Data and analysis plan

Define populations methods and sensitivity

State the denominator before examining response. Restricting analysis to evaluable participants can induce selection. Retain uncertainty in the outcome and in any comparison basis, and limit causal claims accordingly.

Flow diagram

Steps

Resolve data quality and derivations

Version source, endpoint derivation and finalized datasets.

Define analysis populations

Distinguish enrolled, treated, evaluable and safety sets; avoid summing overlapping groups.

Confirm estimands and comparison basis

Link endpoint definitions and any benchmark evidence.

Specify estimation and decision methods

Document uncertainty, assumptions and multiple endpoint rules.

Handle missing outcomes and intercurrent events

Prespecify primary assumptions and strategies.

Plan sensitivity and output shells

Separate planned specifications from completed results and record deviations.