A flexible cure rate model for spatially correlated survival data based on generalized extreme value distribution and Gaussian process priors

A shared spatial model for multivariate extreme-valued binary data with non-random missingness

Clinical studies and trials on periodontal disease (PD) generate a large volume of data collected at various tooth locations of a subject. However, they present a number of statistical complexities. When our focus is on understanding the extent of extreme PD progression, standard analysis under a generalized linear mixed model framework with a symmetric (logit) link may be inappropriate, as the binary split (extreme disease versus not) maybe highly skewed. In addition, PD progression is often hypothesized to be spatially-referenced, i.e. proximal teeth may have a similar PD status than those that are distally located. Furthermore, a non-ignorable quantity of missing data is observed, and the missingness is non-random, as it informs the periodontal health status of the subject. In this paper, we address all the above concerns through a shared (spatial) latent factor model, where the latent factor jointly models the extreme binary responses via a generalized extreme value regression, and the non-randomly missing teeth via a probit regression. Our approach is Bayesian, and the inferential framework is powered by within-Gibbs Hamiltonian Monte Carlo techniques. Through simulation studies and application to a real dataset on PD, we demonstrate the potential advantages of our model in terms of model fit, and obtaining precise parameter estimates over alternatives that do not consider the aforementioned complexities.

Generalized cure rate model for infectious diseases with possible co-infections

This research mainly aims to develop a generalized cure rate model, estimate the proportion of cured patients and their survival rate, and identify the risk factors associated with infectious diseases. The generalized cure rate model is based on bounded cumulative hazard function, which is a non-mixture model, and is developed using a two-parameter Weibull distribution as the baseline distribution, to estimate the cure rate using maximum likelihood method and real data with R and STATA software. The results showed that the cure rate of tuberculosis (TB) patients was 26.3%, which was higher than that of TB patients coinfected with human immunodeficiency virus (HIV; 23.1%). The non-parametric median survival time of TB patients was 51 months, while that of TB patients co-infected with HIV was 33 months. Moreover, no risk factors were associated with TB patients co-infected with HIV, while age was a significant risk factor for TB patients among the suspected risk factors considered. Furthermore, the bounded cumulative hazard function was extended to accommodate infectious diseases with co-infections by deriving an appropriate probability density function, determining the distribution, and using real data. Governments and related health authorities are also encouraged to take appropriate actions to combat infectious diseases with possible co-infections.

A Bayesian piecewise survival cure rate model for spatially clustered data

This paper proposes a Bayesian hierarchical cure rate survival model for spatially clustered time to event data. We consider a mixture cure rate model with covariates and a flexible (semi)parametric baseline survival distribution for uncured individuals. The spatial correlation structure is introduced in the form of frailties which follow a Multivariate Conditionally Autoregressive distribution on a pre-specified map. We obtain the usual posterior estimates, smoothed by regional level maps of spatial frailties and cure rates. A simulation study demonstrates that the parameters of the models with spatially correlated frailties have smaller relative biases and MSE than the ones obtained using simple frailty models. We apply our methodology to Hodgkin lymphoma cancer survival times for patients diagnosed in the state of Connecticut.