Bridging the Gap Between Treatment and Accountability: DR-TB Outcomes and Governance Signals in Rural Eastern Cape,...

Background: Monitoring treatment outcomes is central to evaluating drug-resistant tuberculosis (DR-TB) programmes and ensuring accountability, particularly in high-burden settings such as South Africa. However, routine surveillance data are often constrained by small sample sizes and incomplete reporting. The increasing proportion of outcomes classified as "not evaluated" raises critical concerns about the validity and interpretability of reported treatment success rates. This study examines longitudinal trends in DR-TB outcomes in two rural Eastern Cape facilities (2020-2024), conceptualising routine outcomes as governance signals reflecting health system performance and data management capacity, rather than clinical performance alone. Methods: A retrospective analysis of aggregated DR-TB treatment outcomes from 2020 to 2024 was conducted using routine programme data. Outcomes included treatment success, failure, loss to follow-up (LTFU), death, and "not evaluated." Proportions were estimated using Wilson 95% confidence intervals, and year-to-year differences were assessed using chi-square tests, with cautious interpretation due to small sample sizes. Bounding analyses were applied to assess the sensitivity of treatment success estimates to alternative assumptions about incomplete outcome reporting. Results: Treatment success declined over time for both rifampicin-resistant TB (RR-TB) and multidrug-resistant TB (MDR-TB), with a significant decrease observed in MDR-TB (80.9% in 2021 to 38.5% in 2024; p = 0.007). Concurrently, the proportion of outcomes classified as "not evaluated" increased substantially from 2023 onward, reaching 25.8% in RR-TB and 30.8% in MDR-TB by 2024 (p < 0.001). Mortality remained persistently high in RR-TB, while MDR-TB exhibited higher levels of treatment failure and LTFU. Analyses for pre-XDR-TB and XDR-TB were limited by small sample sizes. Bounding analyses demonstrated that treatment success estimates were highly sensitive to incomplete reporting of outcomes. Conclusions: Declining DR-TB treatment success in this setting is closely associated with increasing incomplete outcome reporting, suggesting that observed trends may reflect challenges in surveillance and governance systems rather than clinical deterioration alone. Interpreting routine DR-TB outcomes as governance signals provides a novel systems-based perspective for programme evaluation. Strengthening data completeness, patient tracking, and surveillance systems is essential to improve the reliability of programme indicators and support effective TB control in resource-constrained rural settings.