Allowing for stratification in sample size planning of two-arm trials with continuous or binary outcome: Overview and...

Area of Science:

• Biostatistics
• Clinical Trial Design
• Epidemiology

Background:

• Clinical trials frequently involve heterogeneous populations with covariates defining strata.
• Stratification is crucial for controlling variability and ensuring balanced treatment group comparisons.
• Current sample size planning often neglects stratification, potentially impacting trial power and efficiency.

Purpose of the Study:

• To systematically review statistical methods for sample size calculation in stratified two-arm trials.
• To compare the efficiency of different power calculation and sample size determination methods for stratified designs.
• To provide evidence-based recommendations for sample size planning in stratified clinical trials.

Main Methods:

• Review of statistical literature on hypothesis testing and power/sample size calculations for stratified parallel group designs.
• Comparative analysis of the efficiency of various sample size planning methods under standard assumptions.
• Derivation of algorithms and closed-form formulae for sample size determination where applicable.

Main Results:

• Identified a range of statistical procedures for stratified comparisons with available power calculation methods.
• Demonstrated varying efficiencies among different sample size planning approaches for stratified trials.
• Established a basis for informed recommendations on sample size determination.

Conclusions:

• Effective sample size planning in stratified trials requires explicit consideration of stratification.
• The reviewed methods offer practical tools for researchers designing stratified two-arm studies.
• Adopting appropriate sample size strategies enhances the statistical power and reliability of stratified clinical trials.