THURSDAY, February 19, 2004
Time: 12:20 - 1:20 PM
Room Constant 1052

Russell Stocker
Department of Statistics
University of South Carolina
Columbia, SC 29208

A general class of parametric models for recurrent event data given by Peña and Hollander (2003) is considered. The class of models includes many of the models found throughout the literature. These include the imperfect repair model of Brown and Proschan (1983), the general Cox proportional hazards model, and the general repair model of Dorado, Hollander, and Sethuraman (1997). The model is based on using a flexible multiplicative intensity process. It allows the researcher to take into account the effect of interventions applied to a unit after an event of interest has occurred by perturbing the baseline intensity process with an “effective age” process. The influence of outside factors through the use of a link function with covariates is included. The inclusion of a frailty component is allowed for modeling unobservable random effects. Estimation schemes are given under the cases in which a frailty is assumed not to exist and when it is assumed to be present. When a frailty is not present the limiting distribution of the properly standardized parameter estimators are Gaussian. The sampling distribution properties are examined through computer simulation. When a frailty is assumed to exist the EM algorithm is utilized to obtain parameter estimators. The recurrence of bladder cancer data given by Wei et al (1989) is analyzed using the class of models both under the assumption that there is a frailty and that a frailty is not present.