SummaryBackgroundDirectly observed treatment (DOT) has been the standard of care for tuberculosis since the early 1990s, but it is inconvenient for patients and service providers. Video-observed therapy (VOT) has been conditionally recommended by WHO as an alternative to DOT. We tested whether levels of treatment observation were improved with VOT.MethodsWe did a multicentre, analyst-blinded, randomised controlled superiority trial in 22 clinics in England (UK). Eligible participants were patients aged at least 16 years with active pulmonary or non-pulmonary tuberculosis who were eligible for DOT according to local guidance. Exclusion criteria included patients who did not have access to charging a smartphone. We randomly assigned participants to either VOT (daily remote observation using a smartphone app) or DOT (observations done three to five times per week in the home, community, or clinic settings). Randomisation was done by the SealedEnvelope service using minimisation. DOT involved treatment observation by a health-care or lay worker, with any remaining daily doses self-administered. VOT was provided by a centralised service in London. Patients were trained to record and send videos of every dose ingested 7 days per week using a smartphone app. Trained treatment observers viewed these videos through a password-protected website. Patients were also encouraged to report adverse drug events on the videos. Smartphones and data plans were provided free of charge by study investigators. DOT or VOT observation records were completed by observers until treatment or study end. The primary outcome was completion of 80% or more scheduled treatment observations over the first 2 months following enrolment. Intention-to-treat (ITT) and restricted (including only patients completing at least 1 week of observation on allocated arm) analyses were done. Superiority was determined by a 15% difference in the proportion of patients with the primary outcome (60% vs 75%). This trial is registered with the International Standard Randomised Controlled Trials Number registry, number ISRCTN26184967.FindingsBetween Sept 1, 2014, and Oct 1, 2016, we randomly assigned 226 patients; 112 to VOT and 114 to DOT. Overall, 131 (58%) patients had a history of homelessness, imprisonment, drug use, alcohol problems or mental health problems. In the ITT analysis, 78 (70%) of 112 patients on VOT achieved ≥80% scheduled observations successfully completed during the first 2 months compared with 35 (31%) of 114 on DOT (adjusted odds ratio [OR] 5·48, 95% CI 3·10–9·68; p<0·0001). In the restricted analysis, 78 (77%) of 101 patients on VOT achieved the primary outcome compared with 35 (63%) of 56 on DOT (adjusted OR 2·52; 95% CI 1·17–5·54; p=0·017). Stomach pain, nausea, and vomiting were the most common adverse events reported (in 16 [14%] of 112 on VOT and nine [8%] of 114 on DOT).InterpretationVOT was a more effective approach to observation of tuberculosis treatment than DOT. VOT is likely to be preferable to DOT for many patients across a broad range of settings, providing a more acceptable, effective, and cheaper option for supervision of daily and multiple daily doses than DOT.FundingNational Institute for Health Research.