A Cautionary Note on the Use of a Common ICC in the Analysis of a Two-Arm Cluster Randomized Trial for Cancer Prevention Studies

• Activity Number: 497
• Type: Contributed
• Date/Time: Wednesday, August 7, 2013 : 8:30 AM to 10:20 AM
• Sponsor: WNAR
• Author(s): Sheng Wu*+ and Catherine Crespi and Weng Kee Wong
• Companies: UCLA and UCLA and Department of Biostatistics, UCLA
• Keywords: Cluster randomized trials ; Cancer screening ; Intraclass correlation coefficient ; Correlated binary data

Abstract

The intraclass correlation coefficient (ICC) is a fundamental parameter of interest in cluster randomized trials (CRTs) as it can greatly affect statistical power. Using data from community-based cancer prevention trials with primary outcome of self-reported cancer screening, we compare common methods of estimating the ICC in CRTs and show that ICC point estimates from different methods can be quite different, although confidence intervals generally overlap. The ICC varied substantially by study condition in some studies, suggesting that the common practice of assuming a common ICC across all clusters in the trial may be questionable. A simulation study confirmed pitfalls of erroneously assuming a common ICC. We recommend that investigators use sample size and analysis methods that allow the ICC to vary by study condition.