A structured framework for adaptively incorporating external evidence in sequentially monitored clinical trials

Area of Science:

• Biostatistics
• Clinical Trial Design
• Bayesian Inference

Background:

• Sequential monitoring in clinical trials is crucial for ethical and efficient trial conduct.
• Traditional methods often lack flexibility in incorporating external data.
• Bayesian approaches offer a robust framework for adaptive trial designs.

Purpose of the Study:

• To develop a flexible Bayesian framework for sequential monitoring that incorporates external data.
• To introduce an adaptive monitoring prior that dynamically weighs evidence.
• To simplify prior elicitation by linking it to trial hypotheses and external data.

Main Methods:

• A Bayesian framework utilizing generalized normal distributions for monitoring priors.
• Incorporation of external data by mixing skeptical and enthusiastic priors with adaptive weighting.
• Application to single-arm and two-arm randomized controlled trials, including retrospective analysis.

Main Results:

• Demonstrated the utility of the framework in pediatric trial examples using adult trial data.
• Illustrated prospective and pre-specified use of external data in monitoring.
• Showcased adaptive prior's ability to adjust based on observed data consistency.

Conclusions:

• The proposed Bayesian framework offers a flexible and efficient method for sequential monitoring.
• Adaptive integration of external data enhances clinical trial design and decision-making.
• This approach simplifies prior elicitation and allows for more informed trial conduct.