Abstract
Background Social vulnerability is strongly associated with tuberculosis (TB) indicators like cure rate. By addressing key social determinants, social protection policies such as Brazil’s Bolsa Família Programme (BFP), a governmental conditional cash transfer, may play a role in TB control. Evidence is consolidating around a positive effect of social protection on TB outcomes, however methodological limitations prevent strong conclusions. This paper uses a quasi-experimental approach to more rigorously evaluate the effect of BFP on TB cure rate.
Methods & Findings The data source was Brazil’s TB notification system (SINAN), linked to the national registry of those in poverty (CadUnico) and the BFP payroll. Propensity scores (PSs) were estimated from a complete-case logistic regression using covariates from this linked dataset, informed by a directed acyclic graph. Control patients were matched to exposed patients on the PS and the average effect of treatment on the treated (ATT) was estimated as the difference in TB cure rate between matched groups (n = 2167). The ATT was estimated as 10·58 (95% CIs: 4·39, 16·77). This suggests that 10·58% of the TB patients receiving BFP who were cured would not have been cured had they not received BFP. The direction of this effect was robust to sensitivity analyses performed and the PS matching broadly improved balance, although missing data limited the sample size.
Conclusions This work is the first quasi-experimental evaluation of social protection in wide-scale practice on TB outcomes. It demonstrates a positive effect of conditional cash transfers on TB cure rate consistent with existing work, suggesting changes to policy and future research on increasing access to social protection for TB patients who remain uncovered by the programme.
Authors’ Contributions
DJC, DB, and RhD conceived and designed the study. DJC analysed and interpreted data. DJC and DB drafted the manuscript. All authors provided intellectual contributions, reviewed, and approved the final version of the report.