Using dichotomized survival data to construct a prior distribution for a Bayesian seamless Phase II/III clinical...

Area of Science:

• Clinical Trial Design
• Biostatistics
• Pharmaceutical Research

Background:

• Master protocols facilitate simultaneous evaluation of multiple treatments or disease subgroups within a single trial framework.
• Seamless study designs integrate multiple clinical phases (e.g., Phase II/III) into a unified trial structure.
• Existing research on Bayesian approaches within master protocol designs, particularly seamless variants, is limited.

Purpose of the Study:

• To propose and evaluate Bayesian operationally seamless Phase II/III designs.
• To compare the performance of these Bayesian designs against traditional operationally frequentist designs.
• To utilize distinct endpoints for each stage: binary for Phase II and time-to-event for Phase III.

Main Methods:

• Development of Bayesian operationally seamless Phase II/III designs.
• Phase II employs a binary endpoint for arm selection using posterior (futility) and predictive (selection) probabilities.
• Phase III utilizes a time-to-event endpoint, incorporating Phase II results as prior distributions for the Bayesian model.

Main Results:

• Bayesian operationally seamless designs demonstrated the capability to approximate the performance of inferentially seamless designs.
• Simulations indicated an increase in statistical power for Bayesian designs compared to their operationally frequentist counterparts.
• The proposed Bayesian framework effectively integrates interim results into subsequent trial phases.

Conclusions:

• Bayesian operationally seamless designs offer a viable and potentially more powerful alternative for adaptive clinical trials.
• These designs can enhance efficiency and statistical power in Phase II/III integrated trials.
• Further investigation into Bayesian methods for complex master protocol designs is warranted.