A0244
Title: A bivariate time-varying copula joint model for longitudinal measurements and time-to-event data
Authors: Zili Zhang - The University of Manchester (United Kingdom) [presenting]
Abstract: A bivariate time-varying copula joint model, which models the repeatedly measured longitudinal outcome at each time point with the survival data, jointly by both the random effects and bivariate copulas, is proposed. A regular joint model normally supposes some subject-specific latent random effects or classes shared by the longitudinal and time-to-event processes. They are assumed to be conditionally independent, given these latent random variables. Under this assumption, the joint likelihood of the two processes can be easily derived, and the unobservable latent random variables naturally introduce the association between them and heterogeneity among the population. However, because of the unobservable nature of these latent variables, the conditional independence assumption is difficult to verify. Therefore, a bivariate time-varying copula is introduced into a regular joint model to account for the cases where there could be an extra association between the two processes which the latent random variables cannot capture. The proposed model includes a regular joint model as a particular case when the correlation function under the bivariate Gaussian copula is constant at zero. Simulation studies and dynamic predictions of survival probabilities are conducted to compare the performance of the proposed model with the regular joint model, and a real data application on the Primary biliary cirrhosis (PBC) data is performed.
