Abstract Body


Digital adherence technologies have gained traction recently with medication monitors showing improved adherence. Data on the impact on long-term clinical outcomes is lacking. We conducted a cluster-randomized trial to measure outcomes amongst persons with drug-susceptible TB (DS-TB) supported by medication monitors and differentiated care (intervention) versus standard care (SoC). Participants had improved adherence in the intervention versus SoC arms. Here we compare 18-month clinical outcomes by study arm.


DS-TB (≥2 years) persons were enrolled from 18 primary health clinics in three provinces (Gauteng, KwaZulu Natal, Western Cape) in South Africa. Intervention arm participants had visual/audio reminders for medication intake with monitoring and support depending on the number of missed doses/week. SoC arm received monitors in silent mode to document adherence. Participants were followed-up with sputum (culture) at treatment end and 18 months. Unfavorable outcome was defined as: on treatment – treatment failure, lost to follow-up, death, culture-positive at 6 months or MDR diagnosis; or recurrence to 18 months.


We enrolled 2727 participants, reporting on 2657 participants: 38% female, median age 36 years, and 53% HIV-positive. Of 2070 participants (587 had outcome undefined), 20.9% (432/2070) had an unfavorable outcome. By arm, unfavorable outcomes were similar (geometric means: 22.3% in SoC versus 17.1% in intervention clusters). The risk ratio was 0.78 (0.53- 1.16), adjusting for age, sex, TB diagnosis, ethnic group, education, marital status, HIV/ART status, and province. The effect of the intervention appeared to be greater among females and those in Gauteng and Western Cape as shown in the figure attached.


Although adherence was improved, there does not seem to be a difference in unfavorable outcomes in persons with DS-TB in the intervention versus SoC arms. Although these interventions are less likely to show an impact on clinical outcomes in routine settings, the effect on adherence is important and warrants continued use and evaluation of these technologies. Adaptation of these technologies to cater for those on both TB and HIV treatment is required.

Overall effect of unfavorable outcome by subgroup