Abstract:
|
Bayesian adaptive designs can help address challenges in pediatric clinical trials including small sample size and slow accrual rate. We applied an innovative Bayesian adaptive design method to look at a re-design of a published Phase III pediatric trial incorporating two prior beliefs: skeptical and enthusiastic. The early success criteria were based on skeptical prior and the early futility criteria were based on enthusiastic prior extrapolated from historical adult trial. We also constructed alternative designs with only one prior belief. To identify a preferred design, the stopping boundaries were tuned to control for type I error, and operating characteristics including power, expected sample size and duration were evaluated via simulation. When performing an increasing number of interim analyses, the innovative Bayesian adaptive design outperforms alternative designs as it allows more reduction in sample size and duration while providing sufficient power. Therefore, planning a Bayesian adaptive design for pediatric trial with both skeptical and enthusiastic priors can be an efficient way for early trial stopping, thus potentially saving time and money for trial conduction.
|