Sample size calculation for clinical trials analyzed with the meta-analytic-predictive approach.

Area of Science:

• Biostatistics
• Clinical Trial Design
• Bayesian Methodology

Background:

• The meta-analytic-predictive (MAP) approach is a Bayesian method for integrating historical control data into new clinical trials.
• Its goal is to improve statistical power and reduce the necessary sample size for new studies.
• Previous methods for sample size calculation using MAP relied on effective sample size (ESS), but its validity is debated.

Approach:

• A novel Monte Carlo simulation approach is proposed for calculating sample size in new trials utilizing historical data with the MAP method.
• Control parameters are sampled from the MAP prior, not used as point estimates, to fully leverage historical information.
• This method derives statistical power and required sample size by simulating new trial data based on sampled control parameters and the MAP prior.

Key Points:

• Addresses the questionable validity of using prior effective sample size (ESS) in MAP-based sample size calculations.
• Proposes a straightforward and generic Monte Carlo approach for sample size determination.
• Demonstrates the approach using real-life data from three studies with diverse outcomes.

Conclusions:

• The proposed Monte Carlo method provides a robust and direct way to calculate sample size for MAP analyses.
• This approach effectively utilizes available historical data to optimize new clinical trial design.
• The method is shown to be applicable across various study types and outcome measures.