Analysis of Multivariate Survival Data Under Semiparametric Copula Models with/Without Measurement Error

• Activity Number: 477
• Type: Topic Contributed
• Date/Time: Wednesday, August 12, 2015 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Wenqing He* and Grace Yi and Naisyin Wang
• Companies: University of Western Ontario and University of Waterloo and University of Michigan
• Keywords: Multivariate survial analysis ; Measurement error ; Copula model ; Semiparametric

Abstract

There has been extensive research on univariate survival data with covariate measurement error. However, there is limited attention on the impact of covariate measurement error on analysis of multivariate survival data. In this paper, we discuss semiparametric linear transformation marginal models for multivariate survival data, with covariate measurement error incorporated. A nonparametric approach is employed to approximate the marginal transformation function, and a two stage maximum likelihood method and a working three stage estimation are developed. The impact of misspecification of the joint copula model on estimation of covariate effects is investigated, and a simulation-based method is explored to correct for measurement error effect on parameter estimation. Extensive simulation studies are conducted to assess the performance of the proposed methods for a variety of scenarios, and a real study is analyzed with the proposed methods.