The effect of incarceration on TB treatment outcomes.

BACKGROUND: TB notifications in Latin American prisons have more than doubled over the past two decades; however, treatment outcomes and their determinants among incarcerated individuals in this region are not well understood.METHODS: Newly diagnosed drug-susceptible TB cases reported to Brazil´s Information System for Notifiable Diseases (Sistema de Informação de Agravos de Notificação, SINAN) between January 2015 and December 2017 were included. Multivariate logistic regression was used to assess socio-economic and clinical factors associated with treatment success among incarcerated individuals.RESULTS: Incarcerated individuals (n = 17,776) had greater treatment success than non-incarcerated individuals (n = 160,728; 82.2% vs. 75.1%; P < 0.0001), including after adjusting for demographic and clinical risk factors (adjusted odds ratio aOR 1.27, 95% CI 1.19-1.34). These differences were partially mediated by increased use of directly observed therapy among incarcerated individuals (DOT) (61% vs. 47%; P < 0.001), which was associated with greater efficacy in the incarcerated population (aOR 2.56 vs. aOR 2.17; P < 0.001). DOT was associated with improved treatment success among incarcerated subpopulations at elevated risk of poor outcomes.CONCLUSION: TB treatment success among incarcerated individuals in Brazil is higher than non-incarcerated individuals, but both fall below WHO targets. Expanding the use of DOT and services for socially and medically vulnerable individuals may improve outcomes in carceral settings.

Macrodetermined racial inequalities in diagnostic testing among tuberculosis patients in Brazil.

OBJECTIVES: To estimate the association between tuberculosis (TB) patients' race and patients' access to diagnostic testing in Brazil. In addition, we evaluated if the associations could be explained by a geographic codistribution between racial groups and diagnostic testing. STUDY DESIGN: It is a cross-sectional study based on secondary data from a national surveillance system of new TB cases diagnosed in 2015. METHODS: We evaluated the association between TB patients' race (independent variable) and the HIV testing and TB mycobacterial culture providing (dependent variables) with logistic regression models. We used multilevel models to consider different geopolitical levels (region, state and municipality). In addition, we used conditional logistic regressions matched by health-care unit. All models were adjusted by individual covariates associated with the outcomes. RESULTS: Compared with non-Afro-Brazilian patients, Afro-Brazilian patients had significantly lower odds to have had HIV testing [odds ratio (OR): 0.72; 95% confidence interval (CI): 0.69-0.75] and mycobacterial culture performed (OR: 0.74; 95% CI: 0.71-0.77). However, these statistically significant negative associations between Afro-Brazilian racial category and testing disappeared when patients were considered as nested in geopolitical contexts or matched for health-care unit. CONCLUSIONS: Afro-Brazilian TB patients had lower probability to have HIV test and mycobacterial culture performed. However, these associations seem to be macrodeterminated by the geographic distribution of both racial groups and diagnostic testing. Our findings can support the formulation of public policies aiming to mitigate regional disparities as a strategy to improve racial equity in access to healthcare. The approach presented can be applied in a range of scenarios to identify disparities, localize its source and support decision-making.

Prevalence and screening of active tuberculosis in a prison in the South of Brazil.

SETTING: Tuberculosis (TB) remains a challenge in Brazil, particularly among prison inmates. OBJECTIVE: To assess TB prevalence by active case finding in a public prison in southern Brazil. DESIGN: Prison inmates were screened for TB using the presence of cough and chest X-ray (CXR) from October 2014 to August 2016. Presence of cough, irrespective of duration, and abnormal CXRs were further investigated using laboratory tests. RESULTS: Of 10 326 inmates screened, 196 had confirmed TB (1898/100 000 inmates screened). At the first screening, 1759 inmates presented with cough only, 16 of whom had TB; among those with only abnormal CXR (n = 1273), 92 had TB. Xpert was positive in 155 patients, and negative in 15; these results were confirmed using culture. The remaining 26 patients did not undergo Xpert testing and were confirmed using microscopy (27%), culture (42%) or both (31%). CONCLUSION: The combined use of symptom screening (cough) and CXR was much more effective in maximising TB yield than using either method alone. If patients presenting with cough alone had not been investigated, 10% of TB patients would have been missed; if those with abnormal CXR but no cough had not been investigated, 51% of TB patients would have been missed. We detected high TB prevalence in this prison by using active case finding.

Factors associated with the drug-resistant tuberculosis incidence rate in Brazil.

SETTING: The selection and transmission of drug-resistant strains of tuberculosis (TB) hamper disease control. OBJECTIVE: To identify health conditions, demographic and socio-economic factors, as well as individual factors associated with the incidence of drug-resistant TB (DR-TB), in Brazil at the municipal level. DESIGN: This was an ecological study covering Brazilian municipalities that had reported DR-TB cases in 2014. Associations were evaluated in a multilevel analysis using negative binomial regression. RESULTS: In the multilevel model, males (incidence rate ratio [IRR] 2.6, 95% confidence interval [CI] 2.3-2.93) and Black persons (IRR 1.82, 95%CI 1.61-2.05) had a higher risk of DR-TB. Compared with those aged 60 years, persons aged 15-59 years (IRR 1.72, 95%CI 1.40-2.11) also had a higher risk. The following contextual factors were associated with the incidence rate (IR) of DR-TB: proportion of previously treated patients (IRR 1.27, 95%CI 1.1-1.5), acquired immune-deficiency syndrome IR (IRR 1.13, 95%CI 1.02-1.25), Municipal Human Development Index (IRR 0.72, 95%CI 0.64-0.81) and the Gini coefficient (IRR 0.86, 95%CI 0.78-0.95). CONCLUSION: We have identified individual and contextual factors associated with the incidence of DR-TB. Our results may help improve integrated public health interventions aimed at controlling DR-TB in Brazil.