A tutorial on sample size calculation for multiple-period cluster randomized parallel, cross-over and stepped-wedge...

Area of Science:

• Biostatistics
• Clinical Trial Design
• Epidemiology

Background:

• Sample size calculations for cluster randomized trials (CRTs) must account for within-cluster correlation.
• Correlation complexity increases with repeated measurements over time and participant types (cross-sectional vs. cohort).
• Existing methodologies for complex CRTs, particularly those with multiple measurement periods, may lack accessibility for many researchers.

Purpose of the Study:

• To provide a tutorial on sample size calculation for cluster randomized designs.
• To emphasize methodologies for designs with multiple measurement periods.
• To introduce the Shiny CRT Calculator, a web-based tool for accessible sample size calculations.

Main Methods:

• The tutorial covers sample size calculations for cross-sectional and cohort designs.
• It addresses various within-cluster correlation structures and cluster heterogeneity in treatment effects.
• The Shiny CRT Calculator computes power or precision based on cluster size or number of clusters for diverse designs.

Main Results:

• The methodology accommodates parallel, cluster cross-over, and stepped-wedge designs.
• The Shiny CRT Calculator supports a wide range of designs and correlation structures.
• Examples illustrate the methodology and the calculator's flexibility.

Conclusions:

• Accessible sample size calculation is crucial for effective cluster randomized trial design.
• The Shiny CRT Calculator enhances the practical application of complex sample size methodologies.
• Researchers can now more easily determine appropriate sample sizes for various CRTs, particularly those with repeated measures.