Statistical analysis plan

Specify estimation missingness and sensitivity

Match the estimator to the estimand and randomization. Address missing outcomes separately from strategies for intercurrent events. Sensitivity analyses probe assumptions for the same estimand; supplementary analyses may target a different question.

Flow diagram

Steps

Confirm estimands and analysis populations

Link endpoints, intercurrent-event strategies and prespecified populations.

Specify primary and secondary methods

Define effect measures, uncertainty, models, covariates and assumptions.

Apply the testing strategy

Link endpoint families and prespecified multiplicity or decision rules.

Specify missing-outcome handling

Document primary assumptions, methods and implementation.

Plan sensitivity and subgroup analyses

Assess alternative assumptions for the same estimand and distinguish supplementary analyses targeting different questions. Prespecify subgroup interactions and avoid interpreting separate significance tests as subgroup differences.

Specify result shells and provenance

Retain planned methods separately from completed output and record deviations.