Abstract

Background Residential segregation is considered a social determinant of health, but there is limited evidence of its impact on tuberculosis (TB). We investigated the associations between municipality-level income and racial segregation and TB treatment outcomes in Brazil. Methods We studied nationwide registries of new TB cases between 1 January 2010 and 31 December 2019. TB treatment was dichotomised as unfavourable (ie, loss to follow-up, modification of treatment regimen, treatment failure and death) and favourable (ie, cured/treatment completion). We assessed individuals' municipality-level income and racial segregation (ie, dispersion of household heads earning ≤half versus those earning >half minimum wage; and of household heads identifying as black or brown/mixed race (Pardo/a) versus white). Logistic regression adjusted for sociodemographic and clinical variables was used to estimate the OR of experiencing an unfavourable treatment outcome associated with segregation overall and by self-identified race/ethnicity. Results Individuals living in highly economically and racially segregated municipalities (fifth versus first quintiles) were more likely to have an unfavourable TB treatment outcome (income segregation: adjusted OR 1.34 (95% CI 1.31 to 1.37); racial segregation: 1.13 (0.94 to 1.36)). Living in municipalities of higher income segregation (third, fourth and fifth quintiles) was associated with higher unfavourable TB treatment outcomes in all self-identified racial groups (fifth quintile: white 1.25 (0.96 to 1.64); black 1.42 (1.15 to 1.74); brown/mixed 1.37 (1.20 to 1.56); Asian=1.30 (1.00 to 1.69) and Indigenous 1.37 (1.00 to 1.87)). Conclusions Living in highly income and racially segregated environments is associated with unfavourable TB treatment outcomes for all self-identified races in Brazil. TB programmes should account for segregation as a barrier to TB treatment completion.