Power considerations for the win ratio: A rank-based simulation approach.

Area of Science:

• Biostatistics
• Clinical Trial Design

Background:

• The win ratio is an innovative statistical method for assessing clinical trial efficacy.
• Complexities like censoring, ties, and correlations complicate win ratio study design.
• Accurate sample size and statistical power are crucial for successful clinical trial planning.

Purpose of the Study:

• To develop a flexible, simulation-based approach for clinical trial design using the win ratio.
• To support accurate determination of sample size and statistical power in trials employing the win ratio.

Main Methods:

• A simulation-based strategy was developed, leveraging the link between the win ratio and rank distributions.
• The method was adapted to incorporate administrative censoring, ties, and correlations.
• A simulation study evaluated the approach's type I error and statistical power.

Main Results:

• The proposed approach effectively preserves type I error rates.
• Simulations confirmed the method samples according to the specified win ratio.
• The approach provides valuable guidance on statistical power relative to sample size.
• The `winratiopss` R package is available to implement this methodology.

Conclusions:

• This simulation strategy presents a novel and adaptable method for informing clinical trial designs that utilize win ratio analysis.