Residuals in Survival Models with cure fraction
Diagnostico. Resíduo de Cox-Snell. Fração de cura; Tempo de vida
Cure fraction survival models represent a generalization of classical survival models, which
allow the occurrence of a fraction of cured individuals in the population studied. Two well-
known parametric cure fraction survival models are the mixture cure models and the
promotion time models. After the fit of a model to a dataset, it is important to validate the
assumptions and the residuals analysis is an important technique for this purpose. Although
residuals analysis is widely used to assess the fit of usual survival models, there are few
residuals proposed for survival data with cure fraction. This work aims to study a recent
proposal of the literature that evaluates, through Cox-Snell type residues, the global
adjustment in mixture cure models and extend this proposal to promotion time model. We
considered a simulation study to evaluate the performance of the proposed residuals. An
application is considered to illustrate the procedures studied.