Win-loss parameters for right-censored event data, with application to recurrent events

Area of Science:

• Biostatistics
• Clinical Trials
• Survival Analysis

Background:

• The win ratio is a popular method for comparing multiple, prioritized events in clinical cohort studies.
• Existing literature lacks robust inference methods for win ratios with censored event data, especially for recurrent events.
• Recurrent events, such as hospitalizations, are common in clinical studies and require specialized analysis.

Purpose of the Study:

• To propose and evaluate a novel statistical inference method for win-loss parameters, specifically for death and recurrent event outcomes.
• To address the gap in analyzing prioritized, censored, and recurrent event data in clinical research.
• To provide a statistically sound approach for win ratio calculation in the presence of censoring and recurrent events.

Main Methods:

• Development of statistical inference for win-loss parameters considering death and a recurrent event.
• Application of the method under independent right-censoring assumptions.
• Simulation studies to assess the small sample properties and accuracy of the proposed variance formula.

Main Results:

• The proposed inference method demonstrates accurate variance estimation, even with small sample sizes in simulations.
• The method is successfully applied to a real-world dataset from a randomized clinical trial.
• The study validates the utility of the new method for analyzing complex clinical event data.

Conclusions:

• The developed statistical inference method provides a reliable tool for analyzing death and recurrent events in clinical trials with censored data.
• The findings contribute to the advancement of statistical methodologies for win ratio analysis in biostatistics.
• The method offers a practical solution for researchers dealing with prioritized, recurrent, and censored outcomes.