Joint analysis of longitudinal and interval-censored failure time data

Abstract

Methods for joint analysis of longitudinal measures and right-censored survival outcomes have received much attention in the literature. However, in clinic and epidemiology research, the event of interest is only examined at the inspection time, resulting in the interval-censored failure time data. In this article, we construct a joint model for the longitudinal outcomes and the failure time, which is subject to interval censoring. We account for the dependence between the longitudinal measures and failure time through the incorporation of random effects. The semiparametric regression model with random effects and the Cox frailty model is used for the longitudinal outcome and the failure time, respectively. Spline functions are applied to approximate the infinite-dimensional baseline mean function and baseline cumulative hazard function in the joint model. The EM algorithm is used to find the maximum likelihood estimators and the estimated standard deviations are computed from the log observed-data likelihood. Extensive simulation studies are conducted to assess the performance of the proposed estimators. Simulation results show that the proposed method performs quite well. Finally, we illustrate the proposed approach by analyzing the chronic kidney disease.