Competing Risk Data Subject to Interval Censoring in Survival Analysis

• Activity Number: 573
• Type: Contributed
• Date/Time: Wednesday, August 1, 2012 : 2:00 PM to 3:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Grace Liu*+ and Lejia Lou
• Companies: Johnson & Johnson and Penn State University
• Keywords: Survival analysis ; Competing Risk ; Cumulative Incidence Function (CIF) ; cause specific hazard functions (CSHF) ; counting process and the martingale central limit theory ; unconstrained full likelihood (UFL) and constrained full likelihood (CFL)

Abstract

In the presence of competing risks, the traditional survival methods like Kaplan-Meier may cause bias. Gray's (1988) first introduced a test for comparing the Cumulative Incidence Function (CIF) for handling the competing risk factor, which is based on comparing the weighted averages of the cause specific hazard functions (CSHF). Nyangweso et al (2000) discussed the parametric method, i.e. Naïve estimator of CIF for handling the interval censored data with the competing risk presented. Their works is the extension of Prentice et al., (1978), which is direct modeling of CIFs. By directly specifying the parametric models for CIFs, the estimations are obtained via maximizing the unconstrained full likelihood (UFL) and constrained full likelihood (CFL). In this presentation, we propose a nonparametric method for estimating the CIF for an event of interest in the presence of the competing risk. A real data example will be used to compare the estimates by applying the proposed method as well the methods discussed above and the Kaplan-Meier approach to demonstrate the importance of appropriately estimating the cumulative incidence of an event of interest in the presence of competing risk