The Impact of Directly Observed Therapy on Successful Tuberculosis Treatment: A Statistical Analysis of National...

Directly Observed Therapy (DOT) is a recommended modality for tuberculosis (TB) treatment, yet is underutilized in many settings. Although DOT has been evaluated in clinical trials, less is known about its effectiveness in routine programs. We estimated the contribution of DOT to TB treatment success in Brazil.

Using data from Brazil's National Disease Notification System for 2015-2018, we fit regression models estimating the relationship between DOT coverage and individual treatment success, leveraging municipality-level variation in DOT usage to avoid confounding due to nonrandom DOT receipt at the individual level. Models were adjusted for demographic and clinical covariates, using generalized estimating equations to account for correlation of outcomes within municipalities. Fitted models were used to project the impact of expanded DOT coverage, with robustness tested using alternative specifications.

The study included 278 007 individuals from 4783 municipalities. Average DOT coverage was 45.3%, with significant geographic variability. Treatment success was 86.9% with DOT and 73.2% without DOT. DOT coverage was associated with a 1.46 (95% confidence interval: 1.38, 1.55) odds ratio (OR) of treatment success. ORs ranged from 1.11 to 2.52 under alternative specifications, consistently indicating a positive impact. Scenario analyses indicated that reducing DOT nonutilization in each municipality by 50% would reduce the proportion of unsuccessful treatment outcomes by 7.1% (6.1, 8.2), equivalent to an average 5.5 (4.7, 6.3) percentage point increase in success probability for individuals newly receiving DOT.

In this routine programmatic setting, higher DOT utilization was associated with better treatment outcomes, underscoring the importance of adherence-supported interventions for TB care.