Keywords: Bayesian adaptive design, simon two stage design, small sample size
Challenged by the difficulty of recruitment in certain patient population and disease feature, the struggle of small sample size is one of the major obstacles in conducting clinical trials in certain areas such as rare diseases and pediatric population. There is substantial needs for a scientifically sound, executable, and interpretable study design for clinical trials challenged by small sample size, especially in circumstance when a randomized study design is not feasible. In this poster, we propose a novel Bayesian adaptive approach motivated by Simon two-stage optimal design1. Compared to the traditional Simon two-stage design, which only provides decision rules at two times during a clinical trial, our proposed Bayesian adaptive approach will provide decision rules at multiple time points. By observing the posterior probability of success, the study could stop at any sample size according to the pre-specified success or futile criteria, which could reduce sample size and potentially save patients from being exposed to an inefficacious treatment. Simulation studies are conducted to compare the Bayesian adaptive design with a standard design under varying success and rutile criteria; also the robustness of the design under the different the priors will be addressed. .
1 Simon, Richard. "Optimal two-stage designs for phase II clinical trials." Contemporary Clinical Trials 10.1 (1989): 1-10.