Online Program Home
My Program

Abstract Details

Activity Number: 540 - SPEED: Clinical Trial Design, Longitudinal Analysis, and Other Topics in Biopharmaceutical Statistics
Type: Contributed
Date/Time: Wednesday, August 1, 2018 : 11:35 AM to 12:20 PM
Sponsor: Biopharmaceutical Section
Abstract #332743
Title: Sample Size Estimation for Stratified Cluster Randomized Trials with Binary Outcomes
Author(s): Lee Kennedy-Shaffer* and Michael David Hughes
Companies: Harvard University and Harvard University
Keywords: cluster randomized trials; binary outcomes; stratification; intracluster correlation coefficient; design effect; sample size

The effects of stratification on the power of clinical trials with binary outcomes are more complex and less well understood than the effects for trials with continuous outcomes. We propose a generalized estimating equations-based method for analyzing stratified cluster randomized trials with binary outcomes. Formulae for determining the required sample size for such trials and for determining the ratio of the sample size for the stratified trial to the sample size for a comparably powered unstratified trial are identified. The key parameters that must be estimated and simplifying assumptions that may be made to determine the sample size in these cases are identified and the effect of these parameters on the sample size is described. Illustrative examples of these formulae and settings, particularly appropriate for cluster randomized trials with rare outcomes, where stratification can substantially reduce (by as much as 20%) the required sample size are provided. These methods are easy to use, also hold for individually randomized trials, and allow more precise sample size estimation and appropriate determinations of the benefits of stratification for trials with binary outcomes.

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

Back to the full JSM 2018 program