Predictors of treatment outcomes in adults with drug–sensitive Tuberculosis in Maharashtra, India: A retrospective study

Tuberculosis (TB) remains the leading infectious cause of death worldwide, with India accounting for nearly one-fourth of global TB cases. Ni-kshay, the country's digital case-based TB notification platform is rich in data pertaining to the continuum of care of TB patients. This study aims to develop a standardized analytical approach to programmatic data to identify predictors of unfavourable treatment outcomes and mortality among adult drug-sensitive TB patients at the state level for Maharashtra during 2021 and 2022. Two separate analyses were undertaken comparing treatment success with: (1) unfavourable outcomes (death, treatment failure, loss to follow-up, regimen change, or not evaluated); and (2) mortality. Age group, gender and pre-treatment weight were entered together in a single logistic regression model, and each remaining risk factor was entered into a separate model alongside those three covariates, to compute adjusted odds ratios (aOR). Records with no entry for diabetes, HIV, tobacco or alcohol were retained as a "not recorded" or "unknown" category. Collinearity across variables was assessed using Cramer's V and variance inflation factors. The final cohort included 323,124 cases for unfavourable outcome analysis and 315,579 cases for mortality analysis. Increasing age, male gender, lower body weight, known HIV and diabetes comorbidities, tobacco and alcohol consumption, and "unknown" status for behavioural risks and comorbidity status were significantly associated with increased odds of both unfavourable outcomes and mortality. This study highlights the utility of programmatic data in identifying high-risk TB patients and offers a reproducible analytic framework.