Modeling cure fraction with frailty term in latent risk: a Bayesian approach
In this paper, we propose a flexible cure rate model with frailty term in latent risk, which is obtained by incorporating a frailty term in risk function of latent competing causes. The number of competing causes of the event of interest follows negative binomial distribution and the frailty variable follows power variance function distribution, in which includes other frailty models such as gamma, positive stable and inverse Gaussian frailty models as special cases. The proposed model takes into account the presence of covariates and right-censored survival data suitable for populations with a cure rate. Besides, it allows quantifying the degree of unobserved heterogeneity induced by unobservable risk factors, in which is important to explain the survival time. Once the posterior distribution has not close form, Markov chain Monte Carlo simulations are considered for estimation procedure. We performed several simulation studies and the practical relevance of the proposed model is demonstrated in a real data set.