A0288
Title: Estimating the cure rate in a mixture cure model using presmoothing
Authors: Maria Amalia Jacome Pumar - Universidade da Coruna (Spain) [presenting]
Ana Lopez-Cheda - University of A Coruna (Spain)
Samuel Saavedra - University of A Coruna (Spain)
Abstract: The mixture cure model in survival analysis has received large and growing attention in the last few decades. For diagnostic and prognostic purposes, an important aspect of the mixture cure model is the estimation of the probability that an individual is cured, that is, belongs to the cured component of the population. There is no reason to believe that the cure rate is always monotone, let alone that it is logistic. Hence, we consider the nonparametric estimator for the cure rate function. It is given by the Kaplan-Meier survival estimator (or the generalized Kaplan-Meier estimator, with covariates) evaluated at the largest uncensored time. Presmoothing has been shown to improve survival function estimation, by replacing the indicators of no censoring with some preliminary nonparametric estimator of the conditional probability of uncensoring. The effect of presmoothing in the estimation of the cure rate will be studied, and the resulting methods will be applied to a real medical database.
