Official websites use .gov
A .gov website belongs to an official government organization in the United States.

Secure .gov websites use HTTPS
A lock ( ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites.

Methods: Mind the Gap

Webinar Series

Implications of Informative Cluster Size for the Design and Analysis of Cluster Randomized Trials

September 27, 2024, 11:00 am EDT
Brennan Kahan, Ph.D.
Brennan Kahan, Ph.D.

University College London

View the Webinar

About the Webinar

Cluster randomized trials involve randomizing groups of participants, such as schools, hospitals, or villages, between different interventions. Because participants belonging to the same cluster are often correlated, statistical methods, such as mixed-effects models or generalized estimating equations, are required to account for this correlation during analysis. However, it is being increasingly recognized that these methods may be biased when outcomes or treatment effects differ between larger and smaller clusters (i.e., informative cluster size). This talk focuses on the implications of informative cluster size for cluster randomized trials, including choice of estimand, choice of analysis method, and ways to evaluate assumptions. 

About Brennan Kahan

Dr. Brennan Kahan is a Principal Research Fellow at the Medical Research Council Clinical Trials Unit at University College London. Dr. Kahan’s research program is focused on developing methods to improve the design, analysis, and reporting of randomized trials. His major research interests include using estimands to improve the interpretability and validity of randomized trials; evaluating statistical methods for complex trial designs, such as cluster, non-inferiority, factorial, and rerandomization trials; and improving the reporting of randomized trials. His recent methods research includes the development of the Consolidated Standards of Reporting Trials extension for factorial trials, developing estimands and robust estimators for cluster randomized trials, and evaluating and comparing estimation methods for different intercurrent event strategies.

Last updated on