# Review an ITC feasibility workbook

Independently check each stable ID against its cited source. Confirm that every directed contrast has a defined population, endpoint, timepoint, effect measure, direction, and target estimand. Check the network for a real shared comparator before calling a scenario anchored.

For population adjustment, inspect whether each claimed effect modifier is measured, harmonized, supported by the source data, and appropriate to the effect scale. Check that conditional and marginal effects are not treated as interchangeable. For MAIC, request weight and effective-sample-size diagnostics; for STC, request model and standardization diagnostics. For unanchored scenarios, identify all unmeasured or non-overlapping prognostic variables and effect modifiers as potential blockers.

Return a table of defects with the relevant IDs, evidence, consequence, and required action. Recommend `proceed`, `restrict`, or `stop` only as an advisory for the named human decision owner; never approve the analysis automatically.