Spatial frailty survival model for infection of tuberculosis among HIV infected individuals.


Srinivasan, R.; Ponnuraja, C.; Rajendran, S.; Venkatesan, P.


SSRG International Journal of Medical Science; 2017; 4; 5-10.

Abstract: Capturing spatial variation is important while studying disease survival in different geographical regions where regions has its own risk factor associated with disease. To account this risk factor, frailty effect was introduced in this model that captures correlation and variation between neighboring locations using conditionally autoregressive (CAR) prior in Bayesian parametric survival model for studying dual infection of tuberculosis and HIV. National Institute for Research in Tuberculosis data on tuberculosis and HIV were used in this study. Monte Carlo Markov Chain (MCMC) technique was used to estimate the parameter. WinBUGS software was used for Bayesian Survival model estimation. The result of the study revealed that the spatial frailty model accounts higher heterogeneity along with weight at baseline was one of the significant factors associated with death and conclude that there were unmeasured covariates and risk factors influencing death in the Chennai regions.

Keywords: Bayesian spatial; Survival; CAR; Weibull; random effects

Back to List of publications / Home