- This event has passed.
Seminar on Sunday, 21 October 2018 at 2 pm
October 21, 2018 @ 2:00 pm - 3:30 pm
Title: Joint Modeling of Longitudinal Response and Time-To-Event Data Using Conditional Distributions: A Bayesian Perspective
Speaker: Arindom Chakraborty, PhD
Assistant Professor
Department of Statistics
Visva-Bharati University, India
Abstract: Over last twenty or more years a lot of methodological development and clinical application of joint models of longitudinal and time-to-event outcomes have come up. In these studies patients are followed until an event, such as death occurs. In most of the articles, using subject specific random effects as frailty, the dependency of these two processes has been established. In this article, we propose a new joint model that consists of a linear mixed effects submodel for longitudinal data and an accelerated failure model for the time-to-event data. These two submodels are linked together by not only latent random process, but also by the conditional distributional assumption. This model will capture the dependency of the time-to-event on the longitudinal measurements more directly. Using standard priors, a Bayesian method is developed for estimation. All computations based on the Bayesian inference via MCMC is implemented using OpenBugs. Our proposed method is evaluated by a simulation study which shows the efficiency of the conditional model over two other models: joint models with local independence and independent models. One clinically motivating data on Duchenne muscular dystrophy (DMD) syndrome and a popular data on AIDS are also analyzed.
