Sample Size Requirements to Test Subgroup-Specific Treatment Effects in Cluster-Randomized Trials

Area of Science:

• Biostatistics
• Clinical Trial Design
• Health Services Research

Background:

• Cluster-randomized trials (CRTs) are widely used for healthcare delivery interventions.
• Existing sample size methods for CRTs primarily address treatment effect heterogeneity, not subgroup-specific effects.
• There's a growing need for methods addressing subgroup-specific effects due to health equity considerations.

Purpose of the Study:

• To develop formal sample size and power analysis methods for testing subgroup-specific treatment effects in parallel-arm CRTs.
• To provide analytical insights into the variances and covariance of subgroup-specific treatment effect estimators.
• To facilitate power calculations for both omnibus and intersection-union tests.

Main Methods:

• Developed analytical methods for sample size and power calculations in parallel-arm CRTs with continuous outcomes and binary subgroup variables.
• Derived formulas for variances of subgroup-specific treatment effect estimators and their covariance.
• Validated methods through a simulation study and illustrated with the Umea Dementia and Exercise (UMDEX) CRT.

Main Results:

• The variances and covariance of subgroup-specific treatment effects are characterized as weighted averages of overall and heterogeneous treatment effect variances.
• The proposed methods provide explicit requirements for achieving desired power for omnibus and intersection-union tests.
• Simulation results show good correspondence between empirical and predicted power.

Conclusions:

• The developed methods offer a formal approach to sample size and power calculations for subgroup-specific treatment effects in CRTs.
• These methods are essential for adequately planning CRTs that aim to assess health equity.
• The findings support robust trial design for evaluating interventions across different participant subgroups.