Data analysis
Data were analysed based on a “as randomized, complete-case” approach, where those missing an AD outcome were excluded. Groups were analysed for the primary outcome as randomised, regardless of adherence to study allocation. Sensitivity per-protocol analyses were conducted based on adherence data from questionnaires and diaries. Using parent-reported emollient data from questionnaires at 2, 4 and 8 weeks, participants in the IG were included in the per-protocol analysis if they reported at least once-daily emollient use at each of the three time-points. CG participants were included if they reported emollient use of <4 days per week at each time-point. In the diary per-protocol analysis, participants in the IG were included if they recorded emollient use on ≥90% of days in the 8-week recording period, equating to >6 days a week. CG participants were included if they used emollient on ≤43% of the days (<4 days a week). Additional sensitivity analyses were conducted including those in the IG who reported emollient use on ≥4 days a week and those who in the CG who used emollient on<4 days3 days a week to more closely align with the adherence definition used in the BEEP study10.