BAYESIAN MODELLING OF TUBERCULOSIS PREVALENCE IN SOUTH AFRICA. AN OVER-DISPERSION STUDIES
Davies Abiodun Obaromi, Samuel Olorunfemi Adams
This is an open access article distributed under the Creative Commons Attribution License CC BY 4.0, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited
Introduction: To account for over-dispersion in the impact of population density, number of schools, and average household size on tuberculosis prevalence in South Africa’s Eastern Cape Province, this study compared robust count regression models with a Poisson regression model. Methods: The prevalence of tuberculosis in the province of Eastern Cape in South Africa in 2022 served as the basis for this study’s data. Various models, including the Poisson regression model; Summed up Poisson and Negative Binomial Relapse were used within the sight of over-scattering. Results: It was seen that Poisson Relapse models couldn’t deal with over-scattering present in the South Africa’s Eastern Cape Region Tuberculosis dataset. Modeling with Negative Binomial Regression (NBR) and Generalized Binomial Regression (GBR) was used to solve the over-dispersion issue because both approaches can accommodate the dispersion parameter. Based on both selection criteria, the Negative Binomial model performed slightly better than the Generalized Poisson model. The findings indicated that population density has a positive and significant impact on tuberculosis prevalence in South Africa, that the number of schools in the province has a negative but significant impact, and that the average household has no significant impact. Conclusions: The Eastern Cape Province of South Africa’s government ought to intensify its efforts to combat the spread and prevalence of tuberculosis by identifying, isolating, and treating tuberculosis early in the population and in schools.