The feasibility, acceptability, and preliminary impact of real-time monitors and SMS on tuberculosis medication adherence in southwestern Uganda: Findings from a mixed methods pilot randomized controlled trial

We conducted a pilot randomized controlled trial among patients initiating treatment for drug-sensitive tuberculosis (TB). Participants received real-time electronic adherence monitors and were randomized (1:1:1) to: (i) daily SMS (reminders to TB patients and notifications to social supporters sent daily for 3 months, then triggered by late or missed dosing for 3 months); (ii) weekly SMS (reminders to TB patients and notifications to social supporters sent weekly for 3 months, then triggered by late or missed dosing for 3 months); or (iii) control (no SMS). Feasibility was mainly verified by the technical function of the intervention at Month 6. The primary outcome was percent adherence as ascertained by the real time monitor. Quantitative feasibility/acceptability data were summarized descriptively. Percentage adherence and adherence patterns were assessed and compared by linear regression models. Qualitative acceptability data was collected through interviews and analyzed using content analysis. Among 63 participants, the median age was 35 years, 75% had no regular income, and 84% were living with HIV. Feasibility was demonstrated as most of the daily [1913/2395 (80%)] and weekly [631/872 (72%)] SMS reminders to TB patients were sent successfully. Also, most of the daily [1577/2395 (66%)] and weekly [740/872 (85%)] SMS notifications to social supporters and adherence data (96%) were sent successfully. Challenges included TB status disclosure, and financial constraints. All patients perceived the intervention to be useful in reminding and motivating them to take medication. Median adherence (IQR) in the daily SMS, weekly SMS, and control arms was 96.1% (84.8, 98.0), 92.5% (80.6, 96.3), and 92.2% (56.3, 97.8), respectively; however, differences between the intervention and control arms were not statistically significant. Real-time monitoring linked to SMS was feasible and acceptable and may have improved TB medication adherence. Larger studies are needed to further assess impact on adherence and clinical outcomes.