Process Evaluation of Pragmatic Cluster-Randomized Trials of Digital Adherence Technologies for Tuberculosis Treatment Support: A Mixed-Method Study in Five Countries

Digital adherence technologies (DATs) could improve the person-centeredness of tuberculosis (TB) treatment. DATs are found to be acceptable, though evidence of their effectiveness is varied. Our objective was to understand the fidelity of DAT interventions within five cluster-randomized trials. Two DATs (smart pillbox, medication labels) were assessed, with real-time adherence data available to healthcare providers (HCPs) on a digital platform in Ethiopia, the Philippines, South Africa, Tanzania, and Ukraine. A framework assessed four components of implementation: inputs (training, support, mobile access), processes (SMS, home visits, platform usage), outputs (DAT engagement, manual dosing), and outcomes (people with TB (PwTB)-HCP relationship). Fidelity was evaluated by quantitative indicators, and content analysis of qualitative sub-studies supplemented some indicators. Engagement with DATs was high among PwTB. Pillbox users showed high levels of sustained engagement (box opening), with digitally recorded doses ranging from 82% to 91%. Differences were observed in login frequency by HCPs to the adherence platform. In Ethiopia, Tanzania, and Ukraine, there was at least one login to the platform on 71% of weekdays per facility compared with the Philippines and South Africa at 42% and 52%, respectively. Intervention fidelity varied among countries, suggesting a need for future work on optimizing implementation.