Bayesian sample size calculations in phase II clinical trials using informative conjugate priors

Area of Science:

• Biostatistics
• Clinical Trial Design
• Bayesian Statistics

Background:

• Bayesian methods are increasingly used for phase II clinical trial analysis.
• Existing Bayesian sample size calculations often rely on diffuse prior distributions.
• Tan and Machin previously focused on sample size calculations using diffuse priors.

Purpose of the Study:

• To extend Bayesian sample size calculations for phase II clinical trials.
• To incorporate informative prior distributions into sample size determination.
• To allow researchers with varying prior beliefs (optimistic/pessimistic) to influence sample size decisions.

Main Methods:

• Utilized informative prior distributions in Bayesian sample size calculations.
• Employed various strategies for selecting informative priors.
• Methods for prior selection included using the mean, median, or mode of conjugate priors.

Main Results:

• Demonstrated that informative priors can be effectively integrated into sample size calculations.
• Showcased how different prior selection methods (mean, median, mode) lead to varying sample sizes.
• Provided a framework for researchers to actively participate in sample size decisions based on their prior knowledge.

Conclusions:

• Informative priors offer a flexible approach to Bayesian sample size determination in phase II trials.
• The choice of prior distribution and selection method impacts the resulting sample size.
• This methodology empowers researchers to tailor sample size based on specific prior beliefs and trial objectives.