Modelos de sobrevivência na presença de eventos recorrentes e longa duração

Abstract

In this thesis it is proposed to analyze recurrent event data, recurrent event data with cure fraction and recurrent event data with censoring and competing causes. For the recurrent event data analysis it is proposed a multiple time scale survival model, which includes several particular cases. For recurrent event data with a cure fraction we consider a multiple time scale survival models embedded on a mixture cure fraction modeling. It is also proposed a general model to survival data in presence of competitive causes. In this case, it is assumed that the number of competitive causes follows a generalized negative binomial distribution. While, for the time of occurrence of each cause, a Weibull and a log-logistic distribution were considered. Simulations studies were conducted for every proposed model in order to analyze the asymptotical properties of the estimation procedures. Both, maximum likelihood and Bayesian approaches were considered for parameter estimation. Real data applications demonstrate de use of the proposed models.