ObjectivesThis study aimed to evaluate the safety of dichoptic balanced binocular viewing (BBV) for amblyopia in children, plus feasibility, adherence, acceptability, trial methodology and clinical measures of visual function.DesignWe carried out an observer-masked parallel-group phase 2a feasibility randomised controlled trial.SettingTwo study sites, a secondary/tertiary and a community site.ParticipantsWe enrolled 32 children aged 3–8 years with unilateral amblyopia who had completed optical adaptation where indicated. 20 children attended the 16-week exit visit (retention 63%).InterventionsChildren were randomised to BBV (movies customised to interocular acuity difference at baseline) for 1 hour a day (active intervention) or standard management as per parental choice (part-time occlusion or atropine blurring, control). All interventions were used at home, daily for 16 weeks.Primary outcome measure‘VacMan suppression test’ of interocular balance at 16 weeks from randomisation. Secondary outcome measures: feasibility outcomes (recruitment and retention ratios, adherence with the allocated intervention); safety outcomes at other time points (changes in prevalence of diplopia, manifest strabismus, suppression/interocular balance on a range of tests); efficacy outcomes (clinical measures of visual function, such as best-corrected visual acuity, BCVA). Outcome measures were identical to those planned in the protocol.ResultsPrimary outcome: At baseline, values for the interocular balance point were higher (indicating greater suppression of the amblyopic eye) in the occlusion group than in the BBV group. These values shifted downwards on average for the occlusion group, significantly decreasing from baseline to week 16 (t8=4.49, p=0.002). Balance values did not change between baseline and week 16 for the BBV group (t9=−0.82, p=0.435). At 16 weeks, there was no statistical difference in interocular balance/suppression change over time between the two arms. The difference at follow-up between the arms, adjusted for baseline, was −0.02 (95% CI −0.28 to 0.23, p=0.87). Feasibility: We prescreened 144 records of potentially eligible children. Between 28 October 2019 and 31 July 2021, including an interruption due to the COVID-19 pandemic, 32 children were screened and randomised (recruitment rate 22%), 16 to BBV and 16 to standard treatment. 20 children attended the 16-week exit visit (retention 63%). Mean adherence with BBV as proportion of viewing time prescribed was 56.1% (SD36) at 8 and 57.9% (SD 30.2) at 16 weeks. Mean adherence with prescribed occlusion time was 90.1% (SD 19.7) at 8 and 59.2% (SD 24.8) at 16 weeks.Secondary safety/efficacy outcomesOne child in the BBV arm reported transient double vision, which resolved; two reported headaches, which led to withdrawal. BCVA improved from mean 0.47 (SD0.18) logMAR at randomisation to 0.26 (0.14) with standard treatment, and from 0.55 (0.28) to 0.32 (0.26) with BBV. Outcomes at 16 weeks did not differ between treatments.Participant experienceFamilies were generally positive about BBV, but families found both patching and BBV difficult to integrate into family routines.ConclusionsRecruitment rates indicate that a future phase 3 trial will require multiple sites or a longer enrolment period. Retention and adherence rates were lower than anticipated, which will influence future study designs. Dichoptic treatment may be equal to occlusion treatment in safety and efficacy; headaches may lead to discontinuation. Integration into family routines may constitute a barrier to implementation.Trial registration numberNCT03754153.