Adverse events among persons with TB using in-person vs. electronic directly observed therapy.

We evaluated patient safety within a randomized crossover trial comparing electronic directly observed therapy (eDOT) to in-person DOT (ipDOT) in persons undergoing TB treatment in New York City, NY, USA. Participant symptoms, symptom severity, and clinical management were documented. We assessed adverse event reports (AERs) by DOT method during the two-period crossover. Using Cox proportional-hazards mixed-effects models, we estimated the adjusted hazard ratio (aHR) of participants reporting an adverse event (AE) vs. not reporting an AE. Of 211 participants, 57 (27.0%) reported AEs during the two-period crossover; of these, 54.4% (31/57) were reported while using eDOT vs. 45.6% (26/57) while using ipDOT. Controlling for study group and period, the aHR for eDOT vs. ipDOT was 0.98 (95% CI 0.49-1.93). Although statistically not significant, the wide confidence intervals suggest that a significant association cannot be entirely ruled out. Gastrointestinal symptoms were most frequently reported (42.1%, 24/57). AER types and severity did not differ significantly by DOT method. Days from symptom onset to medical attention was similar across DOT methods (median: 1.0 day, IQR 0.0-2.0). No participants switched DOT methods due to AERs or monitoring concerns. Further evaluation to ascertain whether AERs differ when patients use eDOT vs. ipDOT is warranted.