Clustered trial analysis specification

Specify multilevel estimation and sensitivity

Use a model or cluster-level analysis compatible with the target weighting and randomization. With few clusters, justify degrees of freedom or small-sample corrections rather than relying on large individual counts.

Flow diagram

Clustered trial analysis specificationF1.S1Choose estimand-compatible modelF1.S2Respect randomization andstratificationF1.S3Specify small-sample inferenceF1.S4Handle missingness and unequalsizesF1.S5Plan sensitivity and reportingF1.N1Cluster analysis planEstimandModelWeightingSmall-samplemethodMissingnessPrimary effectAlternativeweightingCluster dropoutsensitivityF1.N2Assumptions and rationaleF1.N3Method and protocol evidenceF1.S0 Start flowClustered trial analysisspecificationF1.S999 End flowEnd flow

Steps

Choose estimand-compatible model

Specify cluster-level, multilevel or marginal estimation and weighting.

Respect randomization and stratification

Account for allocation restrictions and prognostic covariates as planned.

Specify small-sample inference

Justify degrees of freedom or variance corrections with few clusters.

Handle missingness and unequal sizes

Assess cluster dropout, informative sizes and individual loss.

Plan sensitivity and reporting

Document effect estimates, uncertainty and cluster/participant denominators.

Clustered trial analysis specification

End flow

Tables and supporting information

Cluster analysis plan

Cluster analysis planEstimandModelWeightingSmall-samplemethodMissingnessPrimary effectAlternativeweightingCluster dropoutsensitivity

Do not substitute independent-person standard errors.

Assumptions and rationale

Large participant numbers do not compensate for very few independent randomized clusters.

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