Bayesian Analysis of Tuberculosis Cases in Bolgatanga Municipality, Ghana–West Africa

Abstract Tuberculosis (TB) represents a major threat to global health security and is currently the world most deadly infectious disease. TB is noted to affect people of all age groups and exposure to the disease has been shown to exhibit variation in the risk of infection across different factors. This study was therefore carried out to identify possible disease associations of some risk factors in relation to time until onset of TB among TB infected individuals in the Bolgatanga Municipality of the Upper East Region of Ghana. To this end, secondary data on time until onset of TB was obtained from the Bolgatanga Regional Hospital of the Upper East Region of Ghana and used for statistical analysis. In order to obtain appropriate underlying disease associations, the Bayesian approach to the modelling of survival time was employed in the statistical analyses. Thus, different parametric models including; skew normal, lognormal, shifted lognormal, exponential, Weibull and Fréchet models were fitted to the TB data within the Bayesian framework. The Weibull was the overall best fitted model. Based on the estimation results of the Bayesian Weibull regression model; individuals with no education, individuals with diabetes and males were identified as high-risk groups with shorter time frame for the onset of TB in the Bolgatanga Municipality. The study was concluded by indicating the need for Ghana Health Service to institute targeted interventions aimed at preventing and controlling the TB with special emphasis to the high-risk groups in the Bolgatanga Municipality.