Sample size calculations for stepped wedge and cluster randomised trials: a unified approach

Area of Science:

• Clinical Trials Methodology
• Biostatistics
• Epidemiology

Background:

• Stepped wedge cluster randomized trials (SW-CRTs) are increasingly used in health research.
• Accurate sample size calculations are crucial for the validity and efficiency of SW-CRTs.
• Comparing SW-CRTs with other cluster randomized trial designs is essential for optimal study planning.

Purpose of the Study:

• To clarify and illustrate sample size calculations for cross-sectional SW-CRTs.
• To present a unified framework for comparing the efficiencies of competing trial designs.
• To provide practical guidance for designing stepped wedge studies.

Main Methods:

• Summarized design effects for SW-CRT, parallel cluster randomized trial (CRT), and CRT with before and after observations (CRT-BA).
• Developed new formulas for determining required cluster size based on the number of clusters.
• Illustrated calculations with examples and provided practical guidance.

Main Results:

• The most powerful study design depends on the intracluster correlation coefficient (ICC) and cluster size.
• For small ICCs, parallel CRTs are generally more efficient.
• For large ICCs, SW-CRTs are more efficient and can be an alternative when CRTs are infeasible.

Conclusions:

• A unified approach facilitates comparison of competing designs.
• Trialists can use this framework to select the most efficient design for their specific scenario.
• Informed design decisions enhance the power and feasibility of cluster randomized trials.