Optimal Adaptive Phase III Design with Interim Sample Size and Dose Determination

Abstract

Adaptive designs have been proposed to improve clinical trial efficiency. Examples include selecting a treatment arm OR determining a final sample size based on interim data. On the other hand, optimal designs have been proposed to provide robust power against the uncertainty of a planned effect size. In this paper, we propose an optimized 2-stage phase III clinical trial design that combines all these techniques to offer the opportunity of dose selection and sample size determination based on the first stage data with strict type I error rate control and robust power across an effect size interval.