Abstract Background Adherence to antituberculosis therapy is an important determinant of treatment outcome in rifampicin-resistant tuberculosis (RR-TB). Understanding adherence to contemporary treatment regimens in routine care is needed to support implementation in TB programs. We aimed to characterize temporal adherence patterns among people receiving oral treatment for RR-TB. Methods We conducted a prospective observational cohort study at a referral TB hospital in South Africa. People ≥15 years with pulmonary RR-TB starting an oral 9-12-month regimen were included. Treatment adherence was measured using a digital pillbox during ambulatory care and with directly observed therapy during hospital care. The primary outcome was proportion of adherence days through nine months. Latent class group-based trajectory modelling was used to identify temporal adherence patterns. Results 209/248 (84.3%) participants had assessable adherence data from the digital pillbox or directly observed therapy. Overall median adherence was 82% (IQR 63-98) with combined measures, and 72% (IQR 51-92) with digital pillbox only. Four distinct adherence patterns were identified. Adherence was 93-100% in the first month. Two groups, representing 136 (65.1%) individuals, had small reductions in adherence over time, separated by higher and lower early adherence. In the other two groups, there was a 50% reduction in adherence by months three (48/209, 23.0%) and six (25/209, 12.0%), respectively. Lower adherence over time was associated with having exclusive ambulatory care, treatment with the shorter regimen only, and age 40 years. Conclusions Treatment adherence declined over time in distinct temporal patterns. Group characteristics could identify individuals who may benefit from enhanced treatment support.
Kuhlin et al. (Mon,) studied this question.