Particular Challenges of Sequential Analysis in Cluster-Randomized Trials

• Activity Number: 598
• Type: Contributed
• Date/Time: Wednesday, August 7, 2013 : 2:00 PM to 3:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Abigail Shoben*+
• Companies: The Ohio State University
• Keywords: cluster randomized trials ; sequential monitoring ; clinical trials

Abstract

Cluster randomized trials (CRTs) are becoming increasingly common in public health research. Examples include randomizing counties to receive smoking cessation treatments, randomizing schools to receive exercise promotion materials, and randomizing pharmacy practices to receive an educational intervention. In a CRT, individual outcomes are correlated, and statistical tests for the treatment effect must account for the correlation among individuals from the same cluster. In individually-randomized clinical trials, ethical and financial considerations frequently dictate the use of group sequential methodology to allow the trial to be analyzed prior to all planned data collection; similar considerations might suggest their use in CRTs. However, additional considerations have limited the use of standard group sequential methodology in CRTs, such as estimating the information growth within and between clusters and a limited number of clusters at interim analysis. In this talk, we review these challenges and provide recommendations for future studies.