Online Program

A Study on Cured Fraction Models with Unobserved Heterogeneity (frailty) in Survival Data (306647)

Keywords: Cure rate model, Frailty model, Gamma frailty distribution, cured fraction model, long term survivors, mixture cure models

This paper consider modeling a right censored time to event data with certain fraction in the population cured, while the remaining not cured having unobservable information under the purview of Diabetic Retinopathy Study (DRS) data. We extend the mixture cure model by incorporating frailty among the fraction under risk and propose a maximum likelihood estimation method for the model. The research looks into the Gamma and Inverse Gaussian frailty mixture models and examines the performances of the proposed models to the standard cure model. We apply the models to the Diabetic Retinopathy Study (DRS) dataset.