Individual Submission Summary
Share...

Direct link:

An Introduction to Frailty Models for Multivariate Survival Data with Application to Martial Dissolution and Retirement

Sat, August 11, 10:30 to 11:30am, Philadelphia Marriott Downtown, Floor: Level 5, Salon G

Abstract

A frailty model is a random effect survival model, where the random effect has a multiplicative effect on the baseline hazard. Frailty models can handle multivariate (dependent) failure times, such as clustered or repeatedly measured survival times, and recurrent events, which refers to the same type of event that occurs to an individual multiple times. In this article, we introduce a joint frailty model, and discuss its underlying assumptions. The model is illustrated by an analysis of a Canadian national survey data with observations of two recurrent events in mid and later life: Marital dissolution/widowhood and retirement. Our analysis shows that the risks of these events are related: the timing of retirement changes with the risk of union dissolution. We provide new insights into studying two simultaneous survival processes and present a joint frailty model that can produce unbiased and efficient parameter estimates.

Author