Use this URL to cite or link to this record in EThOS:
Title: On the generalized time dependent logistic family of survival models
Author: Al-Tawarah, Yasin.
ISNI:       0000 0001 3409 2124
Awarding Body: University of Keele
Current Institution: Keele University
Date of Award: 2004
Availability of Full Text:
Access from EThOS:
The main goal of this thesis was to develop the Generalized Time Dependent Logistic (GTDL) family of survival models introduced by (MacKenzie, 1996). A Logistic Proportional Hazard parametric model (LPH) and three Logistic Non-proportional Hazards parametric models (LAL, LAH and CTDL) have been developed. The mathematical derivation of these wholly parametric models is given and their key properties described. An advantage of these parametric models, is that the survivor function is of closed form and when the regression parameter, {3, is zero, the underlying models have testable parametric forms. This led to the investigation of the application of these models to analyzing the type of interval censored data which arises in longitudinal randomized controlled trials. A simulation study was conducted to investigate how inference on the treatment parameter ({3) might be compromised by the use of the mis-specified likelihood when the "true" likelihood obtained. Overall, the findings show for such sampling schemes that use of the true likelihood is to be preferred. Model checking methods including simple plotting techniques, analysis of strata and analysis of residuals were investigated for the four models. The properties of each method are discussed in detail. The utility of the four models was investigated by analyzing survival data from the Northern Ireland lung cancer study and the statistical findings and their clinical interpretation are discussed. The goodness of fit of these models was compared with the non-parametric Kaplan & Meier estimator and the semi-parametric Cox model. For these data, the LPH model was recommended as the first model of choice followed by the LAL, CTD Land LAH models .
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available