Comparison of lotions, creams, gels and ointments for the treatment of childhood eczema: the BEE RCT

Pragmatic randomised trial

The aim of the trial was to compare the effectiveness and acceptability of four commonly used types of emollients for the treatment of childhood eczema.

The objectives of the trial were to compare the four types of emollients at both 16 weeks (medium term) and 52 weeks (long term) in respect of:

• parent-reported eczema symptoms

• objective assessment of eczema signs

• quality of life for the child

• impact of eczema on the family

• adverse effects

• parent satisfaction with study emollient

• frequency and quantity of study emollient and other emollient use

• use of topical corticosteroids

• number of well-controlled weeks.

Nested qualitative study

The aim of the nested qualitative study was to complement, explain and aid understanding of the quantitative findings regarding the delivery/receipt of the intervention, its acceptability and perceived or experienced benefits or harms.

The objectives of the nested qualitative study were to:

• explore facilitators of or barriers to use and follow up with participants who stopped treatment early

• explore carers’ and children’s experiences of study emollient use and their views about perceived effectiveness and/or acceptability of study emollients

• contextualise the trial findings, as an aid to interpreting the results and their potential impact on clinical practice.

Identification of potentially eligible children

Participant recruitment was via GP surgeries in the three clinical research network (CRN) areas of West of England (Bristol), Wessex (Southampton) and East Midlands (Nottingham and Lincoln).

Participant eligibility criteria

To be eligible, children and the responsible adult had to meet the following inclusion criteria:

• Child

  • aged between 6 months and 12 years

  • have eczema diagnosed by an appropriately qualified health-care professional (registered doctor, nurse or health visitor)

  • mild eczema or worse (POEM score of > 2 points within previous 28 days). At the point of randomisation, at least two POEM scores were available for all participants – one recorded at the time of expression of interest and one recorded at baseline. The expression of interest POEM score was used to determine eligibility and for randomisation; a further POEM score was recorded if the baseline visit was > 28 days later.

• The person giving consent must

  • have parental responsibility for the participant

  • be willing to use the randomly allocated emollient type as the only leave-on emollient for 16 weeks.

Exclusion criteria for the child and the responsible adult were as follows:

• Child

  • known sensitivity to study emollients or their constituents

  • participating in another research study currently or in the last 4 months

  • any other known adverse medical or social circumstance that would make invitation to the study inappropriate (as determined by GP practice staff).

• The person giving consent

  • was unable to give informed consent

  • had insufficient written English to complete outcome measures.

Participant recruitment

Participant follow-up

Parents were telephoned by an unmasked member of the research team with their allocation after randomisation, and told to collect their corresponding prescription from their GP surgery. This was followed up with another telephone call 7 days later to confirm that they had received the study emollient and started using it.

Thereafter, with the exception of the week 16 visit, follow-up was remote, by means of parent-completed questionnaires (see Chapter 2, Data collection, Follow-up questionnaires).

The final week 16 follow-up visit was on 17 February 2020. The trial finished after the last participant’s week 52 follow-up questionnaire was returned.

Participant withdrawals

Participants were free to withdraw at any time, without any consequences for their usual care or follow-up. We analysed any data already collected and obtained their electronic medical record (EMR) data, unless the participant expressly withdrew their consent prior to the database being locked. Participants who actively withdrew were asked to give reason(s).

Participant engagement

Parents were thanked for their time in the study with a £10 voucher at the baseline visit and a £10 voucher on completion of the week 16 visit, and a final £10 voucher was sent on completion of the week 52 survey.

At the baseline visit, children were offered a small thank you of a bee soft toy or a bouncy ball. In addition, children were given an A4 bee colouring template to colour in/decorate while the researcher talked to their parent. Once complete, families were encouraged to share pictures with the BEE research team. Submitted colourings were shared via the study website and Twitter (Twitter, Inc., San Francisco, CA, USA; www.twitter.com), and used in dissemination of the study findings, for example in presentations. At the week 16 visit, children were given a certificate to congratulate them on their participation in the study.

Newsletters were sent out to families via e-mail approximately three times per year and e-cards were sent at Christmas. Newsletters contained information on recruitment, reminders on how to complete the questionnaires, seasonal information relating to eczema, and profiles of research team members.

Parents were encouraged to follow the study Twitter account, which charted progress and shared children’s coloured-in bees. Finally, a study website was maintained, containing study information and links to relevant documents.

Primary outcome

The primary outcome was the Patient Orientated Eczema Measure (POEM) score, measured weekly for 16 weeks. POEM scores were calculated by summing the seven questions, with a possible range of 0–28 points.

The POEM is a patient-reported outcome measure that can be completed by proxy (carer report) and captures symptoms of importance to parents and patients over the previous week. 23 It demonstrates good validity, repeatability and responsiveness to change,24,25 and was favoured as the main outcome by patient contributors.

Secondary outcomes

The following secondary outcomes (time period) were collected and/or calculated:

1. Eczema symptoms (POEM), once every 4 weeks for 52 weeks.

2. Eczema signs [Eczema Area Severity Index (EASI)] at 16 weeks. The total EASI score is a weighted sum of the four EASI scores for each body area (head and neck, upper limbs, trunk and lower limbs). The weights allocated to each body area differ according to the age of the child (≤ 7 or ≥ 8 years). The final EASI scores range between 0 and 72.

3. Masking of researcher at the week 16 visit, using the Bang blinding index. 26

4. Use of study emollient/topical corticosteroids.

5. Participant quality of life [disease-specific Atopic Dermatitis Quality of Life (ADQoL)27 and generic Child Health Utility 9-Dimension (CHU-9D)28,29 scores], coded according to the developer’s instructions. The CHU-9D is validated for children aged ≥ 7 years. The pilot versions for younger children were used with additional guidance notes.

6. Impact of the participant’s eczema on the family [Dermatitis Family Impact (DFI) questionnaire]. 30 The DFI was obtained by summing responses to the 10 items to give a score of 0–30 points.

7. Satisfaction with study emollient.

8. Adverse events – localised skin reactions, slips and falls, and unplanned hospital admissions.

9. Number and quantity of study emollients and other emollients used.

10. Proportion of well-controlled weeks between weeks 1 and 16.

All the above were collected by parent-completed questionnaires, except b, which was assessed by a masked researcher; c, which was obtained by researcher-completed questionnaire at the week 16 visit; i, which was obtained from participants’ EMRs; and j, which was derived from POEM scores, in which each week was classified as well-controlled (POEM score ≤ 2) or not (POEM score > 2), with the proportion of weeks with well-controlled symptoms calculated as the number of well-controlled weeks divided by the number of weeks with non-missing POEM scores.

The format of questions for d was changed in month 12 of recruitment from a numbered (day 1, day 2, etc.) to a named day (Monday, Tuesday, etc.), to improve ease of completion (see Chapter 3, Parent and public involvement during the study).

Items in the final questionnaire asked about trial participation, specifically about following directions on study emollient use.

With a view to carrying out economic analyses in the future, we also asked questions about personal costs related to eczema and health-care consultations (4-weekly).

Baseline visit

Once consent was received, the following baseline data were collected:

• participant contact details and demographics

• eczema history and treatment, POEM score, DFI score, parent opinions about different types of emollients, quality of life (ADQoL and CHU-9D scores)

• skin assessment by researcher (EASI score and UK diagnostic criteria for atopic dermatitis).

Parents choosing to complete online were offered a laminated aide memoire to enable them to keep a note of their child’s daily emollient use in-between completing the surveys.

Follow-up questionnaires

Participants completed follow-up questionnaires weekly for 16 weeks, then 4-weekly until 52 weeks. Weekly questionnaires comprised POEM and questions about the use of eczema treatments (study emollient, other leave-on emollients and topical corticosteroids) and adverse events. Parents were asked to report eczema treatment use over the previous week, with the first week commencing the Monday after randomisation.

In addition to the above, the 4-weekly questionnaires asked about consultations with health-care professionals (health visitor, pharmacist, dermatologist, and dermatology nurse) and costs (out-of-pocket expenses for eczema-related purchases, private/alternative treatments, travel costs to appointments).

Quality-of-life measures (ADQoL, CHU-9D) were administered at 6, 16 and 52 weeks, and the impact of the condition on the family (DFI) was determined at 16 and 52 weeks. Participants were asked about their satisfaction with their study emollient at 16 weeks.

Parents received e-mail and/or text reminders to complete the questionnaires, with telephone calls if needed.

Follow-up visit

The follow-up visit was scheduled for 16 weeks (± 10 days) after the baseline visit. Participants’ skin was assessed, usually in their own home, by a masked researcher. To maximise data collection, from month 10 of recruitment, researchers were asked to collect POEM scores at 16 weeks as well as EASI scores. When both parent-completed and researcher-collected week 16 POEM scores were available, the scores were compared and sensitivity analysis undertaken (see Statistical methods, Primary outcome, Sensitivity analyses including imputation of missing data).

Electronic medical record review

Consultation and prescription data (from 4 weeks prior to and 52 weeks after randomisation for each participant) were extracted from participants’ EMR remotely, using reports written for SystmOne (The Phoenix Partnership, Leeds, UK) and EMIS Web (EMIS Health, Leeds, UK). After the last participant from each GP surgery had completed their week 52 follow-up questionnaire, their practice was sent the appropriate electronic search strategy, instructions and list of patients recruited into the study. It was not possible to obtain data on quantity (in grams or millilitres) for items prescribed in SystmOne practices.

Data management and validation

Each participant was assigned a trial participant identification number, which was used on questionnaires and the electronic database. All data relating to participants were stored securely in locked cabinets and/or password-protected file stores. Patient identifiers were separate from clinical data. The trial database was locked on 11 December 2020.

Data for a random sample of 10% of participant questionnaires were checked for quality purposes.

Characteristics of non-study patients

We compared the age and sex of study participants with non-study patients using means and SDs (age) and frequencies and proportions (sex). The purpose of this was to investigate whether or not the study participants differed from those who were excluded at earlier stages of the recruitment and screening process.

Baseline characteristics

Baseline sociodemographic and clinical characteristics of study participants were described by treatment group. Any imbalances informed additional adjustment of the primary analyses as appropriate. Continuous variables were summarised using the mean and SD [or median and interquartile range (IQR) if the distribution was skewed] and categorical data were summarised as frequencies and proportions.

Primary outcome

Primary statistical analyses between the randomised groups were conducted on an intention-to-treat basis, defined as analysing participants as randomised, regardless of the adherence to their allocated group and without imputation for missing data. For the primary outcome, linear mixed models (weekly observations, level 1; nested within participants, level 2) were used to explore whether or not there were differences in mean POEM scores between treatment groups after adjusting for baseline scores and all stratification and minimisation variables used in the randomisation.

This approach allowed incomplete cases (i.e. participants who did not complete all of their weekly scores) to contribute to the analysis. Therefore, all participants who contributed at least two POEM scores (one baseline and another between weeks 1 and 16) were included in the primary outcome analysis.

Pairwise comparisons were conducted to identify which intervention groups differed and were presented as mean differences with 95% confidence intervals (CIs) and p-values. To account for multiple testing, we used a modified alpha of 0.0083 (0.05/6 pairwise comparisons equivalent) as a threshold when interpreting p-values.

Secondary outcomes

Analyses of secondary outcomes were conducted on an intention-to-treat basis, according to the data type and frequency of recording. Continuous outcomes measured at multiple time points (POEM, ADQoL, DFI and CHU-9D scores) were analysed similarly to that described in Primary outcome above.

The EASI scores measured at baseline and 16 weeks were analysed using a linear regression model, adjusting for baseline values when available. A sensitivity analysis for the collection of EASI scores at baseline and 16 weeks by the same and different researchers was also undertaken.

The EASI and DFI scores at follow-up were found to be highly skewed and contained values of 0; therefore, scores were transformed by taking the natural log of the score plus 1. The results of these analyses are presented as the ratio of the geometric means of the two groups being compared.

Patterns of the use eczema treatments over the primary outcome period (weeks 1–16) are summarised using descriptive statistics. The number of days in a week that participants used their allocated emollient was analysed using a mixed-effects Poisson regression model adjusting for all stratification and minimisation variables. A mixed-effects negative binomial model was used for the number of days in a week that participants used a non-study emollient.

Parental satisfaction with the study emollient at 16 weeks was analysed using an ordered logistic regression model adjusting for all stratification and minimisation variables.

The proportion of weeks with well controlled symptoms was analysed using a linear regression model adjusting for stratification and minimisation variables.

Impact of COVID-19

On 11 March 2020, the UK Government issued public health guidance recommending that residents wash hands with soap and water, or use hand sanitiser to protect against COVID-19 transmission. 33 A concern was raised by the TMG that increased hand washing and use of sanitiser gels might worsen eczema symptoms and reduce the effectiveness of the intervention.

To explore the possible impact this guidance had on symptoms, a sensitivity analysis was conducted on the repeated-measures POEM scores, before and after the handwashing advice was issued. A binary variable classifying the follow-up period as pre- or post-COVID-19 handwashing advice was generated. A linear mixed model [weekly observations (level 1) nested within participants (level 2)] including a COVID group interaction term was used to explore whether or not the differences in mean POEM scores between treatment groups differed between the periods before and after handwashing advice. The model also adjusted for baseline POEM scores and variables used in the randomisation.

Analysis of safety

Descriptive analysis of safety end points are presented by allocation. The number of events and the number of participants having at least one are tabulated.

Design

We conducted semistructured interviews with parents of participants from each trial group at 2–4 weeks (hereafter termed week 4 interviews) and 16 weeks after randomisation. The week 4 interviews aimed to capture initial experiences with, and opinions of, the study emollient, along with any early deviations from allocation and the reasons for change. The week 16 interviews aimed to capture the longer-term experiences of using the allocated emollients or reasons for stopping or changing, along with parents’ future intentions around emollient use.

Recruitment and consent of parents/assent of children for interviews

All parents received brief information about the qualitative interview study at the time of recruitment into the trial and were asked for consent to be contacted for interview. Following the sampling framework, potential interviewees were invited to take part by means of an e-mailed invitation letter with accompanying qualitative study information sheet (see Report Supplementary Materials 7 and 8).

Informed written consent was received from the parent (see Report Supplementary Material 9). At the discretion of the parent and the preference of their child, children (usually aged ≥ 7 years) using the emollient were invited to participate in the interview and, if they agreed, written assent (see Report Supplementary Material 6) was obtained from the child.

Recruitment stopped when there was agreement that inductive thematic saturation had been reached for the main themes. 34,35

Sampling

Our original design was cross-sectional, in that we anticipated sampling different participants at the week 4 and week 16 time points. However, we revisited a small number of participants from week 4 at week 16. Their data were analysed, as with all the data, cross-sectionally. We sampled from the characteristics of the participants but primarily interviewed parents. At the preference and with the assent of parents/carers and their children, children also sometimes participated in the interviews.

At week 4, we sampled parents across the four trial groups, with variation according to recruiting centre (Bristol, Southampton and Nottingham), age of child, eczema severity (mild, moderate or severe, determined by categorised baseline POEM score)36 and length of use of study emollient, to include some who had stopped using their allocated treatment (determined from patient contact and information from Clinical Studies Officer/Research Nurses). As secondary sampling criteria we included parents/carers whose children represented a range of different ethnicities to capture the experiences of using emollients for different skin types.

The week 16 interviews were conducted from each group of the trial as soon as possible after participants’ primary outcome had been collected or their week 16 visit conducted by the Clinical Studies Officer, whichever was later. Sampling criteria were the same as the week 4 interviews, with the addition of participant intentions regarding future use of the allocated emollient (e.g. intending to continue, stop or switch to another type of emollient, or already stopped).

Qualitative data collection

Data were collected by an experienced qualitative researcher based in Bristol (ESu) using semi-structured face-to-face or telephone interviews. For the week 4 interviews, participants in the Bristol area were interviewed face to face in their home on all but one occasion. Interviews with participants from the other research sites and the repeat interviews were conducted via telephone.

The interview process followed an agreed study protocol using NHS Research Ethics Committee approved topic guides. The topic guides (see Appendix 3, Topic guide) were informed by the qualitative study aims, the feasibility study16 and the wider literature. They were used flexibly to allow unanticipated issues to emerge and to incorporate new topics that were developing from ongoing analysis. For example, following discussion between the qualitative research team and the chief investigator at analysis meetings, and also as a result of input from the study patient and public involvement (PPI) group, some additional prompts were added during the week 4 interviews. We used a range of prompts building on the topics covered with parents/carers, to enable the child to contribute their views and experience.

The week 4 interviews focused on the initial acceptability of the assigned emollient to families; their prior experiences and beliefs about emollient use; their experiences of early use of the assigned emollient; whether, and how, their use of emollient was consistent with, or differed from, recommended use and anticipated adherence to the emollient during the trial. We elicited barriers to and facilitators of use, particularly examining these with participants who stopped using the emollient early or switched to another type of emollient.

At the week 16 interviews, participants reflected on their emollient use over the full trial period to that date. The topics covered were similar to the week 4 interviews, with additional prompts relating to their overall experience, acceptability and effectiveness of the assigned emollient, and planned future use of emollients. These interviews were also informed by themes emerging from the week 4 interviews.

When parents and children were interviewed together, telephone interviews were conducted via speakerphone. A subtopic guide was created (see Appendix 3, Topic guide), adapting questions to enable children to participate.

All interviews were audio-recorded using an encrypted digital recorder. The qualitative researcher (ESu) met the senior qualitative co-applicants (ARGS, JB) on a fortnightly basis to share progress and discuss issues arising in interviews regarding process and content.

Qualitative data analysis

The interviews were transcribed verbatim by a University of Bristol (Bristol, UK) member of staff and fully anonymised by Eileen Sutton. Coding and data management were aided by NVivo12 software (QSR International, Warrington, UK).

Analysis took place alongside data collection in an iterative process, to allow insights from earlier interviews to shape future interviews. 37 Analysis was led by the qualitative researcher (ESu) with detailed input from the senior qualitative researchers (ARGS, JB), who read and independently conducted preliminary coding on a subset of transcripts from both time points (week 4 and week 16). The chief investigator (MJR) also contributed, but only had sight of redacted transcripts and data excerpts to preserve masking.

Using a thematic approach drawing on Braun and Clarke,34 analysis incorporated a combination of deductive (based on the qualitative study aim) and inductive (data-driven) coding strategies, from which a preliminary coding framework was developed (see Appendix 3, Coding framework). The coding framework was repeatedly refined through discussion within the qualitative team and incorporated feedback from the PPI panel, the TMG and the TSC. The codes used for the week 4 and week 16 interviews were broadly the same, with the addition of some codes for the week 16 data. Eileen Sutton flexibly applied the coding framework to the whole data set, with Alison R G Shaw, Jonathan Banks and Matthew J Ridd checking a subset of the coding for coherence and completeness. Alison R G Shaw, Jonathan Banks and Matthew J Ridd read and independently coded a subset of interview transcripts.

Following coding of the whole data set, the coded data were examined and compared both across and within trial groups. Themes and subthemes were identified and refined through continual comparison of data elements with each other in an iterative manner. A narrative summary of the findings from the interviews was produced (by ESu, JB and ARGS), attending to areas of divergence and convergence in the data sets and the different perspectives represented.

Parent and public involvement before the study

The design of the study to answer the James Lind Alliance question was informed by face-to-face discussions and a ‘tweet chat’ with parents of children with eczema who supported the COMET study, the feasibility trial that preceded this study. We also undertook an online survey of 176 parents/carers of children with eczema, which we have previously reported. 17

Parent and public involvement during the study

Co-applicant Amanda Roberts, a mother of children with eczema and who also has the condition herself, fully participated in study development, delivery and dissemination planning. She brought her experience from involvement in this capacity on previous primary care eczema trials (CREAM,40 BATHE41) by attending TMG, lay advisory group and other ad hoc meetings as appropriate.

We convened a group of parents of children with eczema who met six times over the course of the trial (in person, except for the final meeting owing to COVID-19 restrictions). We also involved members of the UK Dermatology Controlled Trials Network patient panel, mainly by e-mail but also through two online meetings towards the end of the project. Members had lived experience of different skin diseases, but were not necessarily parents of children with eczema.

The TSC included lay contributors, initially Abbi Gutierrez (2017–8) and then Sariqa Wagley (2019–21). They commented on trial progress and findings and provided oversight from a parent perspective.

The remit of different parent and public contributors was agreed at the beginning of the study, aligned with the aim and objectives above. Their collective focus was (but not restricted to) assisting with protocol development and design of all patient-facing study materials, identifying potential problems and helping to troubleshoot unanticipated difficulties, and helping with dissemination and planning pathways to meaningful impact.

Support for parent and public involvement

All parent and public activity was supported by a dedicated co-ordinator, Julie Clayton. She liaised with public contributors, facilitated meetings, and dealt with queries and administration of payments. Dr Clayton developed role descriptions and terms of reference for the public advisory group, and provided support and training opportunities (e.g. FutureLearn Research Ethics and Clinical Research in Health Care online courses).

All public contributors were sent copies of the newsletter, which was sent to study participants on an approximately 4-monthly basis.

Other public engagement activity

The lead trial team undertook three public engagement events, led by Shoba Dawson and Anna Gilbertson, supported by two public engagement grants from the NIHR Schools for Primary Care Research. These took place in different venues close to communities that were under-represented in health research and involved two University of Bristol (Bristol, UK) medical students (Jonathan Chan and Alisha Bhanot) in planning and delivering the events.

The first event was in October 2019, at Barton Hill Settlement in East Bristol, as part of ‘Fun Palaces’, a national initiative designed to support people to co-create cultural and community events across the UK and worldwide, with local communities co-developing and hosting events. It involved art and other activities designed to engage with parents and children about eczema. Co-applicant Amanda Roberts attended and shared her experiences of living with eczema, being a mother of children with eczema and being involved in research as a public contributor.

The second event was in January 2020, at St. Werburgh’s Community Centre, East/Central Bristol, in partnership with a South Asian community group. It was more structured, with short presentations and plenty of opportunities for attendees to ask questions. Community group organisers acted as interpreters throughout the event when needed.

The third event was in September 2020 during National Eczema Awareness Week, ‘hosted’ by East Bristol Children’s Centre (a group of four children’s centres); owing to COVID-19 restrictions, this event was online and so was advertised more widely.

In addition to regular tweets via the study Twitter account (@bee_study), we published a series of blogs by different team members on the study website.

Parent and public involvement before the study

Discussions with parents face to face in groups and online confirmed that the question ‘Which emollients are the most effective and safe in treating eczema?’ remained highly relevant. They also supported comparing the four main types of emollient (lotion, cream, gel or ointment) for treating eczema but that a ‘no emollient’ group would not be acceptable. Parents thought that the evaluation should be driven by patient-reported outcomes. They also advocated for qualitative research to understand the trade-offs described between effectiveness and acceptability.

We discussed issues of masking with public contributors, including the influence of brand and packaging on use and the perceived effectiveness of different products. In common with participants in the feasibility trial and respondents to our public pre-grant application survey,17 public advisory group members and their children had used many different emollients. Although some disliked the ‘medicalised’ nature of the emollients and may have initially valued products that look more attractive and ‘cosmetic’, they also told us that the proof was in their use – that is, whether it helped with eczema symptoms and did not cause any harm. Attempts to mask users to the intervention would have required repackaging or overpackaging of products, and public contributors were concerned about the potential effects of on the usability and portability of the emollients. For these, and the other reasons listed in Chapter 10, Strengths and limitations, Pragmatic randomised trial, Internal validity, we decided not to try to mask participants to the emollients.

Parent and public involvement during the study

Co-applicant Amanda Roberts attended most TMG meetings and acted as a ‘critical friend’, giving sufficient challenge and support as needed over time. She also liaised with patient stakeholder groups, such as the National Eczema Society (London, UK) and the Nottingham Support Group for Carers of

Children with Eczema (Nottingham, UK). Online, she helped to promote the study through her @eczemasupport Twitter feed (with over 6500 followers).

We initially established a group local to the lead centre of four parents of children with eczema, who were variously involved in the feasibility study and in developing the BEE study application and had expressed interest in continuing involvement. We expanded membership of the group to nine, and recruited new members at later stages, as some parents were unable to continue for the full duration of the study owing to work and other commitments. For example, new contributors joined following public engagement work in 2019/2020.

Other public engagement activity

Only 10, predominantly white, families attended the first public engagement ‘Fun Palaces’ event. Discussions with attendees revealed a low awareness of the different types of emollients available or how to use them safely and effectively. After the event, we shared our experiences and lessons learned with the public advisory group and their suggestions (reducing the number of activities and adopting a more structured approach) shaped the second event. This was attended by approximately 20 people, primarily of South Asian origin, and led to new members joining the public advisory group. More detail on both the events can be found in a subsequently published paper. 42

The third, 1-hour Zoom (Zoom Video Communications, San Jose, CA, USA), event was attended by 16 people, including student nurses, mothers and grandmothers of children with eczema, and a health visitor. Eight attendees returned online evaluation forms, with an overall score of 4 out of 5, and statements that the event was either ‘useful’ or ‘very useful’. One responded that the event was ‘[v]ery well planned and executed. Was left in awe of [those] who are doing so much for children suffering eczema. The practical tips were invaluable.’

By June 2021, we had published four blogs on the study website: ‘Including the views and experiences of parents and children in a clinical trial’ [Eileen Sutton, May 2019, URL: https://capcbristol.blogs.bristol.ac.uk/2019/05/20/including-the-views-and-experiences-of-parents-and-children-in-a-clinical-trial-the-best-emollient-for-eczema-bee-study/ (accessed 10 September 2022)], ‘Finding the best moisturiser for eczema’ [Zoe Wilkins, May 2020, URL: https://capcbristol.blogs.bristol.ac.uk/2020/05/20/finding-the-best-moisturiser-for-eczema-the-impact-research-can-have-on-everyday-lives/ (accessed 10 September 2022)], ‘Why does the type of moisturiser matter to a child with eczema?’ [Sue Davis-Jones, July 2020, URL: https://capcbristol.blogs.bristol.ac.uk/2020/07/29/why-does-the-type-of-moisturiser-matter-to-a-child-with-eczema-a-research-nurses-perspective/ (accessed 10 September 2022)], and ‘A for planning, BEE for impact’ [Amanda Roberts, November 2020, URL: https://www.bristol.ac.uk/primaryhealthcare/researchthemes/bee-study/blogs/ (accessed 10 September 2022)]. Further blogs to accompany the publication of the trial results are planned.

Primary outcome

Across the treatment groups, mean (SD) POEM scores at baseline were lotion 8.7 (5.2) points, cream 9.3 (5.3) points, gel 9.8 (5.4) points and ointment 9.5 (6.0) points, with a range of 0 (no eczema symptoms) to 28 (very severe eczema symptoms) points. By week 16, scores had fallen slightly and the mean (SD) POEM scores at 16 weeks were lotion 5.7 (4.5) points, cream 6.7 (5.1) points, gel 6.7 (5.8) points and ointment 6.1 (5.9) points.

No differences in mean POEM scores were seen between the different groups over the 16-week primary outcome period (repeated-measures analysis of POEM scores, p = 0.765; see Table 8). (POEM scores by week and group are given in Appendix 2, Table 34). There was also no evidence of a difference in mean POEM scores over the first 16 weeks of the study between treatment groups in any of the pairwise comparisons (see Table 8).

TABLE 8 - Differences between treatment groups in weekly POEM scores over weeks 1–16 analysed using a linear mixed model

Adjusted for baseline scores and all stratification and minimisation variables used in the randomisation.

	Allocated emollient (number of participants included in model/number of participants randomised)				Univariate difference in mean POEM (95% CI)	Adjusteda difference in mean POEM (95% CI)	p-valuea
	Lotion (n = 131/137)	Cream (n = 137/140)	Gel (n = 130/135)	Ointment (n = 126/138)
Mean POEM score (SD) (points)	6.79 (5.14)	7.62 (5.37)	7.45 (5.81)	7.04 (6.13)			0.765
Pairwise comparisons
Lotion vs. cream					0.87 (–0.27 to 2.00)	0.42 (–0.48 to 1.32)	0.360
Lotion vs. gel					0.84 (–0.31 to 1.99)	0.17 (–0.75 to 1.09)	0.718
Lotion vs. ointment					0.42 (–0.74 to 1.58)	–0.01 (–0.93 to 0.91)	0.983
Cream vs. gel					–0.02 (–1.16 to 1.11)	–0.25 (–1.15 to 0.65)	0.586
Cream vs. ointment					–0.45 (–1.59 to 0.70)	–0.43 (–1.34 to 0.48)	0.354
Gel vs. ointment					–0.42 (–1.59 to 0.74)	–0.18 (–1.11 to 0.75)	0.704

‘Per-protocol’ analysis

Participants were consistent in returning POEM scores over the follow-up period, with at least one POEM score reported monthly by 458 out of 550 (83.3%) participants. In contrast, only 276 (50.2%) participants reported emollient use every month, with 152 (27.6%) reporting emollient use on 60% of days (see Appendix 2, Table 38).

When restricting the analysis of the primary outcome to these 152 ‘per-protocol’ participants, there was no evidence of a difference in mean POEM scores over the first 16 weeks between treatment groups or in any of the pairwise comparisons (p = 0.238; see Appendix 2, Table 39).

Sensitivity analyses

Subgroup analyses

We planned four subgroup analyses to assess whether or not treatment group differences were modified by different baseline characteristics: parental expectation of effectiveness, age, eczema severity and meeting the UK diagnostic criteria for atopic dermatitis. The findings of these analyses are presented in Table 17. There was no evidence that treatment effects differed by any of the subgroups studied.

Adverse events

Patient-reported adverse reactions are presented in Tables 18 and 19. Table 18 presents the total number and type of adverse reactions, and Table 19 presents the number of patients and type of adverse reactions. Overall, 37% (205/550) of participants reported at least one adverse reaction, but there was no evidence that the proportion of children reporting adverse reactions in the first 16 weeks differed by treatment group (p = 0.794).

The most reported adverse reactions were ‘application site reactions’. Overall, the ‘top five’ reported problems were worsening of eczema (19.6% of adverse reactions and 24% of participants), itching (17.5% and 23%, respectively), dryness (15.0% and 16%, respectively), redness/inflammation (14.2% and 20.0%, respectively) and stinging (13.1% and 16%, respectively).

Skin infections appeared to be more common with gels and ointments (3% of participants) and slips or falls with creams (3%) and ointments (3%). However, it should be noted that the number of events (12 skin infections in 12 participants and 12 slips or falls in nine participants) is small.

There were no significant adverse events.

Study lotions

Most interviewees felt that in terms of effectiveness the lotion they were allocated was ‘about the same’ as their pre-study emollient and had ‘maintained’ rather than improved the skin:

Well I’d say his skin generally is probably about the same … I don’t know whether it maybe is taking slightly longer for the dry patches to clear up, but I just make sure that I moisturise him regularly.

B04, Mild, ≥ 7 years, SL, week 4

One parent reported a decrease in the number of flare-ups while using their study lotion, but most felt that this and other eczema-related symptoms had not lessened:

Doesn’t always necessarily get rid of the eczema but sometimes it just stops it getting any worse, stops her getting any flare ups.

N06, Mild, 0–7 years, SL week 16

His scratching is probably about the same to be honest.

S03, Mild, 0–7 years, SL, week 4

Some parents noted that they were having to apply the lotion more frequently than emollients of other types for it to be effective:

I feel like if we just applied it twice a day, I don’t know that his skin would be quite as good … it’s gone so quickly and then his skin gets quite – with the cream his skin just, it didn’t feel rough again. It would feel rough again before you applied it again in the evening but not a couple of hours after. You do a nappy change couple of hours after and you think that’s already getting a bit dry and rough again.

SO3, Mild, 0–7 years, week 4

Only one of the parents interviewed reported a noticeable improvement with their child’s skin while using their study lotion, with both the parent and child confirming this at the week 16 interview:

I’d say it was better with the new one.

B06 child, Mod, ≥ 7 years, SL, week 16

Yeah, it seems to have improved, yes, ‘cos there were a couple of quite bad sore patches, quite oozy. Now the worst ones seem to have cleared up completely.

B06, Mod, ≥ 7 years, SL, week 16

One of the children believed that their skin had improved with the lotion, whereas their parent saw it as of a similar efficacy:

I think it’s better than the one I had before. Because the rest of my itchy spots have gone.

B04 child, Mild, ≥ 7 years, SL, week 4

Well I’d say his skin generally is probably about the same.

B04, Mild, ≥ 7 years, SL, week 4

Most participants who felt that lotion was equivalent or better than their pre-study emollient intended to keep using the lotion after the 16-week study period.

Adverse effects such as stinging on application of lotion were reported by some, but this tended to be an issue when applying to a ‘raw’ or ‘fresh’ patch of eczema:

It’s only like sometimes when this one [points to raw patch of eczema on arm] is like worse, then I put it on and it stinged.

B06 child, Mod, ≥ 7 years, SL, week 16

Flare-ups, itching and worsening of eczema were also reported. Some participants with these symptoms carried on using their study lotion and the situation improved. However, for others, the problems they experienced led them to stop using their lotion. Its perceived lack of effectiveness, for a number of families, was linked to it being too ‘thin’ and unable to sufficiently moisturise dry skin:

It was so thin and watery, it had no effect whatsoever and we had to go … can’t remember if we went to the doctors and got a different like better emollient ‘cos it just wasn’t effective. I just don’t think it was effective enough to keep a flare-up at bay.

B16, Mod, ≥ 7 years, SL, week 16

Two of the participants who had reported issues with using their study lotion and had stopped using it revealed existing preferences that may have had an impact on their assessment, although one had never used a prescribed lotion previously:

It was like she had the first-time round. When we got randomised to that Certraben [lotion] I thought ‘Oh no!’ [laughs], I’ll give it a go but I’ve done this before so I expected the same kind of results [itching, child scratching until skin bled] which did happen.

N02, Mod, 0–7 years, SL, week 4

I’d already got a bit of a preconceived idea of what it was going to be like … [but] I wanted to give it a go because again ‘cos it was a prescribed one as well, it was potentially different to the shop-bought lotions so I was happy to see if it made any difference.

B15, Sev, 0–7 years, SL, week 16

Study creams

Some participants reported that their study cream was more effective than their previous emollient. One parent was dubious about changing from their pre-study emollient and had not previously used the cream they were allocated, but reported that they were very happy and saw it as a positive change:

Still using it, yeah, and I prefer it to Diprobase [pre-study lotion]. I don’t know, it just seems to have a better effect on his skin so I don’t know if it’s ‘cos it’s a little … well I think it’s oily by comparison but, yeah, he seems to have done really well with it … I was really worried, sort of got used to Diprobase ‘cos we’ve tried a few different ones and ‘cos we’d got used to Diprobase, I was a bit worried about changing it but it changed for the positive actually.

S07, Mod, ≥ 7 years, (study cream) SC, week 16

Another parent had been wary about the effectiveness of their study cream when they were first using it, describing their child’s skin as ‘sandpaper rough’, but they persevered and reported that ‘her skin is better, it seems to be working well’. They recognised that the time of year may also have had an impact:

I don’t know if it’s the cream, I don’t know if it’s the sun, I don’t know if it’s the warmth but her skin has suddenly become really good so we actually went over the 16 weeks and there was times in that I was thinking this is not doing anything to suddenly really good.

B12, Mod, 0–7 years, SC, week 16

Reductions in flare-ups and symptoms were also reported:

Certainly in terms of patches of like flary-uppy eczema we’ve hardly had any at all, even when she’s kind of been away with the Brownies [Girlguiding, London, UK] and stuff where we know she’s perhaps not as thorough about applying [emollient] as she could be. Normally that would be a trigger for a flare-up and that hasn’t happened at all, so yeah, it’s a lot better … It felt a bit like we were fighting a losing battle of just sorting one flare-up out and then you were getting another one whereas now we’re hardly – I mean you’ve not had a bad flare since you’ve been on it have you? It’s been brilliant.

N03, Mod, ≥ 7 years, SC, week 4

I fill in those weekly surveys and you’re surprised how quickly that goes, but probably maybe once or twice we’d had to use hydrocortisone but it’s never been really bad so we haven’t had cracked or bleeding and it’s gone quite well.

S07, Mod, ≥ 7 years, SC, week 16

One family reported that despite experiencing issues at first, they persevered with their study cream because they were committed to the trial, and things had improved. Similarly, another felt that it was important to give the skin time as it adjusts to a different emollient (see also Study participation and emollient use, Using the study emollient for longer – ‘persevering’):

I think just recently we’ve not been putting it on so much so we’ve been weaning off a bit ‘cos his skin’s gradually improving. I think the first few weeks his skin was quite bad with the Epimax [study cream] but its slowly getting better. It’s adjusting to a different cream [generic].

B05, Mod, 0–7 years, SC, week 4

A few were neutral, some reporting that their study cream was maintaining but not improving their child’s skin, that it ‘kept it at bay’ or was ‘keeping things going’:

I’d say it’s kept it at bay … I can remember as a child the creams never do away with the eczema completely, but it helps soothe it. It’s one of those things that happens.

S05, Mod, 0–7 years, SC, week 4

Stinging on application was reported by a number of participants. Although this was mainly related to applying the cream to ‘raw’ patches of eczema, it did have an impact on families’ ongoing use of cream. One family was in the process of arranging to see their GP to request an alternative emollient as they felt that in addition to stinging on application the cream was ineffective, leading to a worsening of eczema that had then become infected:

It stings where the main part is on my arm, so like the bend.

B10 child, Mod, ≥ 7 years, SC, week 4

So it’s never stung before when we’ve put previous creams [generic] on. So this one’s done straight away and some days it’s a lot worse than others. I think it depends how raw her eczema is … ‘cos I filled out my form and said there was just this really gross thing on her arm and it kind of burst and a load of yellow pus came out and then there was something else there and my husband sort of give it a bit of a like a … hard white lump just came out.

B10, Mod, ≥ 7 years, SC, week 4

Three families who had already stopped using their study cream at the time of the interview reported ‘stinging’ and symptoms such as ‘itching’ and the appearance of ‘spots all over my body’. Another parent reported that it could not control their child’s eczema. One parent reported several attempts at using it, deciding to stop before the week 4 interview as their child’s skin ‘… went all bobbly and he was really, really itching’ (S01, Sev, ≥ 7 years, SC, week 4).

Another of these parents reported that they did not expect the cream to be effective owing to prior negative experiences, but was surprised at how rapidly their child’s skin had deteriorated:

So as soon as I was told that was the one I knew that it wouldn’t work anyway but I didn’t expect it to – I think within 2 or 3 weeks his skin was worse than it had ever been … the Zerobase [study cream] obviously wasn’t good enough to then control it and it went nuts …

S08, Mild, 0–7 years, SC, week 16

One participant had to use topical corticosteroids for the first time in a while and another reported that they had to use these more frequently over the 4 weeks:

During the first week. It was quite quick unfortunately. Yeah, so we had to go back to using steroids which we hadn’t used for quite a while.

B05, Mod, 0–7 years, SC, week 4

It was wrecked.

B05 child, Mod, 0–7 years, SC, week 4

Study gels

The study gel was unusual in that participants who were happy with this emollient tended to report using it alongside other treatments, including other emollients and topical corticosteroids, particularly on certain areas of the body:

It were brilliant to be honest with you. It’s the Isomol gel [study gel]. So it were brilliant to be fair. It’s a combination, we’ve been using a combination if I’m honest, we’ve been using the Isomol gel with the Oilatum [non-study cream], with a bit of the [bath emollient]. To be honest with you I believe it’s a combination of those three things that’s cleared her eczema up. She’s not had a breakout since I’d say the second week that we started the study.

N09, Sev, 0–7 years, (study gel) SG, week 16

Like I say, our magic cream [generic] is the steroid … to maintain it well. But when it flares up, because he’s just finished Year 6 SATs [Standard Assessment Tests] as well and the weather’s been changing, so I think all that as well, we still need actually use steroid cream [generic] unfortunately in combination … but I think of the normal regular ones Isomol [study gel] seems to work.

S04, Mod, ≥ 7 years, SG, week 16

We carried on using it because on the skin that wasn’t broken it was fine and it seems to do a decent enough job. I would say it’s just as good as the Cetraben [pre-study lotion] that we’ve been using. It seems to give the same kind of level of moisturising and its kinder to his clothes, it’s not really greasy, it seems to soak in well. So we carried on using it and then just literally would put something else on his face and hands when he needed it.

B18, Mild, 0–7 years, SG, week 16

A number reported adverse effects including stinging, particularly when applied to broken skin, as well as soreness and redness. Some persevered despite adverse effects and the skin improved:

Initially when we started applying, he almost straightaway got a slight flare up and we were a bit concerned … but we persevered and it was fine.

N07, Mod, 0–7 years, SG, week 16

Others continued but reported negative effects:

It’s got like a bit more of a cracked, bleeding.

S04 child, Mod, ≥ 7 years, SG, week 4

It stings, sometimes, you know, she cries and she says, ‘oh I want my ointment back’.

S04, Mod, ≥ 7 years, SG, week 4

Two families who had already stopped using their study gel at time of interview reported that it was not effective in treating eczema, with one reporting an immediate adverse reaction to using it:

I didn’t think it made any difference to her eczema. It didn’t, you know, she still had like really red dry bits, so I don’t think it really helped, considering she was putting it on probably … twice a day, so sometimes even more.

B14, Mod, ≥ 7 years, SG, week 16

Terribly from the off to be honest … by the end of that day it was red and really bad and within a couple of days she was screaming and crying in pain whenever we were putting it on her … so we were probably only on it 3 or 4 days I think.

B07, Sev, 0–7 years, SG, week 4

A further participant continued using their study gel until week 16 but decided to change as they had not found it effective in hydrating the skin:

Most of it was the fact that his body had gone back to being like the sandpapery effect when he was a baby … no it had been like 9, 10 years.

S11, Mild, ≥ 7 years, SG, week 16

Study ointments

Some parents and children reported an improvement using study ointment with a decrease in symptoms, such as itching, and a reduction in flare-ups:

We currently use an [study] ointment and we have used it and since then it has cleared up beautifully, like round her neck and on her arms you can barely tell she’s [got eczema] … massively [improved].

B11, Mod, ≥ 7 years, SO, week 16

The chest and the back of her neck has cleared up a lot. Obviously it’s still there and noticeable but it’s not as bad as it was before.

B08, Mild, 0–7 years, SO, week 4

It was quite good. It helped.

S09 child, Sev, ≥ 7 years, SO, week 16

One of the children thought that their study ointment was more effective than their pre-study emollient, but their mother disagreed, believing it to be comparable. Nevertheless, this parent conceded that they preferred using it as it was easier to apply (see Acceptability of emollients, Study ointments). A further parent was unsure that it was as effective as their pre-study emollient:

I don’t think it’s worse, but I don’t think it improves it ‘cos she does still get the flare up of the very fine bumps that comes with her eczema. I don’t know if it’s as effective.

B09, Mild, 0–7 years, SO, week 4

Adverse effects such as stinging on application, redness and itching were reported by some along with a worsening of eczema symptoms:

Within a couple of days I would say he was starting to have a flare … He did say it was stinging but whether that’s stinging or just that tingly sensation that you get from a different emollient or that his skin was particularly sensitive that day, though … he never complains about the cream I would say. I would say that is a big difference.

B17, Mod, 0–7 years, SO, week 16

We had to stop using it after two days because she just broke out in a horrible, itchy rash all over her body.

N01, Mod, 0–7 years, SO, week 4

All the families who had stopped using their study ointment at the time of their interview reported experiencing adverse effects and a worsening of eczema, although some had persevered with using it for the 16-week period but then changed back to their previous emollient:

So I thought it felt lovely and I was hoping it would work because when we put it on it felt really lovely on her skin when you were putting it on, but then it just seemed to make her skin worse so her skin went really red with it.

N05, Mod, 0–7 years, SO, week 16

Not too great. Like I said its one of the ones with the texture that she hates and it’s … spread her eczema now, it’s kind of all over her back, all over her belly and its down on her bum and things where she’s never had it before.

S06, Sev, ≥ 7 years, SO, week 4

One of these parents who changed after 16 weeks indicated an existing preference:

[Research nurse] had obviously gone through them with me and she said were there any that you wouldn’t do and I said at the time, I said, ‘well I don’t get on with the ointment so I wouldn’t be particularly pleased to get that, but that I would do whichever one that we got’.

B17, Mod, 0–7 years, SO, week 16

Temperature change/heat was mentioned by three of those who had stopped using their study ointment at time of interview and one whose skin had improved (see also Acceptability of emollients, Study ointments and Participant characteristics and perceived effectiveness and acceptability):

It was OK for a few weeks, first three weeks, 2 to 3 weeks it was fine, then I think with the temperature change as well ‘cos it dropped quite a lot … between October and November it’s dropped quite a lot. I didn’t really notice till [son] said, ‘my skin’s itching again,’ … I don’t know whether it’s the temperature change or – you don’t know do you?

B02, Mod, ≥ 7 years, SO, week 4

Study lotions

The study lotion was generally reported as easy to apply and as cooling to the skin:

It goes on well, it’s just a bit thin, a bit runny …

B01, Sev, ≥ 7 years, SL, week 4

I quite like it … and it’s cold.

B01 child, Sev, ≥ 7 years, SL, week 4

It is cold which [I] quite like that.

B01, Sev, ≥ 7 years, SL, week 4

It was reported to soak into the skin well, which was a plus for most participants, but some felt that the lotion was too thin and watery to sufficiently moisturise the skin and preferred an emollient that left a visible residue or protective barrier. As previously noted (see Perceptions of effectiveness of emollients, Study lotions), participants reported that they needed to apply it more often or use a greater amount:

I think the lotion that he’s got now soaks in a lot better. Obviously it’s a lot thinner so … specially the dry patches that he’s got at the moment, I’ll put some on, just a little bit, rub it in and almost make sure it’s soaked in and then I tend to put some more on whereas with the thicker one you could put one lot on, you could kind of see it all and know it was going to stay on, although back then I think probably sometimes it rubbed off on clothes anyway.

B04, Mild, ≥ 7 years, SL, week 4

It’s gone so quickly and then his skin gets quite – with the cream his skin just, it didn’t feel rough again. It would feel rough again before you applied it again in the evening but not a couple of hours after. You do a nappy change couple of hours after and you think that’s already getting a bit dry and rough again.

S03, Mild, < 7 years, SL, week 4

I think I had to use it more because it wasn’t moisturising his skin as much as I’d hoped it would.

B15, Severe, 0–7 years, SL, week 16

Participants generally found the smell of their study lotion to be acceptable.

Study creams

There were differing opinions on how well the study cream was absorbed into the skin, with some reporting that it was easy to apply and absorbed well and others reporting that it took a while to ‘rub in’. The time taken for the study cream to dry was an important factor for some in the context of busy family life:

The Epimax [study cream] I found isn’t sticky, it seems to absorb quite quickly, ‘cos even – Daddy can put it on you and then say I’m sorting out rooms for bedtime then I can come in and feel her arm and I don’t know if he’s put it on or not.

B12, Mod, 0–7 years, SC, week 16

It’s almost on the skin for a while. You can see it and it doesn’t quite absorb as quickly as maybe the Balneum [pre-study cream] did so maybe need to wait a while to put his pyjamas on or his clothes on.

B05, Mod, 0–7 years, SC, week 4

It’s pretty hard to rub in and it takes like 3 or 4 minutes to rub in.

B10, Mod, ≥ 7 years, SC, week 4

Participants described the consistency of their study cream as ‘thick’ and ‘gloopy’, one child commenting that their cream was ‘disgusting’. Those who saw it in these terms also described it as being difficult to apply:

If I was honest and had to describe it [laughs] I would say it’s really, really like, oh, paint, what’s the paint, the thick … it’s like gloss paint. It is like, you know, the really thick stuff, so not the one-coat.

B10, Mod, ≥ 7 years, SC, week 4

It’s quite sort of gloopy and I don’t know whether applying it is that easy whereas with the Balneum [pre-study cream] it was a lot smoother and more easier to soak in.

B05, Mod, 0–7 years, SC, week 4

The thickness of the cream was thought by some to be a good attribute as it could provide a protective barrier on the skin, or it meant that less needed to be applied in comparison with other types of emollients they had used, particularly lotions:

I think it seems to be slightly thicker so she’s still putting the same amount on that she would have done with the Cetraben [pre-study lotion] but it’s actually more appropriate, so you don’t actually need as much.

N03, Mod, ≥ 7 years, SC, week 16

The smell of their study cream was not a significant issue, with just two families commenting that ‘it smells like paint’ or that ‘it doesn’t smell great’.

Study gels

A number of participants described the study gel as absorbing into the skin quickly, which was seen as a positive attribute:

I think just the ease of the emollient, you know, being absorbed by the skin is quite helpful because when it’s … just sitting on top is not something very attractive to use.

S02, Mod, ≥ 7 years, SG, week 16

However, one of the children reported that ‘white bits’ remained visible on the skin despite rubbing it in well.

Some participants felt that as it was thinner, they needed to use more, in comparison with creams and ointments they had previously used. Others reported that they were not happy with the way the gel made their skin feel, commenting that it ‘felt cold’, ‘like clay’ or was a little ‘sticky’:

Well it kind of stuck a little bit when I like put my clothes on and when you put it on it was like all cold on your skin and it felt really cold a little bit.

B14 child, Mod, ≥ 7 years, SG, week 16

It’s like clay. I don’t really like the feeling and like what it does to my skin. It kind of makes my skin sticky and then it feels weird, when I put it on my skin inside my skin it feels like really weird. When I itch or when it becomes itchy it does start to sting as well.

S04 child, Mod, ≥ 7 years, SG, week 4

Participants had less experience of using a gel prior to the study and some were surprised at its appearance, expecting it to be translucent rather than white, or thinner, and these pre-conceptions may have had an impact on their views:

It’s quite creamy, it’s called a gel for some reason … I thought a gel was like – I thought it was translucent.

S02, Mod, ≥ 7 years, SG, week 4

I always thought a gel should be a bit more like watery content … But I think its content is almost exactly the same as Zeroveen [non-study cream], which … is a cream so, yeah, but fine, if they want to call it a gel its fine. When I found out it’s a gel, because I had never heard of this product before, try the gel, let’s see what happens. But as far as I was concerned, as I said because gels tend to be more watery and like thinner, thought it might not be as effective at hydrating his skin.

N07, Mod, 0–7 years, SG, week 16

Participants generally found the smell of their study gel to be acceptable.

Study ointments

The perceived thickness of ointment meant that, for some, it ‘seems to last longer on the skin’ and therefore could provide a protective barrier:

So, I thought it felt lovely and I was hoping it would work because when we put it on it felt really lovely on her skin when you were putting it on.

N05, Mod, 0–7 years, SO, week 16

Participants generally felt that as it was thick, their study ointment took longer to absorb into the skin than other types of emollient they had used. Some found this difficult in the context of a busy family life, whereas others found ways of coping if they found their ointment effective:

It does absorb into the skin though not as quickly and I found it was particularly difficult because – so on a morning we need to be out of the house by 07.20, quite early, three of us. Dad goes at 05.30 so he’s long gone. It was adding extra time to – I had to get [son] up early ‘cos the cream it sunk in, you turn around and its sunk in, the ointment takes what felt like quite a lot longer so I almost had to put a towel, do him head to toe, let him lie on the towel while I was getting … baby ready or whatever and then come back, check, and then if we were putting steroid on then we’ve got to apply the steroids and then go off again and then OK, right, well you can get dressed now.

B17, Mod, 0–7 years, SO, week 16

It makes a huge difference how well it absorbs and with him, when his skin’s really bad … like getting into the habit of putting a towel on the bed, putting the telly on, a programme that he likes, and then after he’s had a nice warm soak in the bath, drying him off sort of partially, then putting it all over him and just letting him lie on the towel with nothing on while it soaks in before I put his pyjamas on. It is all about having a technique, about how we put it on, when we put it on.

B13, Mild, ≥ 7 years, SO, week 16

However, the ‘greasy’, ‘oily’ or ‘waxy’ qualities led to problems with clothing for some participants as the ointment stuck to, or stained, clothing:

I was going to say it’s nice, isn’t it, it doesn’t sting your legs, she hasn’t complained ever about it and so that’s, you know, really good from her point of view. I suppose the worst thing about it is the fact that it does mark her clothes.

B11, Mod, ≥ 7 years, SO, week 16

Obviously it’s a bit slippery on the kind of flooring that we’ve got and the fact that you have to put clothes on over the top of it because of what it is and they do kind of stick so when it’s a hot day she’s got clothes like stuck to her.

S06, Sev, 0–7 years, SO, week 4

It’s worked really well. I think she finds it a bit strange ‘cos the texture, ‘cos its quite waxy isn’t it, and she’s used to creams so she’s a bit like, ‘oh, what’s that on me’.

B08, Mild, 0–7 years, SO, week 4

Participants reported issues with ointment and the weather or temperature. One parent raised concerns about using it in hot weather. There were also problems highlighted for cold temperatures, whereby the emollient became thicker and harder to apply:

I have to be like mindful, especially because the weather’s so hot, so like ‘cos I cover her skin with it so it’s like I don’t want to use it on her just before she goes in the sun ‘cos its very oily and I don’t know if it’s going to have an impact on the skin.

B09, Mild, 0–7 years, SO, week 4

Participants also disagreed on the ease of application of their study ointment, and again, temperature change/heat were important factors here:

They get quite thick so it’s hard, especially during winter when it’s like cold so they get harder … [laughs] so you kind of almost melt them or get them somewhere that’s quite soft because otherwise when you apply it to the skin in winter it’s almost like a brick … so like now during the warmer times it tends to be a bit more softer but during winter it’s like rock hard so applying it is quite tricky actually, ‘cos getting it out and like I’ll use – I’ll put it in the airing cupboard ‘cos its gets a bit warm so it softens a little bit. Other than that it’s very difficult to get out.

B09, Mild, 0–7 years, SO, week 4

As noted earlier, a parent and child disagreed on effectiveness (see Perceptions of effectiveness of emollients, Study ointments), but the parent preferred using it as it was easier to apply:

It’s better than the old one.

B13 child, Mild, ≥ 7 years, SO, week 16

I don’t think there’s a change but I don’t think that’s a bad thing ‘cos I think that we were happy with the Hydromol [pre-study ointment]. Hydromol was maintaining like an optimum level for him so it just seems to have maintained what I felt was already working to some degree … do you know actually there is improvement, not necessarily in his skin but I think more like what [son] was saying about its easier to apply, I think that’s where the improvement is.

B13, Mild, ≥ 7 years, SO, week 16

There was a tension around ointment use, more than other emollients, between perceived effectiveness and acceptability. The perception of it as a barrier, as good for the skin, was set against the difficulty in applying it and problems associated with it sitting on the skin more than other emollients:

To be honest it’s not the type of emollient that I’d go for again because it’s just too greasy.

N01, Mod, 0–7, SO, week 4

It’s not very nice to have all over your hands ‘cos it doesn’t – like with the cream, you put the cream on your hand and I can rub it into him say and then I can rub my hands together and it’s gone but with the ointment, ‘cos it’s got like a, I don’t know if it is, but it feels like it’s almost water repellent, it’s got that kind of texture … it feels more like a barrier cream. It’s got that Vaseline kind of feel about it.

B16, Mod, ≥ 7 years, SL, week 16

Epaderm [pre-study ointment] does help her so I can cope with skanky pyjamas for it [laughs].

B07, Severe, 0–7 years, SG, week 4

The latter parent reverted to using their pre-study emollient ointment in favour of the gel they were allocated because it was more effective, despite the effect it had on the child’s clothing. A few participants mentioned the ‘oily’ or slightly unpleasant smell of their study ointment, but this did not put off those who found it effective.

Decisions to continue with study emollient

For some participants there was a clear improvement, and it was therefore an easy decision for them to carry on using their allocated emollient:

So it seems to be keeping it under control better … the fact that it seems to be helping [was a] good incentive, so we weren’t counting down the weeks until the end of the study, having to decide whether or not it was worth carrying on … there was nothing really to think about.

B06, Mod, ≥ 7 years, SL, week 16

The data indicated that effectiveness was the primary driver of decision-making. This was evident in cases in which parents recognised the improved acceptability of their assigned emollient but were unable to keep using it because it did not control eczema in the way that their previous emollient had done:

It went on really nicely and she was quite excited about having this new cream [generic] … It just didn’t solve, yeah, if anything it made it so worse. I think if it had just kept it the same I would have probably carried on using it but I felt it was making it worse and so couldn’t then carry on.

N05, Mod, 0–7 years, SO, week 16

However, the data also highlighted the value placed on acceptability by participants, especially when the effectiveness of the emollient was negligible in comparison with emollients they had previously used:

Actually, there is improvement, not necessarily in his skin but I think like [son] was saying about it’s easier to apply, I think that’s where the improvement is. It’s easier to apply, it’s not as sticky … and uncomfortable as the other one.

B13, Mild, ≥ 7 years, SO, week 16

Trading off did not just happen at the 16-week point, for those that stopped using the study emollient at an earlier stage also had to grapple to obtain balance between effectiveness and acceptability:

I remember he liked the smell and I remember he liked the feel and him saying I like the smell of it. But in terms of effectiveness, I don’t think we remember but it just didn’t do anything … it just wasn’t rich enough. I don’t know, just didn’t have enough moisture in it.

B16, Mod, ≥ 7 years, SL, week 16

Decisions to discontinue study emollient

Some participants had persisted with their study emollient owing to their commitment to the study, but admitted that they reverted to their preferred pre-study emollient as soon as they reached the 16-week point, with the daughter commenting here that primary outcome period felt long:

We used it for the whole study … and the day after the study finished she went back to the other one [Zeroveen cream].

B14, Mod, ≥ 7 years, SG, week 16

This is really funny, I know from the BEE study that 16 weeks is a really, really, really, really long time [laughter].

B14 child, Mod, ≥ 7 years, SG, week 16

This family had an older sibling with very severe eczema and had found that Zeroveen (non-study cream) had worked best for them so reverted to using this for both of their children.

Another family who decided to change at weeks 16 explained that they had experienced issues with their study ointment and had reverted to using a combination of emollients that they already had at home. This was partly from convenience, but mainly because they were still searching for an effective emollient:

It really just seemed to flare-up and her arms seemed to go really red and ever drier … I just thought I can’t carry on just for the sake of the study and let it get worse … I’ve got three on the go, so I’ve got Aveeno, Cetraben and the Child’s Farm are kind of in the bathroom and in her bedroom and whatever’s nearest I get on her ‘cos I think well its better than nothing … I kind of hope that one of them works and I mean like I say if someone could tell me what the magic one was then I’d just use that but it’s – and I would pay anything for it, I don’t even mind what it costs but I just – I kind of think well I’ll use it till it finishes, you know, until I’ve finished it and then maybe I’ll buy a different one and see if that has a better effect … so that’s the way we’re going to play it I think.

N05, Mod, 0–7 years, SO, week 16

Another participant had decided to swap based on reasons related to acceptability, preferring the type of container of their pre-study emollient (see Emollient containers), as their study gel had not proved to be more effective than their pre-study emollient.

As noted in Acceptability of emollients, a number of participants had already stopped using their study emollient at the time of interview. Most had stopped at the early stages of the study, even before the week-4 interview. Those who stopped at this early stage tended to do so because of adverse reactions such as stinging, worsening of eczema and/or increasing flare-ups. Some stopped as soon as they experienced difficulties, whereas a few paused and then tried again (see also Study participation and emollient use, Using the study emollient for longer – ‘persevering’).

Severity of eczema

We found no clear preference for a particular type of emollient across the different levels of severity (mild, moderate and severe). However, participants whose children had very dry and/or rough skin tended to prefer an emollient with a thicker consistency, such as a cream or an ointment, because they were seen as better at keeping the skin hydrated or moisturised. All types of emollient were reported to sting on application to cracked or broken skin but there were slightly more reports of this occurring while using a lotion or gel.

Age of child

Parents reported that because it was easier to regularly apply emollients in younger children, for example at nappy change times; consequently, they may have been applied more frequently. As they became older, some children were less co-operative, with one parent describing it as a ‘battle’to persuade their child to have the emollient applied. With increasing independence and becoming more self-conscious of their body, some of the older children had started to apply their own emollients, but parents varied in how far they trusted their child to do this. Parents reported allowing this ‘if his skin is under control’ or stepping in if they thought their child was ‘struggling’:

He’s nearly seven so he doesn’t want mum putting cream [generic] on his legs and his back but if I look at it and think it’s bad I’ll do it, but generally he does it himself.

S10, Sev, 0–7 years, SL week 16

Some of the children in the older group reported a slight preference for types that were lighter and more easily absorbed, and hence quicker to apply, with two of these mentioning a preference for non-study emollients in a spray format, which made them easy to transport for use at school or when participating in physical activity.

Ethnicity

Our sample included 10 participants from different ethnic backgrounds (see Tables 20 and 21). Overall, we found that there was no clear pattern around emollient preference when taking ethnicity into consideration. There was evidence that participants related their child’s eczema and the way that an emollient worked or did not work to their ethnicity and skin type. In the example below, the participant explained that they had found ointments, including their study ointment, to be generally more effective for their child’s ‘skin type’ (white and black Caribbean), relating this to their ethnicity:

That would be an ointment I think … I mean everyone’s different but in our experience … like I’ve got friends who’ve got children who’ve got eczema and I don’t know if it makes a difference as well … in terms of skin type. I think personally potentially, ‘cos as a child I had eczema but my eczema’s completely different to his, I think ‘cos his mix is different to mine and he’s sort of a bit more Afro-Caribbean than I am.

B13, Mild, ≥ 7 years, SO, week 16

Regular use of study emollient

At the baseline visits participants were provided with an emollient information sheet (see Report Supplementary Material 11), which included information on how emollients should be applied. Participants were directed to use their study emollient at least twice per day, and during the interviews they compared this usage with that prior to participation.

Many participants reported that this was similar to their usual routine, but there were also some who acknowledged that having to fill in their study surveys had acted as a reminder to help them to apply their emollient regularly, rather than using it for an exacerbation of eczema:

We probably haven’t been as diligent in terms of applying it on a daily basis as we have been with the study … because we said we would, we’ve probably only used the [pre-study emollient] when we’ve needed to, when he’s had really dry skin.

N04, Mod, ≥ 7 years, SG, week 4

Participating in the study had reoriented some parents’ approach to emollient use. Many reflected on their experiences of using different emollients to find one that suited them best:

I think the problem with chopping and changing all the time is that you don’t really know … you don’t give it a chance and that’s why I didn’t really want to just try the study moisturiser for a couple of days and go ‘it’s not working’.

B15, Sev, 0–7 years, SL, week 16

Some participants commented that more regular application may have contributed to an improvement in their child’s skin as before participation they could sometimes be ‘slack’ or ‘lazy’:

Probably, like religiously and I think because we’ve had the form to fill in and keep an eye on it, I think it’s definitely helped us manage it better … If you were to look at him now and compared to last summer … he was at school in long-sleeved tops and trousers because of his eczema and he didn’t want to have sun cream on, he’s actually in shorts and t-shirt this year and it’s actually under control.

S07, Mod, ≥ 7 years, SC, week 16

There were also participants who reported that participating in the study had improved their knowledge of eczema management and treatments by engaging with the study information about emollients:

I’ve only ever had that from the BEE study actually, advice about baths and things like that, about how to put on [emollients]. That’s the first time I’d ever heard of that.

S07, Mod, ≥ 7 years, SC, week 16

However, there were also cases in which participants deviated from best practice and/or study guidance. Few participants adhered to instructions for using a spoon to avoid contamination from the hands when applying their study ointment (see Emollient containers).

Using the study emollient for longer – ‘persevering’

Committing to using the study emollient for 16 weeks led some to see benefits in using regularly over a longer period. In the quote below a parent describes the emollient’s effectiveness in equivocal terms but recognises a benefit in sticking with an emollient:

I don’t think it’s improved things massively but I am going to continue using it because I think, like I was saying before about not giving things a long enough go, I think especially with the winter coming up and chances are he’s probably going to get – it’s probably going to get worse for him a little bit, so we’ve ordered more of it … on prescription to continue using it. I mean apart from the fact that it’s thin and it’s more difficult for him to apply, it’s certainly not a bad [emollient].

S10, Sev, 0–7 years, SL, week 16

This parent, like many, had not found the study lotion to be a significant improvement but it had made enough of an impact to outweigh the negative aspects associated with acceptability and this was a pattern found across the study emollients.

Commitment to the study encouraged participants to use their emollient for a longer period and ‘persevering’ with the same emollient, despite experiencing periods of worsening eczema during the 16 weeks. Overall, participants could see a positive impact:

I think if I wasn’t on the study when it was really bad I would have gone back to the doctor and possibly seen if we could try something else so yes, so I don’t know whether the persevering and having it a set length of time has done us good … so this is the longest we’ve ever tried something … maybe just to persevere [laughs] and to give it like – giving something 6 weeks isn’t long enough, maybe 16 weeks isn’t long enough, maybe like 6 months, a year, you know, even though you’ve got rough times through that.

B12, Mod, 0–7 years, SC, week 16

I think because it was so bad I don’t think we ever gave it long enough in terms of, you know, we’d get this [emollient] and then we’d go back to the doctors and they’d say, ‘oh it’s not really working so change it,’ and I think in hindsight … and also based on this experiment that he’s done with the cream [generic] and we’ve obviously had it for, what are we … so we would never have given [an emollient] this long to work.

S10, Sev, 0–7 years, SL, week 16

Persisting with an emollient was novel for participants who had previously tended to swap frequently. Participation in the trial had contributed to a change whereby they recognised it was important to continue with their study emollient for longer, and sticking with an emollient beyond the main study period of 16 weeks gave an opportunity to assess its performance throughout the year:

I think it would be really interesting to see if we give it more time whether it would improve things for him even just a little bit, if his legs weren’t dry all the time or they weren’t … so they feel rough quite a lot as well. If it stopped maybe just a little bit of that I think it’s worth a go.

S10, Sev, 0–7 years, SL, week 16

Pragmatic randomised trial

In a primary care trial comparing four types of emollient for the treatment of eczema in children aged 6 months to 12 years, we found no evidence of a difference in eczema symptoms between lotions, creams, gels or ointments over the primary outcome period of 16 weeks. In our prespecified subgroup analyses, effectiveness did not vary by parent expectation, participant eczema severity, age or whether or not they met the UK diagnostic criteria for atopic dermatitis. Similarly, our secondary outcomes and per-protocol and sensitivity analyses showed no evidence of a difference between treatments. Over one-third of participants reported at least one adverse reaction, mainly application site reactions, with a similar number of problems across all emollient types. Reported daily use of allocated emollient and topical corticosteroids was similar for all four groups. At 16 weeks, overall satisfaction and intention to continue using the study emollient was highest for lotions and gels, whereas opinion on cream and ointments was more divided.

Nested qualitative study

The nested qualitative study was highly informative in understanding the findings of the main trial. No clear ‘winners’ or ‘losers’ were identified, yet problems were reported with all the different types of emollient.

We explored how families got on with emollients mainly through the lenses of perceived effectiveness and acceptability. Some interviewees positioned these aspects in opposition to each other, with an emollient judged to be useful but difficult to use or apply, whereas others deemed their emollient to be both effective and acceptable. However, for most participants, opinion was somewhere in between – for example, the study emollient might be as effective as their previous emollient but not stain clothing, making it more acceptable.

In terms of effectiveness, a number of participants using creams and ointments felt that their child’s eczema improved whereas those using lotions and gels tended to report their emollients as only controlling or maintaining their child’s eczema. Peculiarly to gel, a number of users found it useful only when using it on certain areas of the body or alongside other emollients (or sometimes topical corticosteroids).

Regarding acceptability, participants reported a wider range of views (positive and negative) on the acceptability of creams and ointments than for lotions and gels. A key consideration was the absorbency. Creams and ointments ‘sat’ on the skin, which for some was a positive, acting as a protective barrier, keeping out environmental factors that could irritate the skin, requiring fewer applications, hence reducing the time devoted to eczema management. However, others disliked the staining of clothes, and complained that the higher viscosity of these emollients made them more difficult to apply and uncomfortable for the child. In contrast, lotions and gels were described as lighter emollients that were easier to apply and ‘disappeared’ into the skin. Some parents thought that this indicated moisturising properties, whereas others perceived this as not as giving sufficient protection and so applied them more frequently.

Participants also reported factors, notably climate, which affected severity of their child’s eczema and the perceived effectiveness and acceptability of an emollient. The ways in which ointments were perceived and used were linked to the weather, with complaints of them being difficult to apply in colder weather and concerns about skin safety in hot sunny conditions.

As a consequence of taking part, some parents reported improved knowledge about emollients and how to use them, which led to their more regular use. This change was partly attributed to the ‘emollient information sheet’, given to all participants at baseline and partly from trial-related activities. Parents variously said they were more likely to ‘persist’ with a study emollient, more than they might normally, because they were taking part in a study, and study questionnaires served both as a reminder to use the study emollient and also as a self-monitoring tool, enabling parents to see the benefit of regular emollient use.

Pragmatic randomised trial

To the best of our knowledge, this is the first head-to-head randomised trial of the four main types of emollients as a leave-on treatment for childhood eczema.

Nested qualitative study

To our knowledge, this is the first qualitative study nested in a trial comparing the four main types of emollients for treating childhood eczema, the analysis of which was completed before the quantitative data were available. The purpose of the nested qualitative study was not to make overall recommendations in terms of effectiveness or acceptability. Instead, it was to explore and better understand the quantitative findings, with which it is consistent.

Our sampling framework encompassed a wide range of participant characteristics across the four trial groups. We included those who continued to use and those who had stopped using their study emollient at different time points and questioned participants regarding their intentions moving forward after the primary outcome period. Nevertheless, we recognise that owing to the wide range of our sampling criteria, the numbers within each category were small.

We also endeavoured to capture the views of children during the paired interviews. The level of their contribution varied considerably, so we are unable to draw firm conclusions regarding differing views and opinions with their parents and carers, but we have noted where their views on effectiveness and acceptability diverged. When considering our findings, it is important to recognise the potential impact of research participation on emollient use.

Data were collected by means of face-to-face and telephone interviews. We recognise that these modes of interviewing can shape the depth of data generated. 57 However, we did not find any substantive differences between the two modes of interviewing in terms of depth of data, although telephone interviews were more challenging when including children.

Parents and public involvement

A considerable strength of our research was our involvement of parent and public contributors. We listened to, and incorporated, many of their suggestions in the design of study recruitment materials and questionnaires, as well as the interpretation and development of key messages from the trial results. The commitment of our public co-applicant throughout the study was particularly valuable, providing continuity of memory across the life of the study.

Clinical guidelines and formularies

There are many guidelines and review articles about on how eczema should be managed. Although emollient use, including as maintenance therapy, is recommended by all guidelines, there is minimal consensus between them on which emollient(s) children with eczema should use, or the frequency or quantity that should be applied. 58 In 2016–7, Chan et al. 13 identified 102 unique emollient formularies across all 209 ICBs and seven local health boards (LHBs) in England and Wales. They generally supported the use of the four main types (recommended at least one cream, 99%; ointment, 98%; lotion, 85%; and gel, 85%) but poor agreement over which emollient should be used ‘first line’, with three of the ‘top five’ being ointments (White Soft/Liquid Paraffin 50/50, Emulsifying Ointment BP, Hydromol ointment), followed by Dermol 500 Lotion and Cetraben Cream. The justification for multiple different, conflicting, formularies is unclear and reflects the prior lack of good research in this area.

Randomised trials comparing emollients in eczema

The best synthesis of existing evidence to date is the 2017 Cochrane review by van Zureen et al. ,22 which included 77 trials, comprising 6603 participants, evaluating the effectiveness of emollients. The majority of these trials (70/77) were classified as ‘unclear’ or at ‘high’ risk of bias. The authors were unable to conclude whether some of the emollients or their ingredients were better than others, as most head-to-head comparisons had been evaluated in single studies, which generally had small sample sizes. It did, however, support the value of emollient compared with no emollient or placebo, with statistically (but not clinically) significant difference in reduction of disease severity, something that our study does not add to, because all participants received an active treatment.

We are not aware of any on-going trials like the BEE study, comparing commonly used emollients in a head-to-head fashion in children recruited from a primary-care setting. Of trials included in the Cochrane review, the study of 80 children aged 1–12 years with mild to moderate eczema by Hlela et al. 59 is most similar. They found no differences between emulsifying ointment, cetomacrogol, white petroleum jelly and glycerine/petroleum, although the results are difficult to interpret because another trial (comparing two different emollients as soap substitutes) is reported in the same paper. Since the Cochrane review literature searches were done, the closest comparable published study is COMET,15 which was the feasibility forerunner to this main trial. Children who took part in it had eczema, were aged 1-month to < 5-years, and were randomised to Aveeno lotion, Diprobase cream, Doublebase gel or Hydromol ointment for 12 weeks. It was not powered to compare the effectiveness of the interventions, but eczema severity improved in all groups. 16

Djokic-Gallagher et al. 54 compared Doublebase Dayleve gel (similar to Doublebase gel, with the addition of povidone, a film forming agent) and Diprobase cream in double-blind, concurrent bi-lateral comparison in 34 women with eczema. They compared their effectiveness using corneometry, a non-invasive method for accurate determination of skin hydration by measuring changes in electrical capacitance of the stratum corneum, and physical acceptability by completion of an unvalidated patient questionnaire (likeability, willingness to use again and preference). Hydration was greater with the gel than the cream, the clinical significance of which is unknown, and participants appeared to prefer the gel. Tiplica et al. 60,61 randomised 335 children aged 2–6 years with mild to moderate atopic dermatitis to Dexeryl^® [(V0034CR) Pierre Fabre Dermatology, Castres, France] cream, Atopiclair^® (Alliance Pharmaceuticals Limited, Chippenham, UK) cream or no emollient. Children in emollient groups had fewer flares, with Dexeryl cream appearing to perform better than Atopiclair cream.

Other studies of emollient awareness and satisfaction

Parents’ awareness of the different types of emollients was highest for creams and lowest for gels. We did not ask specifically about their understanding of the purpose of emollients or how they worked, but poor knowledge has been associated with low or no treatment use. 62 In a survey63 of parents of 350 children with mostly mild–moderate eczema attending a hospital dermatology clinic in China (2014–15), only 50.0% knew that moisturisers can restore the skin barrier and 15.4% knew that emollients can improve eczema. This was evident in interviews in which participants described how the study information improved their knowledge about emollient use.

An alternative approach to trials, to compare emollients, is the use of self-complete ‘satisfaction’ questionnaires. Unfortunately, many studies adopting this approach have been small53,54 and until recently all have been in adults and used unvalidated tools. 53,54,64–66 Rowley et al. 67 have recently validated a seven-item ‘Emollient Evaluation Questionnaire’, which was completed by participants in the COMET study. 15 Analysing data from parents of 152 children, they reported highest satisfaction with Aveeno lotion and lowest for Hydromol^® ointment, yet the number of participants intending to continue using their emollient was highest in the Hydromol ointment group, which suggests a trade-off between effectiveness and acceptability.

Unwanted and adverse effects

There is limited research into how common emollients are associated with adverse events and what implications this may have for adherence and use of other therapy. In an online survey of 210 patients and carers in 2016, Oakley and Lawson68 reported that 68% had experienced ‘unwanted effects’, of which 71% said that as a consequence they stopped using the leave-on emollient. The most common problem was stinging, followed by greasiness that interfered with everyday life.

‘Unwanted’ or ‘adverse effects’ affect treatment use, and studies over several years have reported low adherence to topical dermatological treatments in general and specifically in eczema. 62 In an online survey of 86 parents of children with mostly moderate or severe eczema, just over half reported taking or applying eczema medications as directed. 69 The top reasons for low adherence included worry about side effects, symptom resolution and perception that the medication was not working: 71% (90/126) of respondents stopped a leave-on emollient because of unwanted effects.

Although over half of the studies (41/77) in the 2017 Cochrane review20 included some data on adverse reactions, detail to enable comparison of the frequency and nature of adverse reactions across different emollients is absent. Bhanot et al. 70 conducted a ‘restricted’ review of published data on adverse events associated with emollient use in eczema. Interpretation of the results and comparison of the emollients across the 24 papers reporting on 29 different emollients was difficult owing to poor reporting and missing data. However, 2–59% of participants experienced treatment-related adverse events, most of which were skin related and mild. The data presented in our study, both in terms of overall frequency and by emollient type, are therefore unique.

Frequency of application and quantity of emollients and topical corticosteroids

Parent-reported use of emollients can be considered as both a mediator (‘Did participants use their allocated emollient?’) and an outcome (‘Is there a difference in the frequency of application of different types of emollients for the same effect?’). Although the findings from our study suggested that lotions and creams require more frequent application for the same effect than gels or ointments, the differences were not statistically significant.

Published data to support the amount of emollient needed to maintain disease control, and how this varies by age, severity and type of emollient, are scarce. Hon et al. ,71 in a study of 67 children with eczema using Cetaphil Moisturizing Cream (supplied by Galderma) correlated the quantity of emollient used with different measures of eczema severity (Nottingham Eczema Severity Scale, SCORing Atopic Dermatitis, Children’s Dermatology Life Quality Index), trans-epidermal water loss and mid-forearm skin hydration. They proposed that the quantity of emollient needed can be based on body surface area (between 73 and 136 g/m²/week) rather than disease severity. ‘Real world’ emollient use in moderate to severe disease is reported to be much lower (median of 17.5 g/day, 25th and 75th percentiles 12 g and 30 g/day, respectively, for children). 72 Choi et al. ’s72 interpretation of this finding is that many patients use emollients reactively rather than proactively. Our findings point to less than daily use, even in the context of a clinical trial that requested participants to apply their emollients ‘twice daily and when required’.

The van Zureen et al. 20 Cochrane review also found low-quality evidence that emollients prolong the time to flare, decrease the number of flares and reduce the amounts of topical corticosteroid needed to control eczema. What causes and how best to define an eczema ‘flare’ are both uncertain63,73 but we did not find any difference in ‘well-controlled weeks’ between the four types of emollient. Although we found no difference in the reported number of days of topical corticosteroid use, there may have been differences in total number of applications or quantity used. Two previous trials, by Giordano-Labadie et al. 74 and Grimalt et al. ,75 reported that the use of emollients decreased the use for topical corticosteroids. This may be desirable in respect of simplifying treatment burden and minimising the risk of adverse effects, but should not distract from the appropriate use of topical corticosteroids to treat inflamed skin.

Qualitative research

There is an established body of qualitative research literature around childhood eczema,76 which looks at the lived experience of eczema for children and families,8,18,77 beliefs around eczema and its causes,78,79 and the management and treatment of eczema. 18,80 Our research cuts across all these but engages mostly with the latter aspect.

Although many qualitative studies among people with eczema have highlighted confusion and concern about different emollient types and some concern about different emollient constituents,76 only one paper has specifically focused on perceptions of emollients among patients or carers with eczema. Santer et al. 52 drew on data from two studies, totalling 54 interviews with carers of children aged ≤ 5-years with eczema. This research identified the same kind of trade-offs that we saw in our work, with participants weighing up effectiveness and acceptability when choosing an emollient. Our results resonate with and add to these findings because we were able to assess experiences across the main four types of emollients.

There has been limited qualitative work with children81–84 and adolescents/young adults85–87 with eczema before. In a recently published study,88 Teasdale et al. interviewed 14 children (predominantly girls, with mild to moderate eczema) aged from 6 to 12 years. 88 They found that applying ‘creams’ (also used generically to describe creams, lotions, gels or ointment type of emollients) was helpful in relieving eczema symptoms. As in our interviews, participants had mixed views about the texture, viscosity and odour of some topical treatments, yet reported using topical treatments even if they disliked its texture or odour, supporting any trade-off favouring effectiveness over acceptability. Participants also wanted treatments that soaked into the skin quickly without having to reapply emollients, for example during the school day.

Emollients for primary prevention of eczema

Finally, ours was a study of emollients for the treatment of eczema, but it should be noted that during its lifetime two large trials (BEEP89 and PreventADALL90) and a recently published meta-analysis91 have found no evidence that their prophylactic use prevents the onset of disease. These findings have been a surprise, because prior pilot studies suggested a relative risk reduction of 50%. Indeed, the Cochrane review91 by Kelleher et al. concluded that their use increases the risk of skin infections and may increase the risk of food allergy. One explanation is that the application of emollients with unclean hands may introduce foods to the immune system via the defective skin barrier they were seeking to enhance. Questions have been raised previously about systemic absorption of some constituents of emollients via the same route,92 with concerns about food allergy induced by oatmeal-based emollients in particular. 93

Medicine management teams, policy-makers and guideline developers

In the UK, emollient formularies are determined by medicine management teams, which in turn are influenced by policy-makers and guideline developers, mainly NICE, the Scottish Intercollegiate Guideline Network (SIGN) and NHS in the devolved nations. They are the primary audience in terms of ensuring that new knowledge is incorporated into guidelines and formularies, to support effective prescribing by clinicians.

Policy-makers (national or local) are always mindful of costs. As all four emollient types were found equally effective but not equally acceptable to everyone, we recommend that they should be available to all children with eczema. However, there is a risk that ‘all equally effective’ may be taken out of context with respect to acceptability, and misused to restrict patient choice to the cheapest type(s) only.

Patients and families

Eczema is common in children yet their care varies across the UK, and emollient formularies in adjacent ICBs and LHBs can make contradictory recommendations. In the absence of consistent clinical advice, the emollient a family uses is often decided through a process of trial and error, which can be confusing and frustrating. Families who relocate may be forced to change from tried and trusted emollient(s). Effectiveness, acceptability (because of properties such as feel or absorbency) and use are interlinked, because a ‘less effective but more acceptable’ emollient used regularly is likely to be more beneficial than an infrequently used ‘more effective but less acceptable’ emollient.

Given that the results of the trial support all four types of emollient for use, the results of the trial are likely to be acceptable to families, as they support patient choice. However, if the results of the trial are used to suggest that all four emollient products are equally effective and there is no need to have all four available on the formularies, then this may cause difficulties for patients and families. It could result in distrust in the research as it is balanced against their own personal experience. Although this ‘end user’ group has little weight on policy decision-making, they may be able to influence policy through support groups.

Support groups (including voluntary sector and patient interest groups)

The main charities in the UK are the National Eczema Society, Eczema Outreach Support, Nottingham Eczema Support Group for Carers of Children with Eczema and the British Skin Foundation (London, UK). These groups consist of patients and health-care professionals, who are supported by experts in, or are specialists themselves, in eczema. Use of topic experts to inform these groups may be an effective strategy to ensure that knowledge is communicated accurately.

Skin specialists (dermatologists and dermatology specialist nurses)

Dermatologists, dermatology specialist nurses and GPs with extended roles in dermatology often influence local formularies that cover primary and secondary care. A change in their prescribing practice based on the study’s findings may be reflected in follow-on prescriptions in primary care. However, the findings may be less applicable to the children they see with more severe eczema and their decision-making may also be restricted by local formularies.

Specialists are likely to be supportive of new evidence, especially from a randomised controlled trial. Reaching this group can be done via scientific journals and special conferences. We discussed our findings with key opinion leaders, or those likely to be approached by the press, prior to publication to aim for consistent messaging. Careful explanation of the findings within the context of previous work can ensure that topic experts are available to support other audiences with accurate translation of new knowledge.

Primary care clinicians (including general practitioners and community pharmacists)

Most children with eczema are diagnosed and managed by their GP surgery. Historically, undergraduate and postgraduate GP training in dermatology has been limited, with skin problems such as eczema overshadowed by other long-term conditions. Community pharmacists sell emollients directly or give advice to families, based on the prescription issued by the GP surgery.

Prescribing clinicians in primary care (GPs, nurse practitioners, physician associates, pharmacists) will have varying interest in the results. They are most likely to be influenced by changes in local prescribing guidelines, which should reflect the evidence such as that from our randomised trial. Some community pharmacists will have relationships with medicines management teams (made up of pharmacists) so may be able to influence decisions.

Manufacturers of emollients

There are many manufacturers of emollients, with the main suppliers in the UK being Alliance Pharmaceuticals Limited (Chippenham, UK), Aspire Pharma Limited (Petersfield, UK), Bayer UK Ltd (Reading, UK), Crawford Healthcare (Knutsford, UK), Dermal Laboratories Ltd (Hitchin, UK), Fontus Health Ltd (Walsall, UK), Galen Limited (Craigavon, UK), Mölnlycke Health Care Limited (Milton Keynes, UK), and Thornton & Ross (Huddersfield, UK). Many advertise directly to parents, with some placing emphasis in their promotional material on claims about clinical effectiveness whereas others focus on cost-effectiveness. All have vested financial interests in their product(s) being portrayed in the best light.

Academics

Academics have an important role in reviewing the basis of the recommendations we make, and in supporting their uptake. Careful explanation of the findings within the context of previous work can support the accurate translation of new knowledge to different audiences.

Breadth approaches (dissemination)

These approaches focus on disseminating the key knowledge findings with target audiences and other stakeholders. Breadth approaches were chosen to target each stakeholder group, but they assume a more linear approach to the way knowledge is shared. However, these approaches have the advantage of reaching a far larger number of individuals than the depth approaches.

Depth approaches (knowledge mobilisation)

The depth approaches were targeted for the key stakeholders identified as having high power in the stakeholder analysis, in terms of future impact on policy and practice. Previous research has illustrated the importance of personal relationships and face-to-face interactions to maximise knowledge sharing and mobilisation with clinicians and policy-makers. Although restrictions related to the COVID-19 pandemic have limited physical meetings, virtual ‘face-to-face’ discussions have been possible.

The depth approaches aimed to target individuals and key organisations to achieve maximum opportunities to share and co-produce knowledge with stakeholders. It was only possible to conduct these on a small scale and at a local level, but it was intended that this learning could then be used to inform the wider-reach breadth approaches at a later stage.

Two theories of knowledge mobilisation (communities of practice and socialisation121 and externalisation, combination and internalisation theory122) were drawn on to inform the approach and strategy for knowledge mobilisation during the study. Both theories explain how knowledge is shared through everyday interaction and highlight where opportunities to introduce new knowledge may lie. By adopting the knowledge mobilisation approaches likely to be most effective for the key stakeholder groups identified (policy-makers and clinicians), the impact of the study findings on practice should be maximised, as well as the opportunities for practice to influence the potential study outputs.