Using Bayesian Evidence to Estimate District‐Wise Tuberculosis Incidence Rates, Missed Cases and Fraction of Treated Cases in Bangladesh

Undetected and untreated tuberculosis (TB) remains a serious public health challenge, which increases mortality and transmission within households and communities. This study estimates district-wise TB incidence rates, missed cases, and the fraction of treated cases in Bangladesh using a Bayesian approach. By integrating data from national TB notifications, TB-related mortality, prevalence surveys, and auxiliary sources in Bangladesh, we developed an analytical model that estimates district-level TB burden from 2008 to 2017. We used a Bayesian modelling approach, treating incidence as a hidden event inferred from other components. We included predictor variables in the analysis at two points in the model-incidence and fraction notified. Here, we estimate the number of missing TB cases across districts, allowing us to identify areas with high proportions of untreated cases. Our findings revealed notable spatial and temporal variation, with some districts showing consistently higher predicted incidence rates and others exhibiting lower detection or treatment coverage. However, treatment coverage showed clear regional disparities, with poorer districts recording substantially lower fractions of treated cases compared to wealthier districts. These findings highlight the need for policymakers to prioritise equitable resource allocation to mitigate the burden of TB in Bangladesh.