Bayesian Adaptive Design for clinical trials with potential subgroup effects
Conference
64th ISI World Statistics Congress
Format: IPS Abstract
Keywords: bayesian, multipletesting
Session: IPS 379 - Innovative Designs and Statistical Inference for Clinical Trials
Thursday 20 July 10 a.m. - noon (Canada/Eastern)
Abstract
Many designs of clinical trials face the challenge of two sources of uncertainty. First, to calculate the statistical power of a clinical trial, traditional design methods require users to provide the value of drug efficacy. However, investigating drug efficacy is the aim of the clinical trial, and hence, its true value is never known. Second, many modern clinical trials have a potential subgroup effect. But, drug developers often are unsure if a predefined subgroup of patients can respond to the drug better than the entire population before the clinical trial is conducted. In this talk, we present a novel Bayesian adaptive design, that correctly handles these two uncertainties to help design a better clinical trial.