A0835
Title: Analyzing bivariate survival data using pseudo-observations
Authors: Sy Han Chiou - Southern Methodist University (United States) [presenting]
Chien-Lin Su - McGill University (Canada)
Feng-Chang Lin - University of North Carolina at Chapel Hill (United States)
Abstract: Copula models have become increasingly popular in various fields as they provide effective tools for modeling correlated responses. In modeling multivariate survival data, copula models offer flexibility by enabling users to specify both the marginal survival functions and the association structure between them. A semiparametric transformation model is used to define the marginal survival functions, and a conditional Archimedean copula is considered to address the associations among different types of survival times. To expedite computation, pseudo-observations for both the marginal survival and association components are introduced, and inference is implemented using generalized estimating equation techniques. Additionally, variable selection and goodness-of-fit tests are explored to aid in the selection of appropriate copula models. The effectiveness of the proposed methods is demonstrated through extensive simulations.
