Joint frailty modelling of time-to-event data with recurrent and terminal events

Shu-Kay Angus Ng (Griffith University)

Wednesday 30th August, 2023 11:00-12:00 Maths 311B

Abstract

We present an innovative perspective on analysing longitudinal data, within a statistical framework of survival analysis of time-to-event recurrent data, in order to elicit the evolution pathway of events of interest. The proposed methodology is based on a joint frailty modelling approach via a generalised linear mixed model (GLMM) formulation to account for the heterogeneous risk of failure and the presence of informative censoring due to a terminal event. The GLMM approach considers two correlated random effects to jointly model the dependence between the hazard rates of recurrent events and the terminal event. Efficient estimation of model parameters is achieved using an extended BLUP plus REML procedures for survival analysis of censored data with recurrent and terminal events. We will demonstrate the capacity of our method using a cancer registry data set of melanoma patients in Australia, regarding the occurrence of secondary primary cancer after the diagnosis of melanoma. We will also present the relative performance of the proposed joint frailty model to a standard frailty model via simulation studies.

Add to your calendar

Download event information as iCalendar file (only this event)