Application of Recursive Adaptive Design with Maximum Two Interim Analysis

• Activity Number: 318 - Adaptive (and Other) Clinical Trial Designs
• Type: Contributed
• Date/Time: Wednesday, August 11, 2021 : 3:30 PM to 5:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Chuhan Zhou* and Chao Li and Vincent Yu
• Companies: Pacira Pharmaceuticals, Inc and Pacira Pharmaceuticals, Inc and Pacira Pharmaceuticals, Inc
• Keywords: Recursive Adaptive Design; Two-stage Combination Test; Overall p-value

Abstract

Recursive Adaptive Design enables flexibility in the number of interim analyses and the choice of stopping boundaries over general adaptive design. At each interim look, the design of the next stage can be determined using all the information accumulated, either a final stage with complete enrollment or a further interim analysis. Accordingly, we propose the maximum two interim analysis design - a recursive application of the two-stage combination test. At first interim, we will evaluate if early stopping criteria are met and then decide if the second interim is necessary before jump to the final stage directly. This requires a prospective sample size adaptation plan to incorporate the treatment effect observed from the first interim. The whole point is to combine p-values from different stages to a single overall p-value using an appropriate p-value function without inflating family-wise Type I error. This paper is intended to address this issue through simulating a series of scenarios about the plausible effect sizes with different variance scales. Family-wise Type I error under each scenario is estimated by the percentage of simulated trials leading to false positive conclusion.