Cluster follow-up and data populations

Track cluster losses and participants

Loss of a cluster and loss of a participant have different consequences. Preserve both identifiers and reconcile counts within their own units; do not add them together.

Flow diagram

Cluster follow-up and data populationsF1.S1Track clusters through follow-upF1.S2Track participant observationsF1.S3Reconcile losses at both levelsF1.S4Define analysis populations andweightsF1.S5Freeze multilevel data definitionsF1.N1Cluster follow-up definitionsLevelObservationunitLoss rulePopulationWeightingClusterParticipantVisitF1.N2Assumptions and rationaleF1.N3Method and protocol evidenceF1.S0 Start flowCluster follow-up and datapopulationsF1.S999 End flowEnd flow

Steps

Track clusters through follow-up

Record delivery, withdrawal and cluster-level outcomes.

Track participant observations

Define follow-up, repeated measures and movement between clusters.

Reconcile losses at both levels

Keep cluster dropout and missing individual outcomes separate.

Define analysis populations and weights

Link cluster membership and estimand-specific weighting.

Freeze multilevel data definitions

Retain cluster, participant and time identifiers with reproducible derivations.

Cluster follow-up and data populations

End flow

Tables and supporting information

Cluster follow-up definitions

Cluster follow-up definitionsLevelObservationunitLoss rulePopulationWeightingClusterParticipantVisit

Preserve hierarchical identifiers; this table stores definitions rather than patient records.

Assumptions and rationale

Loss of a whole cluster is not equivalent to independent loss of the same number of individuals.

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