Successful treatment of tuberculosis depends to a large extent on good adherence to treatment regimens, which relies on directly observed treatment (DOT). This in turn requires frequent visits to health facilities. High costs to patients, stigma and burden to the health system challenged the DOT approach. Digital adherence technologies (DATs) have emerged as possibly more feasible alternatives to DOT but there is conflicting evidence on their effectiveness and feasibility.

Our primary objective is to evaluate whether the implementation of DATs with daily monitoring and a differentiated response to patient adherence would reduce poor treatment outcomes compared with the standard of care (SOC). Our secondary objectives include: to evaluate the proportion of patients lost to follow-up; to compare effectiveness by DAT type; to evaluate the feasibility and acceptability of DATs; to describe factors affecting the longitudinal engagement of patients with the intervention and to use a simple model to estimate the epidemiological impact and cost-effectiveness of the intervention from a health system perspective.

This is a pragmatic two-arm cluster-randomised trial in the Philippines, South Africa, Tanzania and Ukraine, with health facilities as the unit of randomisation. Facilities will first be randomised to either the DAT or SOC arm, and then the DAT arm will be further randomised into medication sleeve/labels or smart pill box in a 1:1:2 ratio for the smart pill box, medication sleeve/label or the SOC respectively. We will use data from the digital adherence platform and routine health facility records for analysis. In the main analysis, we will employ an intention-to-treat approach to evaluate treatment outcomes.