Treatment patterns · Reporting

Sankey / Alluvial Diagram (Treatment Patterns)

Visualizes how patients flow between therapies across lines of treatment, sizing each ribbon by the number of patients making that transition.

Download SVG Open code and full worked example

When to use it

To summarize sequential treatment decisions (1L→2L→3L), switching, and discontinuation in a single figure. Best for 2–4 stages and a small set of drug classes; beyond that, ribbons overlap and a transition matrix/table is clearer.

How to read it

Node height = patients on that therapy at that line; ribbon width = patients moving between two therapies. Conservation holds — flows into a node equal flows out (plus discontinuation). Color ribbons by source to trace cohorts.

Worked example

1000 patients start first-line therapy (600 Drug A, 400 Drug B) and are followed across second and third lines. Transition counts between each pair of therapies (and to discontinuation, 'Stop') size the ribbons.

1L→2L: A→B 230, A→C 150, A→Stop 220, B→B 90, B→C 100, B→Stop 210 (flows into 2L nodes: B 320, C 250, Stop 430).

Result: Of 600 Drug A initiators, 220 (37%) discontinue by second line and 380 switch; the largest single flow is the 430 patients who reach 'Stop' by second line, and conservation checks (A→B 230 + B→B 90 = 320 into the 2L Drug B node).

Produced by

Treatment patterns / lines of therapy Switching, add-on & augmentation

Reference: Gatto NM, Wang SV, Murk W, et al. Visualizations throughout pharmacoepidemiology study planning, implementation, and reporting. Pharmacoepidemiol Drug Saf. 2022;31(11):1140-1152.