Abstract Tuberculosis remains a major public health concern in Brazil, particularly in the Amazon region, where its distribution is shaped by persistent social inequalities and environmental pressures. This study analyzed the spatial and sociodemographic dynamics of tuberculosis in the state of Pará, Northern Brazil, between 2018 and 2022. A total of 26,127 confirmed cases were identified through the Notifiable Diseases Information System and the National Registry of Health Establishments. We applied empirical Bayes smoothing and Kulldorff’s spatial scan statistic to detect high-risk territorial clusters and used multilevel logistic regression to assess individual and contextual predictors of unfavorable treatment outcomes. The disease primarily affected men, young adults aged 15-49 years, individuals with low educational attainment, and those self-identified as of mixed race. Spatial analysis revealed persistent high-risk clusters in the Metropolitan Region of Belém and a progressive expansion into Southeastern and South-Central municipalities. Lower education and Black or mixed-race identity were independently associated with a higher likelihood of unfavorable outcomes. The findings underscore the importance of integrating spatial epidemiology with multilevel modeling to uncover both individual and territorial determinants of tuberculosis. They highlight the need for geographically targeted interventions, the expansion of primary care services, and intersectoral public policies aimed at mitigating structural vulnerabilities in the Brazilian Amazon. Our results provide evidence to support precision public health strategies and enhance local planning in regions of high social and environmental risk.
Sales et al. (Wed,) studied this question.