A Bayesian sequential design with adaptive randomization for 2-sided hypothesis test.

Area of Science:

• Biostatistics
• Clinical Trial Design
• Statistical Inference

Background:

• Bayesian sequential and adaptive randomization designs are increasingly used in clinical trials.
• These designs offer potential benefits in reducing participant numbers and conserving resources.
• Efficiently allocating patients to treatment arms is crucial for trial success.

Purpose of the Study:

• To propose a novel Bayesian sequential design incorporating adaptive randomization rates for 2-arm clinical trials.
• To enhance the efficiency of patient allocation to treatment arms.
• To optimize resource utilization and reduce the overall sample size.

Main Methods:

• Development of a Bayesian sequential design with adaptive randomization rates.
• Algorithms for calculating optimal randomization rates, critical values, and statistical power.
• Sensitivity analysis using varying prior distributions.
• Simulation studies to compare the proposed design with traditional methods.

Main Results:

• The proposed adaptive Bayesian sequential design demonstrated superior performance in simulations.
• Compared to traditional Bayesian sequential designs, it achieved greater statistical power and/or reduced actual sample sizes for a fixed total sample size.
• Application to a real dataset confirmed its ability to further reduce required sample size.

Conclusions:

• The proposed Bayesian sequential design with adaptive randomization offers significant advantages over traditional methods.
• It provides a more efficient approach to patient allocation, leading to improved trial outcomes and resource management.
• This design represents a valuable advancement in optimizing clinical trial methodologies.