We have identified several limitations to the methods or execution of the research in this programme. Workstream 2 delivered three robust systematic reviews, including a Cochrane review. While many novel and useful findings came from the reviews, safety data were often under-reported and follow-up periods were often short, which limited the conclusions we were able to draw from the reviews.
During the trial in workstream 4, young men and fathers were under-represented despite our efforts to tailor recruitment materials to be inclusive and accessible to men. More process measures may have allowed us greater insight into the mechanisms of action of the interventions but may also have adversely affected follow-up due to the questionnaire burden.
The ECO interventions were developed mainly for people with mild-moderate eczema managed in primary care. Modifications may be needed to enhance the relevance to people with severe eczema or those managed in secondary care. Similarly, the interventions were developed and trialled in young people up to 25 years. While this is likely to be relevant to most adults, further research may be needed to optimise and tailor the intervention to the needs of older adults.
In the workstreams 4 and 5 qualitative process evaluation, we were able to purposively sample participants across a range of demographics and geographical areas, with different eczema severities and different levels of intervention usage. We recruited people from a range of ethnic minority groups and people from areas with greater levels of deprivation; however, parents had educational attainment above population averages, so the findings may be less applicable to those with lower education levels.
There were also limitations to the quantitative process evaluation. We encountered challenges with how best to measure time spent on the intervention as it is not known whether people were actively engaging in the intervention page or whether the page was open while they did something else. For the mediation analysis, the primary outcome and potential mediators were measured at the same time point. Ideally, the mediators should have been measured at an intermediate time point, after the use of the intervention and before the measurement of the outcome at 24 weeks.