A Unified Bayesian Decision Rule-Based Approach for Bayesian Design of Clinical Trials Using Historical Data

Chen et al. (Biometrics. 67(3):1163–1170, 2011) proposed a general Bayesian methodology for the design of non-inferiority clinical trials with a focus on controlling type I error and power. More recently, the posterior probability approach, the Bayesian factor approach, and the conditional borrowing approach are further developed. All of these methods can be unified under the general Bayesian decision rule-based framework. In this chapter, we provide a detailed elaboration of the decision-based approach for designing a clinical trial under the Bayesian paradigm. The Bayesian methods are further demonstrated via the designs of a non-inferiority medical device clinical trial as well as a superiority Duchenne muscular dystrophy (DMD) trial, in which multiple historical data sets from previous trials are available.