Joint Modeling of Marked Recurrent Event Data in the Presence of a Terminating Event

• Activity Number: 74
• Type: Contributed
• Date/Time: Sunday, July 29, 2012 : 4:00 PM to 5:50 PM
• Sponsor: Biometrics Section
• Author(s): Yu Ma*+ and Douglas E Schaubel
• Companies: University of Michigan and University of Michigan
• Keywords: Proportional hazards model ; Proportional rates model ; Recurrent event ; Frailty ; Marks

Abstract

Typically in observational studies, it is necessary to account for measured and unmeasured heterogeneity across study subjects. This is often accomplished through model covariates (for measured factors) and frailty variates (to account for unmeasured predictors). We investigate a frequently occurring data structure where the event of interest is recurrent (e.g., hospital admission), marks are observed upon the occurrence of each event (e.g., length of stay) and the recurrent event process may be permanently stopped by a terminating event (e.g., death). The proposed methods utilize a form of hierarchical modeling: a proportional hazards model for the terminating event; a proportional rates model for the conditional recurrent event rate given survival; and a poisson regression modeling approach for the marks, given an event has occurred. Maximum likelihood based estimation is carried out via Gaussian Quadrature technique for integration. Through simulation, the methods are shown to work well for practical sample sizes. Significant biases are detected under model misspecification, especially when heterogeneity is large. We apply the proposed methods to end-stage renal disease data.