Adaptive Multi-Arm Multi-Stage Designs

• Activity Number: 218 - Efficient Designs and Better Decision-Making Strategies in Complex Clinical Trials: Multiple Arms, Multiple Endpoints, and Multiple Stages
• Type: Topic Contributed
• Date/Time: Monday, July 31, 2017 : 2:00 PM to 3:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Cyrus Mehta* and Pranab Ghosh
• Companies: Cytel Inc and Cytel Inc, Boston University
• Keywords: MAMS designs ; Combination p-values ; adaptive ; Sample Size Reestimation ; multiple arms ; enrichment

Abstract

Multi-arm multi-stage (MAMS) designs are designs that compare several intervention arms to a common control arm in a randomized clinical trial with one or more interim analyses at which arms can be terminated either for futility or overwhelming efficacy. There are two approaches for constructing such designs. The p-value combination approach, with closed testing to ensure strong control of type-1 error, is the method that is most frequently used. Recently, however, there has been a great deal of interest in the extension of group sequential methods from two arm trials to multi-arm trials with stopping boundaries derived from error spending functions. In this presentation we will discuss the methodological difference between the two approaches and compare their operating characteristics in various settings including adaptive sample size re-estimation.