A Nonparametric Test for Equality of Survival Medians

• Activity Number: 237
• Type: Contributed
• Date/Time: Monday, August 1, 2011 : 2:00 PM to 3:50 PM
• Sponsor: Section on Nonparametric Statistics
• Author(s): Mohammad Hossein Rahbar*+ and Zhongxue Chen and Sangchoon Jeon and Joseph C. Gardiner and Jing Ning
• Companies: The University of Texas Health Science Center at Houston and The University of Texas Health Science Center at Houston and Yale University School of Nursing and Michigan State University and The University of Texas School of Public Health
• Keywords: Median ; Nonparametric ; Censored Data ; Clinical Trials ; Survival Analysis ; Simulations

Abstract

In clinical trials researchers often encounter testing for equality of survival medians across study arms based on censored data. Even though Brookmeyer-Crowley (BC) introduced a method for comparing medians of several survival distributions, still some researchers misuse procedures which are designed for testing the homogeneity of survival curves. This practice often leads to inflation of type I error. We propose a new nonparametric method for testing the equality of several survival medians based on Kaplan-Meier estimation from randomly right censored data. We derive asymptotic properties of this test statistic. Through simulations we compare the power of this new procedure with that of the log-rank, Wilcoxon, and BC method. Our simulation results indicate that performance of these test procedures depends on the level of censoring and appropriateness of the underlying assumptions. When the assumptions of the log-rank test are met, some of these procedures are more powerful than the Wilcoxon, BC, and our proposed test. However, when the objective is testing homogeneity of survival medians rather than survival curves, our test statistic provides an alternative to the BC test.