A new approach for sizing trials with composite binary endpoints using anticipated marginal values and accounting for...

Area of Science:

• Clinical Trials and Biostatistics
• Statistical Methods in Medical Research

Background:

• Composite binary endpoints are frequently utilized as primary endpoints in clinical trials.
• Accurate sample size determination is critical for the statistical validity of these trials.
• Sample size calculations for composite endpoints necessitate knowledge of component event rates, effect sizes, and their correlations.

Purpose of the Study:

• To investigate the impact of correlation and marginal parameter uncertainty on sample size calculations for composite binary endpoints.
• To propose a robust strategy for sample size determination when correlation is unknown and marginal parameters are uncertain.
• To introduce the CompARE web platform for characterizing composite endpoints and calculating sample sizes.

Main Methods:

• Developed a general strategy for sample size calculation that accommodates unspecified correlation and uncertainty in marginal parameters.
• Evaluated the proposed method through a comprehensive simulation study.
• Utilized a real-world case study to illustrate the application of the method and the CompARE platform.

Main Results:

• Sample size for composite binary endpoints is highly sensitive to the correlation between components.
• Inaccurate prior information on marginal parameters can lead to underpowered trials.
• The proposed strategy effectively addresses uncertainty in correlation and marginal parameters.

Conclusions:

• Accurate sample size calculation for composite binary endpoints requires careful consideration of component correlations and parameter uncertainties.
• The proposed method and the CompARE platform offer valuable tools for clinical trial designers to ensure adequate power.
• Addressing these uncertainties is essential for achieving study objectives and avoiding underpowered research.