Keywords: Bayesian clinical trial, Bayesian barriers,
Bayesian clinical trials in oncology began with non-linear mixed effect modeling in pharmacokinetics and phase I dose-finding designs using the continuous reassessment method (CRM). In recent years, the number of clinical trials applying Bayesian methods continues to increase, but many barriers to their use remain. In consideration of the ongoing development of new clinical trial methods, we update and expand our previously reported institutional experience with Bayesian clinical trials (Biswas et al., Bayesian clinical trials at the University of Texas M. D. Anderson Cancer Center, Clinical Trials 2009) by reviewing MD Anderson (MDA) clinical trials registered from 2009 through 2013. Overall, 283 in 1020 (28%) trials had Bayesian components and were designated as Bayesian trials. Among MDA only and multi-center trials, 56% and 14% were Bayesian trials, respectively. Bayesian trials were more common in Phase I/II trials (34%) than in Phase III/IV (6%) trials. Among Bayesian trials, the most commonly used features were for toxicity monitoring (65%), efficacy monitoring (36%), and dose finding (22%). The vast majority of Bayesian trials used a non-informative prior (86%). A subset of 75 trials applied Bayesian methods as their primary designs. Among them, the most commonly used designs were Bayesian logistic regression model (N=22), CRM (N=20), and adaptive randomization (N=16). In the presentation, we will also discuss common barriers to implementing Bayesian trials and solutions to overcome these obstacles.