A Bayesian sequential design with binary outcome.

Area of Science:

• Biostatistics
• Clinical Trial Design
• Statistical Methodology

Background:

• Sequential designs are crucial for adaptive clinical trials, but traditional Bayesian methods can inflate type I error rates.
• Controlling the type I error rate is paramount in hypothesis testing to avoid false positives.

Purpose of the Study:

• To propose a novel Bayesian sequential design for binary outcomes that effectively controls the type I error rate.
• To introduce a new futility stopping rule to improve trial efficiency.

Main Methods:

• Development of a Bayesian sequential design incorporating an alpha-spending function to manage the type I error rate.
• Algorithms for calculating critical values and statistical power.
• Simulation studies to compare the proposed design with traditional Bayesian sequential designs.
• Sensitivity analysis to assess the impact of prior distribution parameters and sample size.

Main Results:

• The proposed Bayesian sequential design demonstrates greater statistical power than traditional designs with fixed total sample sizes.
• The new futility stopping rule allows for earlier trial termination with a smaller actual sample size.
• Simulations confirm the effectiveness of the alpha-spending function in controlling the type I error rate.

Conclusions:

• The proposed Bayesian sequential design offers an effective approach to control type I error rates in clinical trials with binary outcomes.
• The novel futility stopping rule enhances trial efficiency by enabling earlier cessation when appropriate.
• This methodology provides a more powerful and efficient alternative to traditional Bayesian sequential designs.