Identifying risk factors of aortic valve replacement using frailty models

Abstract

Traditional survival modeling techniques, including the Kaplan Meier estimator, Cox regression and parametric survival models assume a fairly homogeneous population, where variation in survival durations can be explained by a small number of observed explanatory variables. However, in the presence of heterogeneity, frailty models are more appropriate to model survival data by introducing random effects that account for the variability generated from unobserved covariates. This paper presents two types of frailty models. The unshared frailty model assumes that different individuals have distinct frailties, while the shared frailty model assumes that the population can be divided into clusters, where members in the same cluster share the same frailty. Due to their nice mathematical properties, the Gamma and the Inverse Gaussian distributions are the most popular choices for the frailty distribution.