TL7: Bayesian Application in Registration Trials with Confirmatory Secondary Endpoints
*Yi-wen Ma, Janssen Pharmaceutical Companies of Johnson & Johnson 


Drug companies are increasingly in need of clinical trials that incorporate confirmatory secondary endpoint(s) in order to differentiate efficacy and/or safety from marketed drugs to better characterize their compounds. In frequentist framework, it leads to the issues of applying optimal adjustment strategy for multiplicity. In Bayesian framework, it is believed that multiplicity typically is less of a concern. As Gelman et al. (2007) indicated, Bayesian method is more conservative as the credible intervals from Bayesian approach are wider than the classical frequentist confidence intervals. Nevertheless, Bayesian approach helps the interpretation of confidence intervals, and relieve the issues associated with multiplicity adjustments. One critical issue with Bayesian method is how prior distributions can be specified with credibility. More research is needed about the applicability of the Bayesian approach in trials with confirmatory secondary endpoints. We would like to open up discussions on available research and findings, share experiences on operational challenges, successful application of Bayesian method in clinical trials, and regulatory implications and acceptance.