Conference Program Home
  My Program

All Times EDT

Abstract Details

Activity Number: 468 - Statistical Challenges and Novel Methodologies for Analyzing Health Outcomes
Type: Topic Contributed
Date/Time: Wednesday, August 10, 2022 : 2:00 PM to 3:50 PM
Sponsor: Section on Nonparametric Statistics
Abstract #322730
Title: Power Analyses for Stepped Wedge Designs with Multivariate Continuous Outcomes
Author(s): Monica Taljaard and Fan Li and Kendra Plourde*
Companies: Ottawa Hospital Research Institute and Yale School of Public Health and Yale University
Keywords: cluster randomized trial; co-primary endpoints; multivariate linear mixed model; power; sample size estimation; stepped wedge design

Multivariate outcomes are not uncommon in pragmatic cluster randomized trials. While sample size calculation procedures for multivariate outcomes exist under parallel assignment, none have been developed for a stepped wedge design. We present efficient power and sample size procedures for stepped wedge cluster randomized trials (SW-CRTs) with multivariate outcomes that differentiate the within-period and between-period intracluster correlation coefficients (ICCs). To do so, we first extend the existing univariate linear mixed model for cross-sectional SW-CRTs to a multivariate linear mixed model that allows simultaneous estimation of the intervention effects on multiple outcomes. We then derive the joint distribution of the intervention test statistics which can be used for determining power under any specified hypothesis and provide an example using the commonly utilized intersection-union test for co-primary outcomes. Simplifications under a common treatment effect and common ICCs across endpoints and an extension to closed cohort designs are also provided. We present the results of a simulation study to validate the methods and illustrate application using data from a SW-CRT.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2022 program