Estimand specification

Question → population and conditions → outcome → intercurrent-event strategy → summary measure

Flow diagram

Estimand specificationEstimand specificationF1.S1Clinical question and objectiveF1.S2Population and treatment conditionsF1.S3Outcome and assessment windowprimary · unspecifiedF1.S4Intercurrent eventsF1.S5Strategy for each intercurrenteventF1.S6Population-level summary measureF1.S7Estimator and sensitivity analysesF1.N1Intercurrent-event strategiesEventStrategyRationaleData requiredSensitivityEventF1.N2Estimand and estimator assumptionsF1.N3Estimand rationale

Steps

Clinical question and objective

State the target effect or descriptive quantity.

Population and treatment conditions

Define the population and intervention or exposure strategies.

Outcome and assessment window

Define the variable and its assessment window.

Intercurrent events

Identify treatment discontinuation, rescue therapy, death or other relevant events.

Strategy for each intercurrent event

Choose and justify a strategy for each event; distinguish intercurrent events from missing observations.

Population-level summary measure

Define the effect or summary measure that answers the question.

Estimator and sensitivity analyses

Align the estimator and sensitivity analyses to the estimand.

Tables and supporting information

Intercurrent-event strategies

Intercurrent-event strategiesEventStrategyRationaleData requiredSensitivityEvent

Enter study-specific events and strategies; no default strategy is selected.

Estimand and estimator assumptions

Record identification and estimation assumptions separately.

Estimand rationale

Cite the protocol and methodological rationale.

Sources