Inference on Bivariate Survival Data with Interval Sampling Through Kendall's Tau: Testing and Association Measure

• Activity Number: 253
• Type: Contributed
• Date/Time: Monday, July 30, 2012 : 2:00 PM to 3:50 PM
• Sponsor: Biometrics Section
• Author(s): Hong Zhu*+ and Mei-Cheng Wang
• Companies: The Ohio State University and The Johns Hopkins University
• Keywords: Bivariate survival data ; Inverse probability weighting ; Kendall's tau ; Nonparametric estimation ; Dependence ; U-statistic

Abstract

In biomedical applications,interest focuses on occurrence of two or more consecutive failure events and association between event times.Bivariate survival data with interval sampling arise frequently when disease registry commonly collect data with incidence of disease occurring within a calendar time interval. The initiating event is retrospectively confirmed and subsequent failure event is observed during follow-up. Such data represent a non-randomly screened subset of a population and the interval sampling bias needs to be properly adjusted for in analysis. Similar to truncated survival data, the analysis method for this type of data relies on assumption of independence, that is, the disease process does not depend on when the initiating event occurs. This paper proposes a nonparametric test of a weaker but sufficient assumption of quasi-independence based on a coordinatewise conditional Kendall's tau. Further, to quantify dependence between bivariate failure times given quasi-independence,a nonparametric estimator of tau that uses inverse probability weights is developed. Simulations are conducted and application to cancer data are presented for illustration.