A Bayesian Optimal Design Using Natural Conjugate Prior Families in Two-Arm, Randomized Phase II Clinical Trials with Endpoints from Exponential Families

• Activity Number: 127
• Type: Contributed
• Date/Time: Monday, August 4, 2014 : 8:30 AM to 10:20 AM
• Sponsor: Biopharmaceutical Section
• Author(s): Wei Jiang*+ and Jo A. Wick and Matthew S. Mayo
• Companies: University of Kansas Medical Center and University of Kansas Medical Center and Kansas University Medical Center
• Keywords: Optimized design ; Sample size ; Multiple constraints ; Posterior credible intervals

Abstract

An optimal design in two-arm randomized phase II clinical trials was proposed by Mayo et al. (2010), where the total sample size is minimized under multiple constraints on the standard error of the estimated event rates. The previous design is applicable for dichotomous outcomes from a frequentist perspective. In this paper, a Bayesian design using natural conjugate prior families for Phase II clinical trials with endpoints from the exponential families is extended from Mayo's approach. The proposed optimal design minimizes the total sample size under pre-specified constraints on the expected length of posterior credible intervals for both group means and their difference. This design is applicable for trials with fixed or optimal randomization allocation ratios. Examples of method implementation are provided for different types of endpoints in the exponential families.