uncritically to non-US data** — the\nCDM is designed for US claims coding (ICD-10-CM, NDC, HCPCS); applying it to European administrative\ndata or EHR requires remapping and validation. (3) **Citing Sentinel results as causal evidence**\n— Sentinel queries are designed for rapid signal detection and hypothesis generation, not definitive\ncausal inference; a Sentinel signal requires confirmatory study design with full confounder control.\n(4) **Using Sentinel modular programs without CDM alignment** — the programs assume specific CDM\ntable structures; applying them to data not mapped to the Sentinel CDM produces incorrect results.\n(5) **Ignoring the sequential testing framework for repeated analyses** — running standard cohort\nanalyses with multiple looks without a pre-specified sequential test inflates the false-positive\nrate; the CIDA framework and maxSPRT exist precisely to control this. (6) **Over-relying on\nhdPS without clinical review** — the hdPS empirically identifies covariates from claims patterns;\nit can produce technically adequate propensity scores while omitting clinically important confounders\nnot captured in claims codes; clinical expert review of the covariate space remains necessary.\n\n**How it maps to this catalog** — Sentinel methods are the implementation backbone for most of\nthe US claims-based pharmacoepidemiology concepts in this catalog. The **Sentinel CDM** is the\ndata-infrastructure substrate for **claims-analysis** and is the reference for **fit-for-purpose-\ndata-assessment-rwe** (enrollment continuity, observability windows, coding completeness). The\n**ACNU design** implemented in Sentinel modular programs is the same design documented in\n**active-comparator-new-user** and **time-zero-index-date-alignment-rwe** (index-date alignment\nand washout conventions). The **hdPS toolkit** is the large-scale implementation of\n**high-dimensional-propensity-score-hdps-rwe**; Schneeweiss et al.'s original algorithm and\nSentinel's implementation are the canonical references. **Outcome algorithm validation** conducted\nthrough Sentinel validation studies provides PPV/sensitivity benchmarks for\n**claims-outcome-algorithm-ppv-sensitivity-rwe**. The **sequential testing and maxSPRT** approach\nin the CIDA framework corresponds to **maxSPRT-sequential-testing** and **treescan-statistics-rwe**\n(spatial-temporal self-controlled scan statistics) in the methods catalog. **Distributed execution**\nand the federated data model intersect with **ohdsi-cdm** (the OHDSI/OMOP CDM approach, which is\nan alternative but complementary data-model standard). FDA guidance on using the Sentinel System\nfor regulatory-grade RWE is documented in **fda-rwe-framework** and **fda-rwd-ehr-claims** in this\ncatalog's guidelines section; Sentinel is the operational implementation of those frameworks.", "primary_category": "Guideline", "tags": [ "guideline", "active-surveillance", "pharmacovigilance", "fda", "sentinel-system", "common-data-model", "sequential-testing", "hdps", "distributed-database", "drug-safety" ], "aliases": [ "Sentinel System", "FDA Sentinel", "Sentinel CDM", "Sentinel Common Data Model", "CIDA framework", "Sentinel Initiative", "FDA Sentinel System methods" ], "applies_to_study_types": [ "cer_observational", "cohort_retrospective", "active_surveillance", "new_user" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "doi": "10.1056/NEJMp1809643", "url": "https://doi.org/10.1056/NEJMp1809643", "citation_text": "Platt R, Brown JS, Robb M, et al. The FDA Sentinel System — Adding to the Global Evidence Base for Medical Product Safety. New England Journal of Medicine. 2018;379(22):2151-2156.", "year": 2018, "authors_short": "Platt et al.", "notes": "Primary public description of the FDA Sentinel System as a routine post-market safety assessment tool; describes the distributed network, Common Data Model, and active surveillance methods." }, { "role": "explain", "doi": "10.1097/ede.0b013e3181a663cc", "url": "https://doi.org/10.1097/ede.0b013e3181a663cc", "citation_text": "Schneeweiss S, Rassen JA, Glynn RJ, et al. High-dimensional propensity score adjustment in studies of treatment effects using health care claims data. Epidemiology. 2009;20(4):512-522.", "year": 2009, "authors_short": "Schneeweiss et al.", "notes": "Foundational methods paper for the hdPS algorithm implemented in the Sentinel hdPS toolkit; describes empirical covariate identification from claims data patterns for large-scale propensity- score construction." }, { "role": "use", "url": "https://www.sentinelinitiative.org/methods-data-tools/methods", "citation_text": "FDA Sentinel System. Methods and Data Tools — Sentinel Common Data Model, modular programs, CIDA framework, and hdPS toolkit (maintained). sentinelinitiative.org.", "year": 2024, "authors_short": "FDA Sentinel Initiative", "notes": "Official Sentinel Initiative methods page; access point for the CDM specification, modular SAS/R programs, CIDA framework documentation, and published validation studies." } ], "relations": [ { "relation_type": "used_with", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "The Sentinel hdPS toolkit is the primary distributed implementation of the hdPS algorithm; Sentinel validation studies define the reference standard for hdPS application in US claims data." }, { "relation_type": "used_with", "target_slug": "active-comparator-new-user", "notes": "Sentinel modular programs implement the ACNU design by default; Sentinel analyses are the most operationally mature examples of this design at scale." }, { "relation_type": "used_with", "target_slug": "claims-outcome-algorithm-ppv-sensitivity-rwe", "notes": "Sentinel validation studies provide PPV/sensitivity benchmarks for common outcome algorithms in US claims data; these are the reference values for assessing algorithm adequacy." }, { "relation_type": "used_with", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "Sentinel CDM enrollment and observability conventions define the data-adequacy standard for US administrative claims; alignment with CDM conventions is the practical implementation of fit-for-purpose assessment." }, { "relation_type": "used_with", "target_slug": "ohdsi-cdm", "notes": "Sentinel CDM (claims-centric, US) and OHDSI/OMOP CDM (broader, international) are parallel standardisation efforts; studies may align with both or choose based on data source and geography." }, { "relation_type": "see_also", "target_slug": "fda-rwe-framework", "notes": "The FDA RWE Framework guidances are the regulatory policy documents that the Sentinel System operationalises for post-market safety surveillance." }, { "relation_type": "see_also", "target_slug": "fda-rwd-ehr-claims", "notes": "FDA technical guidance on fit-for-purpose RWD from EHR and claims is the policy complement to Sentinel's operational CDM and methods standards." }, { "relation_type": "see_also", "target_slug": "time-zero-index-date-alignment-rwe", "notes": "Sentinel modular programs enforce continuous-enrollment and index-date alignment conventions that prevent immortal-time and time-zero bias." }, { "relation_type": "see_also", "target_slug": "target-trial-emulation", "notes": "Sentinel ACNU analyses can be framed as target-trial emulations; the TTE framework makes the design assumptions explicit and is increasingly used alongside Sentinel conventions." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "journal" ] }, { "slug": "spirit-pro", "name": "SPIRIT-PRO Extension", "short_definition": "Reporting/protocol guideline that specifies the items a clinical trial PROTOCOL must contain when a patient-reported outcome (PRO) is a primary or key secondary endpoint; the PRO-specific extension of the SPIRIT 2013 protocol guidance, developed under the PROTEUS/SPIRIT framework.", "long_description": "**What it is** — The **SPIRIT-PRO Extension (Standard Protocol Items: Recommendations for Interventional\nTrials — Patient-Reported Outcomes)** is a consensus protocol guideline that sets out the minimum content a\nclinical trial *protocol* must include when one or more **patient-reported outcomes (PROs)** are designated as\na primary or key secondary endpoint. Published in JAMA in 2018 (Calvert, Kyte, Mercieca-Bebber et al.) via a\nDelphi consensus of trialists, methodologists, PRO experts, regulators (FDA, EMA-adjacent), patients, and\njournal editors, it extends the parent **SPIRIT 2013 Statement** (Chan et al.) with **16 PRO-specific items**\n— 11 new items plus 5 elaborations of existing SPIRIT items. It is maintained within the SPIRIT/CONSORT\necosystem and the **PROTEUS Consortium** (Patient-Reported Outcomes Tools: Engaging Users & Stakeholders),\nalongside its reporting-stage sibling **CONSORT-PRO**. SPIRIT-PRO governs the *plan*: what a trial commits to\nin advance about which PRO is measured, why, when, how, and how its data will be handled — so that PRO\nendpoints are pre-specified, defensible, and not retrofitted after unblinding. It is a *reporting/protocol-content*\nchecklist, not a risk-of-bias instrument and not a quality score.\n\n**When to use** — Apply SPIRIT-PRO when you are **writing or registering the protocol of an interventional trial\nin which a PRO is a primary or important secondary endpoint** — health-related quality of life, symptom burden,\nfunctional status, treatment satisfaction, or other constructs reported directly by the patient. This is the\ncorrect guideline for a trial protocol manuscript, a registry record (ClinicalTrials.gov / EudraCT) PRO section,\nand the PRO methods underpinning a regulatory labeling claim or an HTA/payer submission that relies on\ntrial-based PRO evidence. Decision rules for choosing the right family member: use **SPIRIT-PRO** for the\ninterventional-trial *protocol*; use **CONSORT-PRO** to report the *completed* trial's PRO results; use the\nbase **SPIRIT 2013** when no PRO is a key endpoint; and use **measurement-property** standards such as\n**COSMIN** and **ISOQOL** to *select and validate the instrument itself* (SPIRIT-PRO assumes you cite, but does\nnot adjudicate, the instrument's validation evidence). SPIRIT-PRO does **not** govern non-interventional or\nobservational studies: a primary claims/EHR/registry PRO study is reported under STROBE/RECORD-PE, planned\nunder HARPER or the ENCePP checklist, and (for a systematic review of PRO studies) registered under PRISMA-P.\n\n**What it requires** — The 16 items force pre-specification of the PRO design elements that otherwise drift:\na clear **PRO research hypothesis** and the **specific PRO concepts/domains** measured (not just \"quality of\nlife\" in the abstract); justification of the **instrument** chosen and its measurement properties in the target\npopulation; the **PRO data-collection schedule** (assessment time points, their rationale, and alignment with\nthe clinical endpoints); the **mode and setting** of administration (paper, ePRO, interview) and steps to\nstandardize it; **who completes** the PRO (patient vs proxy) and how proxy/assisted completion is handled;\na pre-specified **PRO-specific analysis plan** including the handling of **multiplicity** across domains and\ntime points, the **estimand** and how **intercurrent events** (death, treatment discontinuation, disease\nprogression) are addressed for a self-reported outcome; an explicit plan for **missing PRO data** — expected\nmechanisms, minimization strategies, and the primary and sensitivity analytic approaches; and a plan for\n**PRO data monitoring, management, and reporting** (training, real-time completeness checks, dissemination to\npatients). For PRO endpoints the missing-data and intercurrent-event items carry unusual weight: PRO scores are\n*not measured once the patient is too sick to respond or has died*, so attrition is rarely missing-completely-at-random,\nand a credible protocol pre-commits to the estimand and to sensitivity analyses rather than defaulting to\ncomplete-case analysis.\n\n**When NOT to use — limitations and common misapplications** — SPIRIT-PRO is a *reporting/protocol-content*\nchecklist; it is **not a risk-of-bias instrument, not a quality score, and not a license to call a PRO endpoint\nvalid**. Concrete failure modes: (1) **Wrong study type** — applying SPIRIT-PRO to a *non-interventional* or\nobservational PRO study (a claims-, EHR-, or registry-based PRO analysis); those use HARPER/ENCePP at the\nprotocol stage and STROBE/RECORD-PE for reporting, never SPIRIT-PRO, because SPIRIT items assume randomized\nallocation, blinding, and trial assessment schedules that observational data lack. (2) **Wrong family member** —\nciting SPIRIT-PRO to report a *completed* trial (use CONSORT-PRO) or invoking it when no PRO is a key endpoint\n(use base SPIRIT). (3) **Mistaking it for instrument validation** — completing SPIRIT-PRO does not establish\nthat the instrument is reliable, valid, or responsive in the population; that is the job of COSMIN/ISOQOL\nevidence the protocol must *cite*. (4) **Checklist-as-theater** — ticking 16 items while leaving the estimand,\nthe multiplicity strategy, or the missing-PRO-data plan vague defeats the purpose; the value is the\npre-specification, not the page count. (5) **Treating completion as causal/unbiased proof** — a fully\nSPIRIT-PRO-compliant protocol can still produce biased PRO estimates if attrition is informative and the\nanalysis ignores it; transparency of the plan is necessary, not sufficient.\n\n**How it maps to this catalog** — In this repo, SPIRIT-PRO's substantive requirements are implemented by\nconcepts a methodologist can pre-specify against:\n- The PRO endpoint itself, its development, and its measurement properties: **pro-rwe**, **pro-development**,\n and **pro-validation** supply the construct definition and the validity/reliability/responsiveness evidence\n the protocol must justify and cite.\n- The protocol-content discipline (SPIRIT-PRO's administrative and methods items): **study-protocol-or-sap-elements**\n provides the pre-specification and amendment-control habits the checklist demands.\n- The estimand and intercurrent-event items: **estimands-ate-att-intercurrent-events-rwe** and\n **estimand-analysis-traceability-rwe** operationalize the ICH E9(R1) estimand framing — including\n treatment-policy vs hypothetical vs while-alive strategies for death and discontinuation — that a PRO\n primary endpoint forces.\n- The missing-PRO-data items: **missing-data-pattern-table-rwe**, **multiple-imputation-longitudinal-rwe**, and\n **mmrm-repeated-measures-rwe** implement the mechanism characterization, primary analysis, and sensitivity\n analyses the protocol must pre-commit to; **attrition-and-loss-to-follow-up-rwe** characterizes the\n informative-dropout structure typical of PRO follow-up.\n- The multi-domain, multi-time-point structure: **composite-endpoint-construction-rwe** and\n **longitudinal-outcomes-modeling-rwe** inform how multiplicity and repeated measures are handled, and\n **picots-framework-rwe** anchors the PRO hypothesis to population/intervention/comparator/outcome/timing.\n\n**Applied note (PRO data in RWE — what transfers and what does not).** SPIRIT-PRO is written for *interventional\ntrials*, so it does not directly govern a claims/EHR/registry RWE study. But its measurement and missing-data\ndiscipline transfers to real-world PRO collection in **product/disease registries**, **pragmatic trials**\n(often the bridge between SPIRIT-PRO and RWE), and **PRO-enabled cohort studies**: pre-specify the PRO concept\nand instrument with cited measurement evidence (COSMIN/ISOQOL), fix the assessment schedule and administration\nmode, define who reports (patient vs proxy), and — most critically for RWE, where follow-up is visit-driven and\nattrition is heavy and informative — pre-commit to the estimand, the handling of death and disenrollment as\nintercurrent events, and the primary and sensitivity analyses for missing PRO data rather than defaulting to\ncomplete-case scores. For a *primary non-interventional* PRO study, use SPIRIT-PRO only as a measurement\nreference and report under STROBE/RECORD-PE with a HARPER/ENCePP protocol.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "protocol", "patient-reported-outcomes", "pro", "spirit", "proteus", "trial-protocol" ], "aliases": [ "SPIRIT-PRO", "SPIRIT-PRO Extension", "SPIRIT Extension for Patient-Reported Outcomes", "Standard Protocol Items - Patient-Reported Outcomes" ], "applies_to_study_types": [ "pragmatic_trial", "pro_rwe" ], "data_sources": [ "registry", "linked" ], "citations": [ { "role": "introduce", "doi": "10.1001/jama.2017.21903", "url": "https://doi.org/10.1001/jama.2017.21903", "citation_text": "Calvert M, Kyte D, Mercieca-Bebber R, et al. Guidelines for Inclusion of Patient-Reported Outcomes in Clinical Trial Protocols: The SPIRIT-PRO Extension. JAMA. 2018;319(5):483-494.", "year": 2018, "authors_short": "Calvert et al.", "notes": "Canonical SPIRIT-PRO statement defining the 16 PRO-specific protocol items (11 new + 5 elaborations) via international Delphi consensus." }, { "role": "explain", "doi": "10.7326/0003-4819-158-3-201302050-00583", "url": "https://doi.org/10.7326/0003-4819-158-3-201302050-00583", "citation_text": "Chan AW, Tetzlaff JM, Altman DG, et al. SPIRIT 2013 Statement: Defining Standard Protocol Items for Clinical Trials. Annals of Internal Medicine. 2013;158(3):200-207.", "year": 2013, "authors_short": "Chan et al.", "notes": "Parent SPIRIT 2013 statement that SPIRIT-PRO extends; defines the base 33-item trial-protocol checklist." }, { "role": "use", "url": "https://www.equator-network.org/reporting-guidelines/spirit-pro/", "citation_text": "SPIRIT-PRO Extension. EQUATOR Network reporting-guidelines library (maintained checklist, elaboration, and links to CONSORT-PRO and the PROTEUS Consortium).", "year": 2018, "authors_short": "EQUATOR Network", "notes": "Maintained landing page with the checklist in usable formats and links to the reporting-stage sibling CONSORT-PRO and the PROTEUS PRO trial-design resources." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "pragmatic-trial", "notes": "Use at the protocol stage for an interventional/pragmatic trial in which a PRO is a primary or key secondary endpoint." }, { "relation_type": "applies_to", "target_slug": "pro-rwe", "notes": "Not a primary RWE guideline, but its PRO measurement and missing-data discipline transfers to PRO-enabled real-world studies; for primary non-interventional PRO studies use STROBE/RECORD-PE with a HARPER/ENCePP protocol." }, { "relation_type": "see_also", "target_slug": "consort-pro", "notes": "SPIRIT-PRO governs the trial PROTOCOL; the completed trial's PRO results are reported with CONSORT-PRO." }, { "relation_type": "see_also", "target_slug": "spirit", "notes": "SPIRIT-PRO extends the base SPIRIT 2013 protocol checklist with PRO-specific items; use base SPIRIT when no PRO is a key endpoint." }, { "relation_type": "used_with", "target_slug": "pro-validation", "notes": "The protocol must cite measurement-property evidence (validity, reliability, responsiveness) for the PRO instrument; SPIRIT-PRO assumes but does not adjudicate this evidence." }, { "relation_type": "used_with", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "A PRO primary endpoint forces an explicit estimand and a strategy for intercurrent events (death, discontinuation, progression) that stop the patient from reporting." }, { "relation_type": "used_with", "target_slug": "study-protocol-or-sap-elements", "notes": "Supplies the pre-specification and amendment-control discipline SPIRIT-PRO's administrative and methods items require." }, { "relation_type": "see_also", "target_slug": "multiple-imputation-longitudinal-rwe", "notes": "Implements the primary and sensitivity analyses for missing PRO data that the missing-data items require." }, { "relation_type": "see_also", "target_slug": "mmrm-repeated-measures-rwe", "notes": "Standard longitudinal model for repeated PRO assessments; supports the multiplicity and time-point analysis plan the protocol must pre-specify." }, { "relation_type": "see_also", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "PRO follow-up attrition is typically informative (sicker patients stop reporting); the protocol must characterize and analyze it rather than default to complete-case scores." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "hta", "journal" ] }, { "slug": "spirit", "name": "SPIRIT (Standard Protocol Items: Recommendations for Interventional Trials)", "short_definition": "Reporting guideline that specifies the minimum protocol content for a randomized/interventional clinical trial, so the trial's design, conduct, and analysis are pre-specified and auditable before enrollment; the protocol-stage companion to CONSORT, maintained within the EQUATOR Network.", "long_description": "**What it is** — **SPIRIT (Standard Protocol Items: Recommendations for Interventional Trials)** is a reporting\nguideline that defines the minimum set of items a *protocol* for a **randomized or otherwise interventional\nclinical trial** should contain, so that the trial's rationale, design, methods, and oversight are completely and\ntransparently pre-specified *before* the trial begins. The original **SPIRIT 2013 Statement** (Chan, Tetzlaff,\nAltman et al.) provided a 33-item checklist plus a figure for the schedule of enrolment, interventions, and\nassessments; it was accompanied by an item-by-item explanation-and-elaboration paper, and the **SPIRIT 2025**\nupdate (Chan et al.) modernized it to incorporate ICH E9(R1) **estimands**, structured **harms** reporting, **open\nscience** (data/code/protocol sharing), **equity/diversity** of trial populations, and a new item set for **AI**\ninterventions. SPIRIT is hosted and maintained as a reporting guideline within the **EQUATOR Network** and is the\nprotocol-stage member of the trial-reporting family: SPIRIT governs what you *commit to in advance*, whereas\n**CONSORT** governs how you report the *completed* trial. Together they form a continuous chain of\npre-specification-to-reporting that journals, registries, ethics committees, and regulators increasingly require.\n\n**When to use** — Apply SPIRIT whenever you are *writing, registering, or appraising the protocol* of an\n**interventional** study — a parallel-group, crossover, cluster, factorial, or adaptive randomized trial — across\nany setting: a regulatory IND/CTA package, an ethics/IRB submission, a funder application, a trial-registry record\n(e.g., ClinicalTrials.gov, EudraCT/CTIS), or a protocol manuscript. In the real-world-evidence space SPIRIT\napplies to the **pragmatic randomized trial** (where eligibility is broad and outcomes are often ascertained from\nEHR, claims, or registries) and the **registry-based randomized trial / RRCT** (where randomization is layered\nonto an existing registry data infrastructure); in **externally-controlled or hybrid designs**, SPIRIT governs the\n*interventional arm's* protocol while a separate observational protocol governs the external real-world comparator.\nDecision rule for the right family member: use **SPIRIT** for the trial *protocol* and **CONSORT** for the\nfinished trial *report*; choose the matching SPIRIT extension by feature — **SPIRIT-PRO** when patient-reported\noutcomes are a focus, **SPIRIT-AI** for AI-based interventions, **SPIRIT-Outcomes 2022** for rigorous outcome\nspecification, and SPIRIT extensions for cluster/pilot designs where they exist. SPIRIT is **not** the checklist\nfor a *non-randomized observational* study protocol — that is the province of **HARPER**, the **ENCePP** checklist,\nor **STROBE/RECORD-PE** — and it is **not** for systematic-review protocols, which use **PRISMA-P**.\n\n**What it requires** — SPIRIT compels pre-specification of the protocol elements that otherwise drift or get\nreconstructed post hoc: administrative information (title, trial registration, protocol version and amendment\nhistory, funding, roles, and committees); a scientific **background and rationale** with explicit objectives and\n**hypotheses**; the trial **design** (allocation ratio, framework — superiority/non-inferiority/equivalence); a\n**participants/interventions/outcomes** core (eligibility criteria, intervention description and adherence\nprovisions, and a clearly defined **primary outcome** with measurement timing); **sample size** justification and\nrecruitment plan; the methods that protect against bias — **sequence generation, allocation concealment, and\nblinding**; **data collection, management, and analysis** plans, including how missing data and **intercurrent\nevents** are handled (SPIRIT 2025 frames the analysis around a formal **estimand**); **harms** monitoring; **data\nmonitoring** committee arrangements and interim analyses; and the governance items — research **ethics approval,\nconsent, confidentiality, declaration of interests, and data-access/dissemination** (including open-science\ncommitments to share protocol, data, and code). For RWE-flavored trials these generic items carry specific weight:\na pragmatic or registry-based trial that **ascertains outcomes from EHR/claims/registry data** must, in the\nprotocol, specify the **outcome/phenotype algorithm** and its validity, the data source's **fitness for use** for\nthe endpoint, and how **attrition / loss to follow-up** is defined and analyzed when \"follow-up\" is the\ncontinued presence of data — even though randomization, not design adjustment, is the engine of causal inference.\n\n**When NOT to use — limitations and common misapplications** — SPIRIT is a *reporting* checklist for a *protocol*;\nit is not a risk-of-bias instrument, not a quality score, and not a substitute for sound design. Concrete failure\nmodes: (1) **Wrong guideline entirely** — applying SPIRIT to a *non-randomized observational* RWE study (a claims\nor EHR cohort/case-control study); those protocols follow HARPER, ENCePP, or RECORD-PE, and reporting follows\nSTROBE/RECORD-PE — SPIRIT does not apply. (2) **Protocol/report confusion** — using SPIRIT to report a *completed*\ntrial (use CONSORT) or citing CONSORT for the protocol. (3) **Wrong extension** — ignoring SPIRIT-PRO for a\nPRO-centered trial, or SPIRIT-AI for an AI intervention, leaving the special items unaddressed. (4) **Mistaking\ncompleteness for validity** — a fully SPIRIT-compliant protocol can still describe an underpowered, poorly\nrandomized, or unblinded trial; SPIRIT documents what you plan, it does not certify that the plan is good. (5)\n**Checklist-as-theater** — ticking 33 items while leaving the estimand, the primary outcome, the blinding\nprocedure, or the missing-data plan vague defeats the purpose; the value is genuine pre-specification, not page\ncount. (6) **Registration ≠ compliance** — a registry record is necessary but not the same as a complete protocol;\nkeep the registration synchronized with protocol amendments. (7) **Importing observational machinery\ninappropriately** — a randomized trial does not generally need propensity scores or active-comparator design\nframing; conflating SPIRIT's needs with those of a confounded observational study is a category error.\n\n**How it maps to this catalog** — In this repo, the SPIRIT-relevant concepts cluster around protocol discipline,\nestimands, and the *narrow* points where randomized trials touch real-world data:\n- The pre-specification discipline itself: **study-protocol-or-sap-elements** supplies the version-control,\n amendment, and pre-specified-analysis habits SPIRIT's administrative and methods items demand; **picots-framework-rwe**\n operationalizes the population/intervention/comparator/outcome/timing/setting frame SPIRIT requires for objectives\n and eligibility.\n- The analysis spine (SPIRIT 2025): **estimands-ate-att-intercurrent-events-rwe** implements the ICH E9(R1)\n estimand and intercurrent-event thinking the updated checklist now expects in the statistical-methods item.\n- The design contrast: **target-trial-emulation** is the conceptual mirror image — SPIRIT governs the protocol of\n an *actual* randomized trial, whereas target-trial emulation specifies an observational analysis to *emulate*\n such a trial when randomization is infeasible; reading them side by side clarifies which assumptions\n randomization buys for free.\n- The RWD-ascertainment caveats (only for pragmatic/registry trials that capture endpoints from EHR/claims):\n **diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe** for how an algorithm-defined outcome is specified and\n validated, **attrition-and-loss-to-follow-up-rwe** for defining and analyzing dropout/data-availability, and\n **claims-analysis** for the operational realities of the underlying data source. These apply *because the trial\n borrows a real-world data substrate*, not because the trial is observational.\nNote what does **not** map: confounding-control concepts such as **high-dimensional-propensity-score-hdps-rwe** and\n**active-comparator-new-user** are observational-design tools; randomization, properly executed under SPIRIT, is\nwhat removes confounding, so those concepts are deliberately *out of scope* for a SPIRIT protocol.\n\n**Applied note (claims/EHR/registry RWE).** For a pragmatic or registry-based randomized trial whose endpoints are\nascertained from administrative or registry data, a SPIRIT-compliant protocol should go beyond the generic items:\nname the data source and justify its **fitness for use** for the specific endpoint, pre-specify the **outcome\nalgorithm** (e.g., a validated 1-inpatient-or-2-outpatient phenotype with its time window) and its expected\nsensitivity/positive predictive value, state how **time alignment** between randomization and data capture is\nhandled, and define **loss to follow-up** as loss of *data availability* (disenrollment, registry exit) with a\npre-specified, estimand-aligned analysis for the resulting missingness — so that the trial's real-world\nascertainment layer is as transparent and auditable as its randomization.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "protocol", "randomized-trial", "pragmatic-trial", "equator", "estimands" ], "aliases": [ "SPIRIT", "SPIRIT 2013", "SPIRIT 2025", "Standard Protocol Items: Recommendations for Interventional Trials" ], "applies_to_study_types": [ "pragmatic_trial", "registry_trial" ], "data_sources": [ "ehr", "registry", "claims", "linked" ], "citations": [ { "role": "introduce", "doi": "10.7326/0003-4819-158-3-201302050-00583", "url": "https://doi.org/10.7326/0003-4819-158-3-201302050-00583", "citation_text": "Chan A-W, Tetzlaff JM, Altman DG, et al. SPIRIT 2013 statement: defining standard protocol items for clinical trials. Annals of Internal Medicine. 2013;158(3):200-207.", "year": 2013, "authors_short": "Chan et al.", "notes": "Canonical SPIRIT 2013 statement defining the 33-item trial-protocol reporting checklist and the schedule-of-events figure." }, { "role": "explain", "doi": "10.1136/bmj.e7586", "url": "https://doi.org/10.1136/bmj.e7586", "citation_text": "Chan A-W, Tetzlaff JM, Gøtzsche PC, et al. SPIRIT 2013 explanation and elaboration: guidance for protocols of clinical trials. BMJ. 2013;346:e7586.", "year": 2013, "authors_short": "Chan et al.", "notes": "Item-by-item explanation and elaboration with model protocol language and rationale for each SPIRIT item." }, { "role": "explain", "doi": "10.1136/bmj-2024-081477", "url": "https://doi.org/10.1136/bmj-2024-081477", "citation_text": "Chan A-W, Boutron I, Hopewell S, et al. SPIRIT 2025 statement: updated guideline for protocols of randomised trials. BMJ. 2025;389:e081477.", "year": 2025, "authors_short": "Chan et al.", "notes": "SPIRIT 2025 update adding estimands (ICH E9(R1)), structured harms, open-science, equity, and AI-intervention items." }, { "role": "use", "url": "https://www.equator-network.org/reporting-guidelines/spirit-2013-statement-defining-standard-protocol-items-for-clinical-trials/", "citation_text": "SPIRIT (Standard Protocol Items: Recommendations for Interventional Trials). EQUATOR Network reporting-guidelines library (maintained checklists, templates, and SPIRIT extensions).", "year": 2025, "authors_short": "EQUATOR Network", "notes": "Maintained landing page with the checklist in usable formats and links to SPIRIT extensions (SPIRIT-PRO, SPIRIT-AI, SPIRIT-Outcomes)." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "pragmatic-trial", "notes": "SPIRIT governs the protocol of a pragmatic randomized trial, including outcomes ascertained from EHR/claims/registry data." }, { "relation_type": "applies_to", "target_slug": "registry-trial", "notes": "SPIRIT governs the protocol of a registry-based randomized trial (RRCT) where randomization is layered on a registry data infrastructure." }, { "relation_type": "see_also", "target_slug": "study-protocol-or-sap-elements", "notes": "Supplies the pre-specification, version-control, and amendment discipline SPIRIT's administrative and methods items require." }, { "relation_type": "used_with", "target_slug": "picots-framework-rwe", "notes": "PICOTS operationalizes SPIRIT's objectives and eligibility items (population, intervention, comparator, outcome, timing, setting)." }, { "relation_type": "used_with", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "SPIRIT 2025 frames the statistical-analysis item around an ICH E9(R1) estimand with explicit intercurrent-event handling." }, { "relation_type": "alternative_to", "target_slug": "target-trial-emulation", "notes": "SPIRIT governs the protocol of an actual randomized trial; target-trial emulation specifies an observational analysis to emulate one when randomization is infeasible." }, { "relation_type": "see_also", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Only when a pragmatic/registry trial ascertains endpoints from EHR/claims; the protocol should pre-specify the outcome algorithm and its validity." }, { "relation_type": "see_also", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "For trials with RWD ascertainment, loss to follow-up is loss of data availability; the protocol must define it and a pre-specified missing-data plan." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "journal" ] }, { "slug": "srqr", "name": "SRQR (Standards for Reporting Qualitative Research)", "short_definition": "A 21-item reporting checklist that defines the minimum information a primary qualitative research study should report, applicable across all qualitative methods (interviews, focus groups, ethnography, document analysis, grounded theory, narrative); the broad-scope counterpart to the interview/focus-group-specific COREQ.", "long_description": "**What it is** — **SRQR (Standards for Reporting Qualitative Research)** is a 21-item reporting checklist\npublished by O'Brien, Harris, Beckman, Reed, and Cook in *Academic Medicine* (2014) as a synthesis of prior\nrecommendations for reporting qualitative work. Its purpose is to make a primary qualitative study transparent\nand appraisable: it specifies the minimum content an author should report so that a reader can judge what was\nstudied, how data were generated and analysed, and how the investigators arrived at and warranted their findings.\nCritically, SRQR is *method-agnostic* within qualitative research — it was written to apply across the full\nrange of qualitative approaches (in-depth and key-informant interviews, focus groups, participant observation\nand ethnography, document/text analysis, grounded theory, phenomenology, narrative inquiry, and qualitative\ncomponents of mixed-methods designs). It is listed and maintained as a reporting guideline within the **EQUATOR\nNetwork** library. SRQR is a *reporting* standard, not a methods textbook and not an instrument for scoring\nstudy quality.\n\n**When to use** — Use SRQR when you are reporting (or peer-reviewing, or pre-specifying the reporting of) a\n*primary qualitative study* and you want a single broad standard that fits whatever qualitative method was used.\nIn HEOR and real-world evidence work this most often means the qualitative strand of an evidence package: patient\nor caregiver burden-of-disease interviews, concept-elicitation and cognitive-interview studies that underpin a\npatient-reported outcome (PRO) instrument, qualitative preference or treatment-experience studies feeding a value\ndossier or patient-focused drug development (PFDD) submission, clinician interviews characterising real-world\ncare pathways, or a qualitative substudy nested alongside a registry or claims/EHR analysis. Decision rule for\nchoosing among siblings: use **COREQ** if and only if the study is specifically *interviews and/or focus groups*\nand you want the more granular 32-item instrument that also probes researcher reflexivity and the participant\nrelationship; use **SRQR** when the method is something else (ethnography, document analysis, grounded theory,\nnarrative) or when you want one checklist that spans a multi-method qualitative project. For the *qualitative\ncomponent of a mixed-methods study*, SRQR governs the qualitative strand but the integration of strands is the\ndomain of mixed-methods reporting guidance (e.g., GRAMMS / O'Cathain), not SRQR alone. SRQR does **not** apply\nto quantitative observational RWE — a claims or EHR cohort is reported with STROBE/RECORD-PE (and conducted under\nHARPER/ENCePP), never SRQR.\n\n**What it requires** — The 21 items span the full manuscript and force authors to make their qualitative\nreasoning visible rather than asserted. The substantive domains include: a **title and abstract** that identify\nthe work as qualitative and name the approach; a **problem formulation and purpose** with an explicit research\nquestion and its theoretical/conceptual grounding; a stated **qualitative approach and research paradigm**\n(e.g., grounded theory, ethnography, phenomenology) with rationale, since the approach dictates what counts as\nrigour; **researcher characteristics and reflexivity** — the investigators' backgrounds, assumptions, and\nrelationship to participants and the topic, because the analyst is the instrument in qualitative work;\n**context** of the study; a transparent **sampling strategy** and rationale (purposive, theoretical, maximum\nvariation), with sampling continued to the point of justified sufficiency/saturation; **ethical issues** and\napprovals; concrete **data collection methods, instruments, and units of study** (who/what, where, how many);\n**data processing** (transcription, de-identification, coding, software); a **data analysis** account detailed\nenough to be followed (coding scheme development, who coded, how disagreements were resolved, how themes were\nderived); and **techniques to enhance trustworthiness** (member checking, triangulation, audit trail, negative-\ncase analysis). Findings must be presented with **direct evidence** (quotations or field-note excerpts) linked\nto interpretation, followed by **integration with prior work, limitations, and implications**. The throughline\nis auditability: a reader should be able to trace a finding back to data and to the analytic decisions that\nproduced it.\n\n**When NOT to use — limitations and common misapplications** — (1) **A reporting checklist is not a risk-of-bias\ntool or a quality score.** A fully SRQR-compliant paper can still rest on a thin sample, a poorly justified\napproach, or over-interpreted data; completeness of reporting is necessary, not sufficient, for trustworthiness.\nDo not convert the 21 items into a numeric \"quality\" tally. (2) **Wrong sibling.** Using SRQR for a pure\ninterview/focus-group study where reviewers expect COREQ's reflexivity and relationship items — or, conversely,\nforcing COREQ onto an ethnography or document analysis it was never designed for — is a recognisable mistake.\n(3) **Wrong family entirely.** Applying SRQR to a *quantitative* observational RWE study (a claims cohort, an\nEHR comparative-effectiveness analysis) instead of STROBE/RECORD-PE; SRQR has no items for confounding control,\ntime-zero alignment, phenotype validation, or estimands because qualitative research does not produce those\nestimands. (4) **Mixed-methods confusion.** SRQR covers the qualitative strand only; it does not tell you how to\nreport the integration, sequencing, or joint display of qualitative and quantitative components. (5) **Checklist-\nas-theatre.** Citing \"reported per SRQR\" while burying reflexivity, the analytic process, or supporting quotations\ndefeats the standard; the value is the transparency, not the citation. (6) **Reflexivity skipped.** Omitting\nresearcher characteristics and the relationship to participants — the items that most distinguish qualitative\nreporting from quantitative — while claiming SRQR compliance.\n\n**How it maps to this catalog** — SRQR is the reporting discipline that sits on top of the *qualitative* methods\nin this repo; each implementing concept supplies the design/analysis substance that SRQR then forces you to\nreport transparently:\n- **qualitative-interview** — implements the data-collection, sampling, coding, and reflexivity reporting that\n SRQR's interview-based items demand (and the natural decision point for SRQR-vs-COREQ).\n- **qualitative-ethnographic** — the observation/field-note and context-reporting strand; a case where SRQR\n applies but COREQ does not.\n- **qualitative-synthesis** — qualitative evidence synthesis; SRQR governs reporting of the primary studies it\n aggregates (the synthesis report itself follows ENTREQ/PRISMA-family guidance).\n- **mixed-methods** — SRQR governs the qualitative strand; integration reporting lives with mixed-methods guidance.\n- **pro-development** and **pro-validation** — concept-elicitation and cognitive-interview work that underpins a\n PRO instrument is qualitative and should be reported per SRQR; the resulting instrument's psychometrics are a\n separate (COSMIN-type) matter.\n- **preference-study** and **pro-rwe** — qualitative preference and treatment-experience components feeding value\n and PFDD evidence are SRQR-governed.\n- **burden-of-disease-cost-of-illness** — when the burden evidence includes patient/caregiver qualitative\n interviews, that component is reported under SRQR.\n- **study-protocol-or-sap-elements** — supplies the pre-specification habits that make SRQR's analysis and\n trustworthiness items credible rather than retrofitted.\n\n**Applied note (HEOR / RWE evidence packages).** In a real-world evidence dossier the quantitative core\n(claims/EHR/registry comparative analyses) is reported with STROBE/RECORD-PE and conducted under HARPER/ENCePP —\n*not* SRQR. SRQR governs the qualitative components that increasingly accompany that core: PFDD-style patient-\nexperience interviews, concept-elicitation studies for endpoint development, qualitative preference work, and\nclinician interviews mapping real-world treatment pathways. Reporting that qualitative strand to SRQR (with\nexplicit approach, sampling sufficiency, reflexivity, a traceable analytic process, and quotation-anchored\nfindings) is what lets an HTA body or regulator judge whether the patient voice in the dossier is credible rather\nthan decorative.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "qualitative", "equator", "patient-experience", "mixed-methods", "pro" ], "aliases": [ "SRQR", "Standards for Reporting Qualitative Research", "O'Brien 2014 qualitative reporting standards" ], "applies_to_study_types": [ "qualitative_interview", "qualitative_ethnographic" ], "data_sources": [], "citations": [ { "role": "introduce", "doi": "10.1097/ACM.0000000000000388", "url": "https://doi.org/10.1097/ACM.0000000000000388", "citation_text": "O'Brien BC, Harris IB, Beckman TJ, Reed DA, Cook DA. Standards for reporting qualitative research: a synthesis of recommendations. Academic Medicine. 2014;89(9):1245-1251.", "year": 2014, "authors_short": "O'Brien et al.", "notes": "Canonical SRQR statement defining the 21-item, method-agnostic reporting standard for primary qualitative research." }, { "role": "explain", "doi": "10.1093/intqhc/mzm042", "url": "https://doi.org/10.1093/intqhc/mzm042", "citation_text": "Tong A, Sainsbury P, Craig J. Consolidated criteria for reporting qualitative research (COREQ): a 32-item checklist for interviews and focus groups. International Journal for Quality in Health Care. 2007;19(6):349-357.", "year": 2007, "authors_short": "Tong et al.", "notes": "The narrower sibling standard. COREQ targets interviews and focus groups specifically; SRQR is the broader choice for other qualitative methods. Included to mark the SRQR-vs-COREQ decision." }, { "role": "use", "url": "https://www.equator-network.org/reporting-guidelines/srqr/", "citation_text": "Standards for Reporting Qualitative Research (SRQR). EQUATOR Network reporting-guidelines library (maintained entry with the checklist and the O'Brien 2014 statement).", "year": 2014, "authors_short": "EQUATOR Network", "notes": "Maintained landing page with the SRQR checklist and links to related qualitative reporting standards." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "qualitative-interview", "notes": "Reports a primary interview study; SRQR is the broad-scope choice (COREQ is the interview/focus-group-specific alternative)." }, { "relation_type": "applies_to", "target_slug": "qualitative-ethnographic", "notes": "Reports observation/ethnographic work, a case where SRQR applies but COREQ (interviews/focus groups only) does not." }, { "relation_type": "alternative_to", "target_slug": "coreq", "notes": "COREQ is the narrower 32-item interview/focus-group standard with stronger reflexivity/relationship items; SRQR is the 21-item method-agnostic alternative for other qualitative designs." }, { "relation_type": "see_also", "target_slug": "qualitative-synthesis", "notes": "SRQR governs reporting of the primary qualitative studies a synthesis aggregates; the synthesis report uses ENTREQ/PRISMA-family guidance." }, { "relation_type": "used_with", "target_slug": "mixed-methods", "notes": "SRQR covers the qualitative strand of a mixed-methods study; integration/joint-display reporting is the domain of mixed-methods guidance." }, { "relation_type": "used_with", "target_slug": "pro-development", "notes": "Concept-elicitation and cognitive-interview work underpinning a PRO instrument is qualitative and should be reported per SRQR." }, { "relation_type": "see_also", "target_slug": "pro-validation", "notes": "The qualitative content-validity work behind instrument validation is SRQR-governed; psychometric validation is a separate (COSMIN-type) matter." }, { "relation_type": "see_also", "target_slug": "preference-study", "notes": "Qualitative preference and treatment-experience components feeding value/PFDD evidence are reported under SRQR." }, { "relation_type": "see_also", "target_slug": "pro-rwe", "notes": "Qualitative patient-experience strands accompanying real-world PRO evidence are SRQR-governed." }, { "relation_type": "see_also", "target_slug": "burden-of-disease-cost-of-illness", "notes": "When burden evidence includes patient/caregiver qualitative interviews, that component is reported per SRQR." }, { "relation_type": "used_with", "target_slug": "study-protocol-or-sap-elements", "notes": "Pre-specification of the qualitative approach and analysis plan makes SRQR's analysis and trustworthiness items credible rather than retrofitted." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "journal", "hta", "fda" ] }, { "slug": "stard", "name": "STARD (Standards for Reporting of Diagnostic Accuracy Studies)", "short_definition": "A 30-item EQUATOR-hosted reporting checklist (plus a participant flow diagram) that specifies the minimum content needed to report a study estimating the accuracy of a test against a reference standard; in RWE it governs reporting of claims/EHR phenotype- and outcome-algorithm validation studies that estimate PPV, sensitivity, and specificity.", "long_description": "**What it is** — **STARD (Standards for Reporting of Diagnostic Accuracy Studies)** is a reporting\nguideline whose current version, **STARD 2015**, is a **30-item checklist** plus a recommended\nparticipant **flow diagram** for studies that estimate the *accuracy* of a test (index test) against a\n**reference standard** (\"gold standard\"). Its purpose is to make a diagnostic accuracy study fully and\ntransparently reportable so a reader can judge internal validity (risk of bias), applicability, and the\nestimates themselves (sensitivity, specificity, predictive values, likelihood ratios). It is maintained\nas part of the **EQUATOR Network** library of reporting guidelines; the checklist was first published in\n2003 (Bossuyt et al., a 25-item list) and substantially revised in 2015 (Bossuyt et al.) with a\ncompanion explanation-and-elaboration paper (Cohen et al., 2016). STARD is a *reporting* standard — it\nprescribes what to write, not how to design or appraise the study. In real-world evidence, the natural\n\"diagnostic test\" is a **computable phenotype / claims or EHR algorithm** (e.g., a 1-inpatient-or-2-\noutpatient code rule, an NLP outcome classifier, a registry case definition) and the \"reference\nstandard\" is the validated truth (adjudicated chart review, linked gold-standard registry); STARD is the\nguideline for reporting the *validation study* that estimates how well that algorithm performs.\n\n**When to use** — Apply STARD whenever the study object is the **diagnostic/classification accuracy of a\ntest or algorithm against a reference standard**, and you are reporting it for a peer-reviewed journal, a\nregulatory submission, or an HTA/payer dossier. In RWE this is precisely the **algorithm/phenotype\nvalidation study**: estimating the positive predictive value, sensitivity, specificity, and (where the\ndesign allows) negative predictive value of a claims or EHR case-finding algorithm against adjudicated\nchart review or a linked gold standard. Decision rule for choosing the right guideline: use **STARD**\nwhen the *estimand is test accuracy* (PPV/Se/Sp of an algorithm or biomarker); use **STROBE** or its\npharmacoepidemiology extension **RECORD/RECORD-PE** when the study is an *etiologic/comparative\nobservational study* whose outcome happens to be algorithm-defined; use **HARPER** or the **ENCePP**\nchecklist for the *protocol* of a non-interventional comparative study; use **TRIPOD+AI** (not STARD)\nwhen the index test is a multivariable **prediction model** that outputs a risk score rather than a\nbinary classification validated against a reference standard; and use **STARD-AI** when the index test\nis an AI/ML diagnostic system. A study can need more than one: a comparative-effectiveness paper that\nembeds an algorithm validation sub-study reports the main study under STROBE/RECORD-PE and the\nvalidation sub-study under STARD.\n\n**What it requires** — STARD 2015's 30 items compel reporting of the elements that determine whether an\naccuracy estimate is trustworthy and transferable, organized across title/abstract, introduction,\nmethods, results, and discussion. The substantive methods and results items that matter most for RWE\nvalidation are: an explicit **study question and intended use / target population** (item 1–4); the\n**index test (algorithm) definition** in enough operational detail to reproduce it — exact codes, code\npositions (inpatient vs outpatient), diagnosis fields, time windows, and the version/date of the code\nset (items 10a, 11–13); the **reference standard** and its rationale, including how the truth was\nascertained (adjudication process, blinding) and why it is credible (items 10b, 12b); **participant\nflow** with a diagram and the **eligibility, sampling, and time interval** between index test and\nreference standard (items 5–9, 19, 21); **how indeterminate/missing results and patients lost to the\nreference standard were handled** (items 23, 24, 25); a **2×2 cross-tabulation** of index-test results\nby reference-standard status (item 23); **accuracy estimates with precision** (PPV, sensitivity,\nspecificity, and confidence intervals, item 24); and any **subgroup, threshold, or sensitivity\nanalyses** (item 25). For claims/EHR work the high-leverage requirements are: report the **verification\nscheme** (was the reference standard applied to a sample of test-positives only, or to test-positives\nand a sample of test-negatives — i.e., is this a PPV-only design or a full Se/Sp design, and is there\n**partial-verification / spectrum bias**); report the **time-zero alignment** between the algorithm date\nand the reference-standard date; and report **how the validation sample's spectrum and prevalence relate\nto the target database** so readers can judge transportability of PPV (which is prevalence-dependent).\n\n**When NOT to use — limitations and common misapplications** — STARD is a **reporting checklist, not a\nrisk-of-bias instrument and not a quality score**; the appraisal tool for diagnostic accuracy studies is\n**QUADAS-2**, and completing STARD says nothing about whether the study was *well designed* — only\nwhether it was *fully described*. Concrete failure modes: (1) **Wrong guideline for the question** —\nusing STARD to report a comparative observational drug study because its outcome is algorithm-defined;\nthat study is STROBE/RECORD-PE, and STARD covers only the embedded validation sub-study, if any. (2)\n**Using STARD for a prediction model** — a multivariable risk score validated by discrimination and\ncalibration belongs under TRIPOD+AI, not STARD. (3) **Checklist-as-theater** — page-numbering all 30\nitems while leaving the algorithm codes, the reference-standard adjudication process, or the\nverification scheme vague defeats the purpose; the value is reproducibility, not the completed grid. (4)\n**Reporting PPV as if it were portable** — PPV is prevalence-dependent, so a PPV validated in one\ndatabase or era does not transfer to a target cohort with different case mix; STARD requires the spectrum\nand prevalence information that makes this judgeable, and omitting it is a common, consequential lapse.\n(5) **Ignoring partial verification** — validating only test-positives yields PPV but *cannot* yield\nsensitivity or specificity; reporting Se/Sp from a test-positive-only sample is a misapplication STARD's\nflow and 2×2 items are designed to expose. (6) **Confusing \"STARD-compliant report\" with \"valid\nalgorithm\"** — a fully STARD-compliant paper can still describe an algorithm too inaccurate to use; the\nchecklist surfaces that, it does not prevent it.\n\n**How it maps to this catalog** — In this repo, STARD is the *reporting standard* for the validation\nconcepts; each requirement is implemented by a concept a methodologist can build and pre-specify against:\n- The core RWE \"index test\" / algorithm definitions STARD asks you to report reproducibly:\n **diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe** (code rules, positions, time windows),\n **ehr-phenotyping-algorithms-rwe** and **outcome-algorithm-construction-rwe** (algorithm\n construction), and **procedure-identification-and-measurement-in-claims-ehr**.\n- The validation study itself and its accuracy estimands: **algorithm-validation** and\n **claims-outcome-algorithm-ppv-sensitivity-rwe** operationalize the PPV/Se/Sp/2×2 results items;\n **endpoint-adjudication-chart-review-rwe** implements the reference-standard / adjudication items\n (10b, 12b) STARD requires you to describe.\n- The consequences of imperfect accuracy STARD's estimates feed into:\n **misclassification-bias-correction-rwe** and **external-adjustment-validation-substudy-bias-\n correction-rwe** use the validated Se/Sp/PPV to correct downstream effect estimates, and\n **quantitative-bias-analysis-toolkit-rwe** propagates that uncertainty.\n- Applicability/transportability of a prevalence-dependent PPV across databases:\n **medicare-ffs-ma-commercial-claims-differences-rwe** (case mix and coding-intensity differences),\n **fit-for-purpose-data-assessment-rwe**, and **database-feasibility-attrition-funnel-rwe** (the\n participant-flow item in a data context). General data handling lives in **claims-analysis**.\n\n**Applied note (claims/EHR/registry RWE).** A STARD-compliant validation of a claims case-finding\nalgorithm should state the intended use and target database; give the exact algorithm (codes, positions,\ndiagnosis fields, time windows, code-set version) so it is reproducible; describe the reference standard\n(e.g., two-physician adjudicated chart review with disagreement resolution) and blinding; show a flow\ndiagram and the index-to-reference time interval; present the 2×2 table; report PPV with a confidence\ninterval and — only if test-negatives were also verified — sensitivity and specificity; declare any\npartial-verification design explicitly; and characterize how the validation sample's prevalence and\nspectrum compare to the analytic cohort, because the PPV you report will not transfer to a database with\na different case mix without that context.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "diagnostic-accuracy", "algorithm-validation", "phenotype", "equator", "rwe" ], "aliases": [ "STARD", "STARD 2015", "Standards for Reporting of Diagnostic Accuracy Studies", "Standards for Reporting of Diagnostic Accuracy" ], "applies_to_study_types": [ "diagnostic_accuracy", "algorithm_validation" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "doi": "10.1136/bmj.h5527", "url": "https://doi.org/10.1136/bmj.h5527", "citation_text": "Bossuyt PM, Reitsma JB, Bruns DE, et al. STARD 2015: an updated list of essential items for reporting diagnostic accuracy studies. BMJ. 2015;351:h5527.", "year": 2015, "authors_short": "Bossuyt et al.", "notes": "Canonical STARD 2015 statement defining the 30-item checklist and recommended participant flow diagram; the version in current use." }, { "role": "explain", "doi": "10.1136/bmjopen-2016-012799", "url": "https://doi.org/10.1136/bmjopen-2016-012799", "citation_text": "Cohen JF, Korevaar DA, Altman DG, et al. STARD 2015 guidelines for reporting diagnostic accuracy studies: explanation and elaboration. BMJ Open. 2016;6(11):e012799.", "year": 2016, "authors_short": "Cohen et al.", "notes": "Item-by-item explanation and elaboration with rationale and good-reporting examples for all 30 STARD 2015 items." }, { "role": "explain", "doi": "10.1373/49.1.1", "url": "https://doi.org/10.1373/49.1.1", "citation_text": "Bossuyt PM, Reitsma JB, Bruns DE, et al. Towards complete and accurate reporting of studies of diagnostic accuracy: the STARD initiative. Clinical Chemistry. 2003;49(1):1-6.", "year": 2003, "authors_short": "Bossuyt et al.", "notes": "Original 2003 STARD statement (25 items) establishing the initiative; superseded for reporting by STARD 2015 but useful historical context." }, { "role": "use", "url": "https://www.equator-network.org/reporting-guidelines/stard/", "citation_text": "STARD (Standards for Reporting of Diagnostic Accuracy Studies). EQUATOR Network reporting-guidelines library (maintained checklist, flow diagram, and extensions including STARD-AI).", "year": 2015, "authors_short": "EQUATOR Network", "notes": "Canonical maintained landing page with the checklist and flow diagram in usable formats and links to STARD extensions." } ], "relations": [ { "relation_type": "see_also", "target_slug": "algorithm-validation", "notes": "STARD is the reporting standard for an algorithm/phenotype validation study; algorithm-validation implements the accuracy estimands (PPV/Se/Sp) STARD requires you to report." }, { "relation_type": "see_also", "target_slug": "claims-outcome-algorithm-ppv-sensitivity-rwe", "notes": "Operationalizes the 2x2 table and PPV/sensitivity/specificity results items for claims outcome algorithms validated against a reference standard." }, { "relation_type": "see_also", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "The RWE \"index test\" - STARD requires the code rules, code positions, and time windows be reported in reproducible detail (items 10a, 11-13)." }, { "relation_type": "see_also", "target_slug": "endpoint-adjudication-chart-review-rwe", "notes": "Implements the reference-standard / adjudication items (10b, 12b) - how the truth is ascertained and blinded." }, { "relation_type": "used_with", "target_slug": "misclassification-bias-correction-rwe", "notes": "STARD-reported sensitivity/specificity/PPV feed quantitative correction of misclassified exposures or outcomes in the downstream analytic study." }, { "relation_type": "used_with", "target_slug": "external-adjustment-validation-substudy-bias-correction-rwe", "notes": "A STARD-reported validation sub-study supplies the accuracy parameters used to externally adjust the main effect estimate." }, { "relation_type": "see_also", "target_slug": "medicare-ffs-ma-commercial-claims-differences-rwe", "notes": "PPV is prevalence-dependent; case-mix and coding-intensity differences across payers determine whether a STARD-validated algorithm's PPV transports to the target database." }, { "relation_type": "see_also", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "Algorithm accuracy (the STARD estimand) is a central component of judging whether a data source is fit for the intended analytic purpose." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "journal", "fda", "ema", "hta" ] }, { "slug": "start-rwe", "name": "STaRT-RWE", "short_definition": "A structured, tabular template for transparently planning and reporting hypothesis-evaluating real-world-evidence studies on treatment effects, fixing in advance the implementation detail (design, time anchors, exposure/outcome/covariate operationalization, and analysis) that free-text protocols routinely leave ambiguous.", "long_description": "**What it is** — **STaRT-RWE (Structured Template for Planning and Reporting on the Implementation of Real World\nEvidence Studies)** is a reporting-and-planning template, published in *BMJ* in 2021 by Wang, Schneeweiss, and an\ninternational group of pharmacoepidemiologists, methodologists, regulators (FDA), and HTA/industry scientists. Its\ndefining feature is **tabular pre-specification**: rather than prose, it forces every implementation decision into\nstructured tables and figures — a design diagram with explicit time anchors (cohort entry, exposure assessment,\ncovariate look-back, washout, follow-up, outcome ascertainment), code-list and algorithm tables, and an analysis\nspecification — so that an independent team could reproduce the study from the document alone. It is a companion to,\nnot a competitor of, the prose-protocol templates: it pairs with the ISPE/ISPOR **HARPER** harmonized protocol\ntemplate (which carries the same implementation logic in narrative form) and feeds the downstream reporting checklists\n(**RECORD-PE**, **STROBE**, **ENCePP**). STaRT-RWE is maintained as a community good-practice instrument rather than\nby a single standards body; it is referenced in FDA real-world-evidence and EMA/ENCePP transparency expectations.\n\n**When to use** — Use STaRT-RWE for **hypothesis-evaluating, treatment-effect** non-interventional studies in\nroutinely collected data: active-comparator new-user cohorts, comparative effectiveness/safety studies, target-trial\nemulations, and post-authorization safety/effectiveness studies (PASS) built on claims, EHR, registry, or linked\ndata. It is appropriate at three moments — the **protocol/planning** stage (lock the design before data access), an\n**amendment/audit** stage (document deviations against the locked specification), and the **reporting** stage\n(publish the completed tables as a manuscript appendix, an HTA dossier annex, or a regulatory submission artifact).\nDecision rules for choosing it over siblings: use STaRT-RWE (or HARPER) — *not* PRISMA-P — when the object is a\n**primary** RWE study rather than a systematic review of studies; use STaRT-RWE's structured *tables* alongside\nHARPER's *narrative* protocol (they are designed to interlock, not substitute); and use STaRT-RWE for the\n*implementation specification*, then report the finished study against **RECORD-PE/STROBE** (item-level reporting\nchecklists) and, for EU PASS, the **ENCePP** checklist. For descriptive/disease-natural-history or\nhypothesis-generating studies, STaRT-RWE's treatment-effect scaffolding is heavier than needed.\n\n**What it requires** — STaRT-RWE enforces the implementation domains where unforced errors actually occur in RWE:\n(1) a **design figure with explicit time-zero and all assessment windows**, which surfaces immortal-time and\nlook-back/look-forward errors before they happen; (2) **data-source fitness-for-use** — provenance, capture, lags,\nlinkage, and the rationale that the source can measure exposure, outcome, and confounders well enough for the\nquestion; (3) **exposure, outcome, and covariate operational definitions as code-list/algorithm tables** with\nwindows, settings (inpatient/outpatient counts, e.g. 1-IP/2-OP rules), grace periods, and — for outcomes —\n**phenotype/algorithm validation** (PPV/sensitivity and the validation source); (4) **eligibility and cohort\nconstruction** with an attrition table from source population to analytic cohort; (5) the **estimand** — target\npopulation, treatment strategies, and handling of **intercurrent events** (switching, discontinuation, death) under\nan ITT-like or per-protocol contrast; (6) **confounding control** — the covariate set, the adjustment method\n(propensity-score or high-dimensional PS, matching/weighting), and balance diagnostics; (7) **missing data and\nloss-to-follow-up/attrition** handling; and (8) a pre-specified **sensitivity and quantitative-bias analysis** plan\n(alternative windows, negative controls, E-value). It also requires **version-controlled code lists and parameters**\nso the specification is auditable.\n\n**When NOT to use — limitations and common misapplications** — STaRT-RWE is a **transparency template, not a\nvalidity guarantee, a risk-of-bias instrument, or a quality score**. Concrete failure modes: (1) **Template-as-theater**\n— filling every cell while the design is biased; a perfectly tabulated immortal-time error is still an immortal-time\nerror. Completing STaRT-RWE does **not** make an observational estimate causal or unconfounded; it makes the design\n*legible* so reviewers can judge it. (2) **Confusing it with a critical-appraisal tool** — STaRT-RWE describes what\nwas done; bias is graded with **ROBINS-I**, and reporting completeness with **RECORD-PE/STROBE**. Do not cite a\ncompleted STaRT-RWE table as evidence of low risk of bias. (3) **Wrong instrument for the object** — using it for a\nsystematic review (that is PRISMA-P/PRISMA 2020) or for an RCT protocol (SPIRIT). (4) **Wrong scope** — applying its\ntreatment-effect machinery to a purely descriptive or hypothesis-generating study, where it adds friction without\nprotecting against the relevant errors. (5) **Specification drift** — locking a template and then silently deviating;\nthe value is in pre-specification *plus* documented amendments, not the blank form. (6) **Stopping at STaRT-RWE for\nreporting** — it specifies implementation; journals and regulators still expect the item-level **RECORD-PE/STROBE**\nreporting checklist and, for EU PASS, **ENCePP**.\n\n**How it maps to this catalog** — Each STaRT-RWE requirement is implemented by a concept in this repo:\n- **Time-zero and design figure** → **target-trial-emulation** (specify the hypothetical trial and align follow-up at\n a defensible time zero) and **active-comparator-new-user** (the new-user washout + active-comparator structure that\n fixes time zero and curbs confounding by indication).\n- **Exposure/eligibility construction in routine data** → **active-comparator-new-user** and **claims-analysis**\n (NDC/fill-date/days-supply exposure, enrollment requirements, MA-vs-FFS capture caveats).\n- **Outcome/covariate definitions and phenotype validation** →\n **diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe** (1-IP/2-OP rules, time windows, position/setting, PPV).\n- **Estimand and intercurrent events** → **estimands-ate-att-intercurrent-events-rwe** (ATE/ATT, ITT vs per-protocol,\n switching/discontinuation/death handling).\n- **Confounding control** → **high-dimensional-propensity-score-hdps-rwe** (proxy selection and PS adjustment when\n key confounders are unmeasured).\n- **Attrition and loss to follow-up** → **attrition-and-loss-to-follow-up-rwe** (CONSORT-style flow, informative\n censoring).\nRead STaRT-RWE as the **specification layer** that ties these concepts together; each catalog concept supplies the\noperational depth a single STaRT-RWE cell only summarizes.\n\n**Applied note (claims/EHR/registry RWE).** For a claims-based active-comparator new-user safety study, the\nSTaRT-RWE design table makes the high-leverage choices explicit and checkable in one view: the continuous\nenrollment + drug-free washout that establishes incident-user status, time zero set at the first qualifying fill\n(not at diagnosis — the classic immortal-time trap), covariates measured only in the pre-index window feeding a\nhigh-dimensional PS, an attrition table from source population to matched cohort, the estimand with its\nswitching/discontinuation rule, and a sensitivity row (washout length, grace period, negative-control outcome,\nE-value). In EHR or registry data the same template forces declaration of order-vs-administration exposure capture,\nlinkage to fills, and explicit observation windows so that visit-driven, potentially informative loss to follow-up\nis handled rather than ignored.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "protocol", "rwe", "pharmacoepidemiology", "transparency", "template" ], "aliases": [ "STaRT-RWE", "Structured Template for Planning and Reporting on the Implementation of Real World Evidence Studies", "StaRT-RWE template" ], "applies_to_study_types": [ "new_user", "active_comparator_new_user", "cer_observational", "target_trial_emulation", "claims_analysis", "ehr_study", "cohort_retrospective" ], "data_sources": [ "claims", "ehr", "registry", "linked", "multi-database" ], "citations": [ { "role": "introduce", "doi": "10.1136/bmj.m4856", "url": "https://doi.org/10.1136/bmj.m4856", "citation_text": "Wang SV, Pinheiro S, Hua W, et al. STaRT-RWE: structured template for planning and reporting on the implementation of real world evidence studies. BMJ. 2021;372:m4856.", "year": 2021, "authors_short": "Wang et al.", "notes": "Canonical STaRT-RWE statement defining the tabular planning-and-reporting template, including the design figure with explicit time anchors and the code-list, algorithm, and analysis tables." }, { "role": "explain", "doi": "10.1002/pds.5507", "url": "https://doi.org/10.1002/pds.5507", "citation_text": "Wang SV, Pottegård A, Crown W, et al. HARmonized Protocol Template to Enhance Reproducibility of hypothesis evaluating real-world evidence studies on treatment effects (HARPER): a good practices report of a joint ISPE/ISPOR task force. Pharmacoepidemiology and Drug Safety. 2023;32(1):44-55.", "year": 2023, "authors_short": "Wang et al.", "notes": "The narrative protocol-template companion to STaRT-RWE; HARPER carries the same implementation logic in prose and is designed to interlock with the STaRT-RWE tables for protocol writing." }, { "role": "use", "doi": "10.1136/bmj.k3532", "url": "https://doi.org/10.1136/bmj.k3532", "citation_text": "Langan SM, Schmidt SA, Wing K, et al. The reporting of studies conducted using observational routinely collected health data statement for pharmacoepidemiology (RECORD-PE). BMJ. 2018;363:k3532.", "year": 2018, "authors_short": "Langan et al.", "notes": "Item-level reporting checklist for the completed study; use downstream of STaRT-RWE to verify reporting completeness for journal/regulatory submission. STaRT-RWE specifies implementation, RECORD-PE/STROBE report it." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "active-comparator-new-user", "notes": "STaRT-RWE provides the implementation-specification structure (time zero, washout, exposure/outcome definitions, estimand, analysis) for active-comparator new-user studies." }, { "relation_type": "applies_to", "target_slug": "cer-observational", "notes": "Use STaRT-RWE to pre-specify and report comparative effectiveness/safety observational studies in routine data." }, { "relation_type": "applies_to", "target_slug": "target-trial-emulation", "notes": "STaRT-RWE's design figure and time anchors operationalize the hypothetical-trial protocol of a target-trial emulation." }, { "relation_type": "applies_to", "target_slug": "claims-analysis", "notes": "STaRT-RWE forces explicit code lists, enrollment requirements, and data-fitness documentation for claims-based studies." }, { "relation_type": "used_with", "target_slug": "target-trial-emulation", "notes": "Specify the hypothetical trial and a defensible time zero; STaRT-RWE's design figure is the table that records it." }, { "relation_type": "used_with", "target_slug": "active-comparator-new-user", "notes": "The new-user washout + active-comparator structure is the canonical content of the STaRT-RWE design figure and eligibility table." }, { "relation_type": "used_with", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Implements the outcome/covariate algorithm tables (1-IP/2-OP rules, time windows, setting, PPV validation) that STaRT-RWE requires." }, { "relation_type": "used_with", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Supplies the estimand and intercurrent-event handling (ITT vs per-protocol, switching/discontinuation/death) that the STaRT-RWE analysis specification must declare." }, { "relation_type": "used_with", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "Implements the confounding-control specification (proxy selection, PS adjustment, balance diagnostics) summarized in the STaRT-RWE analysis table." }, { "relation_type": "used_with", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Implements the attrition flow and informative-censoring handling that STaRT-RWE's eligibility and follow-up tables require." }, { "relation_type": "complements", "target_slug": "claims-analysis", "notes": "Provides the operational depth (exposure capture, enrollment, MA-vs-FFS caveats) behind STaRT-RWE's data-source-fitness and exposure cells." }, { "relation_type": "see_also", "target_slug": "picots-framework-rwe", "notes": "PICOTS frames the question that the STaRT-RWE eligibility, exposure, comparator, and outcome tables then operationalize." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "hta", "journal" ] }, { "slug": "stratos", "name": "STRATOS (STRengthening Analytical Thinking for Observational Studies)", "short_definition": "A methodological initiative — not a single checklist — that produces topic-group guidance documents for the design and analysis of observational studies. Nine topic groups cover foundational areas including selection of variables and functional forms, missing data, measurement error, survival analysis, causal inference, and high-dimensional data. STRATOS supplements reporting guidelines (STROBE) with the statistical analysis guidance those checklists deliberately omit.", "long_description": "**What it is** — **STRATOS (STRengthening Analytical Thinking for Observational Studies)** is an\ninternational collaborative initiative that produces methodological guidance documents — not a single\nunified checklist — for the statistical analysis of observational studies. It was founded by\nWilli Sauerbrei, Patrick Royston, and colleagues and formally introduced in a 2014 *Statistics in\nMedicine* paper. STRATOS is organised into **nine topic groups (TGs)**, each producing technical\nguidance on a different analytic dimension: *TG1 — Selection of variables and functional forms*;\n*TG2 — Missing data*; *TG3 — Measurement error and misclassification*; *TG4 — Study design\n(matching)*; *TG5 — Causal inference, estimands, and target parameters*; *TG6 — High-dimensional\ndata*; *TG7 — Survival analysis*; *TG8 — Prediction modelling (overlaps with TRIPOD/PROBAST)*;\nand *TG9 — Meta-analysis of individual participant data*. Each TG publishes tutorials, primers, and\npeer-reviewed papers in statistics and biomedical journals. Unlike **STROBE** (which tells you what\nto *report* about your analysis) or **GRADE** (which evaluates evidence certainty), STRATOS tells\nyou *how to do the analysis correctly* — it fills the gap between the reporting checklist and\nthe textbook. The initiative is maintained at stratos-initiative.org and publications span BIOM,\n*Statistics in Medicine*, *American Journal of Epidemiology*, and *BMJ*.\n\n**When to use** — Reference STRATOS when you need authoritative, peer-reviewed guidance on *how* to\nconduct a specific analysis in an observational study, beyond what STROBE or RECORD require you to\nreport. Concrete use cases: (1) **Variable selection and functional forms (TG1)** — when building a\nmultivariable model and choosing between a priori variable sets, stepwise selection, or penalised\nregression; or when deciding whether to model continuous variables linearly or with splines/fractional\npolynomials. (2) **Missing data (TG2)** — when designing an imputation strategy (multiple imputation\nvs. single imputation vs. complete-case analysis) and checking whether the missing-at-random\nassumption holds. (3) **Measurement error (TG3)** — when exposure or covariate measurement error\nin claims, EHR, or survey data is likely to bias effect estimates. (4) **Causal inference (TG5)**\n— when translating a PICO into a causal diagram (DAG), selecting an estimand, or using propensity\nscores, g-estimation, or instrumental variables. (5) **Survival analysis (TG7)** — when choosing\nbetween Cox regression, accelerated failure time models, or competing-risk models and reporting\nassumptions. (6) **High-dimensional data (TG6)** — when hdPS, LASSO, or machine-learning variable\nselection is used in a claims or EHR study. Decision rule: STRATOS is a methods *guidance* resource,\nnot a reporting checklist; use it alongside (not instead of) STROBE, RECORD-PE, HARPER, or TRIPOD —\nthose checklists require disclosure of what was done, STRATOS explains how to do it well.\n\n**What it requires (checklist domains)** — STRATOS does not enforce a single checklist; instead, each\ntopic group issues guidance with *recommended analytical practices* and *common pitfalls*. Across TGs\nthe guidance collectively enforces these analytical disciplines:\n*TG1 (Variable selection and functional forms)*: pre-specification of the analysis strategy before\nlooking at outcome-predictor associations; avoiding data-driven selection of covariates without\npenalty or regularisation; and modelling continuous predictors with flexible forms (restricted cubic\nsplines or fractional polynomials) rather than arbitrary categorisation. *TG2 (Missing data)*:\ncharacterising the missing-data mechanism (MCAR, MAR, MNAR) before choosing a strategy; using\nmultiple imputation under MAR with a correctly specified imputation model (including the outcome); and\nconducting sensitivity analyses under MNAR assumptions. *TG3 (Measurement error)*: quantifying the\nlikely direction and magnitude of exposure-measurement error; using regression calibration, SIMEX,\nor probabilistic bias analysis where error is non-differential but substantial; and reporting\nsensitivity analyses. *TG5 (Causal inference)*: explicit specification of the **target estimand**\n(ATE, ATT, or per-protocol); use of **directed acyclic graphs (DAGs)** to guide covariate selection\nand identify confounders vs. mediators; and reporting the assumptions required for causal\ninterpretation. *TG7 (Survival analysis)*: checking proportional-hazards assumptions; considering\ncompeting risks; and using complementary log-log or Nelson-Aalen estimates rather than Kaplan-Meier\nwhen hazards are not proportional. *TG8 (Prediction)*: guidance that parallels TRIPOD/PROBAST on\ncalibration, overfitting, and internal validation.\n\n**When NOT to use — limitations and common misapplications** — (1) **Confusing STRATOS with a\nreporting guideline** — STRATOS is a methods guidance initiative, not a checklist authors submit with\nmanuscripts; journals require STROBE or RECORD, not a STRATOS sign-off. (2) **Using STRATOS in\nisolation without specifying which TG** — citing \"STRATOS\" without specifying the relevant topic\ngroup is vague; be specific (e.g., \"we followed STRATOS TG2 guidance on missing data\"). (3)\n**Treating STRATOS as a mandatory standard** — unlike TRIPOD or PRISMA, STRATOS guidance documents\nare educational outputs; they do not carry the same journal-submission or regulatory-submission\nweight as formal reporting checklists. (4) **Over-applying TG1 to causal studies** — STRATOS TG1\n(variable selection) guidance is primarily framed for prediction modelling; causal observational\nstudies should use DAG-guided pre-specified covariate sets per TG5, not algorithmic selection.\n(5) **Missing the TG5 / estimands connection** — many pharmacoepidemiologists use STRATOS TG5\ninformally for DAG reasoning but do not formally cite it; aligning with ICH E9(R1) estimand\nthinking (**ich-e9-r1** in this catalog) provides a complementary and increasingly regulatory-\naccepted framework. (6) **Assuming STRATOS replaces biostatistical peer review** — STRATOS\nguidance documents are written for a wide audience; complex analytic challenges in specific study\ntypes still require biostatistical co-authorship and review.\n\n**How it maps to this catalog** — STRATOS topic-group guidance directly supports the implementation\nof several core catalog methods: *TG1* (variable selection / functional forms) underpins rigorous\n**propensity-score-methods-psm-iptw** construction and the covariate specification step in any\nmultivariable model. *TG2* (missing data) implements the missing-data discipline required by\n**multiple-imputation-longitudinal-rwe** and contributes to **missing-data-pattern-table-rwe**\n(the reporting side). *TG3* (measurement error) connects to **quantitative-bias-analysis-toolkit-rwe**\n(probabilistic bias analysis accounts for differential and non-differential measurement error) and\nto **claims-outcome-algorithm-ppv-sensitivity-rwe** (misclassification of outcomes is a form of\nmeasurement error). *TG4* (design/matching) informs **matching-estimators-rwe** and the design\nchoices in active-comparator new-user studies (**active-comparator-new-user**). *TG5* (causal\ninference) directly operationalises the DAG discipline underlying **dag-framework** and provides\nthe estimand-selection complement to **estimands-ate-att-intercurrent-events-rwe**. *TG6* (high-\ndimensional data) provides the methods basis for **high-dimensional-propensity-score-hdps-rwe**.\n*TG7* (survival analysis) informs time-to-event modelling across catalog concepts including\n**accelerated-failure-time-models**, **competing-risks-methods**, and **kaplan-meier-methods**. *TG8*\n(prediction) bridges to **tripod** and **probast** in this catalog. The reporting obligations that\nSTRATOS guidance helps fulfil are captured by **strobe**, **record-pe**, and **harper** — the\ncompanion reporting checklists in the guidelines section of this catalog.", "primary_category": "Guideline", "tags": [ "guideline", "methods-guidance", "statistical-analysis", "observational-studies", "missing-data", "measurement-error", "causal-inference", "variable-selection", "survival-analysis" ], "aliases": [ "STRATOS", "STRATOS initiative", "STRengthening Analytical Thinking for Observational Studies", "STRATOS TG1", "STRATOS TG2", "STRATOS TG5", "STRATOS TG7" ], "applies_to_study_types": [ "cer_observational", "cohort_prospective", "cohort_retrospective", "cross_sectional" ], "data_sources": [ "claims", "ehr", "registry", "linked", "primary" ], "citations": [ { "role": "introduce", "doi": "10.1002/sim.6265", "url": "https://doi.org/10.1002/sim.6265", "citation_text": "Sauerbrei W, Abrahamowicz M, Altman DG, et al. STRengthening Analytical Thinking for Observational Studies: the STRATOS initiative. Statistics in Medicine. 2014;33(30):5413-5432.", "year": 2014, "authors_short": "Sauerbrei et al.", "notes": "Founding paper introducing the STRATOS initiative, its rationale, organisation into nine topic groups, and the gap it fills between reporting guidelines (STROBE) and methodological guidance for observational study analysis." }, { "role": "explain", "doi": "10.1186/s12874-024-02294-3", "url": "https://doi.org/10.1186/s12874-024-02294-3", "citation_text": "Heinze G, Baillie M, Lusa L, et al. Regression without regrets — initial data analysis is a prerequisite for multivariable regression (STRATOS TG2 / TG8 guidance). BMC Medical Research Methodology. 2024;24(1):210.", "year": 2024, "authors_short": "Heinze et al.", "notes": "STRATOS-aligned guidance on initial data analysis before multivariable regression; illustrates the journal-publication model through which topic groups disseminate actionable analytic guidance (TG1/TG2/TG8 scope)." }, { "role": "use", "url": "https://www.stratos-initiative.org", "citation_text": "STRATOS Initiative. STRengthening Analytical Thinking for Observational Studies — topic group guidance documents, primers, and tutorials (maintained). stratos-initiative.org.", "year": 2014, "authors_short": "STRATOS Initiative", "notes": "Official initiative website; access point for TG-specific guidance documents, primers, and forthcoming tutorials across all nine topic groups." } ], "relations": [ { "relation_type": "used_with", "target_slug": "strobe", "notes": "STROBE requires disclosure of what was done; STRATOS (especially TG1, TG2, TG5) guides how to do it correctly — the two complement each other." }, { "relation_type": "used_with", "target_slug": "record-pe", "notes": "RECORD-PE and STRATOS together cover reporting and analytical rigor for pharmacoepidemiological studies using routinely collected data." }, { "relation_type": "used_with", "target_slug": "dag-framework", "notes": "STRATOS TG5 causal inference guidance is the primary methods resource for DAG-based covariate selection implemented by the DAG framework concept." }, { "relation_type": "used_with", "target_slug": "multiple-imputation-longitudinal-rwe", "notes": "STRATOS TG2 provides the authoritative methods guidance for the missing-data handling practices this concept implements." }, { "relation_type": "used_with", "target_slug": "propensity-score-methods-psm-iptw", "notes": "STRATOS TG1 and TG5 guide variable selection and estimand specification for propensity-score construction." }, { "relation_type": "used_with", "target_slug": "quantitative-bias-analysis-toolkit-rwe", "notes": "STRATOS TG3 measurement-error guidance underpins probabilistic bias analysis for misclassified exposures and outcomes." }, { "relation_type": "used_with", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "STRATOS TG6 (high-dimensional data) provides the methods basis for hdPS variable selection." }, { "relation_type": "see_also", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "STRATOS TG5 and ICH E9(R1) are complementary estimand frameworks; TG5 focuses on causal observational designs while ich-e9-r1 covers trial and RWE estimands." }, { "relation_type": "see_also", "target_slug": "tripod", "notes": "STRATOS TG8 (prediction modelling guidance) complements TRIPOD reporting; together they cover both rigorous analysis and complete reporting of prediction model studies." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "journal", "fda", "ema" ] }, { "slug": "strega", "name": "STREGA (STrengthening the REporting of Genetic Association Studies)", "short_definition": "A STROBE reporting-guideline extension that adds genetics-specific items - genotyping and quality control, Hardy-Weinberg equilibrium, population stratification, multiple testing, haplotypes, and replication - to the minimum content a genetic association study should report; developed by an international workshop (Little, Higgins et al.) and hosted within the EQUATOR Network.", "long_description": "**What it is** — **STREGA (STrengthening the REporting of Genetic Association Studies)** is a\nreporting-guideline *extension of STROBE* (the Strengthening the Reporting of Observational Studies\nin Epidemiology statement) tailored to genetic association studies. It does not replace STROBE's\n22 items; it adds genetics-specific reporting expectations to a subset of them. STREGA was produced\nby an international, multidisciplinary workshop convened in Ottawa (2006) and published in 2009\nsimultaneously across several journals (Little, Higgins, Ioannidis, and colleagues; PLoS Medicine,\nEuropean Journal of Epidemiology, and others). It is maintained as a STROBE extension within the\n**EQUATOR Network** library. Its purpose is narrow and specific: to make the genotyping technology\nand quality control, the handling of population stratification, the treatment of Hardy-Weinberg\nequilibrium, the multiple-testing/multiple-comparison strategy, the modeling of genetic\ncontrasts/haplotypes, and the replication and synthesis of findings *transparent and reproducible*\nin reports of gene-disease and gene-environment association studies. It is a reporting standard, not\nan analysis method and not a study-quality score.\n\n**When to use** — Apply STREGA when you are *reporting* (or peer-reviewing, or registering the\nreporting expectations for) an observational **genetic association study** — a candidate-gene study,\na genome-wide association study (GWAS), or a meta-analysis/synthesis of genetic associations — most\ncommonly built on cohort, case-control, or cross-sectional designs. In the RWE/HEOR world the\nrealistic trigger is **pharmacogenomic and biomarker-association work**: a claims- or EHR-linked\nbiobank study testing whether a genotype predicts drug response, an adverse drug reaction, or a\ndisease phenotype; a registry-plus-genomics study; or any submission/publication where a genotype is\nthe exposure or effect-modifier of interest. Decision rule for the right family member: use **STROBE**\nfor a non-genetic observational study; use **STREGA** when genotype is central and you need the\ngenetics-specific items layered on top of STROBE; use **RECORD / RECORD-PE** when the study is built\non *routinely collected* health data and the data-provenance items dominate (and combine RECORD with\nSTREGA if a routinely-collected-data study also carries a genetic exposure); and use **STROBE-MR** —\nnot STREGA — when the design is a **Mendelian randomization** study using genetic variants as\ninstruments, because the reporting burden there is about instrument validity, not association\nreporting. STREGA is for the *association* report; STROBE-MR is for the *instrumental-variable causal*\nreport.\n\n**What it requires** — STREGA keeps STROBE's backbone (title/abstract, background, objectives,\neligibility, variables, data sources/measurement, bias, study size, statistical methods, participant\nflow, descriptive and outcome data, limitations, generalizability, funding) and adds genetics-specific\nreporting at the points where genetic studies go wrong: (1) **Genotyping and laboratory methods** —\nthe platform/assay, call thresholds, blinding of genotyping to outcome, and the **genotyping error\nrate / call rate / quality-control** procedures, including how SNPs or samples failing QC were\nhandled. (2) **Hardy-Weinberg equilibrium (HWE)** — whether HWE was tested (typically in controls),\nthe method and threshold, and how departures were interpreted (a classic flag for genotyping error or\npopulation structure). (3) **Population stratification** — how confounding by ancestry was addressed\n(e.g., restriction, family-based design, genomic control, principal components / ancestry adjustment).\n(4) **Multiple testing / multiple comparisons** — the number of variants and models tested and the\ncorrection or significance threshold (e.g., genome-wide significance), to constrain false-positive\nreporting. (5) **Modeling of the genetic contrast** — the inheritance model assumed (additive,\ndominant, recessive, genotypic), how haplotypes were inferred, and treatment of gene-gene and\ngene-environment interaction. (6) **Replication and synthesis** — whether findings were replicated in\nan independent sample, and, for pooled work, how between-study heterogeneity and meta-analysis were\nhandled. For pharmacogenomic RWE specifically, the genetics items must be reported *alongside* the\nordinary RWE reporting burden the linked data create — data fitness, the validity of the\nclaims/EHR-defined outcome or drug-response phenotype, time-zero alignment, and confounding control —\nbecause a genotype-outcome association inherits every weakness of the phenotype it is regressed on.\n\n**When NOT to use — limitations and common misapplications** — (1) **It is a reporting checklist, not\na risk-of-bias instrument and not a quality score.** A fully STREGA-compliant paper can still report a\nbadly confounded, underpowered, or unreplicated association transparently; completeness of reporting\nis necessary, not sufficient, for validity. Do not use STREGA to *grade* studies — that is the job of\nappraisal tools (e.g., Newcastle-Ottawa for the underlying observational design). (2) **Wrong design.**\nUsing STREGA for a non-genetic observational study where plain **STROBE** (or **RECORD/RECORD-PE** for\nroutinely-collected data) is the correct standard is over-reach; using STROBE alone for a genetic\nassociation study under-reports the genotyping-QC, HWE, stratification, and multiple-testing items that\nSTREGA exists to force. (3) **Wrong genetic extension.** Applying STREGA to a **Mendelian\nrandomization** study instead of **STROBE-MR** misses the instrument-validity reporting (relevance,\nindependence, exclusion-restriction, pleiotropy) that is the whole point of an MR report. (4)\n**Checklist-as-theater.** Ticking items while leaving the genotyping error rate, the HWE result, the\nstratification adjustment, or the multiple-testing threshold vague defeats the purpose; the value is\nthe specific numbers, not the page count. (5) **Reporting compliance is not causal inference** — a\nSTREGA-complete pharmacogenomic claims study with an unvalidated response phenotype and unaddressed\nancestry confounding is still not a credible causal claim.\n\n**How it maps to this catalog** — STREGA is a reporting layer; in this repo its substantive\nrequirements are implemented (and should be pre-specified and appraised) by these concepts:\n- **Genotyping QC and the validity of the genetic/phenotype measurement**: `algorithm-validation`\n supplies the validation discipline (sensitivity/specificity/PPV) that the genotyping-error-rate and\n phenotype-definition items demand by analogy; `biomarker-defined-cohort-rwe` operationalizes\n constructing a cohort around a genotype/biomarker exposure.\n- **Data fitness for a linked-genomics RWE study**: `fit-for-purpose-data-assessment-rwe` covers\n whether the linked claims/EHR/biobank substrate is adequate for the genotype-outcome question.\n- **The structured question and confounding/stratification frame**: `picots-framework-rwe` declares\n population/exposure(genotype)/comparator/outcome/timing/setting; `baseline-characteristics-and-covariate-balance-rwe`\n supports the ancestry/covariate reporting that population-stratification control requires.\n- **Missing data and synthesis**: `multiple-imputation-longitudinal-rwe` for missing genotype/covariate\n data, and `meta-analysis-obs` for the replication-and-synthesis item when genetic associations are\n pooled across studies/biobanks.\n- **Distinguish from the instrumental-variable cousin**: `instrumental-variables-pharmacoepi-rwe` is\n the analytic basis of Mendelian randomization — when genetics are used as *instruments* rather than\n as the *exposure of interest*, the report belongs under STROBE-MR, not STREGA.\n\n**Applied note (pharmacogenomic claims/EHR/biobank RWE).** For a study linking a genotyped biobank to\nclaims/EHR to test a genotype-drug-response or genotype-ADR association, STREGA forces you to report\nthe genotyping platform, call rate, and error rate; the HWE test in an appropriate reference group;\nthe ancestry-adjustment (principal components) used to defuse population stratification; the\ngenome-wide or candidate-set multiple-testing threshold; the inheritance model; and whether the\nsignal replicated. None of that displaces the ordinary RWE reporting burden of the linked data — the\nvalidity of the algorithm-defined drug-response or outcome phenotype, time-zero alignment, and\nconfounding control — and a credible report must satisfy both layers.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "genetic-association", "pharmacogenomics", "strobe-extension", "gwas", "equator" ], "aliases": [ "STREGA", "STrengthening the REporting of Genetic Association Studies", "STROBE extension for genetic association studies" ], "applies_to_study_types": [ "cohort_prospective", "cohort_retrospective", "case_control", "cross_sectional" ], "data_sources": [ "registry", "ehr", "linked", "multi-database" ], "citations": [ { "role": "introduce", "doi": "10.1371/journal.pmed.1000022", "url": "https://doi.org/10.1371/journal.pmed.1000022", "citation_text": "Little J, Higgins JPT, Ioannidis JPA, et al. STrengthening the REporting of Genetic Association Studies (STREGA) - an extension of the STROBE statement. PLoS Medicine. 2009;6(2):e1000022.", "year": 2009, "authors_short": "Little et al.", "notes": "Canonical STREGA statement; defines the genetics-specific reporting items layered onto STROBE (genotyping QC, Hardy-Weinberg equilibrium, population stratification, multiple testing, haplotypes, replication)." }, { "role": "explain", "doi": "10.1007/s10654-008-9302-y", "url": "https://doi.org/10.1007/s10654-008-9302-y", "citation_text": "Little J, Higgins JPT, Ioannidis JPA, et al. Strengthening the reporting of genetic association studies (STREGA): an extension of the STROBE statement. European Journal of Epidemiology. 2009;24(1):37-55.", "year": 2009, "authors_short": "Little et al.", "notes": "Co-published version of the STREGA statement with extended item-by-item explanation and rationale for each genetics-specific addition." }, { "role": "explain", "doi": "10.1371/journal.pmed.0040296", "url": "https://doi.org/10.1371/journal.pmed.0040296", "citation_text": "von Elm E, Altman DG, Egger M, et al. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. PLoS Medicine. 2007;4(10):e296.", "year": 2007, "authors_short": "von Elm et al.", "notes": "Parent STROBE statement that STREGA extends; STREGA inherits all STROBE items and adds to a subset of them." }, { "role": "use", "url": "https://www.equator-network.org/reporting-guidelines/strobe-strega/", "citation_text": "STREGA (STrengthening the REporting of Genetic Association Studies). EQUATOR Network reporting-guidelines library (maintained checklist and STROBE-extension links).", "year": 2009, "authors_short": "EQUATOR Network", "notes": "Maintained landing page with the STREGA checklist in usable form and its placement among STROBE extensions." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "cohort-prospective", "notes": "STREGA reporting expectations apply to prospective cohort genetic association studies." }, { "relation_type": "applies_to", "target_slug": "cohort-retrospective", "notes": "STREGA reporting expectations apply to retrospective cohort genetic association studies." }, { "relation_type": "applies_to", "target_slug": "case-control", "notes": "STREGA reporting expectations apply to case-control genetic association studies (the dominant design for candidate-gene and many GWAS analyses)." }, { "relation_type": "see_also", "target_slug": "algorithm-validation", "notes": "The genotyping error/call-rate and phenotype-definition items parallel the validation discipline (sensitivity/specificity/PPV) this concept supplies for algorithm-defined measurements." }, { "relation_type": "used_with", "target_slug": "biomarker-defined-cohort-rwe", "notes": "Operationalizes building a cohort around a genotype/biomarker exposure - the construction STREGA reports." }, { "relation_type": "used_with", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "For linked claims/EHR/biobank genomics studies, assess whether the data substrate is adequate for the genotype-outcome question before reporting it under STREGA." }, { "relation_type": "used_with", "target_slug": "picots-framework-rwe", "notes": "Declares the population/exposure(genotype)/comparator/outcome/timing/setting frame the report must specify." }, { "relation_type": "see_also", "target_slug": "baseline-characteristics-and-covariate-balance-rwe", "notes": "Supports the ancestry/covariate reporting that population-stratification control requires." }, { "relation_type": "see_also", "target_slug": "meta-analysis-obs", "notes": "Synthesis target for STREGA's replication-and-synthesis item when genetic associations are pooled across studies or biobanks." }, { "relation_type": "see_also", "target_slug": "instrumental-variables-pharmacoepi-rwe", "notes": "When genetic variants are used as instruments (Mendelian randomization) rather than as the exposure of interest, the report belongs under STROBE-MR, not STREGA." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "journal", "fda", "ema" ] }, { "slug": "strobe-ams", "name": "STROBE-AMS (STROBE Extension for Antimicrobial Stewardship)", "short_definition": "A topic-specific extension of the STROBE reporting checklist that adds antimicrobial-resistance and stewardship items to the reporting of observational epidemiological studies, so that exposure (antimicrobial use), resistance outcomes, and microbiological methods are described transparently enough to be appraised and compared.", "long_description": "**What it is** — **STROBE-AMS (Strengthening the Reporting of Observational Studies in Epidemiology — extension for\nstudies on Antimicrobial resistance and antimicrobial Stewardship)** is a topic-specific extension of the parent\nSTROBE reporting guideline. It was developed by Tacconelli and colleagues (BMJ Open, 2016) through a two-round Delphi\nconsensus and is catalogued in the **EQUATOR Network** library of reporting guidelines. STROBE-AMS does **not** replace\nSTROBE; it layers additional, AMS-specific reporting items onto the 22-item STROBE checklist so that observational\nstudies of antibiotic exposure, antimicrobial resistance (AMR), and stewardship interventions describe their design,\nmicrobiology, and exposure measurement with enough granularity for readers, evidence synthesists, and decision-makers\nto judge validity and pool results. Like all STROBE-family tools, it is a **reporting** instrument — a transparency\ncontract about what an article must disclose — not an instrument for designing the study or grading its quality.\n\n**When to use** — Apply STROBE-AMS when you are **reporting an observational epidemiological study** (cohort,\ncase-control, or cross-sectional) whose core scientific content concerns antimicrobial use, the selection or spread of\nantimicrobial resistance, or the effect of stewardship interventions, and you are writing for a peer-reviewed journal,\nan HTA/payer evidence package, or an evidence-synthesis input. Decision rule for choosing the right STROBE family\nmember: use **plain STROBE** for a generic observational study with no AMS-specific exposure/outcome; use **STROBE-AMS**\nwhen the study's exposure is antimicrobial use or its outcome is resistance/stewardship effectiveness and the\nmicrobiological detail (organism, susceptibility testing, breakpoints, resistance definitions) is load-bearing; and if\nthe study is built on **routinely-collected health data** (claims, EHR, registries, dispensing or microbiology\ndatabases), layer **RECORD** (or RECORD-PE for pharmacoepidemiology) on top, because STROBE-AMS says little about\ndatabase provenance, linkage, code lists, and data cleaning. STROBE-AMS applies to the *report*; it is not a protocol\ntemplate (use HARPER or the ENCePP checklist for that) and not a randomized-trial guideline (use CONSORT for an RCT of a\nstewardship intervention).\n\n**What it requires** — On top of the standard STROBE domains (title/abstract, background, objectives, design, setting,\nparticipants and eligibility, variables, data sources/measurement, bias, study size, statistical methods, descriptive\nand outcome results, limitations, generalizability, funding), STROBE-AMS adds AMS-specific reporting that maps onto the\nsame validity concerns that govern any real-world-data study: (1) **Exposure definition and measurement** — how\nantimicrobial exposure was ascertained and quantified (defined daily doses, days of therapy, dispensing vs\nadministration, look-back windows), the analogue of phenotype/algorithm specification in claims and EHR work; (2)\n**Microbiological methods and resistance definitions** — the organism(s), specimen source, susceptibility-testing\nmethod, breakpoint system (e.g., EUCAST/CLSI) and version, and the explicit rule that classified an isolate as\nresistant, so that the resistance \"outcome\" is reproducible; (3) **Time relationships** — the temporal ordering of\nexposure and resistance, which is the AMS-specific face of time-zero alignment and immortal-time avoidance; (4)\n**Confounding and case-mix** — patient- and unit-level confounders (severity, prior hospitalization, device exposure,\nco-medications) and how they were handled; (5) **Population, denominator, and setting** — the at-risk denominator and\ncare setting that make rates interpretable and transportable. The extension's intent is that two studies reporting the\n\"same\" association can be compared only if exposure metrics, resistance definitions, and denominators are stated\nprecisely — exactly the comparability problem that defeats naive pooling of observational estimates.\n\n**When NOT to use — limitations and common misapplications** — STROBE-AMS is a reporting checklist, with the hard\nlimits that implies. (1) **It is not a risk-of-bias instrument and not a quality score.** A fully STROBE-AMS-compliant\npaper can still be badly confounded; appraise non-randomized AMS studies with **ROBINS-I**, not with checklist\ncompleteness. Do not sum ticked items into a \"quality score\" — STROBE's authors explicitly warn against this. (2)\n**Completing the checklist does not make the study causal.** Transparent reporting of an exposure–resistance\nassociation is necessary, not sufficient; the causal claim rests on design (active comparator, new-user, target-trial\nemulation), not on disclosure. (3) **Wrong family member / wrong layer.** Using plain STROBE where the AMS-specific\nmicrobiology and exposure items are needed under-reports the study; conversely, using STROBE-AMS alone for a study\nbuilt on claims or linked EHR omits the database-provenance, code-list, and linkage reporting that **RECORD/RECORD-PE**\nexist to enforce — these are complementary, not interchangeable. (4) **Wrong design entirely.** STROBE-AMS is for\nobservational designs; a cluster-randomized stewardship trial is reported with CONSORT, and a study *protocol* with\nHARPER/ENCePP. (5) **Checklist-as-theater.** Pointing every item at a page number while leaving the resistance\ndefinition, breakpoint version, or exposure metric vague defeats the purpose — the value is the substantive disclosure,\nnot the cross-reference table.\n\n**How it maps to this catalog** — In this repo, STROBE-AMS's reporting demands are *implemented* (i.e., the underlying\ndesign/measurement work the report must describe) by these concepts:\n- **Exposure/outcome ascertainment** (antimicrobial-use metrics; resistance/infection outcome definitions):\n **diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe**, **claims-outcome-algorithm-ppv-sensitivity-rwe**,\n **algorithm-validation**, and **misclassification-bias-correction-rwe** for quantifying outcome/exposure\n misclassification the report must acknowledge.\n- **Design validity and confounding control** the paper must report: **active-comparator-new-user**,\n **target-trial-emulation**, and **high-dimensional-propensity-score-hdps-rwe**.\n- **Estimand and time structure** (STROBE-AMS's time-relationship and effect-measure items):\n **estimands-ate-att-intercurrent-events-rwe**.\n- **Population, attrition, and generalizability** (denominator, loss to follow-up, transportability of rates):\n **attrition-and-loss-to-follow-up-rwe**, **database-feasibility-attrition-funnel-rwe**, and\n **generalizability-transportability-external-validity-rwe**.\n- **Data-source provenance** when the study is database-built: **claims-analysis** and\n **medicare-ffs-ma-commercial-claims-differences-rwe** — and, at the reporting layer, pair STROBE-AMS with RECORD/RECORD-PE.\n\n**Applied note (claims/EHR/registry RWE).** A claims- or EHR-based study of, say, fluoroquinolone exposure and\nsubsequent resistant Gram-negative infection should, to satisfy STROBE-AMS *and* be appraisable, report the antibiotic\nexposure metric and look-back window (the **diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe** and exposure-window\nlogic), the validated outcome definition with its PPV/sensitivity (**claims-outcome-algorithm-ppv-sensitivity-rwe**,\n**algorithm-validation**), the breakpoint system/version used to define resistance, time-zero alignment and the\ncomparator strategy (**active-comparator-new-user**), confounding control (**high-dimensional-propensity-score-hdps-rwe**),\nthe analytic estimand (**estimands-ate-att-intercurrent-events-rwe**), the enrolment/attrition funnel and at-risk\ndenominator (**database-feasibility-attrition-funnel-rwe**, **attrition-and-loss-to-follow-up-rwe**), and the database\nprovenance, linkage, and code lists (which is where **RECORD/RECORD-PE** and **claims-analysis** carry the load STROBE-AMS\ndoes not).", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "strobe", "antimicrobial-stewardship", "antimicrobial-resistance", "pharmacoepidemiology", "equator", "observational" ], "aliases": [ "STROBE-AMS", "STROBE for Antimicrobial Stewardship", "STROBE extension for antimicrobial resistance and stewardship", "Strengthening the Reporting of Observational Studies in Epidemiology - Antimicrobial Stewardship" ], "applies_to_study_types": [ "cohort_prospective", "cohort_retrospective", "case_control", "cross_sectional" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "doi": "10.1136/bmjopen-2015-010134", "url": "https://doi.org/10.1136/bmjopen-2015-010134", "citation_text": "Tacconelli E, Cataldo MA, Paul M, et al. STROBE-AMS: recommendations to optimise reporting of epidemiological studies on antimicrobial resistance and informing improvement in antimicrobial stewardship. BMJ Open. 2016;6(2):e010134.", "year": 2016, "authors_short": "Tacconelli et al.", "notes": "Canonical STROBE-AMS statement; two-round Delphi consensus deriving and testing the AMS-specific extension of STROBE." }, { "role": "explain", "doi": "10.1371/journal.pmed.0040296", "url": "https://doi.org/10.1371/journal.pmed.0040296", "citation_text": "von Elm E, Altman DG, Egger M, et al. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. PLoS Medicine. 2007;4(10):e296.", "year": 2007, "authors_short": "von Elm et al.", "notes": "Parent STROBE statement that STROBE-AMS extends; defines the 22-item core checklist the AMS items layer onto." }, { "role": "explain", "doi": "10.1371/journal.pmed.1001885", "url": "https://doi.org/10.1371/journal.pmed.1001885", "citation_text": "Benchimol EI, Smeeth L, Guttmann A, et al. The REporting of studies Conducted using Observational Routinely-collected health Data (RECORD) statement. PLoS Medicine. 2015;12(10):e1001885.", "year": 2015, "authors_short": "Benchimol et al.", "notes": "Complementary STROBE extension for routinely-collected health data; pair with STROBE-AMS for claims/EHR/registry AMS studies to report database provenance, linkage, and code lists." }, { "role": "use", "url": "https://www.equator-network.org/reporting-guidelines/strobe-ams/", "citation_text": "STROBE-AMS reporting guideline. EQUATOR Network reporting-guidelines library (maintained landing page with the checklist and links to STROBE extensions).", "year": 2016, "authors_short": "EQUATOR Network", "notes": "Maintained EQUATOR landing page hosting the STROBE-AMS checklist and related STROBE-family extensions." } ], "relations": [ { "relation_type": "is_variant_of", "target_slug": "strobe", "notes": "STROBE-AMS extends the parent STROBE checklist with antimicrobial-resistance and stewardship-specific reporting items; STROBE remains the base requirement." }, { "relation_type": "used_with", "target_slug": "record", "notes": "For AMS studies built on routinely-collected health data, layer RECORD on top of STROBE-AMS to report database provenance, linkage, and code lists." }, { "relation_type": "used_with", "target_slug": "record-pe", "notes": "RECORD-PE adds pharmacoepidemiology-specific reporting (exposure definitions, drug code lists) that complements the antimicrobial-exposure items in STROBE-AMS." }, { "relation_type": "applies_to", "target_slug": "cohort-prospective", "notes": "Use when reporting a prospective cohort study of antimicrobial use, resistance, or stewardship effectiveness." }, { "relation_type": "applies_to", "target_slug": "cohort-retrospective", "notes": "Use when reporting a retrospective/claims- or EHR-based cohort study of antimicrobial exposure and resistance." }, { "relation_type": "applies_to", "target_slug": "case-control", "notes": "Use when reporting a case-control study of risk factors for resistant infection or stewardship outcomes." }, { "relation_type": "applies_to", "target_slug": "cross-sectional", "notes": "Use when reporting a cross-sectional prevalence study of resistance or antimicrobial use." }, { "relation_type": "see_also", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Implements the exposure/outcome-ascertainment detail (antimicrobial-use and resistant-infection definitions) that STROBE-AMS requires authors to report." }, { "relation_type": "see_also", "target_slug": "algorithm-validation", "notes": "Supplies the validation evidence (PPV/sensitivity) for algorithm-defined exposures and outcomes that the report must disclose." }, { "relation_type": "see_also", "target_slug": "active-comparator-new-user", "notes": "Design that controls confounding by indication; STROBE-AMS requires the comparator and time-relationship choices to be reported, but does not make them." }, { "relation_type": "see_also", "target_slug": "target-trial-emulation", "notes": "The design framework whose elements (eligibility, time zero, assignment) STROBE-AMS asks an observational AMS study to describe transparently." }, { "relation_type": "see_also", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "Confounding-control method the report must describe under STROBE-AMS's bias/statistical-methods items." }, { "relation_type": "see_also", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Clarifies the estimand and intercurrent-event handling behind the effect measures STROBE-AMS asks authors to state." }, { "relation_type": "see_also", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Supports the participant-flow, denominator, and loss-to-follow-up reporting STROBE/STROBE-AMS require." }, { "relation_type": "see_also", "target_slug": "generalizability-transportability-external-validity-rwe", "notes": "Informs the generalizability item — how AMS rates and effects transport across settings and denominators." }, { "relation_type": "see_also", "target_slug": "claims-analysis", "notes": "Data-source context for claims-based AMS studies; pair STROBE-AMS reporting with RECORD/RECORD-PE for provenance." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "journal", "hta", "fda", "ema" ] }, { "slug": "strobe-css", "name": "STROBE for Cross-Sectional Studies", "short_definition": "The cross-sectional-design checklist of the STROBE statement — the 22-item (with design-specific items) reporting guideline for observational epidemiology, applied to studies that measure exposure and outcome at a single point in (or window of) time. It is the design-specific sibling of the cohort and case-control STROBE checklists, maintained within the STROBE Initiative and the EQUATOR Network.", "long_description": "**What it is** — **STROBE for Cross-Sectional Studies (STROBE-CS)** is not a separate\nstatement but the *cross-sectional design variant* of the **Strengthening the Reporting\nof Observational studies in Epidemiology (STROBE)** statement (von Elm, Vandenbroucke\net al., 2007). The single STROBE statement covers the three core analytical\nobservational designs — cohort, case-control, and cross-sectional — in 22 items spanning\ntitle/abstract, introduction, methods, results, and discussion; the STROBE Initiative\nand the **EQUATOR Network** then publish *design-specific checklists* so that the items\nwhose wording differs by design (notably item 6 on participant\neligibility/selection, item 12 on statistical methods, and item 13 on the participant\nflow) are stated in cross-sectional terms. STROBE-CS is therefore the checklist you\nhand to authors, peer reviewers, and editors when the study estimates prevalence,\nassociations, or diagnostic/screening quantities from data captured at a single\npoint in time rather than followed forward (cohort) or sampled on outcome status\n(case-control). Its purpose is *reporting transparency* — to make the design,\nsetting, eligibility, variables, statistical handling, and limitations legible enough\nthat a reader can judge validity and a methodologist could in principle reproduce the\nanalysis. It is maintained as a public, freely downloadable checklist within EQUATOR.\n\n**When to use** — Use STROBE-CS whenever you write, review, or register the report of a\n**cross-sectional observational study**: a prevalence survey, a single-timepoint\nassociation study, a serosurvey, a cross-sectional analysis of a registry or claims\nsnapshot, or a cross-sectional baseline analysis carved out of a larger cohort. It is\nthe right checklist for a peer-reviewed journal submission, for the descriptive or\nprevalence component of an HTA/payer dossier, and as the reporting backbone of a\ncross-sectional PASS or non-interventional study report. Decision rule for choosing the\ncorrect family member: use the **cohort** STROBE checklist when participants are\nfollowed forward from an exposure/time-zero to an incident outcome; the **case-control**\nchecklist when sampling is conditioned on outcome status; and **STROBE-CS** when\nexposure and outcome are ascertained together at one point or window with no\nforward follow-up. Critically, **if the cross-sectional study is built on routinely\ncollected health data (claims, EHR, disease/device registries), STROBE-CS alone is not\nenough — add RECORD (RECORD-PE for pharmacoepidemiology)**, which extends STROBE with\nthe database-specific items (code lists, data-cleaning, linkage, validation) that\nreal-world data demand. For a *prospective protocol* of an RWD study use **HARPER** or\nthe **ENCePP checklist**, not a reporting checklist. STROBE-CS governs the *report of a\ncompleted* cross-sectional study; it is not a planning instrument.\n\n**What it requires** — STROBE-CS enforces the substantive reporting domains that make a\ncross-sectional analysis interpretable: an informative, design-labeled **title/abstract**\n(item 1 — \"cross-sectional\" stated explicitly); **rationale and pre-specified\nobjectives/hypotheses** (items 2–3); a transparent **design, setting, and recruitment\nwindow** (items 4–5); **eligibility criteria and the sampling/selection method** with\nthe source population defined (item 6 — the cross-sectional phrasing emphasizes how\nparticipants were selected at the single timepoint); precise **operational definitions\nof every exposure, outcome, predictor, confounder, and effect modifier**, with data\nsources and measurement methods and comparability across groups (items 7–8); explicit\nhandling of **bias, study size, and quantitative variables** (items 9–11); a complete\n**statistical-methods** section covering confounding control, subgroups/interactions,\n**missing-data** handling, sampling-strategy accounting, and **sensitivity analyses**\n(item 12); a **participant flow** with numbers at each stage and reasons for\nnon-participation (item 13); **descriptive, outcome, and main results** with adjusted\nestimates, confidence intervals, and the analytic denominator (items 14–16); a **key-results\nsummary, limitations (direction and magnitude of potential bias), generalizability, and\nfunding** (items 18–22). For real-world cross-sectional data these generic items carry\nspecific weight: data-source **fitness-for-use** and the snapshot's coverage window must\nbe reported under design/setting; **phenotype/algorithm definitions** for prevalent\nconditions (e.g., a 1-inpatient/2-outpatient claims rule) and any **validation** belong\nunder variables/measurement; and because there is no forward follow-up, the report must\nbe explicit that the design supports **prevalence and association, not incidence or\ncausal temporality**, and must surface missingness and selection at the snapshot.\n\n**When NOT to use — limitations and common misapplications** — STROBE-CS is a *reporting*\nchecklist, **not a risk-of-bias instrument, not a quality score, and not a study-validity\ncertificate**. Concrete failure modes: (1) **Treating completion as appraisal** — to\n*grade* the internal validity of a cross-sectional/RWE study use a risk-of-bias tool\n(ROBINS-I, ROBINS-E) or a design appraisal checklist (e.g., JBI for cross-sectional\nstudies, NOS); STROBE tells you *what was reported*, not *whether it was done well*. (2)\n**Checklist-as-theater** — citing \"reported per STROBE\" while leaving the source\npopulation, the algorithm definitions, the missing-data approach, or the sensitivity\nanalyses vague defeats the purpose; the value is the substance behind each item, not the\nticked box or page reference. (3) **Wrong design variant** — using the cross-sectional\nchecklist to report a cohort (forward follow-up) or case-control (outcome-conditioned\nsampling) study mislabels the design and invites a temporality claim the data cannot\nsupport. (4) **Using plain STROBE-CS where RECORD/RECORD-PE is required** — a claims- or\nEHR-based cross-sectional study reported without RECORD's database items (code lists,\nlinkage, data cleaning, validation) is under-reported by current RWE standards. (5)\n**Inferring causation or incidence** — completing STROBE-CS does not convert a single\nsnapshot into evidence of effect or of temporal sequence; reverse causation and\nprevalent-case (length-time) bias remain, and the limitations item exists precisely to\nstate this. (6) **Wrong specialty/topic extension** — surgical observational studies use\n**STROCSS**, nutritional epidemiology uses **STROBE-nut**, molecular epidemiology\n**STROME-ID**, genetic association **STREGA**; do not substitute the generic\ncross-sectional checklist where a topic extension governs.\n\n**How it maps to this catalog** — STROBE-CS's reporting items are operationalized by the\nconcept entries a reviewer can check the report against:\n- The data substrate and its fitness-for-use (design/setting/source items):\n **claims-analysis** and **medicare-ffs-ma-commercial-claims-differences-rwe** describe\n the coverage, denominators, and snapshot limitations that the setting and limitations\n items must surface.\n- Variable/measurement definitions (items 7–8): **diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe**\n and **claims-outcome-algorithm-ppv-sensitivity-rwe** implement the prevalent-condition\n phenotyping and the PPV/sensitivity validation that the report must disclose;\n **procedure-identification-and-measurement-in-claims-ehr** does the same for\n procedure-defined variables.\n- Participant flow and selection (items 6, 13): **database-feasibility-attrition-funnel-rwe**\n and **attrition-and-loss-to-follow-up-rwe** supply the enumerated source-to-analytic\n funnel and missingness accounting the flow item demands.\n- Statistical methods, estimand, and confounding (items 9, 12, 16): **estimands-ate-att-intercurrent-events-rwe**\n and **estimand-analysis-traceability-rwe** make the target quantity and its analytic\n chain explicit; **high-dimensional-propensity-score-hdps-rwe** and\n **active-comparator-new-user** are the confounding-control machinery to report when a\n cross-sectional design is pushed toward an association/comparative estimate (with the\n standing caveat that cross-sectional data cannot establish temporality the way\n **target-trial-emulation** can for longitudinal designs).\nThese concepts implement *what to report*; STROBE-CS specifies *that you must report it,\nand where*.\n\n**Applied note (claims/EHR/registry RWE).** For a cross-sectional prevalence or\nassociation study on a claims/EHR snapshot, satisfy STROBE-CS by: stating the exact\ndata source, version, and snapshot/coverage window (design/setting); defining every\ncondition, exposure, and covariate by an explicit, citable algorithm and reporting any\nPPV/sensitivity validation (variables/measurement); presenting the\nsource-population-to-analytic-sample funnel with reasons for exclusion (participant\nflow); pre-specifying the estimand and confounding-control strategy and reporting\nmissing-data and sensitivity analyses (statistical methods); and stating plainly in the\nlimitations that a single snapshot supports prevalence and association — not incidence or\ncausal temporality — with prevalent-case and selection bias named. Because this is RWD,\nlayer **RECORD/RECORD-PE** on top of STROBE-CS rather than relying on STROBE alone.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "observational", "cross-sectional", "strobe", "equator", "prevalence" ], "aliases": [ "STROBE-CS", "STROBE cross-sectional checklist", "STROBE for cross-sectional studies", "Strengthening the Reporting of Observational Studies in Epidemiology (cross-sectional)" ], "applies_to_study_types": [ "cross_sectional" ], "data_sources": [ "claims", "ehr", "registry", "linked", "multi-database" ], "citations": [ { "role": "introduce", "doi": "10.1371/journal.pmed.0040296", "url": "https://doi.org/10.1371/journal.pmed.0040296", "citation_text": "von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP; STROBE Initiative. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. PLoS Medicine. 2007;4(10):e296.", "year": 2007, "authors_short": "von Elm et al.", "notes": "Canonical STROBE statement and 22-item checklist covering cohort, case-control, and cross-sectional designs; the cross-sectional checklist is the design-specific rendering of these items." }, { "role": "explain", "doi": "10.1371/journal.pmed.0040297", "url": "https://doi.org/10.1371/journal.pmed.0040297", "citation_text": "Vandenbroucke JP, von Elm E, Altman DG, et al. Strengthening the Reporting of Observational Studies in Epidemiology (STROBE): explanation and elaboration. PLoS Medicine. 2007;4(10):e297.", "year": 2007, "authors_short": "Vandenbroucke et al.", "notes": "Item-by-item explanation and elaboration, including the cross-sectional-specific guidance for the eligibility, statistical-methods, and participant-flow items." }, { "role": "use", "url": "https://www.strobe-statement.org/checklists/", "citation_text": "STROBE Initiative / EQUATOR Network. STROBE checklists — cross-sectional study checklist (maintained, downloadable). Strengthening the Reporting of Observational Studies in Epidemiology.", "year": 2007, "authors_short": "STROBE Initiative", "notes": "Maintained landing page hosting the design-specific cross-sectional checklist in usable formats; the operational artifact authors and reviewers actually apply." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "cross-sectional", "notes": "STROBE-CS is the design-specific reporting checklist for cross-sectional observational studies." }, { "relation_type": "see_also", "target_slug": "claims-analysis", "notes": "For cross-sectional studies on claims, the data-source, denominator, and snapshot limitations this concept describes must be reported under STROBE's design/setting and limitations items." }, { "relation_type": "see_also", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Implements the prevalent-condition phenotype definitions that STROBE items 7-8 (variables, measurement) require the report to disclose." }, { "relation_type": "see_also", "target_slug": "claims-outcome-algorithm-ppv-sensitivity-rwe", "notes": "Supplies the PPV/sensitivity validation of algorithm-defined variables that the measurement and limitations items should report." }, { "relation_type": "see_also", "target_slug": "database-feasibility-attrition-funnel-rwe", "notes": "Provides the source-population-to-analytic-sample funnel that STROBE's participant flow item (13) requires for cross-sectional samples." }, { "relation_type": "see_also", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Makes the target quantity explicit for the statistical-methods and main-results items; cross-sectional data constrain the estimand to association/prevalence." }, { "relation_type": "see_also", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "Confounding-control machinery to report under the statistical-methods item when a cross-sectional design is pushed toward an adjusted association." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "journal", "hta", "fda", "ema" ] }, { "slug": "strobe-equity", "name": "STROBE-Equity (STROBE Extension for Health Equity)", "short_definition": "Reporting-guideline extension that adds equity-specific items on top of the base STROBE checklist, specifying the minimum information an observational study should report when health equity is a research question; developed by an international consensus group and maintained within Cochrane Equity / the EQUATOR Network.", "long_description": "**What it is** — **STROBE-Equity (Strengthening the Reporting of Observational Studies in Epidemiology — extension\nfor health equity)** is a reporting-guideline *extension* that supplements the base STROBE checklist with\nequity-specific reporting items for observational studies in which health equity is part of the research question.\nIt was developed by an international consensus group (Dewidar, Shamseer, Welch, Tugwell and colleagues) using\nEQUATOR/Cochrane methodology and published in 2025 as a combined extension checklist and elaboration in JAMA\nNetwork Open; it is maintained as a project of **Cochrane Equity** within the **EQUATOR Network**. Its purpose is\nto make the equity dimensions of an observational study transparent and auditable: who the disadvantaged and\ncomparison groups are, how equity-relevant characteristics were defined and measured, what stratified/effect-\nmodification analyses were planned and conducted, and how findings bear on avoidable, unjust health differences.\nCrucially, STROBE-Equity is an *add-on*, not a replacement — authors complete base STROBE and then report the\nadditional equity items on top of it. The equity dimensions it operationalizes are framed by **PROGRESS-Plus**\n(Place of residence, Race/ethnicity/culture/language, Occupation, Gender/sex, Religion, Education, Socioeconomic\nstatus, Social capital, plus context-specific factors such as age, disability, and other vulnerabilities).\n\n**When to use** — Apply STROBE-Equity when you are reporting an **observational study** (cohort, case-control,\ncross-sectional, including claims/EHR/registry real-world evidence) in which equity is an explicit aim — e.g.,\nthe study examines health differences across socioeconomic, racial/ethnic, geographic, sex/gender, or other\nPROGRESS-Plus strata, or intends its results to inform policy affecting disadvantaged populations. It is the\nappropriate extension for a peer-reviewed observational manuscript with an equity objective, and increasingly for\nthe observational evidence underpinning an **HTA/payer dossier** where equity impact is a deliberative criterion\n(e.g., NICE/ICER distributional and health-inequality considerations). **Decision rule for the right family\nmember:** observational study with an equity question → base **STROBE + STROBE-Equity**; randomized trial with an\nequity question → **CONSORT-Equity**; systematic review with an equity question → **PRISMA-Equity / equity-focused\nPRISMA extension**; observational study with *no* equity question → base **STROBE alone**. STROBE-Equity sits on an\northogonal axis to **RECORD / RECORD-PE** (routinely-collected-data and pharmacoepidemiologic extensions): an\nequity-focused study built on claims or EHR data can and should apply RECORD/RECORD-PE *and* STROBE-Equity\ntogether — they govern different facets of the same report.\n\n**What it requires** — Layered on the base STROBE items, STROBE-Equity compels explicit reporting of the equity\narchitecture of the study. The added items require, in substance: (1) an equity-framed rationale and objectives,\nstating which populations or PROGRESS-Plus dimensions are of interest and why the differences studied are\nconsidered avoidable and unjust; (2) explicit **definition and measurement of equity-relevant characteristics** —\nhow race/ethnicity, socioeconomic position, sex/gender, geography, etc. were operationalized, the data source and\nclassification scheme, and the limits of those measures (a real-world-data crux, since claims and EHR capture\nthese variably and often by proxy); (3) clear specification of the **disadvantaged group(s)** and the comparison\ngroup(s), and the reference category chosen; (4) pre-specified **stratified analyses and equity-relevant effect\nmodification / subgroup analyses**, distinguishing planned from post-hoc, with appropriate handling of multiplicity\nand sparse strata; (5) reporting of **differential attrition, missingness, and selection** across equity strata —\nnot just overall — because loss to follow-up and unobserved person-time are rarely equity-neutral; (6) honest\ndiscussion of **transportability and generalizability** to the disadvantaged populations the study aims to inform;\nand (7) interpretation framed against existing inequities, including potential for the analysis itself to entrench\nor mask them. For claims/EHR/registry RWE, these generic items carry specific operational weight: phenotype and\noutcome algorithms may perform differently across subgroups, time-zero and enrollment requirements can\ndifferentially exclude disadvantaged patients, and confounding control must not absorb the very equity-relevant\nexposures of interest.\n\n**When NOT to use — limitations and common misapplications** — STROBE-Equity is a *reporting* checklist, not a\nrisk-of-bias instrument, not a quality score, and not an analysis method. Concrete failure modes: (1) **Treating\nit as a replacement for base STROBE** — it is an extension; the base 22 items still apply and STROBE-Equity adds\non top. (2) **Wrong design** — using STROBE-Equity for a randomized trial (use **CONSORT-Equity**) or a systematic\nreview (use a **PRISMA equity extension**). (3) **Confusing the checklist with the framework** — **PROGRESS-Plus**\nis the conceptual framework of equity dimensions; STROBE-Equity is the reporting checklist that operationalizes it.\nThe two are complementary, not interchangeable. (4) **Equity-as-theater** — bolting equity items onto a study that\nhas no genuine equity question, or ticking the items while reporting only a single overall effect, defeats the\npurpose; the value is transparent equity analysis, not box-count. (5) **Reporting ≠ doing** — completing the equity\nitems does not make the study causally valid or its subgroup contrasts unconfounded; a fully STROBE-Equity-\ncompliant paper can still rest on a biased design, mismeasured race/SES variables, or an underpowered,\ndata-dredged subgroup. (6) **Mismeasured equity variables passed off as analyzed** — naming a PROGRESS-Plus\ndimension in the checklist while it is captured only by a crude or missing-laden proxy (e.g., area-level income as\nindividual SES) reports the *intent* without the substance.\n\n**How it maps to this catalog** — In this repo, STROBE-Equity's requirements are implemented (or appraised) by\nthese concepts:\n- **Equity-relevant effect modification / subgroup analysis** (added items on stratified analysis): implemented by\n **causal-mediation-effect-modification-rwe**, which supplies the estimand-aware machinery for analyzing and\n reporting heterogeneity across PROGRESS-Plus strata rather than a single average effect.\n- **Estimand and intercurrent-event clarity for subgroup contrasts**: **estimands-ate-att-intercurrent-events-rwe**\n — the equity contrast must specify a target population and estimand, not an undefined \"subgroup effect.\"\n- **Differential attrition and selection across strata** (the equity attrition item): **attrition-and-loss-to-\n follow-up-rwe** and **database-feasibility-attrition-funnel-rwe**, reported *by* equity stratum, not only overall.\n- **Measurement validity of equity-relevant and outcome variables in RWE**: **algorithm-validation**,\n **diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe**, and **claims-outcome-algorithm-ppv-sensitivity-rwe** —\n because phenotype/outcome algorithms can have subgroup-specific performance that distorts equity conclusions.\n- **Transportability/external validity to disadvantaged populations**: **generalizability-transportability-\n external-validity-rwe**.\n- **Design and confounding spine the equity analysis sits on**: **target-trial-emulation**,\n **active-comparator-new-user**, **high-dimensional-propensity-score-hdps-rwe**, and the structured-question\n discipline of **picots-framework-rwe** (which forces the population and comparator to be named before equity\n strata are examined).\nThese design/measurement concepts are *not* STROBE-Equity items in themselves; they are the substantive analyses\nwhose results the equity items demand you report transparently.\n\n**Applied note (claims/EHR/registry RWE).** An equity-focused comparative claims study should report how each\nPROGRESS-Plus dimension was captured (e.g., race/ethnicity from enrollment files of variable completeness,\nindividual vs area-level SES, Medicare-vs-Medicaid-vs-commercial coverage as a confounded marker of access — see\n**medicare-ffs-ma-commercial-claims-differences-rwe** and **claims-analysis**), present the attrition funnel and\nloss to follow-up *split by* equity stratum, validate that outcome/phenotype algorithms perform comparably across\nsubgroups, pre-specify the equity-relevant effect modification it will test, and state plainly how findings\ntransport to the disadvantaged populations they are meant to inform — all on top of, not instead of, the base\nSTROBE checklist.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "strobe", "health-equity", "progress-plus", "observational", "equator", "subgroup-analysis" ], "aliases": [ "STROBE-Equity", "STROBE Extension for Health Equity", "STROBE-Equity extension", "Strengthening the Reporting of Observational Studies in Epidemiology - Equity extension" ], "applies_to_study_types": [ "cohort_prospective", "cohort_retrospective", "case_control", "cross_sectional" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "doi": "10.1001/jamanetworkopen.2025.32512", "url": "https://doi.org/10.1001/jamanetworkopen.2025.32512", "citation_text": "Dewidar O, Shamseer L, Welch V, et al. Improving the reporting on health equity in observational research (STROBE-Equity): extension checklist and elaboration. JAMA Network Open. 2025;8(9):e2532512.", "year": 2025, "authors_short": "Dewidar et al.", "notes": "Canonical STROBE-Equity statement — the consensus extension checklist with item-by-item elaboration and examples; defines the equity-specific items added to base STROBE." }, { "role": "explain", "doi": "10.1371/journal.pmed.0040296", "url": "https://doi.org/10.1371/journal.pmed.0040296", "citation_text": "von Elm E, Altman DG, Egger M, et al. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. PLoS Medicine. 2007;4(10):e296.", "year": 2007, "authors_short": "von Elm et al.", "notes": "Parent guideline that STROBE-Equity extends; the base 22-item checklist still applies and the equity items are reported on top of it." }, { "role": "explain", "doi": "10.1016/j.jclinepi.2013.08.005", "url": "https://doi.org/10.1016/j.jclinepi.2013.08.005", "citation_text": "O'Neill J, Tabish H, Welch V, et al. Applying an equity lens to interventions: using PROGRESS ensures consideration of socially stratifying factors to illuminate inequities in health. Journal of Clinical Epidemiology. 2014;67(1):56-64.", "year": 2014, "authors_short": "O'Neill et al.", "notes": "Defines the PROGRESS-Plus framework of equity dimensions that STROBE-Equity operationalizes for reporting." }, { "role": "use", "url": "https://methods.cochrane.org/equity/projects/strobe-equity", "citation_text": "STROBE-Equity project page. Cochrane Equity (Campbell and Cochrane Equity Methods Group), maintained within the EQUATOR Network — checklist, guidance, and project resources.", "year": 2025, "authors_short": "Cochrane Equity", "notes": "Maintained landing page hosting the checklist and supporting materials; the authoritative source for the current checklist version." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "cohort-prospective", "notes": "Use alongside base STROBE when a prospective cohort study has an explicit health-equity objective." }, { "relation_type": "applies_to", "target_slug": "cohort-retrospective", "notes": "Use alongside base STROBE for retrospective/claims/EHR cohort studies examining equity-relevant strata." }, { "relation_type": "applies_to", "target_slug": "case-control", "notes": "Applies to case-control studies with an equity question; report equity-relevant strata and definitions." }, { "relation_type": "applies_to", "target_slug": "cross-sectional", "notes": "Applies to cross-sectional studies (e.g., prevalence/disparities surveys) reporting equity dimensions." }, { "relation_type": "used_with", "target_slug": "causal-mediation-effect-modification-rwe", "notes": "Implements the stratified-analysis / equity-relevant effect-modification reporting items beyond a single average effect." }, { "relation_type": "used_with", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "The equity (subgroup) contrast must name a target population and estimand; STROBE-Equity requires this be explicit, not an undefined subgroup effect." }, { "relation_type": "used_with", "target_slug": "picots-framework-rwe", "notes": "Structures the population/comparator/timing the equity items must declare before strata are examined." }, { "relation_type": "see_also", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "STROBE-Equity requires attrition and loss to follow-up be reported by equity stratum, not only overall." }, { "relation_type": "see_also", "target_slug": "database-feasibility-attrition-funnel-rwe", "notes": "The attrition funnel should be examined for differential exclusion of disadvantaged subgroups." }, { "relation_type": "see_also", "target_slug": "algorithm-validation", "notes": "Outcome/phenotype algorithms can perform differently across subgroups; equity reporting depends on validated, subgroup-aware measurement." }, { "relation_type": "see_also", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Phenotype algorithm performance may vary by equity stratum, biasing equity conclusions if unexamined." }, { "relation_type": "see_also", "target_slug": "generalizability-transportability-external-validity-rwe", "notes": "STROBE-Equity requires explicit discussion of transportability to the disadvantaged populations the study aims to inform." }, { "relation_type": "see_also", "target_slug": "medicare-ffs-ma-commercial-claims-differences-rwe", "notes": "Coverage type (FFS/MA/Medicaid/commercial) is a confounded marker of access that must be defined and interpreted carefully when used as an equity variable." }, { "relation_type": "see_also", "target_slug": "claims-analysis", "notes": "Operational context for capturing and interpreting equity-relevant variables in administrative claims data." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "journal", "hta" ] }, { "slug": "strobe-me", "name": "STROBE-ME (STROBE Extension for Molecular Epidemiology)", "short_definition": "Reporting guideline that adds biomarker- and biospecimen-specific reporting items to the parent STROBE statement for observational molecular-epidemiology studies; covers specimen collection/handling/storage, laboratory assay validity and reliability, measurement error, and ethics of stored samples. Maintained within the EQUATOR Network.", "long_description": "**What it is** — **STROBE-ME (STrengthening the Reporting of OBservational studies in Epidemiology — Molecular\nEpidemiology)** is a reporting checklist that extends the parent **STROBE** statement to observational studies that\nmeasure **biomarkers in biological specimens** (genetic, epigenetic, protein, metabolite, adduct, exposure, or\neffect markers). It does not replace STROBE; it adds nine biomarker-specific items and amends eight existing STROBE\nitems so that the laboratory and biospecimen layer of a molecular-epidemiology study is reported as transparently as\nthe design and analysis layers. STROBE-ME was published by **Gallo, Egger, Vineis and colleagues in 2011**, released\nsimultaneously across seven journals (the open-access canonical version is *PLoS Medicine* 8(10):e1001117), and is\nhosted and maintained as a STROBE extension within the **EQUATOR Network**. Its purpose is to make the provenance,\nmeasurement, and quality control of every biomarker auditable, so that a reader can judge whether a reported\nexposure–outcome or biomarker–outcome association is real or an artifact of assay error, batch effects, specimen\ndegradation, or differential measurement.\n\n**When to use** — Apply STROBE-ME when reporting an observational study (cohort, case-control, or nested\ncase-control, including biobank- and registry-linked designs) in which **a biomarker measured from a biological\nsample is a study variable** — as exposure, effect, susceptibility, intermediate, or outcome marker. It is the\nappropriate checklist for a peer-reviewed molecular-epidemiology manuscript and for the molecular-epidemiology\ncomponents of a biomarker-qualification or exposure-assessment package submitted to a regulator. Decision rule for\nchoosing the right family member: use the **parent STROBE** for an observational study with no biospecimen\nmeasurement; use **STROBE-ME** when a biomarker is measured and the study is general molecular epidemiology; use\n**STROME-ID** (a separate extension, slug `strome-id`) for molecular epidemiology of *infectious diseases* where\npathogen typing and transmission inference dominate; and use **STREGA / STROBE-MR** for the gene-association and\nMendelian-randomization questions those extensions were written for. If your study uses **routinely collected\nadministrative claims or EHR data with no biospecimen assay**, STROBE-ME is the wrong checklist — use **RECORD** (or\n**RECORD-PE** for pharmacoepidemiology on routinely collected data), or **HARPER**/the **ENCePP** checklist for a\nnon-interventional study protocol.\n\n**What it requires** — STROBE-ME enforces reporting of the biospecimen-and-assay chain that the parent STROBE leaves\nimplicit. The substantive added/amended domains are: (1) **biomarker rationale and choice** — why this marker, what\nbiological construct it stands for, and whether it is a marker of exposure, effect, or susceptibility; (2)\n**biospecimen collection, transport, processing, and storage** — sample type, time and conditions of collection,\nfreeze–thaw history, storage duration and temperature, and any pre-analytical handling that could degrade the\nanalyte; (3) **laboratory methods** — the assay/platform, analytic protocol, **batch design and randomization of\nsamples across batches**, and **blinding of laboratory personnel to case/control or outcome status**; (4)\n**validity, reliability, and reproducibility of the biomarker** — limits of detection/quantification, coefficients\nof variation, intra-/inter-assay reproducibility, and use of internal/external quality-control samples; (5)\n**measurement error and its handling** — reporting how below-limit-of-detection values, missing assays, and\nclassical or differential measurement error are treated, and any **validation substudy or external adjustment**\nused to correct attenuation; (6) **population stratification** when genetic markers are used; and (7) **ethics of\nstored specimens** — consent for storage and future use, biobank governance, and return of incidental/individual\nfindings. The parent STROBE items it amends are framed so that, for a biomarker variable, \"measurement\" (item 8),\n\"bias\" (item 9), and \"limitations\" (item 19) are all answered in terms of *assay* validity, not just design.\n\n**When NOT to use — limitations and common misapplications** — STROBE-ME is a **reporting** checklist, not a\nrisk-of-bias instrument, not a quality score, and not a substitute for design validity. Concrete failure modes:\n(1) **Wrong extension for the design** — using STROBE-ME for a study with no biomarker (use parent STROBE), for an\ninfectious-disease molecular study (use STROME-ID), or for a routinely-collected-data pharmacoepidemiology study\n(use RECORD/RECORD-PE) — and, conversely, citing plain STROBE for a biomarker study, which omits the entire\nlaboratory-reporting layer. (2) **Checklist-as-theater** — ticking the biomarker items while leaving batch design,\nblinding, limit-of-detection handling, or storage history vague defeats the purpose; the value is the specificity,\nnot the page count. (3) **Mistaking reporting for validity** — a fully STROBE-ME-compliant paper can still describe\na study crippled by batch confounding, reverse causation (the biomarker measured *after* disease onset), or\nselection into the biobank; completing the checklist does not make an observational biomarker association causal or\nunconfounded. (4) **Treating it as a substitute for STROBE** — STROBE-ME is an *extension*; the parent items still\napply. (5) **Ignoring measurement-error correction** — reporting a single biomarker value without CVs,\nreliability, or attenuation correction invites regression dilution that no amount of confounding adjustment fixes.\n\n**How it maps to this catalog** — STROBE-ME's biomarker layer maps to a focused set of concepts in this repo; the\noverlap with generic pharmacoepidemiology design machinery is deliberately thin, because STROBE-ME governs *assay\nand specimen* reporting rather than confounding control:\n- The study object: **biomarker-defined-cohort-rwe** is the design STROBE-ME most directly governs; **cohort-prospective**,\n **cohort-retrospective**, and **case-control** (and **nested-case-control**) are the parent designs whose biomarker\n variables it reports.\n- Assay validity and measurement (items 3–5 above): **algorithm-validation** supplies the validity/reliability and\n sensitivity/specificity discipline that STROBE-ME demands of a biomarker assay (the analogue of phenotype-algorithm\n validation); **misclassification-bias-correction-rwe** and **external-adjustment-validation-substudy-bias-correction-rwe**\n implement the measurement-error correction and validation-substudy adjustment the checklist asks authors to report;\n **negative-control-exposures-rwe** operationalizes the residual-confounding/quality-control checks that detect batch\n or measurement artifacts.\n- Generalizability of a biobank sample: **generalizability-transportability-external-validity-rwe** addresses the\n selection-into-biobank limitation STROBE-ME asks authors to acknowledge.\nThese are the lens STROBE-ME pre-specifies for a biomarker study — *not* a claim that hdPS, time-zero alignment, or\nactive-comparator design are STROBE-ME items.\n\n**Applied note (biobank-/registry-linked RWE).** A molecular-epidemiology study that links a biobank to a disease\nregistry or claims data should, under STROBE-ME, report the full pre-analytical history of each specimen, randomize\nsamples across assay batches and blind the laboratory to case status, quantify assay reliability with QC samples and\nCVs, state how below-LOD and missing biomarker values are handled, and — where a single measurement stands in for a\nlong-term construct — report a validation substudy or external adjustment for measurement error. For the *non-biomarker*\nparts of such a linked study (administrative exposure or outcome definitions, attrition, confounding), the\nroutinely-collected-data reporting items live in **RECORD/RECORD-PE**, not STROBE-ME.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "molecular-epidemiology", "biomarker", "strobe", "equator", "laboratory-methods" ], "aliases": [ "STROBE-ME", "STROBE Extension for Molecular Epidemiology", "STrengthening the Reporting of OBservational studies in Epidemiology - Molecular Epidemiology" ], "applies_to_study_types": [ "cohort_prospective", "cohort_retrospective", "case_control" ], "data_sources": [ "primary", "registry", "linked" ], "citations": [ { "role": "introduce", "doi": "10.1371/journal.pmed.1001117", "url": "https://doi.org/10.1371/journal.pmed.1001117", "citation_text": "Gallo V, Egger M, McCormack V, et al. STrengthening the Reporting of OBservational studies in Epidemiology - Molecular Epidemiology (STROBE-ME): an extension of the STROBE statement. PLoS Medicine. 2011;8(10):e1001117.", "year": 2011, "authors_short": "Gallo et al.", "notes": "Canonical open-access STROBE-ME statement; defines the nine added and eight amended biomarker reporting items." }, { "role": "explain", "doi": "10.1111/j.1365-2362.2011.02561.x", "url": "https://doi.org/10.1111/j.1365-2362.2011.02561.x", "citation_text": "Gallo V, Egger M, McCormack V, et al. STrengthening the Reporting of OBservational studies in Epidemiology - Molecular Epidemiology (STROBE-ME): an extension of the STROBE statement. European Journal of Clinical Investigation. 2012;42(1):1-16.", "year": 2011, "authors_short": "Gallo et al.", "notes": "Simultaneous co-publication of the same statement (one of seven journals); useful for item-level rationale." }, { "role": "use", "url": "https://www.equator-network.org/reporting-guidelines/strengthening-the-reporting-of-observational-studies-in-epidemiology-molecular-epidemiology-strobe-me-an-extension-of-the-strobe-statement/", "citation_text": "STROBE-ME (STROBE Extension for Molecular Epidemiology). EQUATOR Network reporting-guidelines library (maintained landing page with the checklist and links to the parent STROBE statement and related extensions).", "year": 2011, "authors_short": "EQUATOR Network", "notes": "Canonical maintained landing page hosting the checklist and links to STROBE and its extensions." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "cohort-prospective", "notes": "Use when a prospective cohort measures biomarkers from biological specimens." }, { "relation_type": "applies_to", "target_slug": "cohort-retrospective", "notes": "Use when a retrospective/biobank-linked cohort measures biomarkers from stored specimens." }, { "relation_type": "applies_to", "target_slug": "case-control", "notes": "Use for case-control and nested case-control molecular-epidemiology studies; report batch design and blinding of the laboratory to case status." }, { "relation_type": "part_of", "target_slug": "strobe", "notes": "STROBE-ME is an extension of the parent STROBE statement; the parent items still apply alongside the added biomarker items." }, { "relation_type": "see_also", "target_slug": "biomarker-defined-cohort-rwe", "notes": "The design STROBE-ME most directly governs - a cohort defined or characterized by a measured biomarker." }, { "relation_type": "used_with", "target_slug": "algorithm-validation", "notes": "Supplies the validity/reliability and sensitivity/specificity discipline STROBE-ME demands of a biomarker assay (the assay analogue of phenotype-algorithm validation)." }, { "relation_type": "used_with", "target_slug": "misclassification-bias-correction-rwe", "notes": "Implements the measurement-error handling STROBE-ME asks authors to report when a biomarker is measured with error." }, { "relation_type": "see_also", "target_slug": "external-adjustment-validation-substudy-bias-correction-rwe", "notes": "Validation-substudy / external-adjustment correction for biomarker measurement error and attenuation." }, { "relation_type": "see_also", "target_slug": "negative-control-exposures-rwe", "notes": "Quality-control checks that help detect batch effects and measurement artifacts in biomarker data." }, { "relation_type": "see_also", "target_slug": "generalizability-transportability-external-validity-rwe", "notes": "Addresses selection-into-biobank and external-validity limitations STROBE-ME asks authors to acknowledge." }, { "relation_type": "alternative_to", "target_slug": "record-pe", "notes": "For routinely-collected-data pharmacoepidemiology with no biospecimen assay, use RECORD-PE/RECORD instead of STROBE-ME." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "journal", "fda", "ema" ] }, { "slug": "strobe-mr", "name": "STROBE-MR (STROBE Extension for Mendelian Randomization)", "short_definition": "Reporting guideline that specifies the minimum items an observational Mendelian randomization (MR) study should report, extending STROBE to cover genetic instruments, the three core IV assumptions, and MR-specific sensitivity analyses; maintained within the EQUATOR Network.", "long_description": "**What it is** — **STROBE-MR (Strengthening the Reporting of Observational Studies in Epidemiology\nusing Mendelian Randomization)** is a reporting checklist that extends the STROBE statement to\n**Mendelian randomization (MR)** studies — observational analyses that use germline genetic variants\n(typically SNPs from GWAS) as **instrumental variables** to estimate the causal effect of a modifiable\nexposure on an outcome. The statement (Skrivankova, Richmond, Woolf et al., *JAMA* 2021) defines the\nitems an MR study must report; a companion **explanation-and-elaboration** paper (*BMJ* 2021) gives\nitem-by-item guidance and worked examples. STROBE-MR is hosted and maintained as a STROBE extension\nwithin the **EQUATOR Network**. Its purpose is to make the genetic instruments, the data sources\n(one-sample vs two-sample, individual-level vs summary-level GWAS), and the assessment of the three\ncore IV assumptions transparent and auditable, so that readers and reviewers can judge whether an MR\nestimate is credible rather than an artefact of pleiotropy, weak instruments, or population structure.\n\n**When to use** — Apply STROBE-MR whenever you are *reporting a Mendelian randomization study*: a\none-sample MR in a single cohort/biobank, a two-sample MR combining exposure and outcome GWAS summary\nstatistics, multivariable MR, MR with summary data from consortia (e.g., a journal manuscript, a\ntriangulation paper supporting a drug-target or biomarker causal claim, or an MR component embedded in\na larger evidence package). Decision rule for choosing the right STROBE family member: use **plain\nSTROBE** for a generic observational cohort/case-control/cross-sectional study; use **STREGA** when\nthe focus is reporting a *genetic association* study (genotyping, HWE, population stratification) that\nis *not* using genotypes as instruments for causal inference; use **RECORD / RECORD-PE** when the\nstudy is built on *routinely collected* health data (claims/EHR/registries) — those govern data\nprovenance, not genetic instruments; and use **STROBE-MR** specifically when germline variants are\ndeployed as *instrumental variables to estimate a causal effect*. The distinction is the analytic\nintent: genetics-as-instrument (STROBE-MR) versus genetics-as-association (STREGA) versus\nroutine-data-provenance (RECORD). STROBE-MR is a reporting guideline for *journal* publication and\nscientific transparency; it is not itself an FDA/EMA submission template, though MR evidence is\nincreasingly cited in regulatory and HTA causal-triangulation arguments and a STROBE-MR-compliant\nreport is the expected substrate for that use.\n\n**What it requires** — Beyond the generic STROBE items (title/abstract, structured introduction with\na pre-specified hypothesis, methods, results, discussion, funding), STROBE-MR enforces MR-specific\nreporting that maps onto the **three instrumental-variable assumptions**: (1) **Relevance** — report\nhow genetic instruments were selected, the source GWAS, genome-wide significance and clumping\nthresholds, instrument strength (F-statistics, variance explained R²), and steps taken against\n**weak-instrument bias**. (2) **Independence (exchangeability)** — report control for population\nstratification (ancestry, principal components, restriction to a single ancestry), and assortative\nmating / dynastic (parental-genotype) effects where relevant. (3) **Exclusion restriction (no\nhorizontal pleiotropy)** — report the biological plausibility of the variant–exposure pathway and the\nbattery of **sensitivity analyses** that probe the no-pleiotropy assumption: MR-Egger (intercept and\nslope), weighted median, weighted mode, MR-PRESSO, leave-one-out, and heterogeneity statistics.\nSTROBE-MR additionally requires explicit reporting of the **data structure** (one-sample vs\ntwo-sample; degree of **sample overlap** between exposure and outcome GWAS, which biases two-sample\nestimates toward the confounded observational association), the **estimand and its scale** (per-unit\nor per-SD change in the genetically-proxied exposure — a lifelong-exposure contrast that is *not* the\nsame as a clinical-intervention effect), harmonization of effect alleles across data sources, and the\nsoftware/packages and versions used. In RWE terms these are the field's analogues of *data fitness for\nuse*, *time-zero/estimand specification*, *confounding control*, and *quantitative sensitivity / bias\nanalysis* — adapted to the genetic-instrument setting.\n\n**When NOT to use — limitations and common misapplications** — STROBE-MR is a **reporting** checklist,\nnot a risk-of-bias instrument, not a quality score, and not a substitute for the IV assumptions\nthemselves. Concrete failure modes: (1) **Wrong extension** — using plain STROBE (which has no items\nfor instrument strength, pleiotropy, or sample overlap) for an MR study, or using STROBE-MR for a\ngenetic-association study that should use STREGA, or for a routine-data observational study that should\nuse RECORD/RECORD-PE. (2) **Checklist completeness mistaken for causal validity** — a fully\nSTROBE-MR-compliant paper can still report a biased estimate: transparent reporting of weak\ninstruments, residual pleiotropy, or substantial sample overlap does not *fix* those problems, it only\nsurfaces them. Completing the checklist does not make the MR estimate causal. (3) **Treating the MR\nestimate as a drug effect** — MR estimates a lifelong genetically-proxied exposure contrast; reporting\nit as if it were the effect of a short-term clinical intervention is a misinterpretation the discussion\nitems exist to prevent. (4) **Checklist-as-theater** — ticking items while leaving instrument-selection\ncriteria, F-statistics, sample-overlap fraction, or the pleiotropy-sensitivity suite vague defeats the\npurpose; the value is the substantive disclosure, not the page count. (5) **Using it as an appraisal\ntool for someone else's study** — to *grade* MR evidence, pair the report with a critical-appraisal/\nrisk-of-bias framework (e.g., ROBINS-style or MR-specific appraisal); STROBE-MR tells you what *should\nhave been reported*, not whether the study is at low risk of bias.\n\n**How it maps to this catalog** — In this repo, STROBE-MR's substantive requirements correspond to\nthese implementing concepts a reader or reviewer can reason against:\n- The instrument-and-causal engine (the relevance / exclusion-restriction items): **instrumental-variables-pharmacoepi-rwe**\n implements the IV logic — instrument strength, the exclusion restriction, and weak-instrument bias —\n that STROBE-MR's genetic-instrument items demand.\n- The causal-structure justification (the independence / no-pleiotropy items): **dags-backdoor-criterion-drug-studies**\n formalizes the directed-acyclic-graph reasoning that distinguishes a valid instrument from a\n pleiotropic or confounded one.\n- The estimand discipline (what the per-SD genetic contrast actually estimates): **estimands-ate-att-intercurrent-events-rwe**\n supplies the language for stating the target estimand and why the MR contrast differs from an\n interventional effect.\n- The sensitivity / quantitative-bias spine: **e-value-sensitivity-analysis**, **empirical-calibration-negative-controls-rwe**,\n and **quantitative-bias-analysis-toolkit-rwe** are the catalog homes for the \"how robust is this to\n unmeasured violations?\" reporting STROBE-MR requires (the MR field's MR-Egger / leave-one-out\n analogues live conceptually alongside these).\n- External validity of the genetic estimate: **generalizability-transportability-external-validity-rwe**\n frames whether a single-ancestry MR estimate transports to the target population.\n- Sibling reporting guidelines (to pick the right one): **strobe** (the parent), and the genetic-\n association vs routine-data distinctions captured by the STREGA and RECORD/RECORD-PE entries in this\n `guidelines` set. These are the lens for the *wrong-extension* failure mode above, not items of\n STROBE-MR itself.\n\n**Applied note (claims/EHR/registry RWE).** Most STROBE-MR studies run on biobank or consortium GWAS\ndata, but the extension is directly relevant to RWE built on **linked biobank–EHR/claims** resources\n(e.g., UK Biobank, *All of Us*, or biobank-linked administrative cohorts). When MR is run inside such a\nresource, the report must still satisfy the data-fitness items RWE reviewers expect: which linked source\nsupplied the *phenotype* (an algorithm-defined outcome from EHR/claims carries the same misclassification\nand PPV concerns as any RWE outcome — see the catalog's outcome-algorithm and phenotype concepts), how\nancestry and relatedness were handled, and what fraction of the exposure and outcome samples overlap.\nAn MR estimate offered to support a drug-target causal claim in an HTA or regulatory triangulation\nargument should be reported to STROBE-MR *and* its underlying real-world phenotypes documented to the\nstandard the rest of this catalog enforces — the two are complementary, not interchangeable.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "mendelian-randomization", "instrumental-variables", "genetic-epidemiology", "strobe", "equator", "causal-inference" ], "aliases": [ "STROBE-MR", "STROBE extension for Mendelian Randomization", "Strengthening the Reporting of Observational Studies in Epidemiology using Mendelian Randomization", "STROBE-MR Statement" ], "applies_to_study_types": [ "mendelian_randomization", "cohort_prospective", "cohort_retrospective" ], "data_sources": [ "registry", "linked", "ehr" ], "citations": [ { "role": "introduce", "doi": "10.1001/jama.2021.18236", "url": "https://doi.org/10.1001/jama.2021.18236", "citation_text": "Skrivankova VW, Richmond RC, Woolf BAR, et al. Strengthening the Reporting of Observational Studies in Epidemiology Using Mendelian Randomization: The STROBE-MR Statement. JAMA. 2021;326(16):1614-1621.", "year": 2021, "authors_short": "Skrivankova et al.", "notes": "Canonical STROBE-MR statement defining the reporting checklist for Mendelian randomization studies." }, { "role": "explain", "doi": "10.1136/bmj.n2233", "url": "https://doi.org/10.1136/bmj.n2233", "citation_text": "Skrivankova VW, Richmond RC, Woolf BAR, et al. Strengthening the reporting of observational studies in epidemiology using mendelian randomisation (STROBE-MR): explanation and elaboration. BMJ. 2021;375:n2233.", "year": 2021, "authors_short": "Skrivankova et al.", "notes": "Item-by-item explanation and elaboration with worked MR examples and rationale for each reporting item." }, { "role": "use", "url": "https://www.strobe-mr.org/", "citation_text": "STROBE-MR (STROBE extension for Mendelian Randomization). STROBE Statement / EQUATOR Network reporting-guidelines library (maintained checklist and extension links).", "year": 2021, "authors_short": "STROBE Initiative / EQUATOR Network", "notes": "Maintained landing page for the STROBE extensions, including the downloadable STROBE-MR checklist." } ], "relations": [ { "relation_type": "is_variant_of", "target_slug": "strobe", "notes": "STROBE-MR extends the parent STROBE checklist with items specific to genetic instruments and the three instrumental-variable assumptions." }, { "relation_type": "used_with", "target_slug": "instrumental-variables-pharmacoepi-rwe", "notes": "MR is instrumental-variable estimation with germline variants; this concept implements instrument strength, the exclusion restriction, and weak-instrument bias that STROBE-MR requires reporting." }, { "relation_type": "used_with", "target_slug": "dags-backdoor-criterion-drug-studies", "notes": "DAG reasoning formalizes the independence and no-pleiotropy assumptions that distinguish a valid genetic instrument from a confounded or pleiotropic one." }, { "relation_type": "see_also", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "STROBE-MR requires stating the estimand and scale; the genetically-proxied lifelong-exposure contrast is not the same as an interventional drug effect." }, { "relation_type": "see_also", "target_slug": "e-value-sensitivity-analysis", "notes": "STROBE-MR's pleiotropy / robustness items are the MR analogue of sensitivity analysis for unmeasured violations of the causal assumptions." }, { "relation_type": "see_also", "target_slug": "empirical-calibration-negative-controls-rwe", "notes": "Negative-control and calibration logic parallels MR's sensitivity battery (MR-Egger, leave-one-out) for detecting residual bias." }, { "relation_type": "see_also", "target_slug": "generalizability-transportability-external-validity-rwe", "notes": "A single-ancestry MR estimate's transportability to the target population is an external-validity question STROBE-MR asks authors to address." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "journal", "fda", "ema", "hta" ] }, { "slug": "strobe-ni", "name": "STROBE-NI (STROBE Extension for Newborn Infection)", "short_definition": "A reporting checklist that extends the STROBE statement with neonatal-infection-specific items (case definitions, denominators, timing of infection, microbiology, follow-up) so that observational studies of newborn sepsis, meningitis, and other neonatal infections are reported transparently and comparably; maintained as a STROBE extension within the EQUATOR Network.", "long_description": "**What it is** — **STROBE-NI (Strengthening the Reporting of Observational Studies in Epidemiology for\nNewborn Infection)** is a reporting guideline that adds neonatal-infection-specific reporting items to the\ngeneric 22-item **STROBE** checklist for observational research. Published by Fitchett et al. in *The Lancet\nInfectious Diseases* in 2016 and developed by an international consensus group (the STROBE-NI working group),\nit is hosted and maintained as a STROBE extension within the **EQUATOR Network**. Its purpose is narrow and\ndeliberate: studies of neonatal sepsis, meningitis, pneumonia, congenital and early/late-onset infection are\nnotoriously hard to compare because authors use incompatible case definitions, ambiguous denominators, and\ninconsistent timing conventions. STROBE-NI forces authors to state, item by item, how cases were defined\n(clinically suspected vs laboratory/blood-culture-confirmed vs autopsy-confirmed), what the at-risk population\nand time origin were (live births, NICU admissions, age windows for early- vs late-onset disease), how\nspecimens were collected and which pathogens were sought, and how infants were followed and lost. It is a\n*reporting* standard layered on top of STROBE — not a new study design, not an appraisal score, and not a\nconduct manual.\n\n**When to use** — Apply STROBE-NI when you are reporting an **observational** study (prospective or\nretrospective cohort, case-control, or surveillance/registry-based analysis) whose primary focus is\n**infection in the newborn period**: incidence or aetiology of neonatal sepsis/meningitis, risk factors,\nantimicrobial resistance surveillance, case-fatality, or related outcomes. It is the appropriate checklist\nfor a peer-reviewed journal manuscript in neonatology, paediatric infectious disease, or global-health\nepidemiology, and for the analysis plan behind surveillance networks and birth-cohort studies. Decision rule\nfor choosing the right family member: use **plain STROBE** for observational studies *not* centred on neonatal\ninfection; use **STROBE-NI** when neonatal infection is the exposure/outcome of interest and the\ninfection-specific items (case definition, denominator, timing, microbiology) are load-bearing; use\n**RECORD / RECORD-PE** when the data are *routinely collected* (claims, EHR, registries) and the\ndatabase-provenance and code-list items dominate; use **CONSORT/SPIRIT** for interventional trials and their\nprotocols. STROBE-NI and RECORD are not mutually exclusive — a routinely-collected neonatal-infection study\ncan sensibly report against both, with RECORD covering data provenance and STROBE-NI covering the\ninfection-definition items.\n\n**What it requires** — On top of the STROBE backbone (title/abstract, background, objectives, design,\nsetting, participants, variables, data sources/measurement, bias, study size, statistical methods, results,\nlimitations, generalisability, funding), STROBE-NI compels neonatal-infection specifics that are exactly the\nplaces RWE-style misclassification creeps in: an **explicit, reproducible case definition** stated separately\nfor clinically suspected vs laboratory-confirmed infection, with the diagnostic algorithm and thresholds\nnamed (this is the phenotype/algorithm-validation problem in neonatal clothing); the **denominator and\ntime-zero convention** (live births, admissions, person-time at risk; age cut-offs distinguishing early-onset\nfrom late-onset disease) so incidence is interpretable; **specimen and laboratory methods** (what was\ncultured/tested, contamination handling, pathogen panel) because culture-confirmation drives case\nascertainment; **gestational-age and birth-weight ascertainment**, which govern both eligibility and\nconfounding; **completeness of follow-up and handling of competing/intercurrent events** — neonatal mortality\nis both an outcome and a competing risk that truncates infection ascertainment; and transparent reporting of\n**missing data and attrition**, which is severe in neonatal cohorts (early death, transfer, discharge before\noutcome). The checklist requires that these be *reported*, making selective or vague definitions visible to a\nreviewer.\n\n**When NOT to use — limitations and common misapplications** — STROBE-NI is a *reporting* checklist; it is\n**not** a risk-of-bias instrument, **not** a quality score, and completing it does **not** make an\nobservational neonatal-infection study unconfounded or causal. Concrete failure modes: (1) **Wrong scope** —\nusing STROBE-NI for observational studies that are not about neonatal infection; those use plain STROBE.\n(2) **Wrong family member** — applying STROBE-NI to a study built on *routinely-collected* data without also\nsatisfying RECORD/RECORD-PE, so data provenance, linkage, and code lists go unreported; or using a reporting\nchecklist at all when an interventional trial demands CONSORT/SPIRIT. (3) **Checklist-as-score** — tallying\ncompleted items as a measure of study *quality* or risk of bias; appraisal of a non-randomized study is the\njob of ROBINS-I, not a completeness checklist. (4) **Definition theatre** — ticking the \"case definition\"\nitem while the underlying microbiological/clinical algorithm is unvalidated (unknown sensitivity/PPV against\nblood culture or autopsy), which is precisely the misclassification STROBE-NI was created to surface.\n(5) **Reporting ≠ conduct** — a fully STROBE-NI-compliant paper can still have an immortal-time problem, an\nill-defined denominator, or uncontrolled confounding; transparency is necessary, not sufficient.\n\n**How it maps to this catalog** — In this repo, STROBE-NI's substantive items are operationalised by these\nconcepts, which tell a reviewer *how* to satisfy each requirement rather than merely report it:\n- The case-definition and pathogen-ascertainment items map to **diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe**\n (constructing and time-windowing a confirmed vs clinical neonatal-infection phenotype) and to\n **algorithm-validation** (estimating the sensitivity/PPV of the case definition against blood-culture or\n autopsy gold standards — the quantitative backbone of an honest \"case definition\" item).\n- The denominator/time-zero, completeness-of-follow-up, and missing-data items map to\n **attrition-and-loss-to-follow-up-rwe** (early neonatal death, transfer, and discharge create severe,\n potentially informative loss) and to **estimands-ate-att-intercurrent-events-rwe** (neonatal mortality as a\n competing/intercurrent event that must be handled in the estimand, not ignored).\n- The generalisability item maps to **generalizability-transportability-external-validity-rwe**, which is\n acute for neonatal infection because aetiology, resistance, and case-fatality differ sharply between\n high-income NICU settings and LMIC community/facility births.\n- When the neonatal-infection study is built on routinely-collected data, **claims-analysis** and the\n data-fitness items it carries become relevant for provenance — but reported under RECORD, with STROBE-NI\n layered for the infection-definition items.\n\n**Applied note (surveillance / registry / linked RWE).** For a neonatal-infection surveillance birth cohort —\ne.g., a facility-based or community birth cohort in an LMIC with linked microbiology and vital records —\nSTROBE-NI compliance means: state the denominator (live births vs admissions) and the early-/late-onset age\nwindows explicitly; report the confirmed-vs-clinical case algorithm *and* its validation against culture or\nautopsy where available; report specimen/contamination handling and the pathogen panel; quantify loss to\nfollow-up from early death, transfer, and discharge and treat it as potentially informative; and handle\nneonatal mortality as a competing event in the estimand rather than as simple censoring. Reporting these makes\nthe difference between an incidence figure a reviewer can interpret and one that is silently incomparable\nacross sites.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "strobe", "strobe-extension", "neonatal-infection", "observational", "equator" ], "aliases": [ "STROBE-NI", "STROBE Extension for Newborn Infection", "Strengthening the Reporting of Observational Studies in Epidemiology for Newborn Infection", "STROBE for Neonatal Infection" ], "applies_to_study_types": [ "cohort_prospective", "cohort_retrospective", "case_control", "disease_registry" ], "data_sources": [ "primary", "registry", "ehr", "linked" ], "citations": [ { "role": "introduce", "doi": "10.1016/S1473-3099(16)30082-2", "url": "https://doi.org/10.1016/S1473-3099(16)30082-2", "citation_text": "Fitchett EJA, Seale AC, Vergnano S, et al. Strengthening the Reporting of Observational Studies in Epidemiology for Newborn Infection (STROBE-NI): an extension of the STROBE statement for neonatal infection research. The Lancet Infectious Diseases. 2016;16(10):e202-e213.", "year": 2016, "authors_short": "Fitchett et al.", "notes": "Canonical STROBE-NI statement; defines the neonatal-infection-specific reporting items added to STROBE." }, { "role": "explain", "doi": "10.1371/journal.pmed.0040296", "url": "https://doi.org/10.1371/journal.pmed.0040296", "citation_text": "von Elm E, Altman DG, Egger M, et al. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. PLoS Medicine. 2007;4(10):e296.", "year": 2007, "authors_short": "von Elm et al.", "notes": "Parent STROBE statement; STROBE-NI extends its 22-item backbone, so it is the base checklist authors must also satisfy." }, { "role": "explain", "doi": "10.1371/journal.pmed.0040297", "url": "https://doi.org/10.1371/journal.pmed.0040297", "citation_text": "Vandenbroucke JP, von Elm E, Altman DG, et al. Strengthening the Reporting of Observational Studies in Epidemiology (STROBE): explanation and elaboration. PLoS Medicine. 2007;4(10):e297.", "year": 2007, "authors_short": "Vandenbroucke et al.", "notes": "Item-by-item explanation and elaboration of STROBE with rationale and examples of good reporting." }, { "role": "use", "url": "https://www.equator-network.org/reporting-guidelines/strobe/", "citation_text": "STROBE statement and extensions (including STROBE-NI). EQUATOR Network reporting-guidelines library: maintained checklists, downloadable templates, and links to STROBE extensions.", "year": 2016, "authors_short": "EQUATOR Network", "notes": "Canonical maintained landing page hosting STROBE and its extensions in usable formats." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "cohort-prospective", "notes": "Use when prospectively following newborns for incidence/aetiology/outcomes of neonatal infection." }, { "relation_type": "applies_to", "target_slug": "cohort-retrospective", "notes": "Use for retrospective neonatal-infection cohorts; report denominator, time-zero, and follow-up completeness." }, { "relation_type": "applies_to", "target_slug": "case-control", "notes": "Use for case-control studies of risk factors for neonatal sepsis/meningitis; report the case definition explicitly." }, { "relation_type": "applies_to", "target_slug": "disease-registry", "notes": "Use for surveillance/registry-based neonatal-infection analyses; pairs with RECORD for data provenance." }, { "relation_type": "see_also", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Implements the STROBE-NI case-definition item — constructing and time-windowing a confirmed vs clinical neonatal-infection phenotype." }, { "relation_type": "see_also", "target_slug": "algorithm-validation", "notes": "Quantifies the case definition's sensitivity/PPV against blood-culture or autopsy gold standards, backing the case-definition item." }, { "relation_type": "see_also", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Implements the completeness-of-follow-up/missing-data items; neonatal loss (early death, transfer, discharge) is often informative." }, { "relation_type": "see_also", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Neonatal mortality is a competing/intercurrent event that must be handled in the estimand, not silently censored." }, { "relation_type": "see_also", "target_slug": "generalizability-transportability-external-validity-rwe", "notes": "Supports the generalisability item; neonatal-infection aetiology, resistance, and case-fatality differ sharply across HIC and LMIC settings." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "journal" ] }, { "slug": "strobe-nut", "name": "STROBE-nut (STROBE Extension for Nutritional Epidemiology)", "short_definition": "Reporting guideline that extends the STROBE Statement with 24 nutrition-specific reporting items for observational studies of diet, food, and nutrient exposures, addressing how dietary assessment, food/nutrient definitions, and energy adjustment are measured and reported; maintained within the EQUATOR Network.", "long_description": "**What it is** — **STROBE-nut (Strengthening the Reporting of Observational Studies in Epidemiology —\nNutritional Epidemiology)** is a reporting-guideline extension that supplements the parent **STROBE\nStatement** with 24 nutrition-specific items (extensions and elaborations of STROBE's core 22-item\nchecklist) for observational studies whose exposures are dietary intake, foods, nutrients, supplements,\ndietary patterns, or nutritional status. It was published by Lachat et al. in *PLOS Medicine* in 2016\nand is hosted and maintained as a STROBE/EQUATOR extension within the **EQUATOR Network**. Its purpose\nis transparency, not validity scoring: it forces authors to make explicit *how diet was measured* (the\nassessment instrument and its validation), *how foods/nutrients were defined and quantified* (food\ncomposition tables, brand/processing detail, supplement capture), *how total energy intake was handled*\nin the analysis (the energy-adjustment model), and *how missing or implausible intake data were treated*\n— the elements that make nutritional-epidemiology results reproducible and comparable across studies but\nthat generic STROBE does not name. STROBE-nut is a *companion* checklist: it is used alongside STROBE,\nnot instead of it.\n\n**When to use** — Apply STROBE-nut when *reporting a completed observational study* (cohort, case-control,\nor cross-sectional, including secondary analyses of existing data) in which a primary exposure, mediator,\nor key covariate is dietary or nutritional. Typical settings: a peer-reviewed nutritional-epidemiology\nmanuscript; the evidence base behind a dietary-guideline review; a real-world cohort linking dietary\nassessment to claims/EHR/registry outcomes; or a nutrition sub-study within a larger pharmacoepidemiology\nor HTA evidence package. Decision rule for choosing the right family member: use the **core STROBE**\nchecklist for any observational study, and *add* **STROBE-nut** when nutrition is central; use\n**RECORD** (or **RECORD-PE** for pharmacoepidemiology) when the study is built on routinely-collected\nhealth data such as claims/EHR — and if a routinely-collected-data study *also* has a nutritional\nexposure, RECORD and STROBE-nut are complementary, addressing the data-source and the diet-measurement\ndimensions respectively. STROBE-nut governs *reporting of a primary study*; it is not a protocol\nchecklist (use SPIRIT/HARPER/ENCePP for those) and not a systematic-review checklist (use PRISMA).\n\n**What it requires** — Layered on STROBE's title/abstract, introduction, methods, results, and discussion\nstructure, the nutrition-specific items demand: (1) **Exposure/dietary-assessment transparency** — the\nnamed assessment method (FFQ, multiple 24-hour recalls, food records, biomarkers), the number and timing\nof measurements, and reference to its **validation/reproducibility** in the study population, with the\nrecognized measurement-error structure of self-reported intake stated rather than ignored. (2)\n**Food/nutrient definition and quantification** — the food-composition database and version used, how\nrecipes/mixed dishes/brands/fortification were handled, supplement and fortificant capture, and the\nunits reported. (3) **Energy and the analytic model for intake** — explicit statement of whether and how\n**total energy intake** was adjusted for (standard, residual, energy-partition, or multivariable models),\nsince this choice changes the estimand and the interpretation of any nutrient effect. (4) **Data fitness\nand plausibility** — handling of implausible energy intakes, under/over-reporting, and the treatment of\nmissing dietary data. (5) **Confounding and effect modification** specific to diet (e.g., the entire\ndietary context, physical activity, total energy). Read for real-world data, these map onto familiar RWE\nobligations: design transparency, data-fitness-for-use, exposure/algorithm definition and validation,\ntime-zero alignment for incident dietary exposure or follow-up, attrition/missing-data accounting, and\npre-specified sensitivity analyses (e.g., excluding implausible reporters).\n\n**When NOT to use — limitations and common misapplications** — STROBE-nut is a *reporting* checklist, not\na risk-of-bias instrument and not a quality score. Concrete failure modes: (1) **Mistaking it for an\nappraisal tool** — a STROBE-nut-complete paper can still be badly confounded or measurement-error-ridden;\nto *appraise* the study use a risk-of-bias/quality tool (e.g., the relevant JBI checklist, NOS, or for\nRWE-on-routine-data, ROBINS-I), not the reporting checklist. (2) **Reporting ≠ causal** — completing the\nchecklist makes the methods transparent; it does not make an observational diet–outcome association\ncausal or unconfounded. (3) **Wrong extension for the design** — using core STROBE alone for a\nnutrition-central study (omitting the diet-measurement items that determine reproducibility), or using\nSTROBE-nut where the dominant methodological issue is the *routinely-collected data source* and **RECORD**\nis what is actually required. (4) **Wrong study object** — applying STROBE-nut to a *trial* (use CONSORT)\nor to a *systematic review/meta-analysis* (use PRISMA); STROBE-nut is for primary observational studies.\n(5) **Checklist-as-theater** — page-number ticking while the FFQ validation, food-composition source, or\nenergy-adjustment model are left vague defeats the purpose; the value is the substance behind each item,\nnot the completed table.\n\n**How it maps to this catalog** — In this repo, STROBE-nut's substantive requirements are *implemented*\nby concepts a methodologist can build and report against:\n- The diet-exposure definition and its validity (assessment-instrument and food/nutrient items) are\n implemented by **algorithm-validation** and, where intake is operationalized from coded data,\n **diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe** and **outcome-algorithm-construction-rwe**.\n- Data-fitness-for-use (the right data to measure diet at all) is implemented by\n **fit-for-purpose-data-assessment-rwe** and **database-feasibility-attrition-funnel-rwe**.\n- Time-zero alignment for incident exposure/follow-up is implemented by\n **time-zero-index-date-alignment-rwe** and **continuous-enrollment-observable-time-rwe**.\n- Attrition and missing/implausible dietary data are implemented by\n **attrition-and-loss-to-follow-up-rwe**, **missing-data-pattern-table-rwe**, and\n **missing-data-trimming-winsorization-rwe**.\n- Confounding control and the energy-adjustment/estimand decision are implemented by\n **dags-backdoor-criterion-drug-studies**, **estimands-ate-att-intercurrent-events-rwe**, and the\n sensitivity machinery in **selection-bias-sensitivity-analysis-rwe** and\n **e-value-sensitivity-analysis** (e.g., excluding implausible reporters; residual-confounding bounds).\n- Descriptive reporting of the sample and exposure distribution is implemented by\n **descriptive-epidemiology-rwe** and **baseline-characteristics-and-covariate-balance-rwe**.\n\n**Applied note (claims/EHR/registry RWE).** Dietary exposure is rarely captured cleanly in routinely\ncollected data, so a real-world nutritional study typically *links* a cohort with validated dietary\nassessment (FFQ/24-hour recall/registry diet module) to claims/EHR outcomes. STROBE-nut then forces the\nlink to be reported honestly: name the dietary instrument and its validation in *this* population, state\nthe food-composition source and energy-adjustment model, and — because the outcome side runs on\n**claims-analysis**-style routinely-collected data — report it under **RECORD** alongside STROBE-nut, with\noutcome phenotypes validated (**claims-outcome-algorithm-ppv-sensitivity-rwe**), time-zero aligned, and\nattrition/missingness on both the diet and outcome sides accounted for. STROBE-nut and RECORD are\ncomplementary, not interchangeable: one governs the diet measurement, the other the data source.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "strobe", "nutritional-epidemiology", "observational", "equator", "dietary-assessment" ], "aliases": [ "STROBE-nut", "STROBE Extension for Nutritional Epidemiology", "STROBE-Nutrition", "Strengthening the Reporting of Observational Studies in Epidemiology - Nutritional Epidemiology" ], "applies_to_study_types": [ "cohort_prospective", "cohort_retrospective", "cross_sectional", "case_control" ], "data_sources": [ "registry", "ehr", "claims", "linked" ], "citations": [ { "role": "introduce", "doi": "10.1371/journal.pmed.1002036", "url": "https://doi.org/10.1371/journal.pmed.1002036", "citation_text": "Lachat C, Hawwash D, Ocké MC, et al. Strengthening the Reporting of Observational Studies in Epidemiology-Nutritional Epidemiology (STROBE-nut): An Extension of the STROBE Statement. PLOS Medicine. 2016;13(6):e1002036.", "year": 2016, "authors_short": "Lachat et al.", "notes": "Canonical STROBE-nut statement defining the 24 nutrition-specific reporting items extending STROBE." }, { "role": "explain", "doi": "10.1371/journal.pmed.0040296", "url": "https://doi.org/10.1371/journal.pmed.0040296", "citation_text": "von Elm E, Altman DG, Egger M, et al. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement: Guidelines for Reporting Observational Studies. PLoS Medicine. 2007;4(10):e296.", "year": 2007, "authors_short": "von Elm et al.", "notes": "Parent STROBE Statement; STROBE-nut extends and elaborates its core 22-item checklist." }, { "role": "use", "url": "https://www.equator-network.org/reporting-guidelines/strobe-nut/", "citation_text": "STROBE-nut (STROBE Extension for Nutritional Epidemiology). EQUATOR Network reporting-guidelines library (maintained checklist and downloadable materials).", "year": 2016, "authors_short": "EQUATOR Network", "notes": "Maintained landing page with the checklist in usable formats and links to related STROBE extensions." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "cohort-prospective", "notes": "Add STROBE-nut to core STROBE when reporting a prospective cohort with a dietary/nutritional exposure." }, { "relation_type": "applies_to", "target_slug": "cohort-retrospective", "notes": "Use for retrospective/secondary-data cohorts when diet or nutritional status is a key exposure." }, { "relation_type": "applies_to", "target_slug": "cross-sectional", "notes": "Use for cross-sectional nutrition studies (e.g., diet-quality prevalence and correlates)." }, { "relation_type": "applies_to", "target_slug": "case-control", "notes": "Use for case-control studies in which dietary intake is the exposure of interest." }, { "relation_type": "see_also", "target_slug": "algorithm-validation", "notes": "Implements the dietary-assessment validation/reproducibility STROBE-nut requires authors to report." }, { "relation_type": "see_also", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "Implements the data-fitness judgment behind whether a data source can measure diet/nutrition at all." }, { "relation_type": "see_also", "target_slug": "time-zero-index-date-alignment-rwe", "notes": "Implements the time-zero alignment STROBE-nut's design-transparency items demand for incident exposure/follow-up." }, { "relation_type": "see_also", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Implements the attrition and missing/implausible dietary-data accounting STROBE-nut requires." }, { "relation_type": "see_also", "target_slug": "selection-bias-sensitivity-analysis-rwe", "notes": "Implements sensitivity analyses (e.g., excluding implausible energy reporters) the checklist expects." }, { "relation_type": "see_also", "target_slug": "claims-analysis", "notes": "When outcomes run on routinely-collected data, report alongside RECORD; STROBE-nut governs diet measurement." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "journal", "hta" ] }, { "slug": "strobe-rds", "name": "STROBE-RDS (STROBE Extension for Respondent-Driven Sampling)", "short_definition": "Reporting guideline that extends STROBE with a respondent-driven-sampling-specific checklist, requiring transparent reporting of seeds, coupons, recruitment chains, network-size measurement, equilibrium/homophily diagnostics, and the RDS estimator used to produce population-weighted prevalence estimates from chain-referral surveys of hidden populations.", "long_description": "**What it is** — **STROBE-RDS (Strengthening the Reporting of Observational Studies in\nEpidemiology for Respondent-Driven Sampling)** is a 2015 reporting-guideline extension of the\nparent STROBE statement, published by White, Hakim, Salganik and colleagues in the *Journal of\nClinical Epidemiology*. It supplies STROBE's 22-item core checklist with RDS-specific reporting\nitems so that a study using **respondent-driven sampling** — a chain-referral, coupon-based\nrecruitment method for **hidden or hard-to-reach populations** (people who inject drugs, men who\nhave sex with men, female sex workers, undocumented migrants) — can be read, appraised, and\nreproduced. RDS is not a convenience sample dressed up: it recruits in waves from a small set of\n**seeds**, tracks who recruited whom via **coupons**, measures each participant's **network/degree\nsize**, and then re-weights the resulting sample with an RDS estimator (RDS-I, RDS-II, or\nSuccessive-Sampling/SS) to approximate a probability-based prevalence estimate. STROBE-RDS exists\nbecause none of that machinery — and the strong assumptions it rests on — is visible in a generic\nSTROBE report. Like other STROBE extensions it is hosted within the **EQUATOR Network** and is a\n*reporting* checklist, not a design or analysis recipe.\n\n**When to use** — Apply STROBE-RDS whenever the *sampling mechanism* is respondent-driven: the study\nstarts from seeds, recruits through peer-distributed coupons, records recruitment chains, and uses an\nRDS estimator to produce population-level prevalence, behavioral, or biomarker estimates. This is the\ncorrect checklist for an RDS-based HIV/HCV bio-behavioral surveillance study reported in a\npeer-reviewed journal, for the prevalence inputs to a global-health or disease-burden model, and for\nany cross-sectional survey of a hidden population where the recruitment is network-driven rather than\nframe-based. Decision rule: choose STROBE-RDS over plain STROBE only when recruitment is\nrespondent-driven; a venue-based (TLS) or facility sample of the same population uses STROBE (or the\nrelevant STROBE extension for the design), not STROBE-RDS. STROBE-RDS governs *reporting* — it is\nused alongside, not instead of, the survey protocol and the statistical analysis plan that pre-specify\nthe estimator and weighting scheme.\n\n**What it requires** — On top of STROBE's core items (title/abstract, background, objectives,\neligibility, variables, data sources/measurement, bias, study size, statistical methods, descriptive\nand outcome data, limitations, generalizability, funding), STROBE-RDS enforces the RDS-specific\nreporting that determines whether the estimates can be believed: (1) **formative assessment** and the\nrationale for choosing RDS for this population; (2) **seed selection** — how many seeds, how chosen,\nand their characteristics, because seed dependence is the dominant threat to RDS validity; (3)\n**coupon management** — number of coupons per recruit, coupon tracking, and the recruitment incentive\nstructure; (4) **recruitment-chain / wave structure** — depth and breadth of the trees, number of\nwaves reached, and convergence behavior; (5) **network/degree-size measurement** — the exact\npersonal-network-size question used (it feeds RDS-II/SS weights directly); (6) **diagnostics** —\n**equilibrium** (whether the sample composition stabilized across waves), **homophily** (in-group\nrecruitment tendency), bottlenecks, and recruitment-tree visualization; (7) **estimator and weighting**\n— which RDS estimator was used (RDS-I/RDS-II/SS), the assumptions invoked (random recruitment, accurate\ndegree report, with-replacement vs finite-population correction), and the software; (8) **uncertainty\nand design effect** — confidence intervals computed with an RDS-appropriate method (bootstrap that\nrespects the tree structure) and the **design effect** versus simple random sampling; and (9)\n**sensitivity analyses** to seed choice, degree-measurement error, and recruitment-bias assumptions.\nFramed in RWE terms, the burden is **fitness-for-purpose of a non-probability sample**: design\ntransparency (the recruitment process), the estimand (a population prevalence/proportion), selection\nand weighting (the RDS estimator *is* the confounding/selection control), and quantitative sensitivity\nanalysis to the assumptions the weights depend on.\n\n**When NOT to use — limitations and common misapplications** — (1) **It is not for standard\nclaims/EHR/registry HEOR RWE.** A retrospective cohort or comparative-effectiveness study in\nadministrative claims, EHR, or a disease registry uses **STROBE**, **RECORD**, **RECORD-PE**, or\n**HARPER** — never STROBE-RDS, which has no items for those data and omits everything those guidelines\nrequire (database provenance, code lists, phenotype validation, time-zero alignment, confounding by\nindication). (2) **It is not for non-RDS samples.** Using STROBE-RDS for a venue-based, facility-based,\nor probability survey misreports the design; conversely, reporting an RDS study with plain STROBE hides\nthe seeds, coupons, network-size question, equilibrium and homophily diagnostics, and the estimator —\nthe canonical misapplication the extension was written to prevent. (3) **It is a reporting checklist,\nnot a risk-of-bias instrument or a quality score.** A fully STROBE-RDS-compliant paper can still rest\non a biased sample: complete reporting of seed dependence, failure to reach equilibrium, or inaccurate\ndegree reports documents the problem, it does not fix it. (4) **Completing the checklist does not make\nthe estimate population-representative or causal.** RDS approximates a probability sample only when its\nassumptions hold; the checklist forces those assumptions into the open so reviewers can judge them.\n(5) **Checklist-as-theater** — ticking items while leaving the network-size question, equilibrium\ndiagnostics, or the estimator choice vague defeats the purpose; the value is the substantive disclosure,\nnot the page count.\n\n**How it maps to this catalog** — STROBE-RDS sits with the cross-sectional, prevalence-estimation\ncorner of this repo, not the comparative-effectiveness corner:\n- The study type it reports: **cross-sectional**.\n- The estimand it targets: **prevalence-point-period-annual-rwe** — RDS exists to produce a\n population prevalence/proportion, and the checklist's estimator/weighting items are how that estimand\n is made defensible from a non-probability sample.\n- The selection-and-weighting machinery: **selection-bias-sensitivity-analysis-rwe** — RDS-II/SS\n weighting and the required seed/degree/recruitment-bias sensitivity analyses are precisely\n quantitative bias/selection analysis applied to a chain-referral sample.\n- External validity: **generalizability-transportability-external-validity-rwe** — seed dependence and\n failure to reach equilibrium are external-validity threats the checklist forces a study to confront.\n- The pre-specification spine: **picots-framework-rwe** (the population/outcome/setting frame the\n formative assessment must declare) and **database-feasibility-attrition-funnel-rwe** as the RWE analog\n of RDS formative assessment and recruitment accounting (seeds → coupons issued → coupons returned →\n eligible → analyzed).\n\n**Applied note (RDS operational depth, with the HEOR caveat).** For an RDS HIV/HCV bio-behavioral\nsurveillance study, the reportable operational chain is: justify RDS in formative work; document seed\ncount and selection; fix coupons-per-recruit and track the recruitment tree; ask a validated\npersonal-network-size question; assess **equilibrium** across waves and **homophily** before trusting\nany estimate; report the chosen estimator (RDS-II or Successive-Sampling), its assumptions, and the\nsoftware; compute CIs with a tree-aware bootstrap and report the **design effect**; and run sensitivity\nanalyses to seed choice and degree misreport. The one-line HEOR caveat that belongs in any catalog\ncross-walk: if your real-world study draws on **claims, EHR, or a registry**, STROBE-RDS is the wrong\ntool — use **STROBE/RECORD/RECORD-PE/HARPER**; STROBE-RDS is reserved for studies whose *sampling* is\nrespondent-driven.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "strobe", "respondent-driven-sampling", "hidden-populations", "prevalence", "cross-sectional", "equator" ], "aliases": [ "STROBE-RDS", "STROBE for Respondent-Driven Sampling", "STROBE extension for respondent-driven sampling studies", "Strengthening the Reporting of Observational Studies in Epidemiology for Respondent-Driven Sampling" ], "applies_to_study_types": [ "cross_sectional" ], "data_sources": [ "primary" ], "citations": [ { "role": "introduce", "doi": "10.1016/j.jclinepi.2015.04.002", "url": "https://doi.org/10.1016/j.jclinepi.2015.04.002", "citation_text": "White RG, Hakim AJ, Salganik MJ, et al. Strengthening the Reporting of Observational Studies in Epidemiology for respondent-driven sampling studies: \"STROBE-RDS\" statement. Journal of Clinical Epidemiology. 2015;68(12):1463-1471.", "year": 2015, "authors_short": "White et al.", "notes": "Canonical STROBE-RDS statement defining the RDS-specific reporting items that extend the parent STROBE checklist for chain-referral surveys of hidden populations." }, { "role": "explain", "doi": "10.1371/journal.pmed.0040296", "url": "https://doi.org/10.1371/journal.pmed.0040296", "citation_text": "von Elm E, Altman DG, Egger M, et al. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. PLoS Medicine. 2007;4(10):e296.", "year": 2007, "authors_short": "von Elm et al.", "notes": "Parent STROBE statement whose 22-item core checklist STROBE-RDS extends; establishes the baseline observational-study reporting items." }, { "role": "use", "doi": "10.1016/j.jclinepi.2019.09.024", "url": "https://doi.org/10.1016/j.jclinepi.2019.09.024", "citation_text": "Avery L, Rotondi N, McKnight C, et al. More comprehensive reporting of methods in studies using respondent driven sampling is required: a systematic review of the uptake of the STROBE-RDS guidelines. Journal of Clinical Epidemiology. 2020;117:68-77.", "year": 2020, "authors_short": "Avery et al.", "notes": "Systematic review of STROBE-RDS uptake documenting which RDS-specific items (seeds, network size, equilibrium, estimator) remain underreported and where compliance is weakest." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "cross-sectional", "notes": "STROBE-RDS reports cross-sectional surveys of hidden populations recruited by respondent-driven (chain-referral) sampling." }, { "relation_type": "used_with", "target_slug": "prevalence-point-period-annual-rwe", "notes": "RDS exists to estimate a population prevalence/proportion; STROBE-RDS's estimator and weighting items make that estimand defensible from a non-probability sample." }, { "relation_type": "used_with", "target_slug": "selection-bias-sensitivity-analysis-rwe", "notes": "RDS-II/SS weighting and the required seed/degree/recruitment-bias sensitivity analyses are quantitative selection-bias analysis applied to a chain-referral sample." }, { "relation_type": "see_also", "target_slug": "generalizability-transportability-external-validity-rwe", "notes": "Seed dependence and failure to reach equilibrium are external-validity threats STROBE-RDS forces a study to report and probe." }, { "relation_type": "see_also", "target_slug": "picots-framework-rwe", "notes": "The formative-assessment and objectives items map to a structured population/outcome/setting frame the study must declare in advance." }, { "relation_type": "see_also", "target_slug": "database-feasibility-attrition-funnel-rwe", "notes": "RWE analog of RDS formative assessment and recruitment accounting (seeds issued -> coupons returned -> eligible -> analyzed)." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "journal", "hta" ] }, { "slug": "strobe-simsurg", "name": "STROBE Extension for Simulation-Based Research (STROBE-SIM)", "short_definition": "An EQUATOR-listed extension of the STROBE statement that adds simulation-specific reporting items to observational (non-randomized) studies of healthcare simulation — including surgical and procedural simulation — so that simulator fidelity, scenario design, instructional context, and outcome measurement are transparent and reproducible.", "long_description": "**What it is** — The **STROBE Extension for Simulation-Based Research (STROBE-SIM)** is the\nobservational-study arm of a pair of reporting guidelines for **health care simulation research\n(SBR)** developed through an INSPIRE (International Network for Simulation-based Pediatric\nInnovation, Research and Education)-led international consensus process and published by Cheng,\nKessler, Mackinnon et al. in 2016. The work produced two coordinated extensions, co-published in\n*Advances in Simulation* and *Simulation in Healthcare*: a **CONSORT** extension for *randomized*\nsimulation trials and a **STROBE** extension for *non-randomized/observational* simulation studies,\neach accompanied by an Explanation-and-Elaboration document with worked examples. STROBE-SIM does\nnot replace the parent STROBE checklist; it **extends 10 of STROBE's 22 items** — title/abstract\n(1), background/rationale (2), variables (7), data sources/measurement (8), statistical methods\n(12), descriptive data (14), main results (16), limitations (19), generalizability (21), and\nfunding (22) — with content unique to simulation: the simulator/manikin and its fidelity, scenario\nstandardization, instructional design (feedback, debriefing, mastery learning), participant/learner\ncharacteristics, and the validity argument for simulation-derived outcome measures. It is a\n*reporting* tool maintained within the EQUATOR Network library: it tells authors, reviewers, and\nreaders what a simulation study must disclose to be appraisable and reproducible. It is not a\ndesign recipe, a risk-of-bias instrument, or a quality score. (The \"simsurg\" slug reflects this\ncatalog's surgical-simulation entry point; the published guideline applies to health care\nsimulation broadly, surgical and procedural simulation included.)\n\n**When to use** — Apply STROBE-SIM **together with the parent STROBE checklist** when you are\n*reporting* (or refereeing) a **non-randomized/observational** study whose intervention, exposure,\nor measurement instrument is a **simulation** — for example, an observational cohort evaluating a\nsurgical simulation curriculum, a cross-sectional validity study of a simulator-based skills\nassessment, or a before-after evaluation of simulation training on procedural performance. The\nnatural home is a peer-reviewed simulation, surgical-education, or medical-education journal.\n**Decision rule for choosing the right family member:** if the simulation study is a *randomized\ntrial*, use the **CONSORT** simulation extension, not STROBE-SIM; if it is observational, use\n**STROBE + STROBE-SIM**. If the study is observational but *not* about simulation, use base\n**STROBE** (or the design/data extension that fits — RECORD/RECORD-PE for routinely-collected\nhealth data, STROBE-MR for Mendelian randomization, STROBE-NI for non-inferiority claims).\nSTROBE-SIM governs *reporting of the completed study*; the *protocol* of a primary study is\npre-specified with SPIRIT (trials) or the relevant primary-study protocol guidance, and a\n*systematic review* of simulation studies is reported with PRISMA — never STROBE-SIM.\n\n**What it requires** — Beyond the full parent-STROBE disclosures (design named in the\ntitle/abstract, setting and time anchors, eligibility and participant flow, explicit variable\ndefinitions, statistical methods including confounding control and missing-data handling, both\nunadjusted and adjusted estimates, and a candid limitations/generalizability discussion), the 10\nSTROBE-SIM extended items compel the simulation-specific reporting that otherwise makes a study\nun-reproducible: a clear statement that the work is simulation-based and the **type of simulation**\n(item 1); the **educational/theoretical rationale** for the simulation intervention (item 2);\nprecise definition of the **simulation intervention, simulator, and fidelity** plus the\ninstructional design — scenario, feedback, debriefing, dosage (items 7-8); the **psychometric/\nvalidity evidence** for simulation-derived outcome measures and rater training (items 8, 12, 16);\nfull description of **participants/learners and the simulation environment** (item 14); and\nexplicit discussion of the **fidelity gap and transfer** — the threat that performance in the\nsimulated setting may not transfer to real patients — under limitations and generalizability\n(items 19, 21). The validity-of-measurement and transfer requirements are the methodological\ncore: a simulation result is only as interpretable as the argument that the simulated task and its\nscoring stand in for the real clinical task.\n\n**When NOT to use — limitations and common misapplications** — (1) **Wrong family member:** using\nSTROBE-SIM to report a *randomized* simulation trial — that is the CONSORT simulation extension's\njob — or, conversely, citing the CONSORT extension for an observational design. (2) **Wrong\nextension for the design:** applying STROBE-SIM to an observational study that has nothing to do\nwith simulation (it belongs to base STROBE/RECORD), or applying base STROBE/RECORD to a simulation\nstudy and thereby omitting fidelity, scenario, and validity-of-measurement reporting. (3)\n**Mistaking it for an appraisal tool:** STROBE-SIM is a reporting checklist, not a risk-of-bias\ninstrument and not a quality score; ticking the items does not certify that the study is valid,\nthat the simulator is fit for purpose, or that simulated performance transfers to patient care.\n(4) **Checklist-as-theater:** listing \"high-fidelity simulator\" without the make/model, fidelity\ndimensions, scenario script, debriefing structure, and rater-reliability evidence satisfies the\nletter and defeats the purpose. (5) **Over-claiming external relevance:** completing the checklist\ndoes not make a simulation-based finding generalizable to real patients; the transfer/fidelity-gap\nlimitation (items 19, 21) must be argued, not asserted.\n\n**How it maps to this catalog** — Stated honestly, STROBE-SIM is **largely orthogonal to the\nroutinely-collected-data (claims/EHR/registry) RWE concepts** that dominate this catalog; its\nnatural neighbors are the **parent `strobe`** statement it extends and the sibling **`consort`**\nfamily it was co-developed with (the randomized-trial arm of the same simulation guideline pair).\nDo not expect concepts such as target-trial emulation, high-dimensional propensity scores, the\nactive-comparator new-user design, or claims phenotype algorithms to implement this guideline —\nthey do not, and forcing that mapping would be the very \"wrong extension / fabricated relevance\"\nfailure mode the guideline warns against. Where genuine overlap exists, it is narrow and\nitem-specific:\n- **Parent reporting backbone:** `strobe` supplies the 22 base items STROBE-SIM extends — always\n cite and complete both.\n- **Generalizability / transfer (items 19, 21):** `generalizability-transportability-external-validity-rwe`\n is the relevant lens for arguing whether simulated performance and the studied learners\n transport to real clinical settings and patient outcomes.\n- **Statistical methods / estimands (item 12), *only if* the simulation study is comparative:**\n `estimands-ate-att-intercurrent-events-rwe` clarifies the target estimand and intercurrent-event\n handling when comparing groups (e.g., curriculum vs control) — relevant when the simulation study\n is observational-comparative, not when it is a single-arm validity study.\n- **Attrition (parent STROBE item 13):** `attrition-and-loss-to-follow-up-rwe` informs the\n participant-flow accounting when learners drop out across simulation sessions or follow-up to a\n patient-care outcome.\n\n**Applied note.** The realistic application is *not* a claims/EHR analysis. Consider a multicenter\nobservational study of a surgical simulation curriculum measuring skill acquisition on a virtual-\nreality laparoscopic trainer and transfer to operating-room performance. STROBE-SIM is what forces\nthe report to state the simulator make/model and fidelity, the standardized scenario and case mix,\nthe debriefing and feedback structure and dosage, the learners' baseline experience, the\npsychometric validity and inter-rater reliability of the scoring instrument, and — under\nlimitations — the fidelity gap and the strength of evidence that simulated performance predicts\nreal patient outcomes. Only if such a study were *also* linked to routinely-collected patient\noutcomes (e.g., registry-based complication rates) would the catalog's RWE data-fitness, time-zero,\nand confounding concepts begin to apply, and then under STROBE/RECORD rather than STROBE-SIM.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "strobe", "simulation", "surgical-education", "medical-education", "equator", "inspire" ], "aliases": [ "STROBE-SIM", "STROBE Extension for Simulation-Based Research", "Health Care Simulation Research STROBE Extension", "STROBE extension for simulation research" ], "applies_to_study_types": [ "cohort_prospective", "cohort_retrospective", "cross_sectional" ], "data_sources": [], "citations": [ { "role": "introduce", "doi": "10.1186/s41077-016-0025-y", "url": "https://doi.org/10.1186/s41077-016-0025-y", "citation_text": "Cheng A, Kessler D, Mackinnon R, et al. Reporting guidelines for health care simulation research: extensions to the CONSORT and STROBE statements. Advances in Simulation. 2016;1:25.", "year": 2016, "authors_short": "Cheng et al.", "notes": "Canonical statement introducing the paired CONSORT and STROBE extensions for health care simulation research, including the 10 extended STROBE items." }, { "role": "explain", "doi": "10.1097/SIH.0000000000000150", "url": "https://doi.org/10.1097/SIH.0000000000000150", "citation_text": "Cheng A, Kessler D, Mackinnon R, et al. Reporting Guidelines for Health Care Simulation Research: Extensions to the CONSORT and STROBE Statements. Simulation in Healthcare. 2016;11(4):238-248.", "year": 2016, "authors_short": "Cheng et al.", "notes": "Co-published companion version with the item-by-item explanation-and-elaboration content and worked examples for the simulation extensions." }, { "role": "use", "url": "https://www.equator-network.org/reporting-guidelines/reporting-guidelines-for-health-care-simulation-research-extensions-to-the-consort-and-strobe-statements/", "citation_text": "Reporting Guidelines for Health Care Simulation Research: Extensions to the CONSORT and STROBE Statements. EQUATOR Network reporting-guidelines library.", "year": 2016, "authors_short": "EQUATOR Network", "notes": "Maintained EQUATOR landing page linking the checklist and both co-published versions." } ], "relations": [ { "relation_type": "is_variant_of", "target_slug": "strobe", "notes": "STROBE-SIM extends the parent STROBE statement, adding simulation-specific content to 10 of its 22 items; the base checklist must be completed alongside it." }, { "relation_type": "applies_to", "target_slug": "cohort-prospective", "notes": "Use with STROBE to report a prospective observational simulation study." }, { "relation_type": "applies_to", "target_slug": "cohort-retrospective", "notes": "Use with STROBE to report a retrospective observational simulation study." }, { "relation_type": "applies_to", "target_slug": "cross-sectional", "notes": "Use with STROBE to report a cross-sectional simulation study (e.g., a simulator-based validity/assessment study)." }, { "relation_type": "alternative_to", "target_slug": "consort", "notes": "For a randomized simulation trial use the CONSORT simulation extension instead; STROBE-SIM is the non-randomized/observational arm of the same co-developed guideline pair." }, { "relation_type": "see_also", "target_slug": "generalizability-transportability-external-validity-rwe", "notes": "Relevant lens for the transfer/fidelity-gap discussion the extended items 19 and 21 require — whether simulated performance transports to real clinical settings and patient outcomes." }, { "relation_type": "see_also", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Applies only when the simulation study is observational-comparative — clarifies the target estimand and intercurrent-event handling for the statistical-methods item." }, { "relation_type": "see_also", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Informs the participant-flow accounting when learners drop out across simulation sessions or follow-up to a downstream outcome." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "journal" ] }, { "slug": "strobe-vet", "name": "STROBE-Vet (STROBE Extension for Veterinary Epidemiology)", "short_definition": "Reporting guideline that extends the 22-item STROBE checklist with veterinary-specific reporting items (clustering, housing/management, animal/herd-level units, ethics, and reporting of the source population) for observational studies in animal health; maintained within the EQUATOR Network.", "long_description": "**What it is** — **STROBE-Vet (Strengthening the Reporting of Observational Studies in Epidemiology —\nVeterinary)** is a reporting-guideline *extension* of the original STROBE Statement, tailored to\nobservational studies (cohort, case-control, and cross-sectional) conducted in veterinary medicine and\nanimal-health research. It does not replace STROBE: it inherits STROBE's 22 generic items and adds or\nmodifies a subset to capture features that recur in veterinary epidemiology — multilevel/clustered data\n(animals nested within pens, herds, flocks, farms, or clinics), the choice and reporting of the unit of\nanalysis (individual animal vs. herd/group), housing and management exposures, production-system context,\nanimal-welfare and ethics reporting, and explicit description of the source population and sampling frame\n(often a convenience or production sample rather than a defined catchment). The Statement (Sargeant,\nO'Connor, Dohoo et al., 2016) was co-published across four journals — *Journal of Veterinary Internal\nMedicine*, *Preventive Veterinary Medicine*, *Zoonoses and Public Health*, and *Journal of Food\nProtection* — and is accompanied by an item-by-item Explanation and Elaboration document (O'Connor et al.,\n2016). It is hosted and indexed as a STROBE extension within the **EQUATOR Network**.\n\n**When to use** — Apply STROBE-Vet when *reporting a completed observational study in animals or animal\npopulations*: prevalence/cross-sectional surveys of herds or flocks, cohort studies of production or\ncompanion animals, and case-control studies of veterinary outcomes, including studies drawing on\nveterinary administrative/production databases, herd registries, abattoir/surveillance data, or\ncompanion-animal clinical records (the veterinary analogue of human claims/EHR/registry data). Decision\nrule for choosing the right STROBE member: use **plain STROBE** for human observational studies; use\n**RECORD / RECORD-PE** for human studies built on routinely-collected health/pharmacoepidemiologic data;\nuse **STROBE-Vet** specifically when the observational study is *veterinary* and you want the field-specific\nitems (clustering, unit of analysis, housing/management, production context) reported. For *controlled\nanimal experiments* (laboratory/pre-clinical in-vivo work) the relevant guideline is **ARRIVE**, not\nSTROBE-Vet, because those are designed experiments, not observational studies.\n\n**What it requires** — STROBE-Vet enforces the STROBE reporting backbone — a structured title/abstract,\nbackground/objectives, an explicit study design named in the methods, setting and eligibility, clearly\ndefined variables (exposures, outcomes, confounders, effect modifiers) with measurement/data sources,\nefforts to address bias, study size and quantitative methods, a participant/animal flow and descriptive\ntable, main results with confounder-adjusted estimates and precision, and an honest limitations and\ngeneralizability discussion — and layers veterinary specifics on top: (1) the **unit of analysis and\nclustering structure** must be stated and carried through the analysis (animal vs. pen vs. herd; how\nwithin-cluster correlation was handled); (2) the **source population, sampling frame, and selection\nprocess** must be described, because veterinary samples are frequently non-random; (3) **housing,\nmanagement, and production-system covariates** that drive exposure and outcome must be reported; (4)\n**case/outcome and exposure definitions** (including diagnostic-test characteristics where outcomes are\ntest-defined — the veterinary parallel of phenotype/algorithm validation) must be explicit; and (5)\n**ethics, welfare, and animal-use approvals** must be reported. Framed against the same data-fitness and\ntransparency problems that human RWE faces, STROBE-Vet pushes the author to make design, data provenance,\ntime alignment, attrition, confounding control, and outcome-definition validity legible to the reader.\n\n**When NOT to use — limitations and common misapplications** — STROBE-Vet is a *reporting* checklist, not\na methodological fix. Concrete failure modes: (1) **Wrong domain** — using STROBE-Vet for a *human*\nobservational study; report human studies with STROBE, or RECORD/RECORD-PE when they use routinely-collected\ndata, or follow HARPER/ENCePP for pharmacoepidemiologic protocols. (2) **Wrong design class** — using\nSTROBE-Vet for a *controlled animal experiment*; designed in-vivo experiments are reported under ARRIVE.\n(3) **Mistaking it for a risk-of-bias or quality-appraisal instrument** — STROBE-Vet does not score study\nquality or grade bias; critical appraisal of non-randomized studies uses tools such as ROBINS-I, and it is\na category error to \"STROBE-score\" a paper. (4) **Checklist-as-theater** — ticking items while leaving the\nunit of analysis, clustering, sampling frame, or outcome definition vague defeats the purpose; the value is\nthe substantive transparency, not the completed grid. (5) **Reporting completeness ≠ internal validity** —\na fully STROBE-Vet-compliant cross-sectional herd study with confounding by management practice is still\nconfounded; the checklist makes the design legible, it does not make the estimate causal.\n\n**How it maps to this catalog** — This catalog's implementing concepts are written for *human* real-world\nevidence (claims/EHR/registry), so they are not literal implementations of a veterinary guideline. They\nare, however, **direct structural analogues**: the design and data problems STROBE-Vet asks you to report\nare the same problems these concepts operationalize, and a veterinary RWE study can borrow the same\nmachinery against animal-health administrative, production, and clinical-record databases. Read each as the\nhuman-side worked example of a requirement STROBE-Vet imposes:\n- Design transparency and a trial-emulation frame for observational comparisons → **target-trial-emulation**\n and **active-comparator-new-user**.\n- Time alignment / index-date definition (the \"when does follow-up start\" item) → **time-zero-index-date-alignment-rwe**.\n- Outcome and exposure definition validity (test- or algorithm-defined cases) →\n **diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe** and **algorithm-validation**.\n- Confounding control, including the management/housing covariates STROBE-Vet flags →\n **high-dimensional-propensity-score-hdps-rwe** and **propensity-score-methods-psm-iptw**.\n- Estimands and how intercurrent events are handled → **estimands-ate-att-intercurrent-events-rwe**.\n- Attrition / loss to follow-up and the participant-flow item → **attrition-and-loss-to-follow-up-rwe**.\n- Generalizability of the (often non-random) sample → **generalizability-transportability-external-validity-rwe**.\n- The administrative-database mechanics themselves → **claims-analysis** as the methodological template.\nTreat these as parallel structure, not as plug-ins: the clustering, animal-vs-herd unit of analysis, and\nproduction-system context are veterinary-specific and have no direct slug here.\n\n**Applied note (veterinary administrative / herd-registry / clinical-record RWE).** A retrospective cohort\nbuilt from a swine-production database or a companion-animal practice-management EHR faces the same fitness-\nfor-use questions as a human claims study: is enrollment/observation continuous, is the outcome definition\nvalidated, is time zero aligned to a real decision point, and is loss to follow-up informative? Report the\nunit of analysis (animal vs. litter vs. herd) and the within-cluster correlation handling explicitly, state\nthe sampling frame and why it may not represent the target population, and pre-specify how diagnostic-test-\ndefined outcomes were validated — exactly the transparency the human-RWE concept entries above demonstrate.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "strobe", "veterinary", "observational", "equator", "clustering" ], "aliases": [ "STROBE-Vet", "STROBE Veterinary Extension", "STROBE Extension for Veterinary Epidemiology", "Strengthening the Reporting of Observational Studies in Epidemiology - Veterinary" ], "applies_to_study_types": [ "cohort_prospective", "cohort_retrospective", "case_control", "cross_sectional" ], "data_sources": [ "registry", "ehr" ], "citations": [ { "role": "introduce", "doi": "10.1111/jvim.14574", "url": "https://doi.org/10.1111/jvim.14574", "citation_text": "Sargeant JM, O'Connor AM, Dohoo IR, et al. Methods and Processes of Developing the Strengthening the Reporting of Observational Studies in Epidemiology - Veterinary (STROBE-Vet) Statement. Journal of Veterinary Internal Medicine. 2016;30(6):1887-1895.", "year": 2016, "authors_short": "Sargeant et al.", "notes": "Canonical STROBE-Vet Statement defining the veterinary extension items. Co-published in 2016 in J Vet Intern Med, Preventive Veterinary Medicine (10.1016/j.prevetmed.2016.09.005), Zoonoses and Public Health, and J Food Protection; the JVIM version is cited here as the primary record." }, { "role": "explain", "doi": "10.1111/jvim.14592", "url": "https://doi.org/10.1111/jvim.14592", "citation_text": "O'Connor AM, Sargeant JM, Dohoo IR, et al. Explanation and Elaboration Document for the STROBE-Vet Statement: Strengthening the Reporting of Observational Studies in Epidemiology - Veterinary Extension. Journal of Veterinary Internal Medicine. 2016;30(6):1896-1928.", "year": 2016, "authors_short": "O'Connor et al.", "notes": "Item-by-item explanation and elaboration with veterinary worked examples for each STROBE-Vet item." }, { "role": "explain", "doi": "10.1371/journal.pmed.0040296", "url": "https://doi.org/10.1371/journal.pmed.0040296", "citation_text": "von Elm E, Altman DG, Egger M, et al. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement: Guidelines for Reporting Observational Studies. PLoS Medicine. 2007;4(10):e296.", "year": 2007, "authors_short": "von Elm et al.", "notes": "Parent STROBE Statement whose 22 generic items STROBE-Vet inherits and extends." }, { "role": "use", "url": "https://www.equator-network.org/reporting-guidelines/strobe-vet/", "citation_text": "STROBE-Vet (Strengthening the Reporting of Observational Studies in Epidemiology - Veterinary). EQUATOR Network reporting-guidelines library (maintained checklist and links).", "year": 2016, "authors_short": "EQUATOR Network", "notes": "Maintained landing page indexing the STROBE-Vet checklist and the explanation-and-elaboration document." } ], "relations": [ { "relation_type": "is_variant_of", "target_slug": "strobe", "notes": "STROBE-Vet inherits STROBE's 22 generic items and adds veterinary-specific reporting items (clustering, unit of analysis, housing/management, production context, animal ethics)." }, { "relation_type": "applies_to", "target_slug": "cohort-prospective", "notes": "Use when reporting a prospective observational cohort in animals or animal populations." }, { "relation_type": "applies_to", "target_slug": "cohort-retrospective", "notes": "Use when reporting a retrospective veterinary cohort, including studies built on production or clinical-record databases." }, { "relation_type": "applies_to", "target_slug": "case-control", "notes": "Use when reporting a veterinary case-control study." }, { "relation_type": "applies_to", "target_slug": "cross-sectional", "notes": "Use when reporting a veterinary cross-sectional / prevalence survey of herds, flocks, or clinics." }, { "relation_type": "see_also", "target_slug": "target-trial-emulation", "notes": "Human-RWE analogue for the design-transparency item; a veterinary observational comparison can be framed as a target-trial emulation." }, { "relation_type": "see_also", "target_slug": "time-zero-index-date-alignment-rwe", "notes": "Human-RWE analogue for the STROBE item requiring an explicit time origin / start of follow-up." }, { "relation_type": "see_also", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Human-RWE analogue for defining and validating test/algorithm-defined outcomes in veterinary data." }, { "relation_type": "see_also", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "Human-RWE analogue for confounding control over the housing/management/production covariates STROBE-Vet asks authors to report." }, { "relation_type": "see_also", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Human-RWE analogue for pre-specifying the estimand and handling intercurrent events." }, { "relation_type": "see_also", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Human-RWE analogue for the participant/animal flow and loss-to-follow-up reporting item." }, { "relation_type": "see_also", "target_slug": "generalizability-transportability-external-validity-rwe", "notes": "Human-RWE analogue for judging external validity of the often non-random veterinary sample." }, { "relation_type": "see_also", "target_slug": "claims-analysis", "notes": "Methodological template for working with administrative databases, applicable to veterinary production/registry data." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "journal" ] }, { "slug": "strobe", "name": "STROBE", "short_definition": "The core EQUATOR-hosted reporting checklist (22 items) for observational epidemiological studies — cohort, case-control, and cross-sectional — that specifies the minimum information a completed study must report so its design, conduct, and analysis are transparent and appraisable; the parent statement for all STROBE extensions.", "long_description": "**What it is** — **STROBE (Strengthening the Reporting of Observational Studies in\nEpidemiology)** is a 22-item reporting checklist that specifies the minimum content a\n*completed* observational study should report across its title/abstract, introduction,\nmethods, results, and discussion. It was developed by an international collaboration of\nepidemiologists, methodologists, statisticians, and journal editors and published in\n2007 as the STROBE Statement (von Elm et al.) with a companion Explanation-and-Elaboration\npaper (Vandenbroucke et al.) that gives the rationale and worked examples for each item.\nSTROBE is maintained as a reporting guideline within the **EQUATOR Network** library and\nis the *parent* statement from which all design- and domain-specific extensions descend.\nIts three core designs — **cohort, case-control, and cross-sectional** — share 18 common\nitems, with 4 items reported differently by design. STROBE is a *reporting* tool: it tells\nauthors, reviewers, and readers what must be disclosed so a study can be understood and\ncritically appraised. It does not prescribe how to design or analyze the study, and it\ndoes not score quality.\n\n**When to use** — Apply STROBE when you are *reporting* (or refereeing, or registering a\nreporting plan for) a primary observational study of one of the three core designs:\nprospective or retrospective cohort, case-control, or cross-sectional. It is the default\nreporting backbone for an observational manuscript in a peer-reviewed journal, the reporting\nappendix of an HTA/payer dossier built on a non-interventional study, and the transparency\nlayer of an FDA/EMA real-world-evidence submission or PASS report. **Decision rule for\nchoosing STROBE vs an extension:** use the *base* STROBE only when no more specific\nextension governs your design or data. If the study uses **routinely-collected health data**\n(claims, EHR, disease/administrative registries, linked databases), the routinely-collected-data\nextension **RECORD** applies, and for pharmacoepidemiology specifically **RECORD-PE** — these\nadd items on database provenance, code lists, data-cleaning, and linkage that base STROBE\ndoes not cover. Use **STROBE-MR** for Mendelian randomization, **STROME-ID** for infectious-disease\nmolecular epidemiology and **STROBE-ME** for molecular epidemiology (biomarkers), **STROBE-NI** for\nobservational studies of newborn infection, **STROBE-RDS** for respondent-driven sampling, and the\nveterinary/nutritional/equity variants where they fit. STROBE governs *reporting of the\ncompleted study*; the *protocol* of a primary RWE study is pre-specified with HARPER,\nStaRT-RWE, or the ENCePP checklist, not STROBE; and a *systematic review* is reported with\nPRISMA, not STROBE.\n\n**What it requires** — STROBE's 22 items compel disclosure of the elements that, when left\nvague, make an observational result un-interpretable. Substantive domains, framed for\nreal-world data: **design transparency** (item 1 — name the design in the title/abstract;\nitem 4 — present key elements of the design early); **setting and time anchors** (item 5 —\nsetting, locations, and the relevant dates of recruitment, exposure, follow-up, and data\ncollection, which in RWD means the index/time-zero definition and the lookback and follow-up\nwindows); **eligibility and participant flow** (items 6 and 13 — sources, selection methods,\nand a numeric account of participants at each stage, the attrition funnel); **variable\ndefinitions** (item 7 — explicit operational definitions of outcomes, exposures, predictors,\nconfounders, and effect modifiers, which for claims/EHR means the phenotype/algorithm logic\nand code lists); **measurement and data sources** (item 8 — sources and methods of\nassessment, including comparability across data sources); **bias** (item 9 — efforts to\naddress potential sources of bias); **study size and quantitative variable handling** (items\n10-11); **statistical methods** (item 12 — all methods including how confounding was\ncontrolled, how subgroups/interactions and missing data were handled, and how loss to\nfollow-up was addressed); **results** (items 13-17 — flow, descriptive data, outcome counts,\nand crucially item 16 — *both* unadjusted and confounder-adjusted estimates with precision,\nplus the confounders adjusted for); and **interpretation** (item 18-20 — key results,\nlimitations including direction and magnitude of potential bias, and cautious generalizability).\nNote what STROBE does *not* itself mandate but a credible RWE study should report alongside it:\nfitness-for-purpose of the data source, phenotype/algorithm validation metrics (PPV/sensitivity),\nestimand and intercurrent-event handling, positivity/overlap diagnostics, and quantitative\nbias analysis — these are the substance that base STROBE's generic items (7, 9, 12, 19) only\ngesture at, and which RECORD-PE and the catalog concepts below make explicit.\n\n**When NOT to use — limitations and common misapplications** — (1) **A reporting checklist is\nnot a risk-of-bias instrument and not a quality score.** STROBE tells you whether a study is\n*described* completely, not whether it is *valid*. Do not sum ticked items into a \"STROBE\nscore\" to rank studies — the developers explicitly warn against this; appraise validity with\nROBINS-I (or ROBINS-E), the Newcastle-Ottawa Scale, or a domain-based tool instead. (2)\n**Completing the checklist does not make an observational study causal or unconfounded.** A\nfully STROBE-compliant paper can still rest on a hopelessly confounded design; transparency\nis necessary, not sufficient. (3) **Using STROBE where RECORD/RECORD-PE is required.** A\nclaims- or EHR-based pharmacoepidemiologic study reported against base STROBE alone will omit\nthe database provenance, code-list, linkage, and data-cleaning items that RECORD-PE exists to\nenforce — a recurring reviewer rejection in regulatory and HTA submissions. (4) **Wrong\nextension for the design** — citing base STROBE for a Mendelian randomization study (needs\nSTROBE-MR) or a routinely-collected-health-data study (needs RECORD/RECORD-PE), or citing STROBE for a\nrandomized trial (CONSORT) or systematic review (PRISMA). (5) **Checklist-as-theater** —\nattaching a completed checklist whose page references point to text that is itself vague (an\n\"outcomes were defined using ICD codes\" with no code list, no validation, no time window)\nsatisfies the box but defeats the purpose; the value is the *content disclosed*, not the\ncompleted grid.\n\n**How it maps to this catalog** — STROBE's generic items become concrete when each is\nimplemented by a catalog concept the author can point the page reference at:\n- **Design transparency / eligibility / time anchors (items 1, 4-6):** the comparative design\n is built with **active-comparator-new-user** and, for the formal causal contrast,\n **target-trial-emulation**; the index date and immortal-time-safe alignment that item 5's\n \"relevant dates\" demand are implemented by **time-zero-index-date-alignment-rwe**.\n- **Variable definitions and measurement (items 7-8):** outcome/exposure operational\n definitions and their code logic are implemented by\n **diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe**, with validity established by\n **algorithm-validation**; the PICOTS spine for these definitions is **picots-framework-rwe**.\n- **Bias and confounding control (items 9, 12, 16):** confounding adjustment that item 16's\n \"confounder-adjusted estimates\" presupposes is implemented by\n **high-dimensional-propensity-score-hdps-rwe**, with balance reporting by\n **baseline-characteristics-and-covariate-balance-rwe**; the estimand and intercurrent-event\n framing behind item 12 is **estimands-ate-att-intercurrent-events-rwe**.\n- **Participant flow and follow-up (items 12-13):** the attrition funnel and informative loss\n to follow-up are implemented by **attrition-and-loss-to-follow-up-rwe**.\n- **Limitations and sensitivity (item 19):** quantitative bias analysis for residual\n confounding is implemented by **e-value-sensitivity-analysis**.\n- **Data fitness (underpins items 5, 8, 19):** **fit-for-purpose-data-assessment-rwe** and the\n data-source operational depth in **claims-analysis** supply the provenance and limitations\n that base STROBE only gestures at and that RECORD-PE makes mandatory.\n\n**Applied note (claims/EHR/registry RWE).** For a claims- or EHR-based cohort, report against\nSTROBE *and* RECORD-PE: item 5's \"relevant dates\" must specify the index/time-zero rule,\nlookback, and follow-up windows; item 6 must give the eligibility/enrollment requirements and\na numeric attrition funnel from source population to analytic cohort; item 7 must publish the\nphenotype algorithm and code lists with validation metrics rather than a bare ICD reference;\nitem 12 must state the confounding-control method (e.g., hdPS) and the estimand; and item 16\nmust present unadjusted and adjusted effect estimates with the covariate set. Treat the\nchecklist as a map from each reporting obligation to a specific, versioned artifact — protocol\nsection, code commit, validation result, or diagnostic plot — not as a final-page formality.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "observational", "equator", "pharmacoepidemiology", "rwe" ], "aliases": [ "STROBE", "STROBE Statement", "Strengthening the Reporting of Observational Studies in Epidemiology", "STROBE 2007" ], "applies_to_study_types": [ "cohort_prospective", "cohort_retrospective", "case_control", "cross_sectional" ], "data_sources": [ "claims", "ehr", "registry", "linked", "primary" ], "citations": [ { "role": "introduce", "doi": "10.1371/journal.pmed.0040296", "url": "https://doi.org/10.1371/journal.pmed.0040296", "citation_text": "von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP; STROBE Initiative. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement: guidelines for reporting observational studies. PLoS Medicine. 2007;4(10):e296.", "year": 2007, "authors_short": "von Elm et al.", "notes": "Canonical STROBE Statement presenting the 22-item checklist for cohort, case-control, and cross-sectional studies." }, { "role": "explain", "doi": "10.1371/journal.pmed.0040297", "url": "https://doi.org/10.1371/journal.pmed.0040297", "citation_text": "Vandenbroucke JP, von Elm E, Altman DG, et al. Strengthening the Reporting of Observational Studies in Epidemiology (STROBE): explanation and elaboration. PLoS Medicine. 2007;4(10):e297.", "year": 2007, "authors_short": "Vandenbroucke et al.", "notes": "Item-by-item rationale with worked examples; the authoritative reference for what each STROBE item requires and why." }, { "role": "explain", "doi": "10.1136/bmj.k3532", "url": "https://doi.org/10.1136/bmj.k3532", "citation_text": "Langan SM, Schmidt SAJ, Wing K, et al. The reporting of studies conducted using observational routinely collected health data statement for pharmacoepidemiology (RECORD-PE). BMJ. 2018;363:k3532.", "year": 2018, "authors_short": "Langan et al.", "notes": "The pharmacoepidemiology extension that supersedes base STROBE for routinely-collected data; explains the database-provenance, code-list, and linkage items STROBE omits." }, { "role": "use", "url": "https://www.strobe-statement.org/", "citation_text": "STROBE Statement — official site (checklists for cohort, case-control, and cross-sectional studies in usable formats, plus the full list of STROBE extensions).", "year": 2007, "authors_short": "STROBE Initiative", "notes": "Maintained landing page with downloadable per-design checklists and links to extensions (RECORD, RECORD-PE, STROBE-MR, STROBE-ME, STROME-ID, and others)." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "cohort-prospective", "notes": "STROBE governs reporting of completed prospective cohort studies." }, { "relation_type": "applies_to", "target_slug": "cohort-retrospective", "notes": "STROBE governs reporting of completed retrospective cohort studies; for claims/EHR cohorts use RECORD-PE alongside it." }, { "relation_type": "applies_to", "target_slug": "case-control", "notes": "STROBE governs reporting of completed case-control studies, including case/control selection and matching items." }, { "relation_type": "applies_to", "target_slug": "cross-sectional", "notes": "STROBE governs reporting of completed cross-sectional studies, including the single assessment-time and prevalence framing." }, { "relation_type": "see_also", "target_slug": "record-pe", "notes": "The pharmacoepidemiology extension of STROBE; required (not optional) when the study uses routinely-collected data, adding database-provenance, code-list, and linkage items." }, { "relation_type": "see_also", "target_slug": "record", "notes": "The routinely-collected-data extension of STROBE for non-pharmacoepi designs." }, { "relation_type": "see_also", "target_slug": "robins-i", "notes": "Risk-of-bias appraisal for non-randomized studies; use this to judge validity — STROBE reports completeness, not bias." }, { "relation_type": "see_also", "target_slug": "harper", "notes": "For the primary-study PROTOCOL (pre-specification), use HARPER/StaRT-RWE/ENCePP; STROBE governs the completed-study report, not the protocol." }, { "relation_type": "complements", "target_slug": "target-trial-emulation", "notes": "Implements the design transparency and causal-contrast framing behind STROBE items 1, 4-6, and 12." }, { "relation_type": "complements", "target_slug": "active-comparator-new-user", "notes": "Implements the comparative design and eligibility/time-zero structure STROBE items 5-7 must disclose." }, { "relation_type": "complements", "target_slug": "time-zero-index-date-alignment-rwe", "notes": "Implements the index/time-zero definition and immortal-time avoidance behind STROBE item 5's relevant-dates requirement." }, { "relation_type": "complements", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Implements the operational outcome/exposure definitions STROBE item 7 requires for claims/EHR studies (code logic, time windows)." }, { "relation_type": "complements", "target_slug": "algorithm-validation", "notes": "Supplies the phenotype/algorithm validation metrics (PPV, sensitivity) STROBE item 7's \"clearly defined\" demands of algorithm-defined variables." }, { "relation_type": "complements", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "Implements the confounding control STROBE items 12 and 16 (confounder-adjusted estimates) presuppose." }, { "relation_type": "complements", "target_slug": "baseline-characteristics-and-covariate-balance-rwe", "notes": "Implements the descriptive and balance reporting behind STROBE items 14-16." }, { "relation_type": "complements", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Implements the estimand and intercurrent-event framing behind STROBE item 12's analysis plan." }, { "relation_type": "complements", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Implements the participant-flow and informative-loss reporting behind STROBE items 12-13." }, { "relation_type": "complements", "target_slug": "e-value-sensitivity-analysis", "notes": "Implements quantitative bias analysis for residual confounding behind STROBE item 19's limitations." }, { "relation_type": "complements", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "Supplies the data-source fitness and provenance underpinning STROBE items 5, 8, and 19 (and made mandatory by RECORD-PE)." }, { "relation_type": "complements", "target_slug": "claims-analysis", "notes": "Supplies the claims data-source operational depth STROBE/RECORD-PE require for variable definitions and limitations." }, { "relation_type": "complements", "target_slug": "picots-framework-rwe", "notes": "Structures the population/intervention/comparator/outcome/timing/setting spine STROBE items 6-7 must declare." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "hta", "journal" ] }, { "slug": "strocss", "name": "STROCSS (Strengthening the Reporting of Cohort, Cross-Sectional and Case-Control Studies in Surgery)", "short_definition": "Surgery-specific reporting checklist for observational studies (cohort, cross-sectional, case-control), a STROBE-derived extension that adds surgical-research items (e.g., learning curve, intervention/technique detail, follow-up of operated patients); maintained by the STROCSS Group and listed in the EQUATOR Network library.", "long_description": "**What it is** — **STROCSS (Strengthening the Reporting of Cohort, Cross-Sectional and Case-Control Studies in\nSurgery)** is a surgery-specific *reporting* checklist for observational research. It is a domain extension of the\ngeneric **STROBE** statement, adapted by the STROCSS Group to capture items that matter when the exposure is a\nsurgical procedure or technique: a clear description of the intervention/operative technique, surgeon experience and\nthe *learning curve*, peri-operative and longer-term follow-up of operated patients, and registration of the study.\nSTROCSS is maintained as a living guideline — the **STROCSS 2024** statement (Rashid et al., *International Journal\nof Surgery*) is the current version, superseding STROCSS 2021 (Mathew et al.) and the earlier 2017/2019 iterations —\nand is indexed in the **EQUATOR Network** reporting-guideline library. Like all EQUATOR checklists, its job is\n*transparency of reporting*: it tells authors, peer reviewers, and readers the minimum that a surgical observational\nmanuscript must disclose so the study can be appraised and, in principle, reproduced. It is not an Avalere/agency\nproduct and carries no regulatory mandate; its authority is editorial (journals require or recommend it).\n\n**When to use** — Apply STROCSS when you are *reporting a completed observational surgical study* — a cohort\n(prospective or retrospective), a cross-sectional study, or a case-control study — in which the exposure, comparison,\nor population is defined by a surgical intervention, operative approach, device implantation, or peri-operative\npathway, and you are submitting to a surgical journal. Decision rule for picking the right member of the STROBE\nfamily: use **plain STROBE** for a general (non-surgical) observational study; use **STROCSS** when the study is\n*surgical* cohort/cross-sectional/case-control; use **CARE** for a single surgical *case report* and a case-series\nreporting tool for a *case series* (STROCSS covers analytic designs with a comparison or denominator, not pure case\nreports/series); use **RECORD / RECORD-PE** when the surgical cohort is built from *routinely-collected health data*\n(claims, EHR, registries) and pharmacoepidemiologic exposures, because those extensions add the data-provenance and\ncode-list items STROCSS does not. STROCSS sits at the *manuscript-reporting* stage; it is not a protocol tool and not\na risk-of-bias instrument.\n\n**What it requires** — STROCSS inherits the STROBE backbone (title/abstract; structured background, objectives, and\nhypotheses; design, setting, and dates; eligibility and selection of participants; clearly defined exposures,\noutcomes, predictors, and confounders; data sources and measurement; bias; study size; quantitative handling of\nvariables; statistical methods including subgroup, missing-data, and sensitivity analyses; a participant-flow\naccount; descriptive, outcome, and other analyses; key results; limitations; interpretation; generalizability; and\nfunding) and layers on surgery-specific items: explicit **registration** of the study, a precise **description of the\nintervention/operative technique** sufficient for replication, **surgeon/operator experience and learning-curve**\nconsiderations, and **follow-up** of the operated population. Read through a real-world-data lens, the demanding\nitems are the design-transparency and measurement items: an unambiguous **time-zero/index definition** (date of the\nindex operation), how the **surgical exposure and any comparator** were ascertained, how **outcomes were defined and\nvalidated**, how **confounding** (including confounding by surgical indication and operator/center effects) was\nhandled, and how **attrition/loss to follow-up and missing data** were reported. STROCSS asks you to *state* these\nthings clearly; it does not tell you the analytic method to use — that is what the catalog concepts below supply.\n\n**When NOT to use — limitations and common misapplications** — STROCSS is a reporting checklist, and the most common\nerrors come from treating it as something it is not. (1) **It is not a risk-of-bias instrument and not a quality\nscore.** A fully STROCSS-compliant paper can still be badly confounded; appraise validity with **ROBINS-I** (or\nNewcastle-Ottawa / JBI tools), not with a STROCSS tick-list. (2) **Completing the checklist does not make an\nobservational study causal** — transparent reporting of a biased comparison is still a biased comparison; STROCSS\nreports the design, it does not fix immortal time, selection on the operated, or confounding by indication. (3)\n**Wrong family member** — using generic STROBE where STROCSS is required (losing the surgical intervention/learning-\ncurve items), or using STROCSS for a routinely-collected-data pharmacoepidemiology study where **RECORD-PE** is the\nappropriate STROBE extension (losing the database, code-list, and data-cleaning items). (4) **Wrong design class** —\nforcing STROCSS onto a case report or uncontrolled case series (use CARE / a case-series tool) or onto a randomized\nsurgical trial (use CONSORT, not an observational-study checklist). (5) **Checklist-as-theater** — listing page\nnumbers against 30-odd items while the operative technique, the index date, the comparator, or the loss-to-follow-up\nremain vague defeats the purpose; the value is the substance disclosed, not the completed grid.\n\n**How it maps to this catalog** — STROCSS states *what must be reported*; these catalog concepts *implement* each\nreporting requirement for surgical RWE built on claims/EHR/registry data:\n- **Design transparency & time-zero**: the index-operation date and aligned follow-up are operationalized by\n **time-zero-index-date-alignment-rwe** and, for a trial-grade comparison of surgical strategies,\n **target-trial-emulation**; a clean comparator/new-initiator structure is **active-comparator-new-user**.\n- **Exposure & procedure ascertainment**: identifying and dating the operation in routine data is\n **procedure-identification-and-measurement-in-claims-ehr**; building the analytic cohort and observable time uses\n **continuous-enrollment-observable-time-rwe**.\n- **Outcome/phenotype definition & validation**: the STROCSS outcome items map to\n **diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe** and **outcome-algorithm-construction-rwe**, with validity\n quantified via **claims-outcome-algorithm-ppv-sensitivity-rwe** and **algorithm-validation**.\n- **Confounding control**: STROCSS asks you to *report* confounder handling; implement it with\n **high-dimensional-propensity-score-hdps-rwe**, **propensity-score-methods-psm-iptw**, and design-stage thinking via\n **dags-backdoor-criterion-drug-studies**.\n- **Estimands & intercurrent events**: making the surgical estimand explicit (and handling reoperation, crossover,\n death) is **estimands-ate-att-intercurrent-events-rwe**.\n- **Attrition & missing data**: the participant-flow and follow-up items map to\n **attrition-and-loss-to-follow-up-rwe** and **database-feasibility-attrition-funnel-rwe**.\n- **Sensitivity / quantitative bias analysis**: the limitations/sensitivity items are implemented by\n **e-value-sensitivity-analysis** and **quantitative-bias-analysis-toolkit-rwe**.\n- **Underlying data substrate**: **claims-analysis** for the claims-specific operational caveats.\n\n**Applied note (claims/EHR/registry RWE).** A retrospective cohort comparing two surgical approaches in administrative\nclaims should, to satisfy STROCSS in substance and not just in form, define the index operation by procedure codes\n(CPT/ICD-10-PCS/HCPCS) with a dated time zero, require continuous enrollment across the baseline and follow-up\nwindows so absence of prior events is observed rather than missing, validate the outcome phenotype (PPV/sensitivity)\nrather than asserting it, control confounding by surgical indication and center/operator volume with a propensity\napproach, pre-state the estimand and how reoperation/death are handled as intercurrent events, report the attrition\nfunnel and loss to follow-up explicitly, and probe residual confounding with an E-value or negative-control\nanalysis. STROCSS forces these disclosures; the catalog concepts above tell you how to do each one.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "surgery", "observational", "strobe-extension", "equator", "cohort", "case-control" ], "aliases": [ "STROCSS", "STROCSS 2024", "STROCSS 2021", "Strengthening the Reporting of Cohort, Cross-Sectional and Case-Control Studies in Surgery" ], "applies_to_study_types": [ "cohort_prospective", "cohort_retrospective", "case_control", "cross_sectional" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "doi": "10.1097/JS9.0000000000001268", "url": "https://doi.org/10.1097/JS9.0000000000001268", "citation_text": "Rashid R, Sohrabi C, Kerwan A, et al. The STROCSS 2024 guideline: strengthening the reporting of cohort, cross-sectional, and case-control studies in surgery. International Journal of Surgery. 2024;110(6):3151-3165.", "year": 2024, "authors_short": "Rashid et al.", "notes": "Current canonical STROCSS statement (2024 update) defining the surgery-specific observational reporting items." }, { "role": "explain", "doi": "10.1016/j.ijsu.2021.106165", "url": "https://doi.org/10.1016/j.ijsu.2021.106165", "citation_text": "Mathew G, Agha R; STROCSS Group. STROCSS 2021: Strengthening the reporting of cohort, cross-sectional and case-control studies in surgery. International Journal of Surgery. 2021;96:106165.", "year": 2021, "authors_short": "Mathew & Agha", "notes": "Prior major STROCSS version (2021), widely cited; useful for item history and the Delphi development process." }, { "role": "explain", "doi": "10.1371/journal.pmed.0040296", "url": "https://doi.org/10.1371/journal.pmed.0040296", "citation_text": "von Elm E, Altman DG, Egger M, et al. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. PLoS Medicine. 2007;4(10):e296.", "year": 2007, "authors_short": "von Elm et al.", "notes": "Parent STROBE statement from which STROCSS is derived; STROCSS adds surgery-specific items on top of this backbone." }, { "role": "use", "url": "https://www.equator-network.org/reporting-guidelines/the-strocss-statement-strengthening-the-reporting-of-cohort-studies-in-surgery/", "citation_text": "STROCSS reporting guideline. EQUATOR Network reporting-guidelines library (maintained landing page with the checklist and version history).", "year": 2024, "authors_short": "EQUATOR Network", "notes": "Maintained EQUATOR landing page with the downloadable checklist and links to STROCSS versions." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "cohort-prospective", "notes": "Use to report a prospective surgical cohort study." }, { "relation_type": "applies_to", "target_slug": "cohort-retrospective", "notes": "Use to report a retrospective surgical cohort study built on claims/EHR/registry data." }, { "relation_type": "applies_to", "target_slug": "case-control", "notes": "Use to report a surgical case-control study." }, { "relation_type": "applies_to", "target_slug": "cross-sectional", "notes": "Use to report a surgical cross-sectional study." }, { "relation_type": "is_variant_of", "target_slug": "strobe", "notes": "STROCSS is a surgery-specific extension of STROBE, adding intervention/technique, learning-curve, and operated-patient follow-up items to the STROBE backbone." }, { "relation_type": "see_also", "target_slug": "record-pe", "notes": "For surgical cohorts built from routinely-collected health data with pharmacoepi exposures, RECORD-PE is the appropriate STROBE extension (adds database, code-list, and data-cleaning items STROCSS lacks)." }, { "relation_type": "see_also", "target_slug": "record", "notes": "Reporting extension for studies using routinely-collected health data; relevant when the surgical cohort comes from claims/EHR/registries." }, { "relation_type": "see_also", "target_slug": "robins-i", "notes": "STROCSS reports the design; appraise risk of bias of the non-randomized surgical study with ROBINS-I (STROCSS is not a risk-of-bias tool)." }, { "relation_type": "used_with", "target_slug": "time-zero-index-date-alignment-rwe", "notes": "Implements the STROCSS time-zero requirement by aligning follow-up to the index operation date." }, { "relation_type": "used_with", "target_slug": "procedure-identification-and-measurement-in-claims-ehr", "notes": "Implements surgical exposure ascertainment (identifying and dating the operation in claims/EHR)." }, { "relation_type": "used_with", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Implements the outcome-definition items with a validated phenotype algorithm." }, { "relation_type": "used_with", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "Implements the confounding-control the STROCSS methods/limitations items require authors to report." }, { "relation_type": "used_with", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Makes the surgical estimand explicit and handles reoperation/crossover/death as intercurrent events." }, { "relation_type": "used_with", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Implements the participant-flow and follow-up reporting items for the operated cohort." }, { "relation_type": "used_with", "target_slug": "e-value-sensitivity-analysis", "notes": "Implements the sensitivity/limitations items by quantifying robustness to unmeasured confounding." }, { "relation_type": "see_also", "target_slug": "target-trial-emulation", "notes": "For a trial-grade comparison of surgical strategies in observational data, emulate a target trial to sharpen the design STROCSS asks you to report." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "journal" ] }, { "slug": "strome-id", "name": "STROME-ID (Strengthening the Reporting of Molecular Epidemiology for Infectious Diseases)", "short_definition": "STROBE extension that specifies the additional reporting items needed when an observational infectious-disease study uses molecular/genotyping data (e.g., pathogen typing, sequencing, phylogenetics) to infer transmission, strain attribution, or resistance; maintained within the EQUATOR Network.", "long_description": "**What it is** — **STROME-ID (Strengthening the Reporting of Molecular Epidemiology for Infectious\nDiseases)** is a 42-item reporting checklist, published as an extension of the STROBE statement by\nField, Cohen, Struelens et al. (Lancet Infectious Diseases, 2014). It does not replace STROBE; it\n*adds* items specific to studies that use molecular data — pathogen genotyping, whole-genome or\namplicon sequencing, phylogenetic/phylodynamic inference — to make claims about transmission chains,\nstrain identity, clonality, or antimicrobial-resistance mechanisms. Its purpose is to make the\nmolecular layer of an infectious-disease epidemiology study transparent and reproducible: which\nisolates were sampled and how representatively, which typing/sequencing platform and bioinformatics\npipeline (with versions) produced the genotypes, how genetic distance was thresholded into\n\"linkage,\" and how phylogenetic and sampling uncertainty was propagated into the epidemiological\nconclusions. It is hosted and maintained as a STROBE extension within the **EQUATOR Network**.\n\n**When to use** — Apply STROME-ID when *reporting* an observational study (cohort, case-control,\ncross-sectional, surveillance, or outbreak investigation) in which **molecular characterization of a\npathogen is part of the inference** — for example, reconstructing a transmission network from\nsequenced isolates, attributing cases to a clone or lineage, estimating the genetic relatedness that\ndefines an outbreak, or linking a resistance genotype to a clinical outcome. It is a journal-stage\nreporting checklist (peer-reviewed manuscript) and is relevant to public-health/agency surveillance\nreports (e.g., ECDC/ENCePP-adjacent molecular surveillance, outbreak post-mortems). Decision rule for\nchoosing the right STROBE family member: use **plain STROBE** for an observational ID study with *no*\nmolecular component; use **STROME-ID** the moment genotyping/sequencing data drive a substantive\nclaim; use **RECORD-PE / HARPER / ENCePP** instead when the study is a routinely-collected-data\npharmacoepidemiology study (drug safety/effectiveness in claims or EHR), even if the outcome is an\ninfection — those guidelines, not STROME-ID, govern that design. (Note the easily-confused sibling\n**STROBE-ME**, a separate molecular-epidemiology extension oriented toward non-communicable/biomarker\nexposures; STROME-ID is the infectious-disease-specific one.)\n\n**What it requires** — On top of every STROBE item, STROME-ID enforces reporting domains that are the\nfailure points of molecular ID studies: (1) **Sampling and representativeness of isolates** — what\nfraction of incident cases were genotyped, how isolates were selected, and the risk that the sampled\nisolates are not representative of the transmission process (the molecular analogue of selection\nbias). (2) **Laboratory and bioinformatics transparency** — the typing/sequencing method, platform,\ncoverage/depth, quality-control thresholds, reference genome, variant-calling and assembly pipeline,\nand **software versions** — so the genotypes are reproducible. (3) **Definition of genetic linkage** —\nthe explicit distance metric and threshold (e.g., SNP cutoff) used to call two isolates \"linked,\" and\nits justification, because that threshold *creates* the transmission network the study then analyzes.\n(4) **Phylogenetic and temporal inference** — the model, assumptions, and **uncertainty** (posterior\nsupport, bootstrap, dating intervals) carried through to the epidemiological conclusion. (5)\n**Case/exposure ascertainment and time** — consistent with STROBE, how cases were defined and how\ntime/follow-up was handled, which in transmission studies interacts with sampling completeness.\nSeveral of these items are *causal-claim* safeguards: a SNP threshold and a phylogeny do not by\nthemselves prove who infected whom, and STROME-ID forces the authors to state the uncertainty rather\nthan assert linkage.\n\n**When NOT to use — limitations and common misapplications** — STROME-ID is a *reporting* checklist,\nnot a risk-of-bias instrument and not a quality score; a fully STROME-ID-compliant paper can still\nreport a biased or non-causal analysis — completing the checklist documents the molecular methods, it\ndoes not validate the transmission inference. Concrete failure modes: (1) **Wrong guideline for the\ndesign** — using STROME-ID for an ID study with no molecular data (use plain STROBE), or for a\ndrug-safety/effectiveness study in claims/EHR where an infection is merely the outcome (use\nRECORD-PE/HARPER/ENCePP). (2) **Confusing the siblings** — using STROME-ID where STROBE-ME (the\ngeneral molecular-epidemiology/biomarker extension) is the right tool, or vice versa. (3)\n**Checklist-as-theater on the molecular items** — stating that isolates were \"sequenced and\nanalyzed\" without reporting depth, QC thresholds, the SNP/linkage cutoff, the pipeline, or version\nnumbers is non-compliance dressed as compliance; the whole point of the extension is that those\ndetails are the analysis. (4) **Treating the SNP threshold as ground truth** — reporting a single\ndistance cutoff as if it deterministically defines transmission, without sensitivity to the threshold\nor to incomplete sampling, defeats the inference the checklist exists to discipline. (5) **Mistaking\nthe checklist for an appraisal of the included evidence** — STROME-ID governs *one* primary study's\nreport, not a synthesis (use PRISMA/AMSTAR families) and not a formal bias appraisal (use a\nnon-randomized RoB tool).\n\n**How it maps to this catalog** — This catalog is pharmacoepidemiology- and HEOR-heavy, so the\n*molecular core* of STROME-ID — phylogenetic/phylodynamic inference, sequencing QC and bioinformatics\npipeline versioning, SNP-distance transmission-linkage thresholds, and pathogen-isolate sampling\nrepresentativeness — has **no direct concept implementation here**, and a reviewer should be told that\nplainly rather than be pointed at wrong-domain pharmacoepi concepts. What *does* map are the generic,\ncross-domain reporting disciplines STROME-ID inherits from STROBE and adapts:\n- Observational design scaffolding the checklist reports against: **cohort-prospective**,\n **cohort-retrospective**, and **case-control**.\n- Case/outcome definition and validation (the epidemiological-ascertainment items): **algorithm-validation**,\n **outcome-algorithm-construction-rwe**, **diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe**, and\n (for record-based ID case-finding) **ehr-phenotyping-algorithms-rwe**.\n- Data-fitness and source-representativeness (the isolate-sampling/representativeness analogue):\n **fit-for-purpose-data-assessment-rwe** and **disease-registry** (surveillance/registry substrates\n for ID studies).\n- Attrition/completeness and signal-finding context: **attrition-and-loss-to-follow-up-rwe** and\n **signal-detection** (outbreak/surveillance case ascertainment).\nUse these concepts to satisfy the *generic* reporting items; satisfy the *molecular* items\n(sequencing QC, linkage thresholds, phylogenetic uncertainty, isolate sampling) directly from the\nSTROME-ID statement and the bioinformatics literature, since this catalog does not yet implement them.\n\n**Applied note (surveillance / registry / record-based ID RWE).** When an infectious-disease study is\nbuilt on routinely collected records (a surveillance registry, EHR-based case-finding, or linked\nlaboratory data), report the case-ascertainment algorithm and its validity (PPV/sensitivity) under the\ncatalog's algorithm-validation and phenotype concepts, document how completely cases were captured and\nwhat was lost (attrition/observable-time), and separately — outside this catalog — report the isolate\nsampling fraction, sequencing depth/QC, the genetic-distance threshold defining linkage, and the\nphylogenetic uncertainty, exactly as STROME-ID requires. Conflating \"we identified cases from records\"\nwith \"we sequenced a representative sample of those cases\" is the central misapplication the extension\nwas written to prevent.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "strobe-extension", "molecular-epidemiology", "infectious-disease", "transmission", "equator" ], "aliases": [ "STROME-ID", "Strengthening the Reporting of Molecular Epidemiology for Infectious Diseases", "STROBE extension for molecular epidemiology of infectious diseases" ], "applies_to_study_types": [ "cohort_prospective", "cohort_retrospective", "case_control", "cross_sectional" ], "data_sources": [ "registry", "ehr", "linked" ], "citations": [ { "role": "introduce", "doi": "10.1016/S1473-3099(13)70324-4", "url": "https://doi.org/10.1016/S1473-3099(13)70324-4", "citation_text": "Field N, Cohen T, Struelens MJ, et al. Strengthening the Reporting of Molecular Epidemiology for Infectious Diseases (STROME-ID): an extension of the STROBE statement. The Lancet Infectious Diseases. 2014;14(4):341-352.", "year": 2014, "authors_short": "Field et al.", "notes": "Canonical STROME-ID statement defining the 42-item checklist and the molecular reporting items that extend STROBE for infectious-disease transmission studies." }, { "role": "explain", "doi": "10.1371/journal.pmed.0040296", "url": "https://doi.org/10.1371/journal.pmed.0040296", "citation_text": "von Elm E, Altman DG, Egger M, et al. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. PLoS Medicine. 2007;4(10):e296.", "year": 2007, "authors_short": "von Elm et al.", "notes": "Parent STROBE statement that STROME-ID extends; STROME-ID is read as STROBE plus the molecular items, not as a standalone replacement." }, { "role": "use", "url": "https://www.equator-network.org/reporting-guidelines/strome-id/", "citation_text": "STROME-ID — Strengthening the Reporting of Molecular Epidemiology for Infectious Diseases. EQUATOR Network reporting-guidelines library (maintained checklist and statement links).", "year": 2014, "authors_short": "EQUATOR Network", "notes": "Canonical maintained landing page hosting the STROME-ID checklist and links to the parent STROBE statement and related extensions." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "cohort-prospective", "notes": "Use when reporting a molecular-epidemiology infectious-disease cohort that follows genotyped cases forward." }, { "relation_type": "applies_to", "target_slug": "cohort-retrospective", "notes": "Use when reporting a retrospective molecular ID cohort reconstructing transmission or strain attribution from archived isolates." }, { "relation_type": "applies_to", "target_slug": "case-control", "notes": "Use for case-control molecular ID studies (e.g., linking a resistance genotype to an outcome)." }, { "relation_type": "see_also", "target_slug": "algorithm-validation", "notes": "Generic epidemiological item STROME-ID inherits from STROBE — the case/outcome ascertainment algorithm and its validity should be reported; the molecular linkage definition is separate." }, { "relation_type": "see_also", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Implements record-based case-finding for ID studies built on routinely collected data; satisfies the case-definition reporting items, not the molecular-typing items." }, { "relation_type": "see_also", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "Cross-domain analogue for judging whether the data source (and isolate sampling) is adequate and representative for the molecular-epidemiology question." }, { "relation_type": "see_also", "target_slug": "signal-detection", "notes": "Surveillance/outbreak case-ascertainment context in which molecular ID studies are often embedded." }, { "relation_type": "see_also", "target_slug": "disease-registry", "notes": "Surveillance/registry substrate frequently underlying infectious-disease molecular studies." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "journal", "ema" ] }, { "slug": "target-trial", "name": "Target Trial Emulation Framework", "short_definition": "A causal-inference design framework that forces an observational analysis to explicitly specify the protocol of the hypothetical randomized trial it is meant to emulate (eligibility, treatment strategies, assignment, time zero, outcomes, causal contrast, analysis plan), so that common self-inflicted biases — immortal time, prevalent-user, and time-zero misalignment — are designed out rather than corrected after the fact.", "long_description": "**What it is** — The **Target Trial Emulation (TTE) framework** is a methodological discipline for\ndesigning and analyzing observational (\"real-world\") studies of treatment effects by first writing\nthe protocol of the *hypothetical randomized trial* you would run if you could, and then emulating\neach of its components in the available data. It was articulated by **Miguel Hernán and James\nRobins** (Harvard) in the 2016 *American Journal of Epidemiology* paper \"Using Big Data to Emulate a\nTarget Trial When a Randomized Trial Is Not Available,\" operationalized in the companion 2016\n*Journal of Clinical Epidemiology* paper on preventing immortal time bias, and crystallized into the\nnow-standard **seven protocol components** in the 2022 *JAMA* \"Target Trial Emulation\" guide:\n(1) eligibility criteria, (2) treatment strategies, (3) assignment procedures, (4) start and end of\nfollow-up (time zero), (5) outcomes, (6) causal contrast (estimand), and (7) the analysis plan. TTE\nhas **no formal maintaining organization** in the EQUATOR/Cochrane/ISPOR sense — it is a causal\nframework defined by these canonical papers, not a checklist owned by a society — though it has been\nwoven into FDA and EMA/ENCePP real-world-evidence (RWE) expectations and into reporting/templating\nguidelines (HARPER, StaRT-RWE) that operationalize it.\n\n**When to use** — Apply TTE whenever you are designing, conducting, or appraising a non-interventional\ncomparative-effectiveness or safety study that aims to estimate the **causal effect of a treatment\nstrategy** from routinely collected data (claims, EHR, registries, linked sources), and especially\nwhen that study supports an FDA/EMA submission, an HTA/payer dossier, a peer-reviewed comparative\nmanuscript, a registered protocol, or a post-authorization safety study (PASS). The decision rule\nversus its siblings: use **TTE** as the causal *design* spine that fixes eligibility, treatment\nstrategies, time zero, the estimand, and the analysis; pair it with **HARPER** or **StaRT-RWE** when\nyou need a structured *protocol template* to document those choices; and report the completed study\nwith **STROBE** (or **RECORD/RECORD-PE** for routinely collected health data). TTE is not a substitute\nfor these — it is the layer that makes the protocol causally coherent before the template is filled or\nthe report is written. For a *regulatory PASS* the **ENCePP checklist** governs procedural\ncompleteness; TTE governs whether the design actually identifies a causal effect. For randomized\n*trial* protocols use SPIRIT, not TTE.\n\n**What it requires** — TTE forces explicit pre-specification of seven things, each of which has a\nreal-world-data failure mode it is designed to prevent. (1) **Eligibility** must be assessable using\nonly information available at or before time zero, with a documented data-fitness-for-use assessment\nof whether the source can actually capture those criteria. (2) **Treatment strategies** must be\nwell-defined and sustained or point-interventions (e.g., \"initiate and remain on drug A\"), not vague\n\"exposure\" definitions. (3) **Assignment** must specify which baseline covariates are needed to make\ntreatment exchangeable conditional on measured confounders — the explicit confounding-control plan\n(propensity scores, high-dimensional proxies, or g-methods). (4) **Time zero** must align eligibility,\ntreatment assignment, and start of follow-up at a single index moment so there is no immortal time\nand no adjustment for post-baseline variables on the causal pathway. (5) **Outcomes** must rest on a\nvalidated phenotype/algorithm with reported operating characteristics. (6) **Causal contrast** must\nname the estimand — intention-to-treat-analogue versus per-protocol, and how intercurrent events\n(discontinuation, switching, death) are handled. (7) The **analysis** must match the estimand\n(PS-based methods for the ITT-analogue; clone-censor-weight or inverse-probability weighting for a\nsustained per-protocol estimand) and report attrition, balance diagnostics, positivity, and\nsensitivity/quantitative bias analyses.\n\n**When NOT to use — limitations and common misapplications** — TTE is a *design framework*, not a\nproof of validity. (1) **\"Specifying a target trial\" does not eliminate unmeasured confounding** —\nemulation aligns time and structure, but the conditional-exchangeability assumption can still fail;\na well-specified target trial built on incomplete covariates is still confounded. (2) **Estimand\nbait-and-switch** — the default emulation yields an ITT-analogue (effect of *initiating* a strategy);\na per-protocol (sustained-adherence) estimand requires clone-censor-weight or IPW for informative\ncensoring. A common error is claiming a per-protocol effect while actually running an\ninitiation-only analysis. (3) **Time-zero misalignment is the dominant practical failure** — grace\nperiods, eligibility-time ambiguity, and \"exposure defined over a period\" all reintroduce immortal\ntime the framework exists to remove. (4) **Not for descriptive, single-arm, or hypothesis-generating\nwork** — TTE presupposes a comparative causal question; forcing it onto disease-burden or\nnatural-history description is misapplication. (5) **Framework-as-theater** — listing the seven\ncomponents in a protocol table without making the causal contrast actually estimable in the data\n(no defensible comparator, no positivity, an unmeasurable eligibility criterion) is box-ticking. (6)\nTTE is a design layer, **not a reporting checklist and not a risk-of-bias instrument**: completing a\nTTE protocol does not discharge STROBE/RECORD-PE reporting or a formal ROBINS-I appraisal.\n\n**How it maps to this catalog** — In this repo, TTE's seven components are implemented by specific\nconcepts a reviewer can hold the protocol against:\n- The framework itself and its worked emulation: **target-trial-emulation**.\n- Treatment-strategy + assignment + time-zero (components 2–4): **active-comparator-new-user** is the\n usual analytic core — new-user + active comparator + time-zero alignment maps directly onto trial\n eligibility and assignment.\n- Confounding control (component 3): **high-dimensional-propensity-score-hdps-rwe** for proxy\n adjustment when key confounders are unmeasured.\n- Causal contrast / estimand and intercurrent events (component 6): **estimands-ate-att-intercurrent-events-rwe**.\n- Outcome and eligibility ascertainment (components 1, 5): **diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe**\n and **claims-analysis** for code-list construction and validation.\n- Follow-up integrity and attrition (component 4, analysis): **attrition-and-loss-to-follow-up-rwe**.\n\n**Applied note (claims/EHR/registry RWE).** In a claims emulation, time zero is the first qualifying\ndispensing (NDC + fill date), eligibility is assessed only in the pre-index enrollment window, and\ncontinuous medical+pharmacy enrollment across that window is required so \"no prior use\" is observed\nrather than missing — Medicare Advantage-only person-time, which lacks fee-for-service claims, must be\nexcluded or it masquerades as a clean washout. In EHR, initiation is the order/administration (prefer\nlinked dispensing to confirm the patient started), and visit-driven capture makes loss to follow-up\npotentially informative. Registries strengthen indication, severity, and adjudicated outcomes but need\nclaims/death-index linkage for complete exposure and censoring. Across all sources, the discipline is\nidentical: fix one time zero, measure covariates only before it, and apply the same outcome and\ncensoring rules to every arm.", "primary_category": "Guideline", "tags": [ "guideline", "framework", "causal-inference", "target-trial", "rwe", "pharmacoepidemiology", "estimand", "time-zero" ], "aliases": [ "Target Trial Emulation", "TTE", "Target Trial Framework", "Hernán-Robins target trial" ], "applies_to_study_types": [ "target_trial_emulation", "cer_observational", "new_user", "active_comparator_new_user", "cohort_retrospective" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "doi": "10.1093/aje/kwv254", "url": "https://doi.org/10.1093/aje/kwv254", "citation_text": "Hernán MA, Robins JM. Using big data to emulate a target trial when a randomized trial is not available. American Journal of Epidemiology. 2016;183(8):758-764.", "year": 2016, "authors_short": "Hernán & Robins", "notes": "Original statement of the target-trial framework — specify the hypothetical trial protocol, then emulate each component in observational data to avoid self-inflicted design biases." }, { "role": "explain", "doi": "10.1001/jama.2022.21383", "url": "https://doi.org/10.1001/jama.2022.21383", "citation_text": "Hernán MA, Wang W, Leaf DE. Target trial emulation: a framework for causal inference from observational data. JAMA. 2022;328(24):2446-2447.", "year": 2022, "authors_short": "Hernán et al.", "notes": "Modern crystallization of the seven protocol components (eligibility, treatment strategies, assignment, time zero, outcomes, causal contrast, analysis) now used as the standard articulation." }, { "role": "explain", "doi": "10.1016/j.jclinepi.2016.04.014", "url": "https://doi.org/10.1016/j.jclinepi.2016.04.014", "citation_text": "Hernán MA, Sauer BC, Hernández-Díaz S, Platt R, Shrier I. Specifying a target trial prevents immortal time bias and other self-inflicted injuries in observational analyses. Journal of Clinical Epidemiology. 2016;79:70-75.", "year": 2016, "authors_short": "Hernán et al.", "notes": "Operationalizes the time-zero discipline — shows how aligning eligibility, assignment, and start of follow-up at a single index date prevents immortal time bias." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "target-trial-emulation", "notes": "TTE is the design framework for studies that explicitly emulate a hypothetical randomized trial from observational data." }, { "relation_type": "applies_to", "target_slug": "cer-observational", "notes": "Provides the causal-design spine for observational comparative-effectiveness studies." }, { "relation_type": "applies_to", "target_slug": "active-comparator-new-user", "notes": "ACNU cohorts are the usual analytic core of a two-drug target-trial emulation." }, { "relation_type": "applies_to", "target_slug": "new-user-design", "notes": "New-user (incident-user) restriction supplies the time-zero alignment the framework requires." }, { "relation_type": "applies_to", "target_slug": "cohort-retrospective", "notes": "Retrospective cohorts built from routinely collected data are the typical substrate for emulation." }, { "relation_type": "used_with", "target_slug": "target-trial-emulation", "notes": "The implementing concept with the worked seven-component emulation, code, and variants." }, { "relation_type": "used_with", "target_slug": "active-comparator-new-user", "notes": "Implements treatment strategies, assignment, and time-zero alignment (components 2-4)." }, { "relation_type": "used_with", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Implements the causal contrast / estimand and intercurrent-event handling (component 6)." }, { "relation_type": "used_with", "target_slug": "high-dimensional-propensity-score-hdps-rwe", "notes": "Implements confounding control via proxy adjustment when key confounders are unmeasured (component 3)." }, { "relation_type": "see_also", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Implements eligibility and outcome ascertainment with validated phenotype algorithms (components 1, 5)." }, { "relation_type": "see_also", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Implements follow-up integrity, attrition reporting, and handling of informative censoring (component 4)." }, { "relation_type": "see_also", "target_slug": "claims-analysis", "notes": "Code-list construction and operational definitions for claims-based emulation." }, { "relation_type": "see_also", "target_slug": "picots-framework-rwe", "notes": "PICOTS structures the question that the target-trial protocol then makes causally estimable." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "fda", "ema", "hta", "journal" ] }, { "slug": "treatment-failure-endpoint-checklist-rwe", "name": "Treatment Failure Endpoint Checklist", "short_definition": "A checklist for defining treatment failure, non-response, loss of response, and proxy treatment-pattern failure endpoints in RWE.", "long_description": "**What it is** - This guideline is the checklist layer for treatment failure, non-response,\nloss of response, and proxy strategy-failure endpoints in RWE. The companion concept explains\nthe endpoint family; this guideline defines what must be specified when the endpoint is used in a\nprotocol, SAP, manuscript, HTA appendix, or regulatory package. The core rule is separation:\nclinical non-response, treatment-pattern proxies, intolerance, access barriers, non-adherence,\ndeath, and loss of observability are different mechanisms and should not be collapsed silently.\n\n**When to use** - Use it when a study endpoint includes discontinuation, switch, add-on,\naugmentation, rescue therapy, dose escalation, hospitalization, biomarker/imaging progression,\nline-of-therapy transition, or clinician-documented non-response as evidence that an initial\nstrategy failed. It is especially important in claims-only studies where the endpoint is a proxy,\nand in linked EHR/registry studies where clinical assessments can validate or refute the proxy.\nUse it at design time because adequate-trial windows, assessment dates, and intercurrent-event\nstrategies determine who is eligible to fail.\n\n**What it requires / checklist domains** - Define the estimand first: whether the endpoint is\nclinical non-response, treatment-pattern failure, loss of response, composite strategy failure,\nor an on-treatment/per-protocol failure event. State the adequate trial window, minimum exposure\nor persistence requirement, first eligible assessment date, and data sources for each component.\nList every component that can trigger failure and retain the triggering component in the analytic\ndataset. Specify how discontinuation, switching, add-on, rescue therapy, dose escalation,\nhospitalization, death, and loss to follow-up are handled as intercurrent events. Report\ncomponent-specific counts and sensitivity analyses that remove weak or ambiguous components.\n\n**When NOT to use - limitations and common misapplications** - Do not label a claims-only\ntreatment change as clinical non-response unless it has been validated against chart, lab,\nimaging, registry, or clinician-assessment evidence. Do not mix non-response with intolerance,\nformulary access, affordability, patient preference, or administrative censoring unless those are\nexplicitly part of the estimand. Do not allow people to fail before an adequate trial could have\noccurred. Do not use next-line therapy as a universal failure definition in therapeutic areas\nwhere planned sequencing or finite therapy is standard. A composite treatment-failure endpoint is\nonly interpretable if the component distribution is visible.\n\n**How it maps to this catalog** - This guideline cross-references\n`treatment-failure-non-response-rwe` for the endpoint concept, `persistence-time-to-discontinuation`\nand `switch-add-on-augmentation-rwe` for treatment-pattern proxies,\n`estimands-ate-att-intercurrent-events-rwe` for intercurrent-event handling,\n`target-trial-emulation` for time-zero and strategy definition, `ecog-performance-status-score-rwe`\nand `real-world-progression-rwpfs-rwe` for oncology response context, and\n`attrition-and-loss-to-follow-up-rwe` for observability and censoring. Use this checklist only\nfor the reporting/implementation gate; the concept supplies the operational content.", "primary_category": "Guideline", "tags": [ "guideline", "checklist", "treatment-failure", "non-response", "endpoints", "intercurrent-events" ], "aliases": [ "treatment failure checklist", "non-response checklist", "proxy treatment failure checklist" ], "applies_to_study_types": [ "claims_analysis", "ehr_study", "registry_linkage", "comparative_effectiveness", "target_trial_emulation" ], "data_sources": [ "claims", "ehr", "registry", "linked" ], "citations": [ { "role": "introduce", "doi": null, "url": "https://pubmed.ncbi.nlm.nih.gov/16514590/", "citation_text": "Andrade SE, Kahler KH, Frech F, Chan KA. Methods for evaluation of medication adherence and persistence using automated databases. Pharmacoepidemiology and Drug Safety. 2006;15(8):565-574.", "year": 2006, "authors_short": "Andrade et al.", "notes": "Automated database methods for persistence and treatment patterns." }, { "role": "explain", "url": "https://database.ich.org/sites/default/files/E9-R1_Step4_Guideline_2019_1203.pdf", "citation_text": "ICH E9(R1). Addendum on Estimands and Sensitivity Analysis in Clinical Trials to the Guideline on Statistical Principles for Clinical Trials. 2019.", "year": 2019, "authors_short": "ICH", "notes": "Intercurrent-event strategy source." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "treatment-failure-non-response-rwe", "notes": "Checklist for the treatment failure / non-response concept." } ], "index_definitions": [], "checklist_items": [ "State whether the endpoint is clinical non-response, proxy treatment failure, strategy failure, loss of response, or a composite.", "Define the adequate-trial window, minimum exposure or persistence rule, and first eligible assessment date.", "List every component that can trigger failure and retain the triggering component in the analytic dataset.", "Separate lack of efficacy from intolerance, non-adherence, access/formulary barriers, death, and loss of observability where the data allow.", "Specify how rescue therapy, discontinuation, switch, add-on, dose escalation, hospitalization, and next line of therapy are handled as intercurrent events.", "Report component-specific counts and sensitivity analyses that remove weak or ambiguous components.", "Do not label claims-only treatment changes as clinical non-response unless validated against chart, lab, imaging, or registry evidence." ], "regulatory_relevance": [ "fda", "hta" ] }, { "slug": "trend", "name": "TREND (Transparent Reporting of Evaluations with Nonrandomized Designs)", "short_definition": "A 22-item reporting checklist for non-randomized evaluations of behavioral and public-health interventions, hosted in the EQUATOR Network; it is the nonrandomized-design analogue of CONSORT and is meant for primary intervention-evaluation reports, not drug/device pharmacoepidemiology.", "long_description": "**What it is** — **TREND (Transparent Reporting of Evaluations with Nonrandomized Designs)** is a 22-item\nreporting checklist developed by the Centers for Disease Control and Prevention (CDC) HIV/AIDS Prevention\nResearch Synthesis (PRS) team and published as the TREND Statement (Des Jarlais, Lyles, Crepaz et al., *American\nJournal of Public Health*, 2004). It is maintained and indexed within the **EQUATOR Network** reporting-guideline\nlibrary. TREND was created to fill a specific gap: CONSORT governs the reporting of randomized trials, but a large\nshare of behavioral and public-health intervention evaluations cannot randomize (ethical, logistical, or\ncommunity-level constraints) and are evaluated with non-randomized designs. TREND is the **nonrandomized-design\ncompanion to CONSORT** for *primary intervention-evaluation reports*. Its purpose is transparency of how the\nintervention, comparison condition, assignment mechanism, and analysis were actually carried out, so readers can\njudge the threats to internal validity that randomization would otherwise have addressed.\n\n**When to use** — Apply TREND when you are reporting a **primary evaluation of a behavioral, social, educational,\nor public-health intervention** that used a non-randomized design — quasi-experimental, pre-post,\ncontrolled-before-after, interrupted time series, or otherwise non-randomly assigned groups. Its native habitat is\nthe peer-reviewed public-health/behavioral-science literature (HIV/STI prevention, harm reduction, vaccination and\nscreening uptake, school- or community-based programs, health-promotion campaigns). Decision rule for picking the\nright checklist: if the intervention evaluation **was randomized**, use **CONSORT** (or its extensions:\ncluster, pragmatic, PRO); if it is a **non-randomized intervention evaluation**, use **TREND**; if you are\nreporting an **observational etiologic/comparative study with no investigator-assigned intervention** (a\ndrug-vs-drug claims cohort, a case-control study, a registry analysis), use **STROBE** and, when the data are\nroutinely-collected health data, **RECORD / RECORD-PE**; for a pharmacoepidemiology *protocol* use **HARPER** or\nthe **ENCePP checklist**; for an intervention *protocol*, use **SPIRIT**. TREND governs the *report*, not the\nprotocol, and the *intervention evaluation*, not the etiologic observational study.\n\n**What it requires** — TREND's 22 items mirror the CONSORT skeleton but add domains that matter precisely because\nthere was no randomization. They compel reporting of: the title/abstract structured to flag the non-randomized\ndesign; the **theory or behavioral model** underpinning the intervention; eligibility and the **setting/location**\nof recruitment; a **detailed description of the intervention and the comparison condition** (content, delivery,\nprovider, dose/intensity, fidelity), which is the heart of TREND and the item most often done badly; explicitly\nstated **objectives, hypotheses, and outcomes** with how and when each was measured; **sample size** justification;\nthe **unit of assignment and the method by which groups were formed** (the non-random assignment mechanism — the\nitem that replaces CONSORT's randomization items and is where confounding/selection threats live); blinding where\nfeasible; the **analytic methods, including methods used to control for confounding** introduced by non-random\nassignment; **participant flow, recruitment, and losses/exclusions at each stage** (attrition); baseline group\ncomparability; numbers analyzed and the basis (e.g., intention-to-treat-like vs as-treated); estimated effects with\nprecision; ancillary analyses; and a discussion that interprets results **in light of the non-randomized design and\nits specific biases**. The substantive emphasis is therefore: intervention/comparator description and fidelity,\nthe assignment mechanism and consequent confounding control, time-zero/intervention-start alignment, attrition\nacross the participant-flow stages, and an honest accounting of internal-validity threats.\n\n**When NOT to use — limitations and common misapplications** — (1) **TREND is a reporting checklist, not a\nrisk-of-bias instrument and not a quality score.** Completing all 22 items does not certify that the evaluation is\ninternally valid or that the estimate is causal; appraisal of a non-randomized study is done with ROBINS-I, and a\nfully TREND-compliant paper can still report a hopelessly confounded comparison. (2) **Wrong checklist for\ndrug/device pharmacoepidemiology** — the single most common misapplication is reaching for TREND (or its scaffold\nframing as a generic \"non-randomized RWE\" guideline) when reporting a claims/EHR comparative-effectiveness or\nsafety cohort. Those are **observational etiologic studies of routinely-collected data**, not investigator-assigned\nbehavioral-intervention evaluations; they require **STROBE + RECORD / RECORD-PE**, with HARPER/ENCePP at the\nprotocol stage — not TREND. (3) **Wrong sibling for a randomized evaluation** — if the behavioral intervention was\nin fact randomized (including cluster- or stepped-wedge designs), use CONSORT/CONSORT-Cluster, not TREND. (4)\n**Checklist-as-theater** — ticking the intervention-description item while omitting dose, fidelity, or the actual\nassignment mechanism defeats the entire purpose, which is to expose exactly the design features randomization would\nhave neutralized. (5) **It does not replace the analysis** — TREND requires that confounding control be *reported*;\nit does not tell you which method to use, and reporting transparency is necessary but not sufficient for a valid\ncausal claim.\n\n**How it maps to this catalog** — TREND is an intervention-evaluation reporting checklist, so only a subset of the\ncatalog's pharmacoepidemiology concepts genuinely implement its items; the honest mapping is narrow on purpose:\n- **Intervention/comparator and assignment, attrition, time-zero (items that DO apply):**\n **estimands-ate-att-intercurrent-events-rwe** sharpens the objective/outcome and intention-to-treat-vs-as-treated\n reporting TREND demands; **time-zero-index-date-alignment-rwe** operationalizes alignment of follow-up at\n intervention start; **attrition-and-loss-to-follow-up-rwe** implements the stage-by-stage participant-flow and\n losses items; **generalizability-transportability-external-validity-rwe** supports the discussion of how far the\n setting-bound estimate transports.\n- **Confounding from non-random assignment:** when the evaluation has a comparison group formed non-randomly,\n **difference-in-differences-staggered-adoption-rwe** is the design most aligned with TREND's quasi-experimental\n territory (pre-post with a comparison series), and **target-trial-emulation** is the explicit framework for\n making the non-random assignment defensible.\n- **What does NOT map (deliberately):** **high-dimensional-propensity-score-hdps-rwe**, **active-comparator-new-user**,\n **diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe**, and **claims-analysis** are STROBE/RECORD-PE\n territory. If you find yourself needing those concepts, you are reporting a database pharmacoepidemiology study\n and TREND is the wrong checklist — route to **strobe**, **record**, **record-pe**, or **harper**.\n\n**Applied note (claims/EHR/registry RWE).** Behavioral and public-health intervention evaluations increasingly read\noutcomes from routinely-collected data — e.g., a community vaccination-uptake program measured against HEDIS\nclaims, or a screening intervention evaluated through EHR-captured test orders and downstream ED visits. In that\nhybrid case TREND still governs the **intervention/comparator description, the non-random assignment mechanism, and\nparticipant flow**, but it does *not* cover the database-specific items those data require: the **diagnosis/outcome\nalgorithm and its validity** (PPV/sensitivity), the **data-source/linkage and completeness** reporting, and the\n**time-window definitions**. Use TREND together with **RECORD / RECORD-PE** for those items; do not assume TREND's\n22 items suffice for an analysis whose outcomes are algorithm-defined in claims or EHR.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "nonrandomized", "behavioral-intervention", "public-health", "quasi-experimental", "equator", "trend" ], "aliases": [ "TREND", "TREND Statement", "Transparent Reporting of Evaluations with Nonrandomized Designs" ], "applies_to_study_types": [ "cer_observational", "cohort_prospective", "cohort_retrospective" ], "data_sources": [ "ehr", "claims", "registry" ], "citations": [ { "role": "introduce", "doi": "10.2105/AJPH.94.3.361", "url": "https://doi.org/10.2105/AJPH.94.3.361", "citation_text": "Des Jarlais DC, Lyles C, Crepaz N; TREND Group. Improving the reporting quality of nonrandomized evaluations of behavioral and public health interventions: the TREND statement. American Journal of Public Health. 2004;94(3):361-366.", "year": 2004, "authors_short": "Des Jarlais et al.", "notes": "Canonical TREND statement; introduces the 22-item checklist as the nonrandomized-design analogue of CONSORT." }, { "role": "explain", "doi": "10.1136/bmj.c332", "url": "https://doi.org/10.1136/bmj.c332", "citation_text": "Schulz KF, Altman DG, Moher D; CONSORT Group. CONSORT 2010 statement: updated guidelines for reporting parallel group randomised trials. BMJ. 2010;340:c332.", "year": 2010, "authors_short": "Schulz et al.", "notes": "The randomized-trial sibling whose item structure TREND mirrors; clarifies which items TREND adapts because assignment is non-random rather than randomized." }, { "role": "use", "url": "https://www.equator-network.org/reporting-guidelines/improving-the-reporting-quality-of-nonrandomized-evaluations-of-behavioral-and-public-health-interventions-the-trend-statement/", "citation_text": "TREND Statement. EQUATOR Network reporting-guidelines library (maintained landing page with the 22-item checklist and links to related guidelines).", "year": 2004, "authors_short": "EQUATOR Network", "notes": "Maintained EQUATOR landing page with the checklist in usable form and cross-links to CONSORT/STROBE." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "cer-observational", "notes": "Use when reporting a non-randomized comparative evaluation of a behavioral/public-health intervention." }, { "relation_type": "applies_to", "target_slug": "cohort-prospective", "notes": "Applicable to prospective non-randomized intervention-evaluation cohorts." }, { "relation_type": "applies_to", "target_slug": "cohort-retrospective", "notes": "Applicable to retrospective non-randomized intervention evaluations." }, { "relation_type": "alternative_to", "target_slug": "consort", "notes": "CONSORT reports randomized trials; TREND is its analogue when the intervention evaluation is non-randomized." }, { "relation_type": "see_also", "target_slug": "strobe", "notes": "For observational etiologic/comparative studies with no investigator-assigned intervention, use STROBE, not TREND." }, { "relation_type": "see_also", "target_slug": "record-pe", "notes": "Drug/device pharmacoepidemiology in routinely-collected data is RECORD-PE/STROBE territory, not TREND." }, { "relation_type": "used_with", "target_slug": "difference-in-differences-staggered-adoption-rwe", "notes": "A common analytic design for the quasi-experimental, comparison-series evaluations TREND reports." }, { "relation_type": "used_with", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Implements TREND's stage-by-stage participant-flow, recruitment, and losses items." }, { "relation_type": "used_with", "target_slug": "estimands-ate-att-intercurrent-events-rwe", "notes": "Sharpens the objective/outcome and ITT-vs-as-treated reporting TREND requires." }, { "relation_type": "see_also", "target_slug": "time-zero-index-date-alignment-rwe", "notes": "Aligns follow-up at intervention start, addressing the time-zero ambiguity non-random assignment introduces." }, { "relation_type": "see_also", "target_slug": "target-trial-emulation", "notes": "Framework for making the non-random assignment of a TREND-reported evaluation defensible." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "journal", "hta" ] }, { "slug": "tripod", "name": "TRIPOD (Transparent Reporting of a multivariable prediction model for Individual Prognosis Or Diagnosis)", "short_definition": "EQUATOR-network reporting guideline for studies that develop, validate, or update a multivariable clinical prediction (prognostic or diagnostic) model. The 2015 statement is a 22-item checklist; TRIPOD+AI (2024) extends it to regression and machine-learning models. It governs how a prediction-model study is reported, not whether the model is unbiased or clinically useful.", "long_description": "**What it is** — **TRIPOD (Transparent Reporting of a multivariable prediction model for Individual\nPrognosis Or Diagnosis)** is a reporting guideline for studies that *develop*, *validate*\n(internally or externally), or *update* a multivariable prediction model intended to estimate an\nindividual's probability of a current condition (a **diagnostic** model) or a future event (a\n**prognostic** model). The original 2015 statement (Collins, Reitsma, Altman, Moons) is a 22-item\nchecklist with a companion item-by-item Explanation & Elaboration paper (Moons et al. 2015);\n**TRIPOD+AI (2024)** is the current, EQUATOR-listed update that harmonises reporting across\nregression and machine-learning models and strengthens items on data sources, sample size,\nfairness, and open code/data. **TRIPOD-Cluster (2023)** is a sibling extension for models\ndeveloped or validated in clustered/multi-database data (e.g., individual-participant-data\nmeta-analysis or federated multi-site analyses). TRIPOD is maintained within the **EQUATOR\nNetwork** alongside its sister appraisal tool **PROBAST** (risk-of-bias) — TRIPOD is the reporting\nchecklist; PROBAST is the risk-of-bias instrument. Its purpose is to make a prediction model\nreproducible and independently assessable: a reader should be able to see exactly which population\nand data produced the model, which predictors entered it, how it was built and tuned, and how well\nit discriminated and calibrated in development and in validation.\n\n**When to use** — Apply TRIPOD whenever the study's *deliverable is a prediction model* — a risk\nscore, a prognostic index, a diagnostic probability rule, or an ML classifier producing\nindividual-level risk — regardless of data source (claims, EHR, registry, linked, or\nprospectively collected). It is the correct checklist for a journal manuscript reporting model\ndevelopment and/or validation, for the model-development component of an HTA/payer submission (e.g.,\na risk-stratification or enrichment tool), and for FDA/EMA packages where a model supports a\ndevice, an enrichment strategy, or a prognostic claim. Decision rules for choosing the right family\nmember: use **TRIPOD+AI (2024)** for any model built or validated with machine-learning methods, or\nwhenever you want the current, most complete item set (it supersedes the 2015 list for new work);\nuse **TRIPOD-Cluster** when development or validation uses clustered/multi-database data;\nuse **TRIPOD** (not STROBE) when the objective is *prediction* even if the data come from an\nobservational cohort. Conversely, if the objective is an *exposure–outcome causal effect*, TRIPOD\nis the wrong guideline — use STROBE/RECORD-PE/HARPER. If the study evaluates a single diagnostic\n*test's* accuracy rather than a multivariable model, use **STARD**.\n\n**What it requires** — TRIPOD's items force documentation of the elements that make a model\ntrustworthy and reproducible: a clear statement of the **prediction objective** (diagnostic vs\nprognostic; development, validation, or update) and the intended use and target population;\n**source of data and study design** (cohort, case-control, registry, routinely collected data) with\nparticipant eligibility, setting, and dates; rigorous specification of the **outcome** (definition,\nhow and when assessed, blinding to predictors) and of all **candidate predictors** (definitions,\ntiming of measurement, and — critically for real-world data — that predictors are measured at or\nbefore the prediction time point, never using future information); **sample size / events-per-variable**\njustification; explicit handling of **missing data** (the imputation model, not a silent\ncomplete-case default); the full **model-building procedure** (predictor selection, functional\nforms, interactions, penalisation/regularisation or ML hyperparameter tuning, and how\noverfitting/optimism was addressed); **internal validation** (bootstrap/cross-validation, optimism\ncorrection) and any **external/temporal/geographic validation**; and **performance reporting** that\nincludes both **discrimination** (e.g., C-statistic/AUC) and **calibration** (calibration plot,\ncalibration-in-the-large and slope) — not discrimination alone — plus, where relevant, clinical\nutility (decision-curve/net-benefit). TRIPOD+AI adds explicit items on data provenance and\nrepresentativeness, fairness/subgroup performance, and availability of code and the full model\n(coefficients or the deployable algorithm). For models built on **claims/EHR/registry** data these\ngeneric items carry teeth: the source must document **fitness-for-purpose** of the data, the\n**phenotype/algorithm** defining outcome and predictor variables (with validation metrics such as\nPPV/sensitivity), the alignment of the **prediction time point** so no post-baseline information\nleaks into predictors, and how attrition and loss to follow-up were handled in a longitudinal\nprognostic model.\n\n**When NOT to use — limitations and common misapplications** — TRIPOD is a *reporting* checklist; it\nis **not a risk-of-bias instrument, not a quality score, and not a guarantee that a model is valid,\nfair, or clinically useful**. Concrete failure modes a reviewer will flag: (1) **Wrong objective** —\nusing TRIPOD for an exposure–outcome *causal/comparative-effectiveness* study; that work belongs to\nSTROBE/RECORD-PE/HARPER, and conversely using STROBE for a prediction-model paper omits the\ndevelopment, validation, and calibration items TRIPOD exists to enforce. (2) **Mistaking it for an\nappraisal tool** — TRIPOD says nothing about whether a model is at low risk of bias; that judgement\nis made with **PROBAST**, and the two are routinely confused. (3) **Stale version** — using the 2015\nlist for a machine-learning model published after 2024 when **TRIPOD+AI** is the expected standard;\nor using single-study TRIPOD where **TRIPOD-Cluster** fits the clustered/multi-database design.\n(4) **Discrimination-only reporting** — quoting an AUC with no calibration plot or slope is a\nclassic TRIPOD violation; a model can discriminate well yet be badly miscalibrated and unsafe to\ndeploy. (5) **Optimism unacknowledged** — reporting apparent (development-set) performance as if it\nwere validated, with no internal validation or optimism correction. (6) **Wrong tool entirely** —\napplying TRIPOD to a single diagnostic test's accuracy (use **STARD**) rather than to a\nmultivariable model. (7) **Checklist-as-theater** — ticking 22 (or the TRIPOD+AI) items while\nleaving the predictor definitions, the model equation, the imputation approach, or the calibration\nevidence vague; completing the checklist does not make the model reproducible, and it certainly does\nnot make a model developed in one database transportable to another.\n\n**How it maps to this catalog** — In this repo, TRIPOD's requirements are implemented by the\nprediction/validation and data-fitness family of concepts (not the causal-inference concepts a\ncomparative-effectiveness study would use):\n- The core development/validation discipline: **prediction-model-validation-recalibration-rwe**\n (internal/external validation, calibration, recalibration) and **predictive-and-causal-ml-models-rwe**\n (the ML modelling and tuning that TRIPOD+AI items govern).\n- Outcome/predictor definition as a measurement-validity problem:\n **diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe**,\n **claims-outcome-algorithm-ppv-sensitivity-rwe**, **ehr-phenotyping-algorithms-rwe**, and\n **algorithm-validation** implement the phenotype-definition-and-validation items.\n- Data fitness and the prediction-time-point spine: **fit-for-purpose-data-assessment-rwe** and\n **time-zero-index-date-alignment-rwe** (so predictors are measured before the prediction point and\n no future information leaks in).\n- External validation / transportability across databases:\n **generalizability-transportability-external-validity-rwe** implements TRIPOD's external-validation\n and (with TRIPOD+AI) fairness/representativeness items.\n- Missing data and attrition items: **missing-data-pattern-table-rwe**,\n **multiple-imputation-longitudinal-rwe**, and **attrition-and-loss-to-follow-up-rwe**.\n- Sample-size justification and reporting visuals: **sample-size-power-precision-rwe** and\n **visualizations-pharmacoepidemiology-rwe** (calibration plots, ROC/AUC, decision curves).\n- The structured-question and pre-specification habits: **picots-framework-rwe** and\n **study-protocol-or-sap-elements**.\n\n**Applied note (claims/EHR/registry RWE).** A prognostic model built in administrative claims —\nsay, a 12-month hospitalisation-risk score — should report, per TRIPOD: the database and its\nfitness-for-purpose; continuous-enrollment and observability windows; the prediction time point and\nproof that every predictor is measured at or before it; the phenotype/algorithm (with PPV) defining\nthe outcome and key predictors; the events-per-variable and missing-data/imputation strategy; the\nfull model (coefficients or deployable algorithm); internal-validation optimism correction; and\nexternal or temporal validation in a *different* database with both discrimination **and** a\ncalibration plot. Reporting only an in-sample C-statistic, omitting the model equation, or skipping\nexternal calibration are precisely the gaps TRIPOD (and TRIPOD+AI) exist to close.", "primary_category": "Guideline", "tags": [ "guideline", "reporting", "prediction-model", "prognostic-model", "diagnostic-model", "validation", "calibration", "machine-learning", "equator" ], "aliases": [ "TRIPOD", "TRIPOD statement", "TRIPOD 2015", "TRIPOD+AI", "TRIPOD-AI", "TRIPOD-Cluster", "Transparent Reporting of a multivariable prediction model for Individual Prognosis Or Diagnosis" ], "applies_to_study_types": [ "algorithm_validation", "cohort_prospective", "cohort_retrospective", "diagnostic_accuracy" ], "data_sources": [ "claims", "ehr", "registry", "linked", "primary" ], "citations": [ { "role": "introduce", "doi": "10.1136/bmj.g7594", "url": "https://doi.org/10.1136/bmj.g7594", "citation_text": "Collins GS, Reitsma JB, Altman DG, Moons KGM. Transparent reporting of a multivariable prediction model for individual prognosis or diagnosis (TRIPOD): the TRIPOD statement. BMJ. 2015;350:g7594.", "year": 2015, "authors_short": "Collins et al.", "notes": "Canonical TRIPOD 2015 statement defining the 22-item reporting checklist for prediction-model development and validation studies (published concurrently across several journals)." }, { "role": "explain", "doi": "10.7326/M14-0698", "url": "https://doi.org/10.7326/M14-0698", "citation_text": "Moons KGM, Altman DG, Reitsma JB, et al. Transparent Reporting of a multivariable prediction model for Individual Prognosis Or Diagnosis (TRIPOD): explanation and elaboration. Annals of Internal Medicine. 2015;162(1):W1-W73.", "year": 2015, "authors_short": "Moons et al.", "notes": "Item-by-item explanation and elaboration with worked examples of good and poor prediction-model reporting; the authoritative companion to the checklist." }, { "role": "explain", "doi": "10.1136/bmj-2023-078378", "url": "https://doi.org/10.1136/bmj-2023-078378", "citation_text": "Collins GS, Moons KGM, Dhiman P, et al. TRIPOD+AI statement: updated guidance for reporting clinical prediction models that use regression or machine learning methods. BMJ. 2024;385:e078378.", "year": 2024, "authors_short": "Collins et al.", "notes": "Current EQUATOR-listed update harmonising reporting across regression and machine-learning models; expected standard for new ML prediction-model work and supersedes the 2015 item list." }, { "role": "use", "doi": "10.1136/bmj-2022-071018", "url": "https://doi.org/10.1136/bmj-2022-071018", "citation_text": "Debray TPA, Collins GS, Riley RD, et al. Transparent reporting of multivariable prediction models developed or validated using clustered data: TRIPOD-Cluster checklist. BMJ. 2023;380:e071018.", "year": 2023, "authors_short": "Debray et al.", "notes": "Sibling extension for models developed or validated in clustered/multi-database data (e.g., IPD meta-analysis, federated multi-site analyses)." }, { "role": "use", "url": "https://www.equator-network.org/reporting-guidelines/tripod-statement/", "citation_text": "TRIPOD statement and extensions. EQUATOR Network reporting-guidelines library (maintained checklists, TRIPOD+AI and TRIPOD-Cluster, downloadable forms and templates).", "year": 2024, "authors_short": "EQUATOR Network", "notes": "Canonical maintained landing page with the checklist in usable formats and links to the TRIPOD+AI and TRIPOD-Cluster extensions and the sister PROBAST risk-of-bias tool." } ], "relations": [ { "relation_type": "applies_to", "target_slug": "algorithm-validation", "notes": "TRIPOD governs reporting when the deliverable is a developed and/or validated multivariable prediction/classification algorithm." }, { "relation_type": "applies_to", "target_slug": "cohort-prospective", "notes": "Use TRIPOD when a prospective cohort is used to develop or validate a prognostic model." }, { "relation_type": "applies_to", "target_slug": "cohort-retrospective", "notes": "Use TRIPOD when a retrospective (e.g., claims/EHR) cohort is used to develop or validate a prediction model; STROBE/RECORD-PE apply only if the objective is causal rather than predictive." }, { "relation_type": "applies_to", "target_slug": "diagnostic-accuracy", "notes": "Applies to multivariable diagnostic prediction models; a single diagnostic test's accuracy is reported with STARD instead." }, { "relation_type": "used_with", "target_slug": "prediction-model-validation-recalibration-rwe", "notes": "Implements TRIPOD's internal/external validation, calibration, and recalibration items." }, { "relation_type": "used_with", "target_slug": "predictive-and-causal-ml-models-rwe", "notes": "Implements the ML model-building, tuning, and optimism-control items that TRIPOD+AI governs." }, { "relation_type": "used_with", "target_slug": "generalizability-transportability-external-validity-rwe", "notes": "Implements TRIPOD's external-validation and (TRIPOD+AI) representativeness/fairness items for models moved across populations or databases." }, { "relation_type": "see_also", "target_slug": "algorithm-validation", "notes": "Outcome/predictor definitions in real-world data must be validated; TRIPOD requires their definition and validation be reported." }, { "relation_type": "see_also", "target_slug": "diagnosis-phenotype-algorithm-1ip-2op-time-window-rwe", "notes": "Phenotype/algorithm logic and its PPV/sensitivity implement TRIPOD's outcome- and predictor-definition items for claims/EHR-based models." }, { "relation_type": "see_also", "target_slug": "fit-for-purpose-data-assessment-rwe", "notes": "TRIPOD requires the data source and its fitness-for-purpose be documented before a model is trusted." }, { "relation_type": "see_also", "target_slug": "attrition-and-loss-to-follow-up-rwe", "notes": "Handling of attrition and loss to follow-up is a TRIPOD reporting requirement for longitudinal prognostic models." }, { "relation_type": "see_also", "target_slug": "visualizations-pharmacoepidemiology-rwe", "notes": "Supplies the calibration plots, ROC/AUC, and decision-curve visuals TRIPOD requires for performance reporting." } ], "index_definitions": [], "checklist_items": [], "regulatory_relevance": [ "journal", "fda", "ema", "hta" ] } ] }