Date of Award
Doctor of Philosophy
Dr. Gerald L. Sievers
Dr. Michael Stoline
Dr. Dennis Pence
Dr. Thomas Vidmar
A model for the natural history of a progressive disease is developed. The model has three disease states and can be expressed as the joint distribution of two survival random variables.
Covariate information is incorporated into the model using the proportional hazards model for the marginal distributions. The model will also accommodate data with observations which are censored on one or both of the survival random variables.
The likelihood function for censored data is exhibited for finding the maximum likelihood estimates of the parameters and their standard errors for testing the effects of the covariates. The method used to obtain these estimates is the maximum likelihood method. Typical epidemiological measures are written in terms of the parameters of this model. Potential application of this model to cancer and heart disease research is discussed.
Yahya, Hilmi F., "A Progressive Disease Model for Doubly-Censred Bivariate Survival Data that Accommodates Covariate Information" (1992). Dissertations. 1986.