R

Setup

Method: Mixture cure models with age and year as continuous covariables In the following models, we assumed a non-linear effect of age at diagnosis both on the survival probability for the uncured and on the probability of cure. However, for the year of diagnosis, we fitted 4 models, which vary depending on linear or non-linear effect assumed for the year of diagnosis on either the survival probability of the uncured or the probability of cure.

This is an example of analysis that can be done using the R-package mexhaz. This package allows to fit flexible hazard model in the framework of overall survival or net survival (i.e. excess hazard), with/wihout time-dependent effects for covariates, with/without non-linear effect and with/without random effect defined at the cluster level. Start The first step is to install the mexhaz package from the CRAN website. It needs to be done only the first time you’re using mexhaz, as the next time, mexhaz will be automatically pre-installed in your R session.