swdpwr: A SAS Macro and An R Package for Power Calculation in Stepped Wedge Cluster Randomized Trials
share
Conditional and marginal models can both be utilized in power calculations for stepped wedge cluster randomized trials. Hussey and Hughes (2007) first provided the design and analysis for this particular type of cluster randomized trials based on the linear mixed effects model for continuous outcomes, along with an approximation to this approach for binary outcomes. Although not accurate for binary outcomes, it has been widely used. To improve the approximation for binary outcomes, Zhou, Liao, Kunz, Normand, Wang, and Spiegelman (2020) and Li, Turner, and Preisser (2018b) have recently published two new methods for stepped wedge designs (SWDs) of binary outcomes. However, these new methods have not been implemented in publicly available software. In this work, we have developed a SAS macro calculate power in SWDs in various settings for both continuous and binary outcomes.
READ FULL TEXT