Premium
Frailty proportional mean residual life regression for clustered survival data: A hierarchical quasi‐likelihood method
Author(s) -
Huang Rui,
Xiang Liming,
Ha Il Do
Publication year - 2019
Publication title -
statistics in medicine
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.996
H-Index - 183
eISSN - 1097-0258
pISSN - 0277-6715
DOI - 10.1002/sim.8338
Subject(s) - proportional hazards model , residual , computer science , hierarchical database model , multilevel model , statistics , regression , likelihood function , accelerated failure time model , econometrics , maximum likelihood , data mining , mathematics , machine learning , algorithm
Frailty models are widely used to model clustered survival data arising in multicenter clinical studies. In the literature, most existing frailty models are proportional hazards, additive hazards, or accelerated failure time model based. In this paper, we propose a frailty model framework based on mean residual life regression to accommodate intracluster correlation and in the meantime provide easily understand and straightforward interpretation for the effects of prognostic factors on the expectation of the remaining lifetime. To overcome estimation challenges, a novel hierarchical quasi‐likelihood approach is developed by making use of the idea of hierarchical likelihood in the construction of the quasi‐likelihood function, leading to hierarchical estimating equations. Simulation results show favorable performance of the method regardless of frailty distributions. The utility of the proposed methodology is illustrated by its application to the data from a multi‐institutional study of breast cancer.