Quantification of Impact of Safety Monitoring on Type I Error and Power of Efficacy Analysis in Phase III Group Sequential Clinical Trial

• Activity Number: 423
• Type: Contributed
• Date/Time: Tuesday, August 2, 2011 : 2:00 PM to 3:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Yanqiu Weng*+ and Wenle Zhao and Yuko Palesch
• Companies: Medical University of South Carolina and Medical University of South Carolina and Medical University of South Carolina
• Keywords: phase III clinical trials ; group sequential analysis ; stopping guidelines ; safety and efficacy ; bivariate binary response ; power

Abstract

In large phase III clinical trials, group sequential analysis is a common statistical approach to monitor efficacy outcome during the trial. For some life-threatening diseases, safety is also monitored even more frequently than efficacy. However, interim analysis for overwhelming efficacy ignores multiple analyses of the safety outcome up to that point, and when safety and efficacy endpoints are highly correlated, how much does interim safety analyses cost on the type I and II error probabilities of the interim efficacy analyses? We use a multivariate normal approximation approach to estimate the probability of stopping for efficacy and/or safety for trials that concurrently monitor binary efficacy and safety endpoints. The estimation results are verified by simulation method. Our study suggests, type I error decreases as efficacy-safety correlation increases. Moreover, while power for efficacy is robust to the misspecification of the magnitude of safety-efficacy correlation, it is vulnerable to the variation of safety profile: power decreases significantly with between-group difference in safety probabilities increases, even if such difference is clinically trivial.