Analyzing longitudinally correlated failure time data : a generalized estimating equation approach

Hasan, Md. Tariqul (2001) Analyzing longitudinally correlated failure time data : a generalized estimating equation approach. Masters thesis, Memorial University of Newfoundland.

[img] [English] PDF - Accepted Version
Available under License - The author retains copyright ownership and moral rights in this thesis. Neither the thesis nor substantial extracts from it may be printed or otherwise reproduced without the author's permission.

Download (2MB)


Multivariate correlated failure time data can be classified into two different groups: structural failure time data and longitudinal failure time data. As compared to the analysis of the structural failure time data, the analysis of longitudinal failure time data has however proven to be difficult, perhaps because of the difficulty in the modeling the true longitudinal correlation structures. In the present thesis, following certain longitudinal correlation models, recently developed for discrete data, we develop three longitudinal correlation models for exponential failure times to deal with such multivariate longitudinal data. Under these three models, we construct the co-variance structures of the martingales of the failure times for both uncensored and censored cases, and use them to develop a generalized estimating equation approach to estimate the parameters of main interest, namely, the hazard ratio parameters. The efficiency loss due to misspecification of the correlation structure is studied for both uncensored and censored cases. As the proposed generalized estimating equation approach use either the underlying true correlation structure for both uncensored and censored cases or a suitable robust correlation structure for the uncensored case, the methodology yields consistent as well as efficient estimators for the hazard ratio parameters. We apply the methodology to a numerical example.

Item Type: Thesis (Masters)
Item ID: 8608
Additional Information: Bibliography: leaves 87-89.
Department(s): Science, Faculty of > Mathematics and Statistics
Date: 2001
Date Type: Submission
Library of Congress Subject Heading: Failure time data analysis

Actions (login required)

View Item View Item


Downloads per month over the past year

View more statistics