The introduction of video-enabled directly observed therapy (video-DOT) for patients with drug-resistant TB disease in Eswatini amid the COVID-19 pandemic - a retrospective cohort study.

BACKGROUND: Video-enabled directly observed therapy (video-DOT) has been proposed as an additional option for treatment provision besides in-person DOT for patients with drug-resistant TB (DRTB) disease. However, evidence and implementation experience mainly originate from well-resourced contexts. This study describes the operationalization of video-DOT in a low-resourced setting in Eswatini facing a high burden of HIV and TB amid the emergence of the COVID-19 pandemic. METHODS: This is a retrospectively established cohort of patients receiving DRTB treatment during the implementation of video-DOT in Shiselweni from May 2020 to March 2022. We described intervention uptake (vs. in-person DOT) and assessed unfavorable DRTB treatment outcome (death, loss to care) using Kaplan-Meier statistics and multivariable Cox-regression models. Video-related statistics were described with frequencies and medians. We calculated the fraction of expected doses observed (FEDO) under video-DOT and assessed associations with missed video uploads using multivariable Poisson regression analysis. RESULTS: Of 71 DRTB patients eligible for video-DOT, the median age was 39 (IQR 30-54) years, 31.0% (n = 22) were women, 67.1% (n = 47/70) were HIV-positive, and 42.3% (n = 30) were already receiving DRTB treatment when video-DOT became available. About half of the patients (n = 37; 52.1%) chose video-DOT, mostly during the time when COVID-19 appeared in Eswatini. Video-DOT initiations were lower in new DRTB patients (aHR 0.24, 95% CI 0.12-0.48) and those aged ≥ 60 years (aHR 0.27, 95% CI 0.08-0.89). Overall, 20,634 videos were uploaded with a median number of 553 (IQR 309-748) videos per patient and a median FEDO of 92% (IQR 84-97%). Patients aged ≥ 60 years were less likely to miss video uploads (aIRR 0.07, 95% CI 0.01-0.51). The cumulative Kaplan-Meier estimate of an unfavorable treatment outcome among all patients was 0.08 (95% CI 0.03-0.19), with no differences detected by DOT approach and other baseline factors in multivariable analysis. CONCLUSIONS: Implementing video-DOT for monitoring of DRTB care provision amid the intersection of the HIV and COVID-19 pandemics seemed feasible. Digital health technologies provide additional options for patients to choose their preferred way to support treatment taking, thus possibly increasing patient-centered health care while sustaining favorable treatment outcomes.

Sustained high fatality during TB therapy amid rapid decline in TB mortality at population level: A retrospective cohort and ecological analysis from Shiselweni, Eswatini.

OBJECTIVES: Despite declining TB notifications in Southern Africa, TB-related deaths remain high. We describe patient- and population-level trends in TB-related deaths in Eswatini over a period of 11 years. METHODS: Patient-level (retrospective cohort, from 2009 to 2019) and population-level (ecological analysis, 2009-2017) predictors and rates of TB-related deaths were analysed in HIV-negative and HIV-coinfected first-line TB treatment cases and the population of the Shiselweni region. Patient-level TB treatment data, and population and HIV prevalence estimates were combined to obtain stratified annual mortality rates. Multivariable Poisson regressions models were fitted to identify patient-level and population-level predictors of deaths. RESULTS: Of 11,883 TB treatment cases, 1302 (11.0%) patients died during treatment: 210/2798 (7.5%) HIV-negative patients, 984/8443 (11.7%) people living with HIV (PLHIV), and 108/642 (16.8%) patients with unknown HIV-status. The treatment case fatality ratio remained above 10% in most years. At patient-level, fatality risk was higher in PLHIV (aRR 1.74, 1.51-2.02), and for older age and extra-pulmonary TB irrespective of HIV-status. For PLHIV, fatality risk was higher for TB retreatment cases (aRR 1.38, 1.18-1.61) and patients without antiretroviral therapy (aRR 1.70, 1.47-1.97). It decreases with increasing higher CD4 strata and the programmatic availability of TB-LAM testing (aRR 0.65, 0.35-0.90). At population-level, mortality rates decreased 6.4-fold (-147/100,000 population) between 2009 (174/100,000) and 2017 (27/100,000), coinciding with a decline in TB treatment cases (2785 in 2009 to 497 in 2017). Although the absolute decline in mortality rates was most pronounced in PLHIV (-826/100,000 vs. HIV-negative: -23/100,000), the relative population-level mortality risk remained higher in PLHIV (aRR 4.68, 3.25-6.72) compared to the HIV-negative population. CONCLUSIONS: TB-related mortality rapidly decreased at population-level and most pronounced in PLHIV. However, case fatality among TB treatment cases remained high. Further strategies to reduce active TB disease and introduce improved TB therapies are warranted.