PS13: Bayesian Methods Driving Transdisciplinary Decision-Making in Confirmatory Clinical Trials

Bayesian methodology is increasingly used in clinical trials to support efficient decision-making, particularly in settings involving historical borrowing, external controls, subgroup inference, and adaptive designs. In recognition of this growing adoption, the FDA has issued guidance outlining expectations for the appropriate use, evaluation, and documentation of Bayesian methods in drug and biologic development. This session will focus on practical implementation of Bayesian approaches in confirmatory and late-phase clinical trials, explicitly framed around the FDA's Bayesian guidance and the needs of cross-functional decision-making in an increasingly digital and data-rich environment. Speakers will discuss where Bayesian methods most effectively support integrated decisions, how to pre-specify priors and borrowing strategies, and how to demonstrate frequentist operating characteristics to support regulatory decision-making. Using real-world case examples, the session will highlight common regulatory considerations that influence decision-making across disciplines, including control of Type I error, robustness to prior–data conflict, sensitivity analyses, and transparency in reporting. Particular emphasis will be placed on scenarios such as borrowing historical placebo data, augmenting concurrent controls with external data, subgroup borrowing, and dose-finding decisions. The session is intended for statisticians, clinical scientists, and regulatory stakeholders seeking to apply Bayesian methods in a way that supports timely, transparent, and defensible decisions, and that is scientifically rigorous, operationally feasible, and aligned with FDA expectations.