Joint spatiotemporal modelling of tuberculosis and human immunodeficiency virus in Ethiopia using a Bayesian hierarchical approach

Abstract

The aim of this paper was to evaluate the distribution of HIV and TB in Ethiopia during four years (2015-2018) at the district level, considering both spatial and temporal patterns. Methods Consolidated data on the count of TB case notifications and the number of patients with HIV for four years, 2015-2018, were provided by the Ethiopian Federal Ministry of Health. The data was analyzed using the Bayesian hierarchical approach, employing joint spatiotemporal modelling. The integrated nested Laplace approximation available in the R-INLA package was used to fit six models, each with different priors, for the precision parameters of the random effects variances. The best-fitting model with the best predictive capacity was selected using the Deviance Information Criterion and the negative sum of cross-validatory predictive log-likelihood.