Adding emollient bath additives to standard eczema management for children with eczema: the BATHE RCT

Article history

The research reported in this issue of the journal was funded by the HTA programme as project number 11/153/01. The contractual start date was in May 2014. The draft report began editorial review in November 2017 and was accepted for publication in April 2018. The authors have been wholly responsible for all data collection, analysis and interpretation, and for writing up their work. The HTA editors and publisher have tried to ensure the accuracy of the authors’ report and would like to thank the reviewers for their constructive comments on the draft document. However, they do not accept liability for damages or losses arising from material published in this report.

Declared competing interests of authors

Matthew J Ridd is a member of the National Institute for Health Research (NIHR) Health Technology Assessment (HTA) General Board. Nick A Francis was a member of the HTA Antimicrobial Resistance Themed Call Board during the study. Amanda Roberts is a member of the HTA General Board. As patient and public involvement co-applicant, Amanda Roberts also received an attendance allowance and travel expenses from the University of Southampton in order to attend meetings. Emma Thomas-Jones reports grants from the NIHR HTA programme during the conduct of the study. Hywel C Williams is Director of the NIHR HTA programme. Paul Little is a member of the NIHR Journals Library Board.

Copyright statement

© Queen’s Printer and Controller of HMSO 2018. This work was produced by Santer et al. under the terms of a commissioning contract issued by the Secretary of State for Health and Social Care. This issue may be freely reproduced for the purposes of private research and study and extracts (or indeed, the full report) may be included in professional journals provided that suitable acknowledgement is made and the reproduction is not associated with any form of advertising. Applications for commercial reproduction should be addressed to: NIHR Journals Library, National Institute for Health Research, Evaluation, Trials and Studies Coordinating Centre, Alpha House, University of Southampton Science Park, Southampton SO16 7NS, UK.

2018 Queen’s Printer and Controller of HMSO

Scientific background

Eczema is very common, affecting > 20% of children at some point during their first 5 years of life. 1 Eczema has a significant impact on quality of life (QoL)2 and can cause distress to affected children and their families because of sleep disturbance, itching and scratching. 3 The term ‘atopic eczema’ (synonymous with ‘atopic dermatitis’) is widely used to denote a clinical phenotype, rather than those who are truly atopic as defined by the presence of immunoglobulin E-specific antibodies to common environmental allergens. In this study, we use the term ‘eczema’ throughout to refer to the ‘atopic eczema’ clinical phenotype, in accordance with the recommended nomenclature of the World Allergy Organization. 4,5

Skin complaints are the second most common reason for general practitioner (GP) consultations in children aged < 5 years. 6 Health and societal costs of eczema care are difficult to estimate as they vary widely by population under study, but eczema is thought to cause a similar economic burden to that for asthma. 7,8

Emollients for the treatment of childhood eczema

Guidelines suggest that emollients form the mainstay of treatment for eczema and should be used regularly by all patients alongside other treatments, such as topical corticosteroids (TCSs), when necessary to treat flare-ups. 9 Emollients are thought to act by providing a protective layer over the skin, decreasing moisture loss and occluding against irritants.

There are three methods of application of emollients: (1) leave-on (directly applied) emollients, where emollients are applied to the skin and left to soak in, (2) soap substitutes, where emollients are used instead of soap or other washing products, and (3) bath emollients (or bath additives), which are oil and/or emulsifiers designed to disperse in the bath. All three approaches are often used together (referred to as ‘complete emollient therapy’). 5,9 In this report, the term ‘bath additives’ rather than ‘bath emollients’ is used to emphasise the differences between the three methods of application (i.e. these three methods differ in their proposed actions and evidence relating to their effectiveness should also be considered separately).

Although there is widespread clinical consensus on the need for leave-on emollients and soap substitutes, there is less agreement regarding the additional benefits of emollient bath additives. 10–13

Bath additives for the treatment of childhood eczema

A previous systematic review has revealed no convincing evidence for the use of emollient bath additives in the treatment of eczema;10,11 available data consist of case series and very small trials. One study14 in which parents were asked to soak one of their child’s arms for 15 minutes a day for 2 weeks in a basin containing water with bath additive found that clinical assessment by blinded observer was worse for the arm that was soaked daily than for the unsoaked arm in eight of the nine children, suggesting the possibility that bath additives may be harmful. No trials of emollient bath additives have been published since 200715 and trial registries reveal no ongoing studies.

In addition to concerns about cost-effectiveness, potential harms from using bath additives include skin irritation and greasier bath surfaces, which can increase the risk of slips and accidents13 (listed in the Summary of Product Characteristics of leading brands). There is also a concern that people who use bath additives in place of leave-on emollients are receiving substandard emollient therapy. 12

The effectiveness of adding bleach bath additives has also not been demonstrated in a small randomised trial, although 5 out of 18 participants in the bleach bath group experienced mild burning/stinging or dry skin. 16 Two small randomised studies17,18 compared ‘bath emollient’ with ‘bath emollient plus antiseptic’ on a range of outcomes, but there were no significant differences between groups, including colony counts of Staphylococcus aureus. 19 For this reason, we chose to exclude bath additives that incorporate an antiseptic, because of the absence of benefit and the possible increased risk of skin irritation. 5,20

In 2007, the Drug and Therapeutics Bulletin noted that the NHS spends > £16M per year on bath additives (at an average cost of £6.29 per item), representing 38% of the total cost of treatments prescribed for preschool children with eczema and matching the proportion spent on emollients for application directly to the skin. 9

Despite the absence of evidence and the possibility of potential harms, bath additives are widely prescribed at a cost of > £23M per year to the NHS in England. 21 Currently, prescribing advice varies widely. An analysis of 216 formularies from Clinical Commissioning Groups in England and Local Health Boards in Wales22 showed that 68% recommended the use of bath additives, 15% allowed their use but did not encourage it, 13% did not mention bath additives and 5% did not recommend their use.

Pragmatic trial design

Pragmatic clinical trials aim to test the effectiveness of an intervention in a real-life setting in order to recruit a study population that is as similar as possible to the population on which the intervention is meant to be used. Whereas an explanatory clinical trial aims to answer the question ‘Can this intervention work under ideal conditions?’, a pragmatic approach seeks to answer the question ‘Does this intervention work under usual conditions?’. 23,24 Features of pragmatic trials include the use of clinically important outcomes and common participant-reported outcomes, long-term follow-up and encouragement of participants to adhere to the intervention only to the extent that would be anticipated in usual care.

Although relatively few pragmatic trials have been carried out in dermatology,25 we felt that a definitive pragmatic clinical trial, including outcomes of relevance to participants and including long-term follow-up, was the most appropriate design to address the question of the effectiveness of bath additives in addition to standard eczema care in everyday care. 5

Blinding

We chose an ‘open-label’ design as it would not be possible to create a convincing placebo for bath additives, which make the bath feel ‘greasy’, and many families of children with eczema will already have experience of using them. We wished to design a trial with a clinical outcome relevant to participants (as below). Ideally, we would also have included an objective assessment of eczema severity carried out by a blinded assessor. However, this would have increased burden for participants as additional face-to-face assessments would have been required, particularly as the relapsing and remitting nature of eczema means that follow-up assessment at a single time point is problematic. As our primary outcome was participant reported and, therefore, unblinded, incurring substantial additional costs for an objective secondary outcome did not seem warranted.

Participant-reported outcome measure

We wished to design a trial with a clinical outcome relevant to participants. In eczema, the appearance of the skin does not always closely reflect symptoms causing a major impact on the child and family, such as sleep disturbance and itch. 26 It was therefore particularly important to design a trial with a validated participant-reported primary outcome. 5

We chose the Patient Oriented Eczema Measure (POEM)27,28 as our participant-reported primary outcome measure. POEM comprises seven questions about eczema symptoms over the previous week that are summed to give a score from 0 (no eczema) to 28 (worst possible eczema). POEM is a patient-reported outcome that can be used by proxy (carer report), demonstrates good validity, repeatability and responsiveness to change29 and is recommended by the National Institute for Health and Care Excellence (NICE)9 and the international Harmonising Outcome Measures for Eczema (HOME) initiative. 30,31

Capturing meaningful outcomes for people with eczema is complicated by the relapsing and remitting nature of the condition. Therefore, gathering information regularly over time is essential to understand disease burden32 and to assess the impact of interventions. We therefore chose to collect POEM scores on a weekly basis from participants, as has been used successfully in other eczema trials. 33 An article reporting on the acceptability and practicality of weekly POEM completion is in preparation. 34

Development of research priority

This trial was funded by the National Institute for Health Research (NIHR) Health Technology Assessment (HTA) programme following a commissioned call advertised in February 2012. The research topic was suggested via the NIHR HTA website topic suggestion form and was approved in December 2011.

The James Lind Alliance Priority Setting Partnership (PSP) for Eczema published its top 10 priority topics in 2012. 35 Even though this call was not advertised directly as a result of the PSP, it does address some of the issues that patients, carers and clinicians highlighted, including priorities around bathing/washing and also around the best ways to use emollients.

Objectives

The objectives were to determine the clinical effectiveness and cost-effectiveness of adding bath emollient to the standard management of eczema in children, which includes regular application of leave-on emollients with use of TCSs as required.

The trial was registered before commencing on 13 December 2013 and the protocol was published online in August 2015. 5

Trial design

The Bath Additives for the Treatment of Eczema in cHildren (BATHE) trial was a pragmatic, randomised, open-label, multicentre, superiority trial with two parallel groups and a primary outcome of long-term control as measured by POEM weekly scores over 16 weeks. Children were randomised in a 1 : 1 ratio to either standard eczema care plus bath additives or standard eczema care only for 12 months (Figure 1).

FIGURE 1.

Participant flow through study. a, At least 4 weeks allowed for clinical letters to be received at surgery and scanned into patient record. CHU-9D, Child Health Utility-9 Dimensions; CRF, case report form; CSRI, client service receipt inventory; DFIQ, Dermatitis Family Impact Questionnaire; NESS, Nottingham Eczema Severity Score; UKDC, UK Diagnostic Criteria for Atopic Dermatitis.

Changes to trial protocol

As recruitment was ahead of target, and following discussion with the Trial Steering Committee, in October 2015 the decision was taken to increase the target sample size from 423 to 491 participants, to allow an analysis by treatment adherence in addition to an intention-to-treat (ITT) analysis. Assuming that 80% of participants were strictly adherent to their allocated treatment, recruitment of an additional 68 participants would be required to retain 90% power for a per-protocol analysis.

In addition, although not requiring an amendment to the trial protocol, concerns were raised that the participant information sheet was overly formal and text heavy and that this might be contributing to a lower than expected response rate. A colourful summary information leaflet was therefore designed, to be added to the patient invitation pack (see Appendix 1). The additional leaflet was included in mailshots from 12 June 2015 (63% of the invitations), but there is no evidence that its inclusion affected either the number of responses or the proportion of positive replies.

Participants

Interventions

Data collection

Parents/carers were able to choose to complete the trial questionnaires either online or by post following discussion at the recruitment appointment, although online questionnaire completion was strongly encouraged. Online questionnaires became available on the seventh day following recruitment and every 7 days thereafter. Notifications were automatically sent by e-mail when the questionnaire went ‘live’ and reminders were sent after 48 hours if it had not been completed. Participants could also opt to receive the notifications and reminders by automated text. Both e-mails and text messages contained a hyperlink to the online questionnaire and participants were required to log in using their e-mail address and password in order to enter data. Of the 9784 times that parents/carers logged in to the website, 5963 logins (61%) were from mobile devices. A further 2909 logins (30%) were from computers or laptops, whereas 912 logins (9%) were from tablets. 34

The 16-week questionnaire, and subsequent monthly questionnaires, remained available to complete online for 28 days. However, in order to encourage timely completion of the primary and secondary outcomes, a reminder letter and paper copy of the 16- and 52-week questionnaires were posted out if the online questionnaire had not been completed within 7 days. A total of 80 reminders were posted at week 16 and 107 reminders were posted at week 52.

Weekly paper questionnaires were printed in booklets of four. The first booklet, covering weeks 1–4, was marked with the day of the week on which the questionnaires should be completed and then handed to the parent at recruitment. Subsequent questionnaires were posted to participants shortly before they were due, together with a prepaid envelope. The data from the paper questionnaires were entered into a secure database and were merged with the data collected online prior to analysis.

Outcomes

Sample size

The sample size was calculated for repeated measures analysis of variance (ANOVA) in weekly POEM scores over the 16-week observation period. Using weekly data from a similar population to that in the Softened Water Eczema (SWET) trial,48 we aimed to detect a mean difference of 2.0 [standard deviation (SD) 7.0] between the intervention and control groups. The published minimal clinically important difference for POEM is 3.0,41,42 but we wished to detect a difference of 2.0 because of the expectation of low baseline POEM scores in a population recruited entirely through primary care, and because we wished to be able to detect this small difference as the intervention is relatively inexpensive and even small effect sizes may be cost-effective. An alpha of 0.05 and a power of 90% with a correlation between repeated measures of 0.70 gives a sample size of 338 participants. Allowing for 20% loss to follow-up gave a total sample size of 423 participants.

We aimed to report a per-protocol analysis in addition to a primary ITT analysis, and early data suggested that approximately 80% of participants in both groups were strictly adherent to treatment allocation. We were concerned that if only 80% of participants were adherent to treatment allocation then we would have usable data for only 270 participants. To get back up to 90% power for this group, we submitted an ethics amendment requesting recruitment of an additional 68 participants, giving a revised target recruitment of 491 participants.

Randomisation

Randomisation was performed using LifeGuide49 software hosted by the University of Southampton and validated by Southampton Clinical Trials Unit. At the time of trial setup, LifeGuide was unable to easily perform block randomisation and the additional programming time would have resulted in delays to the trial. Simple randomisation was therefore used, stratified by centre. Although this can result in imbalances, it was felt that with strata of > 100 participants each, the overall balance between groups would be preserved.

A backup randomisation system was established for occasions when internet access to LifeGuide was not available or when the parent had opted to complete the trial questionnaires on paper. A second set of random treatment allocations was programmed into a Microsoft Excel^® (Microsoft Corporation, Redmond, WA, USA) spreadsheet by the trial statistician. The spreadsheet similarly allocated treatments on a 1 : 1 basis, stratified by recruiting centre, and the treatment allocation was not revealed until the participant had been recruited. A total of 30 randomisations were conducted using this offline method.

Allocation concealment mechanism

Both the online and the offline randomisation procedures were conducted immediately at the end of the recruitment appointment, following completion of the baseline questionnaire. It was therefore impossible for the treatment allocation to be known prior to study entry. Once randomisation was complete, however, the participant, CSO/RN and the participant’s clinical team were all aware of which group the child had been allocated to.

Implementation

Once recruitment and randomisation was complete, the lead GP at the site was notified of the child’s enrolment into the study and their allocated treatment group was noted on a form completed by the recruiting CSO/RN. The form requested that repeat prescriptions be set up for children allocated to the bath additive group and it also recorded the parent/carer’s preferred bath additive, if any. Practices were recommended to add a Read Code to the child’s electronic record to indicate that they were enrolled in a clinical trial and/or to add an electronic alert or other notification to remind clinicians of their treatment allocation. This, however, was not enforced and the focus remained on ensuring that parents/carers were fully engaged with the aims and requirements of the study.

Blinding

Given the nature of bath emollient, which adds a greasy film to water, it was impossible to manufacture a convincing placebo treatment or to blind the participants to their treatment allocation. Participants were therefore fully aware of the treatment regime they were being asked to adhere to and report on. In order to support families throughout their 12-month participation, the trial team were also not blinded to the treatment allocated.

The trial statisticians carrying out the analysis were blind to allocation group and the statistical analyses were independently verified.

Statistical methods

The primary analysis for the total POEM score was performed using a mixed multilevel mixed model (MMLM) framework with observations over time from weeks 1–16 (level 1) nested within participants (level 2) nested within centres (level 3). Unadjusted results are reported, as well as results adjusting for baseline POEM, recruitment region as covariates and any significant confounders. Confounders were defined as variables associated with both the exposure and the outcome that significantly contribute to the multivariable model (defined as a p-value of < 0.05 or by modifying the effect estimate by > 10%). The following variables were identified a priori as possible confounders: child age, child gender, carer age, carer gender, carer education, prior belief in bath emollients, type of emollient used, other medication used and other items used when washing (e.g. soap/shampoo/soap substitute).

The model used all of the observed data and made the assumption that missing POEM scores are missing at random. The model included a random effect for centre (random intercept) and patient (random intercept and slope on time) to allow for between-patient and between-centre differences at baseline and between-patient differences in the rate of change over time (if a treatment/time interaction was significant), and fixed effects for baseline covariates. An unstructured covariance matrix was used.

The assumptions of the normality of the residuals from the fixed part of the model and the normality of the random effects at the cluster level were checked.

For the analysis of secondary outcomes, repeated measures analysis in line with that used for the primary outcome was used for the monthly POEM up to 1 year. 5 For other secondary outcomes, linear regression was used for continuous outcomes if the assumptions were met; otherwise, non-parametric analyses were used. Logistic regression was used for dichotomous outcomes and a suitable count model, as determined by goodness of fit measures, was used for count data. All analyses controlled for stratification variables and potential confounders. Preplanned sensitivity analysis and exploratory subgroup analyses were carried out as set out in the statistical analysis plan.

To test the sensitivity of the results to missing data, a chained equations multiple imputation model was used to impute the missing values. This analysis used 100 imputations and included all the outcomes and covariates included in the adjusted analysis of the primary outcome.

For all models, participants were analysed in the group to which they were randomised, regardless of their adherence to that allocation (ITT analysis). The only exception to this was the per-protocol analysis, for which participants were analysed on the basis of their reported use of bath additives. The reported use of bath additives was collected at 16 weeks and 52 weeks in both groups using the categorical response options ‘every time’, ‘more than half the time’, ‘less than half the time’ and ‘never’. This allows two possible definitions for adherence to the intervention.

Participant recruitment

Invitation packs were sent to the parents/carers of 12,504 children and 1451 responses were received. A total of 1343 responses were returned by post; 832 respondents (62%) were willing to be contacted and 431 (32%) went on to participate in the trial. A total of 108 responses were received electronically; 88 respondents (81%) were willing to take part and 52 (48%) went on to be recruited. A total of 35 (2.4%) of the responses received were returned from opportunistic invitations; 34 respondents were willing to take part and 19 (54%) went on to participate in the trial.

A total of 237 parents/carers declined participation and returned a blank screening form. A further 104 answered ‘No’ to the first question of the UKDC (‘In the last year, has your child had an itchy skin condition?’) and did not fill in any other information about their child’s condition. A total of 188 parents/carers indicated that they were unwilling or unable to take part but did complete the screening questionnaire. Table 1 summarises the information received about these children’s eczema, in comparison with children who went on to take part.

There were 920 replies expressing a willingness to be contacted and these replies also included a completed screening questionnaire. Of these, 229 did not meet the clinical criteria required to enter the trial. Overall, 662 children met the clinical eligibility criteria and were approached regarding participation, of whom 483 entered the trial and 179 were not recruited for the reasons shown in Figure 2.

Figure 2 shows the flow of participants through the screening and recruitment process.

FIGURE 2.

Recruitment flow chart. a, Unwilling to be randomised/change current treatment, n = 58; eczema no longer a problem, n = 31; child mostly showers, n = 16; child refused, n = 2; other/not specified reason, n = 29. b, Did not meet UKDC, n = 65; NESS score of < 6, n = 76; and did not meet either criterion, n = 88.

Figure 3 shows the number of participants who withdrew or who were lost to follow-up during the study period. Four participants formally withdrew during the 16-week primary outcome period, three from the Cardiff centre and one from the Bristol centre; all withdrawals were from the bath additive group. Permission was obtained from three parents to use the data already collected, and one requested that all data collected be removed. A total of 42 participants were lost to follow-up by the time of the 16-week primary outcome (22 in the bath additive group and 20 in the no bath additive group) and a further four 16-week questionnaires were not received from participants who remained enrolled.

FIGURE 3.

The BATHE participant flow chart. a, Reported eczema had not been a problem for some time, although screening form clearly indicated eligibility. b, One parent wished to bathe two children together but felt that bath additive was worsening the symptoms of the child not enrolled in the study; one was finding improvement with new creams and, despite reassurance, did not wish this to affect the study results; one child was withdrawn by their GP after experiencing a rash. CRF, case report form.

A total of 429 (89.6%) participants opted to complete their weekly questionnaire online. Of the remaining 51 participants, 27 (5.6%) requested the paper option at recruitment and 24 (5.0%) were switched from online to paper format after discussion with the study team. Reasons for this were primarily related to technical issues, as some parents/carers became discouraged after having problems logging in to the online database. Login problems were occasionally compounded by a failure to understand the automated nature of the system and the inability to complete POEMs retrospectively. This was particularly problematic for a small number of participants who experienced delays in obtaining their prescribed bath emollients and, therefore, did not ‘start’ participation in week 1. Although attempts were made to telephone all families during the first week, as a matter of courtesy and in order to address any problems, some parents could not be reached and it is not possible to determine how many participants were lost to follow-up because of technical or logistical problems.

The majority of trial participants needed no prompting to complete the trial and only 25% of the trial participants were contacted by telephone or e-mail during the 16-week primary outcome period because of failure to complete the questionnaires.

Recruitment dates

Recruitment took place from November 2014 to May 2016. The original target of 423 participants was reached in March 2016 and permission was obtained to continue recruiting participants up to an increased target of 491 participants as discussed in Chapter 2, Changes to trial protocol. Recruitment was stopped at the end of May 2016 as planned, with a total of 483 participants enrolled (Figure 4). The 52-week follow-up questionnaires and notes reviews of the last recruited participants were completed in June 2017.

FIGURE 4.

Recruitment by centre.

Baseline data

Table 2 shows that, although there were more participants allocated to the bath additive group than the no bath additive group (see Chapter 5, Discussion), the key characteristics were well balanced at baseline.

Numbers analysed

The questionnaire response rate was high, with 76.7% of participants completing questionnaires for > 80% of the time points for the primary outcome (12 out of 16 weekly questionnaires to 16 weeks). There were no marked differences in completeness of the data by randomisation group (see Table 3 and Figure 5).

FIGURE 5.

Questionnaire completion during the 16-week primary outcome period.

Outcomes

Primary outcomes

All 461 participants who had completed at least one POEM following baseline were included in this analysis. The results in Table 4 indicate no statistically significant difference in weekly POEM scores between the two groups over the 16-week period. Of the prespecified potential confounders (child age, child gender, ethnic group, carer age, carer gender, carer education, prior belief in bath emollients, type of emollient used, other medication used, other items used when washing, such as soap/shampoo/soap substitute), only ethnic group, steroid use and soap substitute use were statistically significant and, therefore, retained in the models. After controlling for baseline severity and these confounders, and allowing for the clustering of patients within centres and responses within patients over time, the POEM score in the no bath additive group was 0.41 points higher than in the bath additive group (95% CI –0.27 to 1.10). Unadjusted POEM scores showed < 1 point difference between groups. These differences are not considered to be clinically meaningful, given the published minimal clinically important difference for POEM of 3.0. 41,42

TABLE 4 - Weekly POEM scores during the 16-week primary outcome period

Time period	Treatment group, mean (SD)
	Bath additive	No bath additive
Baseline	9.5 (5.7)	10.1 (5.8)
Week 1	8.3 (5.6)	9.1 (5.9)
Week 2	7.8 (5.5)	8.2 (5.9)
Week 3	7.4 (5.3)	8.5 (5.9)
Week 4	7.6 (6.0)	8.6 (6.1)
Week 5	7.6 (5.9)	8.3 (6.0)
Week 6	7.8 (6.3)	8.4 (5.7)
Week 7	7.5 (6.1)	8.8 (6.1)
Week 8	7.3 (6.2)	8.2 (5.8)
Week 9	7.2 (5.9)	8.1 (5.8)
Week 10	7.3 (5.8)	8.5 (5.8)
Week 11	7.6 (6.1)	8.2 (6.0)
Week 12	7.7 (6.2)	7.9 (5.9)
Week 13	7.1 (5.9)	8.3 (6.2)
Week 14	7.2 (6.3)	7.9 (6.3)
Week 15	7.0 (6.3)	8.4 (6.5)
Week 16	7.1 (6.1)	8.2 (6.3)

There was no statistically significant interaction between treatment and time (interaction term 0.04, 95% CI –0.02 to 0.11; p = 0.204).

Patient Oriented Eczema Measure weekly for 16 weeks

There is some fluctuation in between-group differences in POEM score over the 16-week period, with some results statistically significant and others not. However, all the between-group differences are < 2 points, suggesting that even those that are statistically significant are not likely to be clinically meaningful as they are below the POEM minimal clinically important difference of 3.0 (Tables 4 and 5 and Figure 6). Moreover, it is likely that some significant results may be found because of multiple testing (type I error).

TABLE 5 - Patient Oriented Eczema Measure scores during the 16-week primary outcome period

Adjusted for baseline severity, clustering of participants within centre, ethnic group, steroid use and soap substitute use.

POEM scores	Treatment group, mean (SD)		Univariate difference in mean POEM (95% CI)	Adjusted difference in mean POEMa (95% CI)
	Bath additive	No bath additive
Primary outcome: 16-week repeated measures	–	–	–	–
Over the 16-week primary outcome period (repeated measures)	7.5 (6.0)	8.4 (6.0)	0.90 (–0.03 to 1.83)	0.41 (–0.27 to 1.10)

FIGURE 6.

Mean POEM scores over the 16-week primary outcome period, by group.

Imputed analysis

As a sensitivity analysis, primary outcome was examined using imputed values for all missing weekly data using an individual chained equations multiple imputation model. Table 6 shows that this produced an adjusted difference in mean POEM score between groups of 0.43 (95% CI –0.26 to 1.12). This is a very similar result to our primary analysis.

TABLE 6 - Patient Oriented Eczema Measure scores during the 16-week primary outcome period: results based on 100 imputations

Adjusted for baseline severity, clustering of participants within centre, ethnic group, steroid use and soap substitute use.

Adherence to allocated treatment	Treatment group, mean (SD)		Univariate difference in mean POEM (95% CI)	Adjusted difference in mean POEMa (95% CI)
	Bath additive	No bath additive
Primary outcome: 16 week repeated measures	–	–	–	–
Over the 16 week primary outcome period (repeated measures)	7.4 (6.0)	8.5 (6.0)	0.96 (0.05 to 1.87)	0.43 (–0.26 to 1.12)

Adherence to allocated treatment (per-protocol analysis)

Parent/carer report of adherence to treatment allocation group at 16 weeks recorded 92.7% of participants in the bath additive group using bath additive ‘every time’ (73.8%) or ‘more than half the time’ (18.9%). Similarly, 92.1% of those in the no bath additive group said that they used bath additives either ‘never’ (87.4%) or ‘less than half the time’ (4.7%) (Table 7).

TABLE 7 - Adherence to allocated treatment during the 16-week primary outcome period

Adherence to allocated treatment	Treatment group, n (%)
	Bath additive (N = 233)	No bath additive (N = 191)
Use of bath additives
Every time	172 (73.8)	14 (7.3)
More than half the time	44 (18.9)	1 (0.5)
Less than half the time	15 (6.4)	9 (4.7)
Never	2 (0.9)	167 (87.4)
Number of baths per week	(N = 221)	(N = 176)
1–2	70 (31.7)	54 (30.7)
3–4	74 (33.5)	56 (31.8)
5–6	45 (20.4)	39 (22.2)
≥ 7, n (%)	32 (14.5)	27 (15.3)

We considered two possible definitions of adherence and on neither basis was there a statistically significant difference between the groups (Table 8).

TABLE 8 - Patient Oriented Eczema Measure scores during the 16-week primary outcome period by adherence

Adherence to allocated treatment	Difference in mean POEM (95% CI)
	Univariate	Adjusted
16-week repeated measures model
‘More than half’ or ‘every time’ compared with ‘less than half the time’ or ‘never’	0.35 (–0.58 to 1.28)	0.32 (–0.37 to 1.02)
‘Every time’ compared with ‘never’	0.23 (–0.79 to 1.24)	0.38 (–0.39 to 1.15)

Participants were also asked about adherence at 52 weeks. Table 9 shows that adherence remained high.

TABLE 9 - Adherence to allocated treatment during the 52-week secondary outcome period

Exacerbations	Treatment group, n (%)
	Bath additive (N = 203)	No bath additive (N = 176)
Use of bath additives
Every time	118 (58.1)	9 (5.1)
More than half the time	55 (27.1)	4 (2.3)
Less than half the time	20 (9.9)	18 (10.2)
Never	10 (4.9)	145 (82.4)
Number of baths per week
1–2	69 (36.5)	57 (35.6)
3–4	65 (34.4)	50 (31.3)
5–6	28 (14.8)	29 (18.1)
≥ 7	27 (14.3)	24 (15.0)

Secondary outcomes

Exacerbations

In total, we had exacerbations data for 470 (97.5%) children. Of these, 257 (54.7%) had at least one exacerbation as defined in the protocol (i.e. GP notes review recorded consultations in which there was mention of eczema and topical steroid or TCI was advised or prescribed).

The distributions of the number of exacerbations is skewed and follows an approximately negative binomial distribution (Figure 7). As shown in Table 10, on average, the number of exacerbations was similar between groups. The unadjusted results indicated slightly more exacerbations in the no bath additive group, but in the adjusted model there was no significant difference in the number of exacerbations between groups.

FIGURE 7.

Distribution of exacerbations.

TABLE 10 - Number of exacerbations of eczema over 52-week secondary outcome period

IQR, interquartile range.

Time point	Treatment group		RR exacerbations (95% CI)
	Bath additive	No bath additive	Univariate	Adjusted
Median number of exacerbations (IQR)	1 (0–2)	1 (0–3)	1.33 (1.02 to 1.75)	1.24 (0.96 to 1.60)

Eczema severity over 1 year (from baseline to 52 weeks): monthly Patient Orientated Eczema Measure scores

The difference between groups in monthly POEM scores from baseline to 52 weeks was explored and tended to be non-significant (Table 11 and Figure 8).

TABLE 11 - Mean monthly POEM scores during the 52-week secondary outcome period, by group

POEM scores	Treatment group, mean (SD)
	Bath additive	No bath additive
Baseline	9.5 (5.7)	10.1 (5.8)
Week 4	7.6 (6.0)	8.6 (6.1)
Week 8	7.3 (6.2)	8.2 (5.8)
Week 12	7.7 (6.2)	7.9 (5.9)
Week 16	7.1 (6.1)	8.2 (6.3)
Week 20	6.9 (6.2)	8.7 (6.5)
Week 24	7.3 (6.5)	8.3 (6.7)
Week 28	7.4 (6.4)	8.8 (6.5)
Week 32	7.8 (6.7)	8.7 (7.0)
Week 36	7.2 (6.8)	8.8 (6.8)
Week 40	7.3 (6.4)	8.8 (6.7)
Week 44	6.9 (6.3)	8.2 (6.4)
Week 48	7.4 (6.6)	8.6 (6.9)
Week 52	7.1 (6.2)	8.0 (6.4)

FIGURE 8.

Patient Oriented Eczema Measure scores during the 52-week secondary outcome period, by group.

There was no statistically significant difference between the two groups at 52 weeks, based on monthly POEM scores over the period, and the CIs were well below 2 points (Table 12).

TABLE 12 - Mean POEM scores during the 52-week secondary outcome period, by group

Time point	Treatment group, mean (SD)		Univariate difference in mean POEM (95% CI)	Adjusted difference in mean POEM (95% CI)
	Bath additive	No bath additive
Secondary outcome: monthly repeated measures
Over the 52-week period (repeated measures)	7.3 (6.3)	8.4 (6.4)	0.99 (0.03 to 1.96)	0.75 (–0.05 to 1.55)

Dermatitis Family Impact Questionnaire

The distribution of the DFIQ was very skewed, with almost two-thirds of participants (62.9%) scoring ≤ 4 out of 27 on the scale. Therefore, a non-parametric approach has been used. The quantile regression compares the median values rather than the means. There was no statistically significant difference in the DFIQ at either 16 or 52 weeks (Table 13).

TABLE 13 - Dermatitis Family Impact Questionnaire at 16 and 52 weeks, by group

IQR, interquartile range.

Prescriptions	Treatment group		Univariate difference in median DFIQ (95% CI)	Adjusted difference in median DFIQ (95% CI)
	Bath additive (n = 230)	No bath additive (n = 186)
Median DFIQ at baseline (IQR)	2 (1–6)	3 (1–7)	–	–
Median DFIQ at 16 weeks (IQR)	2 (0–5)	3 (1–7)	1.00 (0.09 to 1.91)	0.29 (–0.57 to 1.14)
Median DFIQ at 52 weeks (IQR)	2 (0–5)	2 (0–6)	0.00 (–0.93 to 0.93)	–0.29 (–1.36 to 0.79)

Health-related quality of life

See Chapter 4, Health economic evaluation, for a discussion of the CHU-9D.

Use of topical corticosteroids and topical calcineurin inhibitors

There were a total of 671 prescriptions for TCSs and 32 prescriptions for TCIs (Table 14). As shown in Figure 9, the distribution is skewed, with 44% children receiving no TCS or TCI prescription and 85% receiving fewer than five prescriptions over the 52-week study period.

TABLE 14 - Prescriptions for TCSs/TCIs, by group

Number of prescriptions	Treatment group
	Bath additive (n = 258)	No bath additive (n = 164)
Total number of TCS/TCI prescriptions	325	346
Median number of TCS/TCI prescriptions	0 (0–2)	1 (0–3)

FIGURE 9.

Prescriptions for TCSs and TCIs.

Introduction

Eczema is a skin condition that is very common in children. The economic implications of eczema in children are described in the introduction (see Chapter 1) of this report and are well documented in the literature. 33 Recommendations by the NICE guideline9 on childhood eczema suggest the use of a ‘complete emollient therapy’ that includes bath emollient (bath additives) in addition to emollient cream and soap substitutes. However, the guideline also highlights the uncertainty from limited evidence on the benefit of including bath additives in this combination of treatments. In addition to the clinical question that this uncertainty generates, there is also an economic question to be addressed. This is even more important in the current economic climate in which NHS resources are extremely limited. Given that some estimates have suggested that bath additives account for more than one-third of the total costs treating eczema in childhood,9,13 the relevance of the economic question becomes crucially important.

Economic evaluations alongside clinical trials (EEACT) provide timely information with high internal validity. 50 When conducting EEACT, the quality of the economic information derived depends on the features of the trial’s design. However, it is generally acknowledged that pragmatic effectiveness trials are the best vehicle for economic studies. 50 For the economic analysis, the pragmatic nature of the trial means that the external validity of the economic results is increased by avoiding protocol-driven biases, such as artificial resource use.

Economic evaluations alongside clinical trials involve the comparative analysis of alternative interventions in terms of their costs and benefits. 51,52 Methodological guidelines for EEACT differ in their recommendations for the most appropriate perspective that should be adopted. As a minimum, it is recommended that analysts adopt a health system perspective for analysis. For England and Wales, this is currently considered to include the NHS and personal social services. 53

The BATHE trial was a multicentre, pragmatic, non-blinded, randomised controlled trial with two parallel arms. The study population for the BATHE trial is described in Chapter 3, Results. Ideally, the economic evaluation would be factored into sample size calculations using standard methods, based on asymptotic normality, or by simulation. 54,55 However, it is common for the sample size of the trial to be based on the primary clinical outcome alone. As a consequence, sample size restrictions necessitate a focus on estimation rather than hypothesis testing for our economic evaluation. 56 The sample size in the BATHE trial was calculated for repeated measures ANOVA in weekly POEM scores over the 16-week observation period. In our protocol and the health economics analysis plan (HEAP) we stated our intention to conduct a cost-effectiveness analysis using the primary outcome, POEM, a cost–utility analysis using utility values obtained from the paediatric CHU-9D and to also report cost per exacerbation avoided. However, the economic results in this study reveal that none of these outcomes showed a clinically important difference. Therefore, our approach was to proceed by reporting our findings and ultimately presenting our economic evaluation in the form of cost–consequences analysis,57 as per our HEAP. Following this, the economic evaluation was conducted from the NHS and Personal Social Services perspective; however, family-borne costs (FbCs), in the form of additional analyses, were also incorporated and the combined results are also reported in this chapter.

Methods

The range of cost and effectiveness outcomes used are described in this chapter. The economic evaluation used resource use, cost and effectiveness outcomes data collected for all of the participants enrolled in the BATHE trial as described in Chapter 3, Results. The time points of the evaluation were the 16- and 52-week follow-up periods used for the trial. The maximum 52-week follow-up period of the trial means that discounting of costs and outcomes was not relevant and was not conducted. The intervention was conducted in primary care settings; however, the setting for the economic evaluation covers both primary and secondary care resource use.

The analyses were carried out using the ITT approach, and individual patient data were estimated for each participant. All economic analyses were performed using Stata^® version 14 (StataCorp LP, College Station, TX, USA) and Microsoft Excel.

Analysis

Our analysis follows a prespecified HEAP. The purpose of the economic analysis was to estimate costs, cost and effectiveness outcome differences associated with the treatment, the variability of differences and whether or not the differences occurred by chance. Our economic results reported here are expressed in terms of net and incremental costs and effectiveness outcomes.

Regression models were used to estimate net and incremental cost and effectiveness outcomes and to adjust for confounders where appropriate. We used MMLMs controlling for baseline POEM and allowing for clustering of patients within centres, thereby following the same process undertaken for the statistical analysis for consistency of results.

Results

The sample population for the economic analysis was 482 participants in total, the same as the statistical analysis. Missing values are indicated in each section below. In our sample, 264 participants were included in the bath additives group and 218 participants in the no bath additives group. The clinical and sociodemographic characteristics of the BATHE participants included in our economic study were well balanced between groups (see Table 2). The sections below report resource use and cost estimates from both data sources (i.e. CSRI and GP NR), and Table 17 presents the unit costs used for the valuation of the resources used.

Family-borne costs

Table 23 presents estimates of the FbCs as reported by parents within the CSRI questionnaire. The FbCs were estimated at baseline and at 16-week and 52-week follow-up, but none of the differences assessed at each time period reached statistical significance. Overall, parents within the no bath additives group annually spent £51.37 (95% CI –£118.49 to £15.74) more on household items for their child’s eczema than parents within the bath additives group (see Table 23), but this difference was not statistically significant.

TABLE 23 - Family-borne costs from the CSRI

Parents were asked to report household items in excess of normal use attributable to their child’s eczema.

Household itemsa	Treatment group, mean cost (£) (SD)		Difference, mean (95% CI)
	Bath additives (n = 264)	No bath additives (n = 218)
Baseline
Clothes	7.09 (63.3)	4.45 (15.5)	2.64 (–5.99 to 11.27)
Food	4.13 (16.9)	5.39 (19.4)	–1.26 (–4.51 to 1.99)
Over-the-counter products	7.15 (15.7)	9.02 (19.9)	–1.87 (–5.06 to 1.31)
Laundry	7.64 (20.5)	8.47 (18.2)	–0.83 (–4.33 to 2.68)
Equipment	1.92 (12.8)	0.83 (12.2)	1.09 (–1.15 to 3.34)
Travel costs	0.78 (6.6)	0.66 (3.3)	0.12 (–0.84 to 1.08)
Complementary medicine	2.56 (37.0)	0.47 (3.4)	2.10 (–2.84 to 7.04)
Other	2.27 (36.9)	0.07 (0.8)	2.20 (–2.72 to 7.12)
Total family-borne extra costs	33.54 (105.6)	29.35 (49.5)	4.19 (–11.09 to 19.47)
16 weeks	n = 236	n = 194
Clothes	3.42 (11.5)	4.38 (16.7)	–0.96 (–3.64 to 1.72)
Food	2.60 (10.1)	4.19 (20.4)	–1.59 (–4.56 to 1.39)
Over-the-counter products	6.23 (14.5)	7.35 (17.2)	–1.12 (–4.13 to 1.89)
Laundry	4.98 (11.7)	8.01 (19.0)	–3.03 (–5.97 to –0.09)
Equipment	0.71 (5.6)	0.86 (5.7)	–0.15 (–1.23 to 0.92)
Travel costs	0.85 (8.9)	0.49 (2.9)	0.35 (–0.96 to 1.67)
Complementary medicine	6.32 (95.5)	0.93 (6.5)	5.39 (–8.11 to 18.89)
Other	0.05 (0.6)	0.23 (2.2)	–0.18 (–0.46 to 0.11)
Total family-borne extra costs	25.15 (112.6)	26.44 (52.7)	–1.29 (–18.55 to 15.97)
52 weeks	n = 203	n = 184
Clothes	11.58 (29.8)	22.15 (99.1)	–10.57 (–24.90 to 3.76)
Food	12.86 (45.6)	26.57 (112.5)	–13.72 (–30.59 to 3.15)
Laundry	20.11 (45.3)	28.70 (73.3)	–15.86 (–35.98 to 4.27)
Over-the-counter products	23.23 (42.3)	39.09 (138.9)	–8.59 (–20.64 to 3.46)
Equipment	2.46 (12.8)	6.99 (47.5)	–4.53 (–11.33 to 2.28)
Travel costs	2.87 (21.7)	4.10 (20.8)	–1.24 (–5.49 to 3.02)
Complementary medicine	8.44 (110.4)	7.93 (48.7)	0.51 (–16.84 to 17.86)
Other	9.24 (98.4)	6.76 (53.0)	2.48 (–13.55 to 18.51)
Total family-borne extra costs	90.93 (276.6)	142.30 (390.1)	–51.37 (–118.49 to 15.74)

Figure 10 highlights the skewed distribution of FbCs, showing that this difference is to a great extent driven by outliers. When these outliers were normalised by assigning mean values to them, the mean difference of the FbCs was reduced to –£40.28 (–£78.78 to –£1.78). Figure 11 shows the normalised data, highlighting the similarities between the two groups.

FIGURE 10.

Family-borne costs. Data sourced from the CSRI. 58

FIGURE 11.

Family-borne costs (normalised outliers). Data sourced from the CSRI. 58

Implications of the results

We aimed to assess whether or not bath additives, when used in addition to standard management versus standard management alone, could provide a cost-effective treatment option. The economic study shows that in children with eczema, the use of bath additives does not provide any additional economic or otherwise benefit. Therefore, bath additives cannot be considered value for money for the NHS. Given the amount spent annually on bath additives, this has important implications for the NHS and decision-makers alike.

Strengths and weaknesses

A strength of this economic study is the use of alternative data sources estimating the health system costs. This allowed us to present a comprehensive resource use profile for children with eczema. We also consider the pragmatic nature of this trial, which reinforces the external validity of our results, to be a strength.

The broad spectrum of the age of the children included in our trial is a limitation when assessing QoL, especially as there are no suitable measures to assess the QoL of very young children. Although we cannot eliminate this limitation, we have reported our results for the full sample alongside the two different groups (< 5 years/> 5 years) to avoid diluting the conclusions drawn by presenting only the full sample, which includes the very young age group and introduces additional uncertainty due to measurement limitations.

Conclusions

Given the finding of no clinical effectiveness of the intervention in this study, it could be argued that a full cost-effectiveness analysis has nothing to offer in addition to what is already presented in this chapter. We have, however, presented our results at length adopting a cost–consequences analysis format, in line with our HEAP. During this trial, we collected comprehensive data for both the costs and QoL, and we believe that these data present a valuable source of information for future studies and decision-making.

Comparing the alternative data sources, the difference between primary and secondary care consultations reported by parents (CSRI) and recorded within the GP electronic records (NR) perhaps shows the different perceptions classifying consultations as being for eczema or not, and this raises important questions for future economic studies. By comparing the two data sources, we also validated our assumption that missing data were missing at random.

Given that emollient bath additives account for more than one-third of the total costs of treating eczema in childhood,9,13 we believe that the relevance of our economic results is essential.

Main findings

This trial found no evidence of any clinical benefit or cost-effectiveness of adding bath additive to the standard management of eczema in children aged between 12 months and 12 years. After controlling for baseline variables, the weekly POEM score over 16 weeks in the no bath additive group was 0.41 points higher than in the bath additives group (95% CI –0.27 to 1.10) on a scale of 0 (clear) to 28 (most severe). The narrow 95% CIs suggest that a clinically important treatment effect is unlikely to have been missed and is well within the meaningful difference of 2.0 between groups that we aimed to detect.

Although not powered to look at subgroups, prespecified exploratory subgroup analyses explored features that could plausibly modify the effectiveness of bath additives including age, baseline severity, use of leave-on emollient, use of TCSs/TCIs, use of soap substitute, frequency of bathing and prior belief in bath emollient. Although most of these showed no significant differences between groups, we could not exclude the possibility that younger children or children bathing ≥ 5 times per week might experience very modest benefit [1.29 (95% CI 0.33 to 2.25) and 2.27 (95% CI 0.63 to 3.91), respectively]. Although the published minimal clinically important difference for POEM is now widely regarded as being 3.0,41,42 at the trial design stage we based our sample size calculation on a difference of 2.0, given that this could be regarded as meaningful for an inexpensive intervention in a population with mild eczema.

Primary analyses were carried out on an ITT basis. We carried out a per-protocol analysis in addition to this in order to explore the effect of adherence to randomised treatment allocation. This allowed an estimate of effectiveness for children whose carers reported that they had regularly used bath additives rather than the effectiveness of being allocated to the bath additive group; this also showed no statistically significant difference between groups.

Relevance to existing literature

Previous systematic reviews have noted the sparse randomised trial evidence for bath additives and have been unable to draw conclusions regarding their effectiveness. 9 Current prescribing guidelines vary, but a recent analysis of 216 formularies in England and Wales showed that 68% recommend their use. 22 This is the largest trial to date exploring the effectiveness of these widely used products.

In a large case series (post-marketing surveillance) of 3566 people,62 most of whom had eczema and were aged ≤ 5 years, the tolerability of the bath additive was stated to be ‘good’ or ‘very good’ for 96.8% of participants. However, few data were presented on other treatments used and the absence of an experimental design meant that the effectiveness of the treatment cannot be inferred.

Although the addition of antiseptics to emollients is used by some for the prevention of infected eczema, there is no robust evidence for this. 63 There are some reports that irritant reactions are more frequent in individuals treated with bath additives with added antiseptics. 20 The trial reported here supports the finding that most bath additives are well tolerated but did not explore the effectiveness or tolerability of bath additives containing antiseptics/antimicrobials or antipruritics.

Strengths

This was an adequately powered randomised controlled trial, with high follow-up rates and good adherence to randomised trial allocations.

As an unblinded trial with a participant-reported measure as the primary outcome, there was a risk of bias in favour of the trial intervention, which is widely prescribed in childhood eczema in the UK. However, a convincing placebo is not possible for emollient bath additives and we wished to design a trial with a clinical outcome relevant to participants. As bath additives are a relatively inexpensive and widely used intervention, we felt that the risk of expectation of benefit leading to systematic bias in outcome reporting was less problematic. Having made this design decision, the finding of no significant difference between groups gives additional reassurance that we are very unlikely to be missing a meaningful benefit from the use of bath additives in childhood eczema.

Limitations

Generalisability

Participants were recruited through primary care in southern England, the west of England and in Wales. As > 90% of eczema is managed in primary care in the UK, these findings would be applicable to most children with eczema. The response rate to letters of invitation from practices, although in keeping with similar studies,67 was relatively low, suggesting that participating families may have been particularly motivated to be involved in research. However, of those who replied that they did not wish to participate, by far the most common reason for this was that their child’s eczema was no longer a problem. It seems likely that many who did not respond would have not returned the reply slip for this reason.

Emollient bath additives are popular eczema treatments with some parents/carers and some prescribers in countries where baths, as opposed to showers, are the norm, although there is international lack of consensus about their role (Professor Masutaka Furue, Kyushu University, Japan, 13 October 2017; Dr Roberto Takaoka, University of São Paulo, Brazil, 12 October 2017; and Professor Eric Simpson, Oregon Health and Science University, USA, 12 October 2017; personal communication).

Implications for health care

This trial provides useful information for parents/carers who are seeking to manage their child’s eczema. Findings from this trial indicate that parents/carers can be advised that adding bath additive to bath water is unlikely to provide additional benefit over standard therapy in childhood eczema. Although we cannot completely exclude the possibility that children aged < 5 years who bath frequently might benefit, this is unlikely to be a clinically meaningful benefit.

Prescribers and policy-makers have already started to look at limiting the use of emollient bath additives in some Clinical Commissioning Group (CCG) areas as a potential cost saving, although an analysis of CCGs (England) and Local Health Boards (Wales) in 2016–17 showed that 68% still recommended the use of bath additives. 22 The findings of this trial would support the suggestion that adding emollients to bath water has no benefit, or minimal benefit at best, and prescribing budgets and parental energies may be best spent on more effective treatments. However, emollient products may be used for more than one purpose and some emollients marketed as bath additives are used as soap substitutes. Our study has not addressed the question of whether or not ‘bath additives’ are effective as soap substitutes, but it has found that advising parents/carers that pouring bath additives into bath water is not effective.

Anecdotal and qualitative evidence (Santer et al. 68 and Dr Miriam Santer, University of Southampton, 2015, personal communication) suggests that some families use emollient bath additives as an alternative to leave-on emollients as a response to child resistance. As there is good evidence for the effectiveness of leave-on emollients,69 our findings suggest that families may be turning to an ineffective treatment (bath additives) because of barriers experienced in adhering to an effective eczema treatment (leave-on emollients). Barriers to regular application of topical treatments, such as leave-on emollients, include the time-consuming nature of such treatments and child resistance. 68 Some strategies to overcome barriers, such as involving the child in treatment, allowing them to choose an emollient that they like, distracting the child during treatment or using rewards, are likely to be more helpful than others, such as physically restraining the child or reducing the frequency of applications. 68 Promoting positive strategies to facilitate regular topical treatment use seems more likely to be helpful than using emollient bath additives as an alternative to leave-on emollients.

It is essential to highlight to prescribers and policy-makers that, although this trial found that pouring emollient bath additives into bath water was not beneficial, the role of emollients as soap substitutes or leave-on emollients has not been explored. Both soap substitutes and leave-on emollients have a much stronger a priori rationale for effectiveness than emollient bath additives and greater clinical consensus around their effectiveness, even though there are relatively few large trials in this area. 69 It is therefore important that policy-makers do not conflate the findings of this study, which relates to a specific method of delivering emollients to the skin via bath water, with the use of emollients as leave-on preparations or as soap substitutes as recommended by the NICE guideline on eczema in children. 9

Recommendations for research

Some participants in this trial were using emollient bath additives as soap substitutes and further research is needed to explore what the best soap substitute is for use in eczema.

Several questions around emollients and washing in eczema that were highlighted in the James Lind Alliance PSP34 remain outstanding, in particular the shared priority from patients and health-care professionals, ‘which emollient is the most effective and safe in treating eczema?’. This is currently being addressed by the NIHR HTA-funded Best Emollient for Eczema (BEE) study (15/130/07). The priority from patients and carers, namely ‘what is the best way for people with eczema to wash: frequency of washing, water temperatures, bath versus shower?’ remains unanswered. The priority from health-care professionals, namely ‘how effective are interventions to reduce skin infections in the management of eczema?’, is also related to the question of bathing and also remains unanswered.

Topic prioritisation

This trial was funded by the NIHR HTA programme as the result of a commissioned call. The research topic had been suggested through the NIHR HTA website topic suggestion form. The NIHR HTA programme has PPI embedded in its prioritisation processes and Boards.

The James Lind Alliance PSP for Eczema34 published its top 10 priority topics in 2012. Even though this call was not advertised directly as a result of the PSP, it addresses issues that patients, carers and clinicians highlighted as an outcome of the PSP, including priorities around bathing/washing and also around the best ways to use emollients.

Patient and public involvement co-applicant

Amanda Roberts has eczema, has two (grown-up) children with eczema and is an experienced PPI representative on a number of bodies, including the Centre for Evidence-Based Dermatology (CEBD) Patient Panel. Amanda is in contact with many carers of children with eczema through running the Nottingham Support Group for Carers of Children with Eczema (NSGCCE).

Trial conduct

Amanda Roberts attended the team training day in 2014 and gave a presentation on patient/parent/carer perspectives to all trial staff. She attended further face-to-face meetings on trial conduct in 2016 and 2017.

Study materials: Amanda Roberts and members of the CEBD Patient Panel were involved in reviewing and revising the patient information sheet, child information sheet, consent form (see Appendix 12), assent form (see Appendix 13) and invitation letter.

Newsletters: Amanda Roberts was involved in providing content for quarterly newsletters to participants’ families and reviewing the newsletters prior to them being circulated.

After the first few months of recruitment, we found that we were getting relatively low response rates. We therefore discussed recruitment at a CEBD PPI meeting in March 2015. Attendees responded that the patient information sheet was too dense and unfriendly. Using this feedback, a short and colourful summary leaflet was designed and included in the invitation pack from July 2015, but we were unable to detect any change in the response rate as a result of the additional information.

The independent Trial Steering Committee, appointed by the NIHR HTA programme, included independent PPI representative Rosemary Humphries.

End-of-study meeting for presentation and discussion on interpretation of findings

We invited additional patient representatives to the end-of-study discussion of findings, including representatives of national patient advocacy groups (NSGCCE, National Eczema Society, Eczema Outreach Scotland). Attendees included Amanda Roberts (PPI co-applicant), another member of the CEBD Patient Panel (a mother of children with severe eczema), the Chief Executive and Director of Services of the National Eczema Society and the PPI member of the BATHE Trial Steering Committee.

Patient representatives raised concerns that the trial results might be used to justify removing bath emollients from prescribing formularies and that some families rely on bath emollients as soap substitutes. Others acknowledged that understanding the likely effectiveness of these products was potentially useful for parents trying to work out how best to ‘live with eczema’. The study team highlighted that the trial design was more likely to result in a false-positive result than a false-negative result, as it was an open trial with a participant-reported primary outcome measure (and was thus open to expectation bias). All felt that it was better for prescribers and commissioners to make prescribing decisions based on high-quality evidence, although they would not wish to see the use of emollient soap substitutes compromised through misinterpretation of trial findings.

The context of these concerns is that CCG formularies are increasingly restricting the availability of emollients in general, and bath additives in particular, and the National Eczema Society reports that patients are having problems accessing preferred products. Others were also concerned that there is an underlying drive to remove all emollients from the prescribing guidelines, which would leave patients and carers in a difficult position and potentially lead to more flare-ups and greater service use. It was agreed that we need to be clear in all communications that emollients should still be available and that this study suggests only that pouring them into the bath does not offer benefit. All agreed that using the term ‘bath additives’ rather than ‘bath emollients’ might avert some confusion.

Discussion and conclusions

Amanda Roberts was an integral member of the team from the first decision to respond to the call and was an enthusiastic, reliable and inspiring member of the team in terms of maintaining the focus of the study on providing answers for parents/carers on one of the key questions in eczema care. Her input was invaluable.

The following quotation illustrates her perspective on her involvement:

My take on what I contributed to the trial was this: everyone on the team was enthusiastic about bearing in mind the needs of the patients and carers – but they had other hats. I reminded people in telephone conferences things – like producing the newsletters etc. – that had been perhaps lost sight of in all the business of the trial.

Reproduced with permission

Input from the CEBD Patient Panel was useful at a number of points, but the other aspect of PPI input that was particularly valuable was obtaining feedback on results from a wide range of stakeholders, particularly PPI, at the end-of-study meeting for discussion of interpretation of findings. This prepared and briefed us for report writing, identifying early on how to communicate messages about trial findings and preparing wider materials for dissemination with trial participants, parents/carers of children with eczema and clinical and commissioning communities.

Contributions of authors

Miriam Santer (Associate Professor in Primary Care Research) contributed to the conception and design of the study, trial oversight (as chief investigator) and interpretation of the data and drafted and critically reviewed the report.

Kate Rumsby (Clinical Trial Manager) contributed to the design of the study, managed the trial, contributed to interpretation of the data and drafted and critically reviewed the report.

Matthew J Ridd (Consultant Senior Lecturer in Primary Health Care) contributed to the conception and design of the study, was responsible for oversight of the trial at the Bristol centre (as principal investigator), contributed to interpretation of the data and critically reviewed the report.

Nick A Francis (Senior Clinical Research Fellow) contributed to the design of the study, was responsible for oversight of the trial at the Cardiff centre (as principal investigator), contributed to interpretation of the data and critically reviewed the report.

Beth Stuart (Senior Research Fellow and Medical Statistician) contributed to the conception and design of the study, was responsible for the statistical analysis plan, carried out the statistical analyses, contributed to interpretation of the data and drafted and critically reviewed the report.

Maria Chorozoglou (Senior Research Fellow and Health Economist) contributed to the design of the study, carried out the health economics analyses, contributed to interpretation of the data and drafted and critically reviewed the report.

Amanda Roberts (Patient Representative) contributed to the conception and design of the study and interpretation of the data and critically reviewed the report.

Lyn Liddiard (Clinical Studies Officer) contributed to the management of the study and to the collection of data and critically reviewed the report.

Claire Nollett (Clinical Studies Officer) contributed to the management of the study and to the collection of data and critically reviewed the report.

Julie Hooper (Trial Administrator) contributed to the management of the study and to the collection of data and critically reviewed the report.

Martina Prude (Clinical Studies Officer) contributed to the management of the study and to the collection of data and critically reviewed the report.

Wendy Wood (Research Adviser in NIHR Research Design Service and former Senior Trials Manager Southampton CTU) contributed to the conception and design of the study, was responsible for oversight of study setup and critically reviewed the report.

Emma Thomas-Jones (Senior Trial Manager) contributed to the design of the study, contributed to interpretation of the data and drafted and critically reviewed the report.

Taeko Becque (Research Fellow and Medical Statistician) carried out statistical analyses and critically reviewed the report.

Kim S Thomas (Professor of Applied Dermatology Research and Co-Director of the Centre of Evidence-Based Dermatology) contributed to the conception and design of the study and contributed to interpretation of the data, and drafted and critically reviewed the report.

Hywel C Williams (Professor of Dermato-Epidemiology and Co-Director of the Centre of Evidence-Based Dermatology) contributed to the conception and design of the study, contributed to interpretation of the data and drafted and critically reviewed the report.

Paul Little (Professor of Primary Care Research) contributed to the conception and design of the study, contributed to interpretation of the data and drafted and critically reviewed the report.

Publications

Santer M, Rumsby K, Ridd MJ, Francis NA, Stuart B, Chorozoglou M, et al. Bath additives for the treatment of childhood eczema (BATHE): protocol for multicentre parallel group randomised trial. BMJ Open 2015;5:e009575. https://doi.org/10.1136/bmjopen-2015-009575

Santer M, Ridd MJ, Francis NA, Stuart B, Rumsby K, Chorozoglou M, et al. Emollient bath additives for the treatment of childhood eczema (BATHE): multicentre pragmatic parallel group randomised controlled trial of clinical and cost effectiveness. BMJ 2018;361:k1332.

Stuart B, Rumsby K, Santer M, Ridd M, Francis N, Chorozoglou M, et al. Feasibility of weekly participant-reported data collection in a pragmatic randomised controlled trial in primary care: experiences from the BATHE trial (Bath Additives for the Treatment of cHildhood Eczema). Trials 2018; in press.

Data-sharing statement

All data requests should be submitted to the corresponding author for consideration. Access to available anonymised data may be granted following review.

Patient data

This work uses data provided by patients and collected by the NHS as part of their care and support. Using patient data is vital to improve health and care for everyone. There is huge potential to make better use of information from people’s patient records, to understand more about disease, develop new treatments, monitor safety, and plan NHS services. Patient data should be kept safe and secure, to protect everyone’s privacy, and it is important that there are safeguards to make sure that they are stored and used responsibly. Everyone should be able to find out about how patient data are used. #datasaveslives You can find out more about the background to this citation here: https://understandingpatientdata.org.uk/data-citation.

Disclaimers

This report presents independent research funded by the National Institute for Health Research (NIHR). The views and opinions expressed by authors in this publication are those of the authors and do not necessarily reflect those of the NHS, the NIHR, NETSCC, the HTA programme or the Department of Health and Social Care. If there are verbatim quotations included in this publication the views and opinions expressed by the interviewees are those of the interviewees and do not necessarily reflect those of the authors, those of the NHS, the NIHR, NETSCC, the HTA programme or the Department of Health and Social Care.