Sample size and testing

Specify scenarios hypotheses and multiplicity

Choose the claim before the calculation. Superiority, noninferiority and equivalence require different success rules; justify margins clinically and from evidence. Account for multiple claims and assumptions about missingness and event frequency.

Flow diagram

Sample size and testingF1.S1Specify estimation and testingobjectivesF1.S2Define hypotheses and successcriteriaF1.S3Plan endpoint families andmultiplicityF1.S4Evaluate sample-size scenariosF1.S5Document calculation method andevidenceF1.S6Select and document designF1.N1Sample-size and power scenariosScenarioOutcome/modelTargeteffectNuisanceassumptionsAlphaPower/precisionN/eventsSourceF1.N2Testing strategyEndpointfamilyRoleHypothesisSuccess ruleOrderingError controlF1.N3Hypothesis specificationF1.N4Design evidenceF1.S0 Start flowSample size and testingF1.S999 End flowEnd flow

Steps

Specify estimation and testing objectives

Choose superiority, noninferiority, equivalence or precision as appropriate; justify margins if used.

Define hypotheses and success criteria

Specify effect scale, direction, null and alternative, alpha and clinically relevant effects.

Plan endpoint families and multiplicity

Specify multiple primary/secondary success rules, testing sequence and error control.

Evaluate sample-size scenarios

Specify outcome distributions, event/follow-up assumptions, allocation and attrition.

Document calculation method and evidence

Record software/version, calculation provenance and verified results.

Select and document design

Record chosen assumptions, feasibility and any planned monitoring.

Sample size and testing

End flow

Tables and supporting information

Sample-size and power scenarios

Sample-size and power scenariosScenarioOutcome/modelTargeteffectNuisanceassumptionsAlphaPower/precisionN/eventsSource

This is an authoring table, not a calculation engine. Leave numerical inputs and outputs blank.

Testing strategy

Testing strategyEndpointfamilyRoleHypothesisSuccess ruleOrderingError control

Support multiple primary and secondary endpoints.

Hypothesis specification

Write the hypotheses for the chosen estimand, design objective and outcome scale; define symbols and assumptions.

Design evidence

Attach evidence for effect sizes, variability, margins, event rates or dropout assumptions.

Sources