Computer Methods and Programs in Biomedicine Update · Published 2025-01-01 · DOI 10.1016/j.cmpbup.2025.100222
Cure models are essential in survival analyses because they take into account the fraction of patients who achieve long-term survival and will not experience disease-related death. In a relative survival framework, these models estimate the excess hazard (i.e. the additional death risk attributable to the disease), while assuming that some patients will not experience such excess risk. In 2002, Phillips and colleagues introduced a class of cure models where background mortality is modelled as a scaled version of the general population’s mortality. This tutorial presents the curesurv R package for parametric cure modelling in the relative survival setting, with and without background mortality rescaling. The package supports both the Weibull mixture cure model and the non-mixture cure model proposed by Boussari and colleagues in 2018, offering flexible tools for assessing long-term survival. The curesurv R package application is illustrated using testicular, breast, and prostate cancer datasets from the French FRANCIM cancer registries. This tutorial provides a practical guide to researchers in epidemiology and health economics for implementing cure fraction models and interpreting long-term survival estimates.
Abstract from DOAJ. Public domain (CC0 1.0).
Read the article at the publisher →