BOP2: Bayesian optimal design for phase II clinical trials with simple and complex endpoints.

Area of Science:

• Clinical Trial Design
• Biostatistics
• Bayesian Inference

Background:

• Phase II clinical trials require adaptive designs to efficiently evaluate drug efficacy.
• Existing Bayesian designs often lack explicit type I error control and flexibility for diverse endpoints.

Purpose of the Study:

• To propose a unified Bayesian optimal phase II (BOP2) design.
• To accommodate simple and complex endpoints within a single framework.
• To ensure type I error rate control and practical implementation.

Main Methods:

• Utilizing a Dirichlet-multinomial model for diverse endpoint types.
• Employing posterior probability evaluation for interim go/no-go decisions.
• Optimizing decisions to maximize power or minimize sample size under the null hypothesis.

Main Results:

• The BOP2 design explicitly controls the type I error rate, bridging Bayesian and frequentist approaches.
• Simulation studies demonstrate higher power and reduced risk of premature termination compared to other Bayesian designs.
• Stopping boundaries are pre-defined, enhancing practical usability.

Conclusions:

• The BOP2 design offers a flexible, unified, and statistically rigorous approach for phase II trials.
• Its features facilitate adoption by researchers and regulatory agencies.
• Freely available software enhances accessibility for implementing this advanced trial design.