smcure: An R-package for estimating semiparametric mixture cure models

Abstract

The mixture cure model is a special type of survival models and it assumes that the studied population is a mixture of susceptible individuals who may experience the event of interest, and cure/non-susceptible individuals who will never experience the event. For such data, standard survival models are usually not appropriate because they do not account for the possibility of cure. This paper presents an R package smcure to fit the semiparametric proportional hazards mixture cure model and the accelerated failure time mixture cure model.

Macro-Economic Factors in Credit Risk Calculations: Including Time-Varying Covariates in Mixture Cure Models

“…This is done non-parametrically using the Breslow-type estimator for S 0 (t) and combining the results of Andersen (1992) and Cai et al (2012a).…”

“…A widespread method for estimating variances in the mixture cure context is bootstrapping (e.g. Peng, 2003;Cai et al, 2012a;Tong et al, 2012). While easy to implement, this method is computationally expensive, especially with big data sets and a slow convergence of the EM-algorithm.…”

“…While using parametric survival distributions in the survival component initially, Kuk and Chen (1992) extended the mixture cure model using nonparametric survival distributions (see also Peng and Dear, 2000;Sy and Taylor, 2000). Cai et al (2012b) introduced the smcure-package in R (R Core Team, 2013) to estimate semi-parametric mixture cure models. This latter version of the mixture cure model was introduced in the credit risk context by Tong et al (2012).…”

A multiple imputation approach for semiparametric cure model with interval censored data

“…There are two components in the mixture cure model: the cure rate component, For right censored data, an EM algorithm was proposed by Peng [14] for fitting the mixture cure model, and an R package "smcure" was contributed by Cai [2] to realize this method.…”

“…where t (i) is the ith ordered event times with t (1) ≤ t (2) ≤ · · · ≤ t (k) , if we have total of k events. d i and n i are number of events and number of cases at risk at time t (i) .…”