Sample size formula for a win ratio endpoint

Area of Science:

• Clinical Trials
• Biostatistics
• Medical Research

Background:

• The win ratio composite endpoint is increasingly used in late-stage clinical trials.
• Current sample size calculations for win ratio endpoints often rely on complex simulations.
• A simple, accessible formula for sample size determination is lacking in the literature.

Purpose of the Study:

• To derive and validate a simple formula for calculating sample sizes for win ratio composite endpoints.
• To compare confidence intervals derived from the new formula with those from patient-level data.
• To provide insights into trial design and power considerations using the win ratio method.

Main Methods:

• A novel sample size formula was developed based on probabilities of patient outcomes and ties.
• The formula's performance was evaluated by comparing confidence intervals with those calculated from patient-level data across 17 endpoints.
• Simulations were conducted to assess the formula's accuracy and utility.

Main Results:

• The formula provides accurate confidence intervals comparable to those requiring patient-level data.
• Simulations demonstrated the formula's effectiveness in sample size calculations.
• The formula offers insights into how hierarchical endpoint inclusion impacts statistical power.

Conclusions:

• The developed formula offers a practical and efficient alternative to complex simulations for win ratio sample size calculations.
• This formula can aid researchers, including non-specialists, in determining appropriate trial sizes.
• The findings support the broader adoption and understanding of the win ratio method in clinical research.