Using Weighted Differences in Hazards as Effect Sizes for Survival Data

Abstract Sensitive to the change in sample sizes, traditional measures such as values of test statistics or p can fail quantify difference survival between populations for time-to-event data. We thereby propose use effect sizes defined weighted differences hazards evaluate two groups. On basis logrank statistic, Gehan–Wilcoxon statistic and Prentice–Wilcoxon we developed three that compare experiences over time period investigation. Estimates these were provided their large behaviors studied. In light Mann–Whitney parameter, presented an size compares entire possible/hypothetical period. Two estimates this constructed. compared proposed illuminated by theoretical studies, simulations real cancer The article help understand are expected have promising applications field analysis.