Mixed Effects Gompertz Model of Clustered Survival Data in Presence of Cure

• Activity Number: 531
• Type: Contributed
• Date/Time: Wednesday, August 7, 2013 : 10:30 AM to 12:20 PM
• Sponsor: Biometrics Section
• Author(s): Chien-Lin Su*+ and A. Adam Ding and Weijing Wang
• Companies: Institute of Statistics,National Chiao-Tung University and Department of Mathematics, Northeastern University and Institute of Statistics, National Chiao-Tung University
• Keywords: Gompertz distribution ; MCEM algorithm ; Mixture cure model ; Random effects

Abstract

We propose to analyze clustered survival data in presence of cure by a mixed-effect Gompertz model. Most literature on analysis of such data adopts the mixture approach which divides the population into two distinct groups of being "susceptible" or "cured". Here we propose a different approach which directly models the overall distribution function by the Gompertz distribution with two parameters as functions of fixed and random covariates. An important feature of the proposed model is that "cure" is only a possibility rather than a deterministic status. Furthermore our model allows that cured individuals only exist in some covariate groups. The maximum likelihood estimates can be obtained via a Monte Carlo Expectation Maximization (MCEM) algorithm. Simulation studies are performed to evaluate the finite-sample performances of the proposed methods. The proposed methods are applied to analyze two real datasets for illustration.