Factorial data and analysis populations

Derive combination-specific and pooled populations

Keep assignment, receipt, adherence and intercurrent events for each factor separately. The four combinations should remain reconstructable from the released dataset.

Flow diagram

Factorial data and analysis populationsF1.S1Reconcile combination assignmentsF1.S2Derive factor-specific receiptF1.S3Define analysis populationsF1.S4Resolve outcomes and missingnessF1.S5Freeze derived data rulesF1.N1Factorial population specificationPopulationAssignment ruleReceipt ruleMissingnessContrastCombinationpopulationFactor AcomparisonFactor BcomparisonInteractionF1.N2Assumptions and rationaleF1.N3Method and protocol evidenceF1.S0 Start flowFactorial data and analysispopulationsF1.S999 End flowEnd flow

Steps

Reconcile combination assignments

Retain randomized factor levels and combination membership.

Derive factor-specific receipt

Record adherence and discontinuation separately for each intervention.

Define analysis populations

Specify combined-arm and main-effect populations without duplicate counting.

Resolve outcomes and missingness

Link deviations and missing observations to each estimand.

Freeze derived data rules

Version combination codes, contrasts and population definitions.

Factorial data and analysis populations

End flow

Tables and supporting information

Factorial population specification

Factorial population specificationPopulationAssignment ruleReceipt ruleMissingnessContrastCombinationpopulationFactor AcomparisonFactor BcomparisonInteraction

Overlapping main-effect populations are not disjoint participant groups.

Assumptions and rationale

A participant may contribute to both factor contrasts; count reconciliation must not add overlapping analysis groups.

Method and protocol evidence

Attach the protocol, methodological sources and versioned software specification supporting these decisions. Record citations and provenance; do not enter participant identifiers.

Sources