swdpwr

swdpwr computes statistical power for stepped wedge cluster randomized trials (SW-CRTs), providing power estimation for cross-sectional and closed-cohort designs with continuous and binary outcomes.

Key Features:

• Design support: Supports cross-sectional and closed-cohort SW-CRT designs and accommodates continuous and binary outcomes.
• Model specification: Allows marginal and conditional model specifications and supports three different link functions.
• Correlation structures: Implements exchangeable, nested exchangeable, and block exchangeable correlation structures and accounts for within-period, between-period, and within-individual intracluster correlations (ICCs).
• Unequal cluster sizes: Permits unequal numbers of clusters per sequence.
• Implementation: Provided as a SAS macro (%swdpwr) and an R package (swdpwr).
• Non-simulation computation: Employs computationally efficient, non-simulation-based methods for power calculation.
• Binary-outcome methodology: Implements recently proposed methods for binary outcomes that improve upon traditional normal approximations in SW-CRTs.
• Design assumptions: Assumes complete designs with equal cluster-period sizes and does not permit decay in between-period ICC.

Scientific Applications:

• SW-CRT design and analysis: Power calculation and study design for stepped wedge cluster randomized trials in public health, clinical trials, and epidemiological research.

Methodology:

Uses non-simulation-based computational methods grounded in recent advancements that improve upon traditional normal approximations for binary outcomes in SW-CRTs; accommodates multiple intracluster correlation parameters and unequal cluster sizes, and assumes complete designs with equal cluster-period sizes without permitting decay in between-period ICC.