Cluster trial endpoints and estimands

Define cluster and individual effects

Choose whether the average effect weights clusters equally or participants equally. Unequal and informative cluster sizes can make these different targets.

Flow diagram

Cluster trial endpoints and estimandsF1.S1Define outcome measurement levelF1.S2Define target treatment effectF1.S3Specify weighting and populationF1.S4Handle cluster and person eventsF1.S5Prioritize endpoints and claimsF1.N1Cluster estimand registerEndpointMeasurementlevelTargetweightingIntercurrenteventsPriorityPrimarySecondarySafetyImplementationF1.N2Assumptions and rationaleF1.N3Method and protocol evidenceF1.S0 Start flowCluster trial endpoints and estimandsF1.S999 End flowEnd flow

Steps

Define outcome measurement level

Specify individual, cluster or repeated outcomes and assessment timing.

Define target treatment effect

Distinguish participant-average from cluster-average effects where relevant.

Specify weighting and population

Specify whether clusters or participants receive equal weight in the target effect. Address unequal or informative cluster sizes and align the estimator with that target.

Handle cluster and person events

Address cluster loss, participant discontinuation and contamination.

Prioritize endpoints and claims

Separate efficacy, safety and implementation objectives.

Cluster trial endpoints and estimands

End flow

Tables and supporting information

Cluster estimand register

Cluster estimand registerEndpointMeasurementlevelTargetweightingIntercurrenteventsPriorityPrimarySecondarySafetyImplementation

Do not conflate cluster-average and participant-average estimands.

Assumptions and rationale

Unequal or informative cluster size can change weighting and interpretation of treatment effects.

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