A behaviour change package to prevent hand dermatitis in nurses working in health care: the SCIN cluster RCT

Moisturisers

Two systematic reviews of the management of occupational dermatitis20,21 have concluded that moisturisers contributed importantly to both prevention and treatment of hand dermatitis at work. The review commissioned by NHS Plus focused on the evidence for managing established occupational dermatitis and found a small body of consistent evidence that hand moisturisers used before and during work improved skin condition in health-care workers with damaged skin,22,23 and concluded that there was sufficient evidence to recommend that skin-care programmes should include the use of hand moisturisers. 24,25

Guidelines produced by the British Occupational Health Research Foundation21 recommended the regular application of emollients to help prevent the development of occupational dermatitis, citing three high-quality studies: a systematic review26 and two randomised controlled trials (RCT). 27,28 One RCT found an improvement in all outcomes, including visible dermatitis. In the other RCT, transepidermal water loss improved among construction workers who used pre- and after-work creams compared with controls, but there was no difference in the workers’ clinically assessed skin condition. 28 Moisturisers also improved skin condition in health-care workers with damaged skin. 29 More recent reviews, including a Cochrane review,30 concluded that there is some evidence to support the use of educational interventions that include moisturisers to prevent occupational hand dermatitis, but, as the evidence came from only a small number of studies, the authors strongly recommended that larger high-quality RCTs in different occupational groups were needed. 30,31 Following these recommendations, Soltanipoor et al. 11 conducted a RCT in the Netherlands that examined the effectiveness of an educational programme together with the provision of moisturisers in dispensers. The dispensers were fitted with electronic usage monitors, allowing the researchers to give feedback to the participants on their moisturiser use. The effectiveness of the intervention was measured from the assessment of skin conditions in participants before and after the interventions. The study found higher levels of self-reported use of hand moisturising cream among nurses before and during shifts in the intervention group than in the control group. However, the self-reported prevalence of hand dermatitis was similar in both study groups at follow-up. Notwithstanding, evidence from the study suggests that electronic tracking of hand cream usage coupled with feedback has a positive impact on skin-care behaviour among health-care workers.

In the experience of the dermatologists and occupational health physicians in this research team, moisturisers are not widely used by health-care workers in the UK. This anecdotal observation is supported by a study of nurses working in intensive care units (ICUs) in Germany that found that only 15% of the 204 respondents reported that they applied moisturising creams after hand-washing and only 2% after skin disinfection with hand rubs. Furthermore, 9% of respondents never applied skin care to their hands and 72% reported that they did not perform final skin care after the last hand wash of the day. 32

Hand cleansing

The use of antibacterial (alcohol-based) hand rubs, with the addition of moisturisers for hand hygiene, reduces the drying and cracking of the skin that commonly results from repeated hand cleansing with soap and water. 33,34 This has the added benefit of also reducing rates of some health-care-associated infections, such as S. aureus bacteraemia and Clostridium difficile infections. 35 Moreover, antibacterial hand rubs have been associated with increased hand-hygiene compliance and reduced rates of nosocomial infection. 17,36 Although the benefits of using antibacterial hand rubs as a safe and effective dermatitis prevention strategy have now been established,2 a notable difficulty faced by occupational health-care workers and dermatologists is to challenge the misconceptions among nurses that antibacterial hand rubs are more damaging to the skin integrity than hand-washing and, therefore, more likely to cause hand dermatitis. 37 There is a misconception among some health-care workers that alcohol-based hand rubs are more irritating and drying than cleaning hands using soap and water; therefore, they are less likely to use them. 38 Accordingly, Stutz et al. 37 argue that hand-hygiene education and training should focus on promoting the use of antibacterial hand rubs as an effective protective behaviour for reducing the prevalence of hand dermatitis in nurses. The implementation of the World Health Organization (WHO)’s guidelines among physicians working in Ireland, which promoted the use of alcohol-based hand rubs by senior management and infection control teams, had a positive impact on changing attitudes and practices towards the use of hand rubs without damaging the epidermis. 39

Hand drying and glove use

The appropriate use of gloves during clinical work and proper drying of the hands after washing is pivotal to good hand hygiene and care, particularly as wet skin is more likely than dry skin to facilitate the transmission of bacteria. A recent review of hand-drying processes,40 which included 12 studies, concluded that paper towels are superior to electric air dryers in drying hands in clinical environments. The ensuing recommendation that paper towels should be used for hand drying in clinical environments, where hand hygiene is vital, was supported by The Royal College of Physicians20 and WHO. 17

Effectiveness of skin-care programmes

Various skin-care programmes that incorporated the provision of hand moisturisers, hand rubs and paper towels have shown a beneficial effect in the prevention of hand dermatitis in health-care workers. 22,23,41–44 However, as John and Kezic45 pointed out, although effective preventative strategies are available, these are often limited by poor understanding of the causes of occupational skin diseases among at-risk workers. Furthermore, others have argued the need to reconsider what is deemed acceptable levels of hand-hygiene compliance, as, for example, a 70% hand-hygiene compliance rate among health-care workers is as effective at reducing infections as 90% compliance. 46

Educational programmes (individual or part of hand health programmes)

Previous studies examining the effectiveness of educational programmes designed to influence and encourage optimal hand-care behaviours among health-care workers22 including ICU nurses47 have yielded only moderately promising results. However, Ibler et al. 43 found that an educational programme coupled with individual counselling focusing on good hand dermatitis prevention behaviours had a positive impact on the secondary prevention of hand dermatitis. The intervention reduced the frequency with which participants washed their hands and increased the frequency of their use of protective gloves during wet work, when compared with their usual treatment. However, this intervention was delivered on a one-to-one basis and so might be difficult to scale up across the health service.

A recent systematic review suggested that educational programmes could benefit from being more strongly informed by psychological theories of behaviour, as programme success relies on employees adopting appropriate preventative and protective behaviours. 31 Furthermore, other studies have also investigated interventions specifically designed to facilitate behaviour change, with an emphasis on improving hand care among health professionals. The Hands4U RCT trial in the Netherlands examined a multifaceted intervention that comprised (1) a leaflet containing evidence-based hand dermatitis prevention recommendations, (2) a participatory working group to explore facilitators of and barriers to adopting recommendations, (3) provision of trained role models (dermacoaches) to encourage and model good hand dermatitis prevention behaviours at a departmental level and (4) provision of education and training to promulgate preventative measures in the workplace. 44 The main outcome measures assessed were self-reported hand dermatitis and adoption of preventative behaviours over a 12-month period. The study found that health-care workers in the intervention group were significantly more likely to engage in positive preventative behaviours to protect themselves from hand dermatitis. That is, the health-care workers reported significantly less hand-washing and more frequent use of hand moisturisers and wearing of cotton under gloves than colleagues in the control arm. However, the intervention group reported a higher prevalence of hand dermatitis than the control group at 12 months. This may have been due to an increased awareness of hand dermatitis as a consequence of the intervention. A notable limitation of this study was the use of self-reported measures to assess hand dermatitis, as opposed to more objective measures. Moreover, no cost-effectiveness analysis of the intervention was undertaken. A prominent observation with many skin-protection programmes is that sustained compliance among at-risk workers is poor and so programmes that promote sustained behaviour change are required. 48

Although there are good reasons to expect that well-designed skin-care programmes would be beneficial for nurses, their clinical effectiveness and cost-effectiveness remain uncertain. Trials to date have been limited by size and the possibility that the control group was aware of the intervention,44 or by a failure to address cost-effectiveness. 49 Furthermore, trials have been hampered by lack of sustainability of behaviour change. Psychological theory suggests that knowledge transfer coupled with a behaviour change programme would be effective in changing the beliefs and behaviours of at-risk nurses in relation to hand care and, if accompanied by provision of hand moisturisers, would lead to a reduction in the prevalence of hand dermatitis.

Inclusion criteria

Included individuals were:

• student nurses who –

  • were first-year student nurses, and

  • had a history of atopic tendency (i.e. had ever had, or received treatment for, asthma, hay fever, eczema anywhere on the body or dermatitis of the hands)

• ICU nurses (including SCBU nurses) who –

  • worked on ICUs (or SCBUs)

  • worked full-time (i.e. ≥ 30 hours per week)

  • had no extended periods of leave of > 3 months in the following 12 months.

Exclusion criteria

The study excluded individuals who:

• were first-year mental health student nurses

• were ICU nurses (or SCBU nurses) working on a part-time basis (i.e. < 30 hours per week), or

• had extended periods of leave of > 3 months planned in the following 12 months.

Student nurses were recruited during their occupational health pre-placement screening appointments in the occupational health departments located at either their local NHS trusts/health boards or their respective universities. ICU/SCBU nurses were recruited on-site at their respective NHS trust/health board. Figures 1 and 2 outline the trial procedures for ICU/SCBU nurses and student nurses, respectively.

FIGURE 1.

Flow chart of the trial procedures for ICU/SCBU nurses. BCP, behaviour change package; GP, general practitioner; SCIN, Skin Care Intervention in Nurses.

FIGURE 2.

Flow chart for the trial procedures for student nurses. BCP, behaviour change package; GP, general practitioner; SCIN, Skin Care Intervention in Nurses.

Questionnaire development and management

Careful consideration was given when designing the study questionnaires to ensure that the study was able to collect reliable measures of TPB variables in relation to hand dermatitis prevention behaviours among the study participants, for example ‘When I start clinical work I intend to apply hand cream before and after work and during each of my breaks’ (intentions), ‘Contacting occupational health if I get symptoms which I think might be hand dermatitis would be worthwhile’ (behavioural belief) and ‘Other nurses at my placement would approve of me using hand rubs, when appropriate for infection control, instead of washing my hands with soap and water’ (subjective norms).

Three study questionnaires were developed and administered at 0, 3 and 12 months during the study. Participant questionnaires comprised a modified version of the Nordic Occupational Skin Questionnaire – 2002 (NOSQ-2002) to measure, before and after randomisation, participants’ self-reported behaviours relating to hand care and dermatitis prevention practices, and activities both inside and outside the workplace. The questionnaires also asked about participants’ rating of their overall health status using the validated EuroQol-5 Dimensions (EQ-5D) questionnaire and their views on the behaviour change package (BCP) (participants in the intervention plus arm of the study only). In addition, the study separately gathered site-level data on the use of moisturising cream (i.e. ad hoc requests for further personal supplies by student nurses and monthly estimates of moisturising creams available in each dispenser on the ICUs and SCBUs).

Local fieldworkers at each participating site made use of individual ‘study packs’ when recruiting participants to the study. Study packs contained each of the research documents (e.g. data collection worksheets, consent forms, questionnaire A booklet, a one-page self-assessment dermatitis questionnaire, business reply envelope) that required completion at the time of recruitment. The second questionnaire booklet (i.e. questionnaire B) was not included in these packs as these were posted directly to participants by the SCIN trial research team based in London. A matching preassigned participant identification number (PIN) was recorded on all research documents contained in each pack and these were numbered sequentially to reflect the order in which they were to be completed. A unique PIN was allocated to each participant as they were recruited to the study. The local fieldworkers then recorded the PIN, participant initials and date of birth on a separate participant registration log sheet as participants were recruited to the study.

Hand photography protocol development

Photographs of hands that were taken by trained personnel using high-resolution digital cameras were considered to be a cost-effective and practical method for collecting data on the presence or absence and severity of dermatitis. Developing a photographic guide to assist with diagnosing hand dermatitis and for use when assessing self-taken photographs of hands (i.e. ‘selfies’) was previously proposed by van der Meer et al. 44 In developing this photographic guide for diagnosing hand dermatitis and assessing its severity, the study team had the following requirements:

• The method had to measure presence or absence of any form of hand dermatitis, as well as severity.

• The method could not involve physical examination of the participants, as that would be logistically very difficult, expensive and likely to result in poor response rates.

• The method had to be objective and not based on self-report as self-reporting tends to over-report hand dermatitis.

• The severity scale needed to be able to distinguish dermatitis towards the milder end of the disease spectrum.

For the purpose of this study, three distinct stages were undertaken when developing the new hand photography method for use in the study. Accordingly, this new method for diagnosing hand dermatitis and assessing its severity offered a multidisciplinary method for diagnosing hand dermatitis and its severity. The method relied solely on dermatologist and research nurse inspection of hand photographs from research participants (in lieu of physical examinations), with comparisons from standardised images contained in Coenraads et al. ’s photographic guide. 71 The stages were:

1. developing a standardised procedure for hand photography

2. a stepwise validation process of rules for the study dermatologists to diagnose and determine the severity of the hand dermatitis

3. training, by a dermatologist, of the research nurse to screen out hand photographs of study participants without dermatitis (‘clear cases’).

The method for taking high-resolution digital photographs of the participants’ hands was developed in conjunction with a medical photographer and is consistent with the views required for the photographic assessment scale, for use in clinical trials, described by Coenraads et al. 71 When participants were unable to return to have their hands photographed by the trained fieldworkers, they were asked to provide, via e-mail, self-taken photographs of their hands, with a stipulation that self-taken photographs were taken against a grey or white background.

Hand dermatitis was assessed through photographs taken of each of the two sides of both hands (i.e. palm and dorsum). The presence of hand dermatitis was based on comparisons made with the standardised images of severity at various stages of diseases, which were contained in Coenraads et al. ’s photographic severity guide. 71 In accordance with this guide, for each combination of sides of the hand, the study dermatologists were required to indicate whether dermatitis was clear (absent), almost clear, moderate, severe or very severe. These four variables (i.e. dermatitis in the right hand at the dorsum, right hand in the palm, left hand at the dorsum and left hand in the palm) were then dichotomised as clear versus almost clear/moderate/severe/very severe. A single binary variable was generated for the presence of dermatitis (i.e. no/yes). Agreement or disagreement on the severity of hand dermatitis was not assessed during the validation process, as it was realised early on in the study that the likelihood of two dermatologists agreeing on the severity grading (five grades) at four different sites was likely to be poor and that perfect agreement on each site was not necessary for this study, which sought to establish a global estimate of hand dermatitis severity. Therefore, a pragmatic view was taken that severity would be defined as the combined score from the two dermatologists. Agreement between the two dermatologists on the binary rating (yes/no) was assessed using Cohen’s kappa statistic. Each participant’s overall severity of hand dermatitis was defined as the most severe combined score from both dermatologists on Coenraads et al. ’s scale from their four hand photographs. The levels of severity of dermatitis, as these were assessed by the two dermatologists, were 1 = clear, 2 = almost clear, 3 = moderate, 4 = severe and 5 = very severe. For each participant positive for dermatitis, a severity score by each of the two dermatologists was defined as the maximum score across the four photographs, then an overall severity of dermatitis score was derived as the average value of the two severity scores given by the two dermatologists. Overall severity was dichotomised using 3 as a cut-off point. Overall severity scores of < 3 (including those without dermatitis) indicated no severe dermatitis, whereas scores of ≥ 3 indicated severe dermatitis. Only two participants had severe dermatitis at baseline and only three participants had severe dermatitis at follow-up. As the prevalence of severe dermatitis was so low, this secondary outcome was not analysed further.

Figure 3 outlines the procedure that the dermatologists and dermatology research nurses used for assessing the hand photographs.

FIGURE 3.

Flow chart for assessing hand photographs. USB, universal serial bus.

Development of the behaviour change package

A key component of the study’s intervention was the BCP, which provided nurses with evidence-based information on good hand dermatitis prevention behaviours and encouraged them to form implementation intentions regarding specific hand-care behaviours, such as when and where they would use hand moisturising cream and when they would regularly check their hands for signs of hand dermatitis. The BCP employed the TBP, supplemented by implementation intentions, as its key theoretical framework. As described previously, the TPB postulates that a person’s behaviour is influenced by their intentions and perceived behavioural control (PBC). For example, whether or not a nurse uses hand moisturiser at the start and end of a shift is influenced by his/her intentions (e.g. ‘I strongly intend to use hand moisturiser at the start and end of a shift’) and his/her PBC (e.g. ‘I find it quite easy to use hand moisturiser at the start and end of a shift’). Intentions are influenced by attitudes towards the behaviour (e.g. ‘I believe using hand moisturiser at the start and end of a shift will be beneficial’) and subjective norms (e.g. ‘the matron wants me to use hand moisturiser at the start and end of a shift’) as well as PBC (as before).

Prior to the development of the BCP, the trial’s health psychologist conducted a focus group meeting with first-year student nurses, exploring their beliefs about the different hand dermatitis prevention behaviours likely to be targeted by the intervention. This work partly used open-ended belief elicitation questions recommended by Ajzen,57 but also explored student nurses’ perceptions of the practicality of performing behaviours frequently and their understanding of the consequences of hand dermatitis. For practising nurses, similar work was performed using an online questionnaire (with open-ended questions), so that the data could be collected at a time that was convenient for each nurse. Importantly, the early work revealed that nurses perceived that applying hand cream more frequently than at the start and end of shifts and before going on breaks was an impractical recommendation; therefore, making such a recommendation would damage the credibility of the BCP. It also allowed the health psychologist to identify misconceptions relating to hand dermatitis, such as alcohol-based hand rubs being more harmful to skin than hand-washing, and to ensure that specific information was included in the BCP to help to challenge these beliefs.

The BCP was designed to target four key hand dermatitis prevention behaviours:

1. using hand moisturising cream

2. using disinfectant hand rubs rather than hand-washing, while remaining in line with infection control guidance

3. using gloves for the shortest time possible while still conforming with infection control guidance

4. checking oneself regularly for hand dermatitis symptoms and seeking support from occupational health if required.

Different versions were developed for student nurses with an atopic tendency and ICU/SCBU nurses, given their differing levels of nursing experience and working environments.

The intervention was intended to be delivered primarily online, as this would allow good scalability if it proved to be effective. A web-based intervention standardises treatment delivery, thereby enhancing intervention fidelity. 72 Moreover, given that nurses work a variety of shift patterns, an online intervention allowed participants to access the BCP at a time that was convenient for them. Given the demands on nurses’ time, the intervention consisted of a single session with two brief follow-up reminder e-mails to access the intervention.

The intervention aimed to increase each of the behaviours by targeting relevant attitudes, subjective norms, PBC and implementation intentions. Specific change techniques73 were selected to target each construct, based on published expert consensus regarding which behaviour change techniques (BCTs) were most effective at changing particular psychological constructs. Where the expert consensus suggested a number of possible BCTs, the ultimate selection was based on feasibility of delivery in an online intervention and likely acceptability to participants. 73

Content of the BCP was intended to reflect current evidence-based guidelines and scientific literature on hand dermatitis prevention and infection control, at the time the intervention was developed, and to be appropriate for nurses working in NHS settings.

Patient and public involvement

The patient involvement in this study differs from that in other studies in that the ‘patients’ are NHS nurses. During the set-up and conduct of the trial, the research team actively consulted with representatives from the nursing (including student) profession. In particular, the trial involved a patient representative, Wendy Taylor, who:

1. ensured that the proposed interventions and data collection tools were acceptable

2. provided her expertise on the planning and management of the trial

3. provided her expertise on the optimum dissemination of the trial results to ensure that they reach the target audience.

Wendy Taylor is a midwife with a history of hand dermatitis acquired during her nurse training. She commented on the draft proposal, was a member of the Trial Steering Committee and attended one of the training workshops for fieldworkers.

During the intervention development stage, the study also conducted focus group sessions involving student nurses and registered nurses for the purpose of seeking their views on hand dermatitis, including prevention strategies, approaches to hand hygiene in the workplace and their preference for hand moisturisers to use during the trial.

Study group 1 student nurses

All student nurses are required to attend an occupational health assessment before starting their clinical work. At participating sites, with permission from the universities concerned, all student nurses in one or more year groups (excluding mental health nursing students) who were due to start their first clinical placement were provided with a participant information sheet by their university or their occupational health department, before or at the time of their mandatory assessment. Those student nurses who had a history of atopic tendency or hand dermatitis were identified by the occupational health department at the assessment from information that they provided in a generic pre-placement health screening questionnaire or during screening conversations the fieldworkers had with potential participants. An occupational health clinician explained to the student nurses that, because of their constitution, they are at increased risk of hand dermatitis. Student nurses who met the inclusion criteria were then invited to participate in the study. The local fieldworker obtained written consent from those nurses who agreed to take part in the study (the lead occupational health practitioner at the study site and the trial manager were available to answer questions when necessary). The consent form asked participants to provide a preferred e-mail address and telephone number to facilitate follow-up, and so that those at intervention plus sites could be sent a link to the online BCP. Paper copies of the BCP were provided at sites where participants were unable to access the online version. One copy of the signed consent form was filed in the nurse’s occupational health notes, one copy was sent to the trial manager and a third copy was given to the participant. Participants were also provided with an information sheet to give to their general practitioner (GP).

Study group 2 intensive care unit nurses (including special-care baby unit nurses)

The investigators, trial manager and lead occupational health clinicians from each site identified all critical care nurses who were suitable for the study. A local occupational health clinician or senior ICU nurse explained to all nurses working on the selected ICUs that they are at increased risk of hand dermatitis because of frequent hand-washing with cleansers and water. Full-time ICU nurses (i.e. those working ≥ 30 hours per week) were provided with a participant information sheet and, when possible, given a week to decide whether or not they wished to participate in the study. Consent to take part was obtained by the fieldworker, who, along with the trial manager, was available to answer any queries about the study. The consent form asked participants to provide a preferred e-mail address to facilitate follow-up, and so that those participants at the intervention plus sites could be sent a link to the online BCP. Paper copies of the BCP were provided at sites where participants were unable to access the online version. Participants were provided with an information sheet to give to their GP.

The study was presented to both study groups as research to assess the causes, consequences and ways of managing hand dermatitis in nurses who are at increased risk, either because of a personal history of atopy disease or because of the type of work that they do. However, to minimise the chance of bias, the nurses were not told that they were in an intervention plus or intervention light group. During the recruitment phase, fieldworkers reinforced the importance of participants being able to commit fully to the study, particularly with respect to the completion of the study questionnaires.

To check that rates of participation did not differ importantly between sites randomised to the two arms of the study, the recruitment process was carefully documented. The total number of eligible student and critical care nurses was recorded, as was the number who were consented to take part. In addition, the numbers and dates of those who dropped out of the study were recorded.

The recruitment period was between September 2014 and December 2015 and the follow-up period was between September 2015 and March 2017.

Intervention plus

The intervention plus in both staff groups centred on a bespoke BCP that targeted the behaviours of appropriate glove use, use of antibacterial hand rubs versus hand-washing, regular use of moisturising cream and contacting occupational health early if hand dermatitis occurred.

The BCP was developed by members of the study team with expertise in dermatology, occupational medicine, nursing and health psychology. As described earlier in this report, the intervention aimed to change relevant attitudes, subjective norms, PBC and implementation intentions. Copies of the BCP interventions screenshots (from the online platform: see Report Supplementary Materials 1 and 2) for the student nurses and ICU nurses are provided as supplementary documents. When participants were asked to form implementation intentions for performing the behaviours in their workplace, a record of each participant’s implementation intentions was generated by the online BCP package and e-mailed to her/him.

If a participant reported being unable to access the online BCP, she/he was posted a paper-based magazine version of the BCP, which reflected the information provided in the online BCP. Participants were asked to read through the material provided and write down their action plans in the spaces provided. The participants were then asked to keep the paper-based BCP in a convenient place so that, as required, they could refer back to it. The BCP was supported by provision of facilities to encourage adherence. These included personal supplies of moisturising cream for at-risk student nurses and provision of (1) optimal equipment for cleaning hands and (2) moisturising cream dispensers on ICUs and SCBUs.

The BCP was made available to ICU/SCBU nurse participants within 12 weeks of their recruitment to the study and for student nurse participants 2 weeks before they started their first clinical attachment. However, for many participants, access to the BCP occurred within a shorter timeframe (i.e. 2–4 weeks from their recruitment). It was actively reinforced over the course of the study by consistent messages on skin care from the local occupational health and control of infection teams, and from local line management. Research has shown that senior role models have important effects on more junior health professionals’ hand-hygiene behaviours,74,75 and it seemed reasonable to assume that this influence would extend to behaviours preventing dermatitis. To facilitate this, when implementing intervention plus, a dialogue was engaged in with local occupational health staff and line managers about the nature and purpose of the study and information was provided to them on the content provided in the BCP to ensure that they promoted consistent messages on skin care.

The online BCP was offered to allow nurses to access it at a time convenient to their schedules, to permit standardisation of the delivery of key information across all intervention plus sites and to reduce the potential burden on occupational health staff. Moreover, it was considered that, if the BCP was found to be effective in this trial, it would be simple to scale up access to the website to deliver the BCP across the country.

In the early stages of the trial, responses from the intermediate questionnaires and electronic user activity data showed that the initial uptake of the BCP was poor for both study groups. The most commonly cited reasons for this were that most participants forgot to, or did not have time to, access the BCP. In response to this, the research team implemented a targeted strategy to further promote and encourage uptake of the BCP. This comprised attaching a leaflet to the intermediate questionnaire that contained the website link and encouragement for participants to access the material; additional specific information in the follow-up reminder e-mails that were sent to participants; and displaying BCP information on posters in prominent locations in clinical areas.

Intervention light

Nurses at the intervention light sites were managed in accordance with what was currently regarded as best practice, with the provision of an advice leaflet about optimal hand care entitled Dermatitis: Occupational Aspects of Management. Evidence-based Guidance for Employees (which was also provided to the intervention plus group) and encouragement to contact their occupational health department early if hand dermatitis occurred. However, the nurses did not receive the BCP or active reinforcement of its messages. Nor were the nurses routinely offered supplies of moisturising cream over and above what was already standard practice at their site.

During the recruitment phase, fieldworkers reinforced to participants the importance of being able to fully commit to the study, particularly with respect to the completion of each of the study questionnaires.

Study group 1 student nurses

At both the intervention plus and the intervention light sites, all first-year student nurses were given the participant information sheet about the study. Student nurses who agreed to take part were invited to complete a consent form, a self-administered baseline questionnaire booklet and a one-page hand dermatitis self-assessment form (t = 0 months). The completed questionnaires were then placed into a sealed business reply envelope and returned to the fieldworker, who forwarded them on to the SCIN trial research team in London. If the student nurses preferred, they could send the completed questionnaires directly back to the SCIN trial team in London. After the two questionnaires had been completed, all student nurses were invited to have their hands photographed as per a standard operating procedure developed for use in the study. Separate consent was obtained from the participating nurses each time their hands were photographed. In addition, the fieldworkers provided participants with oral information about hand dermatitis and with the Dermatitis: Occupational Aspects of Management. Evidence-based Guidance for Employees written leaflet. Participants were also told that they may be sent a link to access a BCP that they should undertake in the week before starting their first clinical attachment. Participants were required to log on to the online BCP package and register as a first-time user. At that time, the nurses would also be sent (by post) three personal tubes of moisturising cream, with guidance on how to request further supplies if needed.

All participants in both the intervention plus and the intervention light sites were encouraged (orally, through the written advice leaflet and by e-mail reminders at 4 and 8 months) to attend their local occupational health department at an early stage if they developed hand dermatitis during the study period. One week after starting their first clinical attachment, all nurses were asked to complete a further short self-administered questionnaire booklet (t = 3 months), covering beliefs and plans regarding dermatitis prevention behaviours. At intervention plus sites, it also asked about participation in, and views on, the BCP. The intermediate questionnaire was sent by post to participants by the SCIN trial research team along with a business reply envelope so that completed questionnaires could be returned. E-mail reminders were also sent to participants containing a positive reinforcement message to encourage ongoing participation in the study. At the intervention sites, the e-mail reminders also reinforced the BCP.

To account for seasonal variations in the prevalence of dermatitis, the final study data collection tools were administered 12 months after questionnaire A. All participants were asked to complete the final self-administered questionnaire booklet and a one-page dermatitis self-assessment form and to have their hands photographed. The questionnaires were sent out in the post to fieldworkers 2 months before the end of the study by the SCIN trial research team. The fieldworkers gave the student nurses the questionnaires at the time that they recalled them to have their hands photographed. The questionnaire C booklet covered clinical attachments undertaken in the previous year; hours worked per week over the previous year; beliefs and plans regarding dermatitis prevention behaviours; participation in, and views about, the BCP (only at intervention plus sites); activities outside work that predispose to hand dermatitis; recent practices regarding use of gloves; recent practices regarding hand cleansing; recent use of moisturising creams; history of hand dermatitis in the previous 12 months (including its investigation and treatment, and any consequent loss of time from work or restriction of duties); and the EQ-5D questionnaire.

At the follow-up time point (i.e. t = 12 months), participants who were unable to attend to have hand photographs taken by a fieldworker were provided with an opportunity to send the SCIN trial research team self-taken photographs, taken with their mobile phones, of their hands. If it was not possible for a student nurse to complete the questionnaire at the time, then they were asked to return the completed questionnaire to the SCIN trial team directly using a business reply envelope.

Information on the number of attendances at occupational health with symptoms of hand dermatitis and requests for extra provisions of emollients was also recorded.

Study group 2 intensive care unit nurses (including special-care baby unit nurses)

At both the intervention plus and the intervention light sites, all nurses who worked on selected ICU/SCBU wards were given the participant information sheet. Those who agreed to participate were asked to complete a consent form and self-administered questionnaire booklet (t = 0 months). This was similar to the baseline questionnaire booklet administered to student nurses, but also included items on current occupation; recent practices regarding the use of gloves; recent practices regarding hand cleansing; recent use of moisturising creams; and any sickness absence or modification of duties during the previous 12 months because of hand dermatitis. At the time of the photograph, ICU/SCBU nurses were asked to complete a one-page dermatitis self-assessment form evaluating whether or not they considered themselves to have dermatitis and the extent to which it interfered with work and hobbies. ICU/SCBU nurses were asked to return the completed questionnaires either to the fieldworkers or (by post) directly back to the SCIN trial research team in London. All ICU/SCBU nurses had their hands photographed as per protocol. All ICU/SCBU nurses, including those at intervention light sites, were provided with the Dermatitis: Occupational Aspects of Management. Evidence-based Guidance for Employees written leaflet. Participants were also told that they may be sent a link to access a BCP in due course that they should use. Participants were required to log on to the online BCP package and register as a first-time user.

At the intervention plus sites, once participants had been recruited, fieldworkers (occupational health clinicians) promoted the importance of optimisation of equipment for hand cleansing and use of moisturising cream. An e-mail was also sent via the lead ICU/SCBU nurses to all staff on the wards, with a link to the BCP. Two weeks after the BCP was offered (or at a similar interval after recruitment in the intervention light sites), participants were asked to complete an intermediate questionnaire about beliefs regarding prevention of dermatitis and (at intervention plus sites) participation in and views on the BCP. Questionnaire B was sent by post to participants by the SCIN trial research team along with a business reply envelope. All participants, at both the intervention plus and intervention light sites, were encouraged (orally, through the written advice leaflet and by e-mail reminders sent out by the SCIN trial research team at 4 and 8 months) to attend their local occupational health department at an early stage should they develop hand dermatitis. E-mail reminders were also sent to participants containing a positive reinforcement message to encourage ongoing participation in the study. At the intervention sites, the e-mail reminders also reinforced the BCP.

To account for seasonal variations in the prevalence of dermatitis, the final study data collection tools were administered 12 months after questionnaire A. All participants were asked to have their hands photographed and invited to complete questionnaire C. At the time of the photograph, ICU/SCBU nurses were asked to complete a one-page dermatitis self-assessment form evaluating whether or not they considered themselves to have dermatitis and the extent to which it interfered with work and hobbies. The questionnaires were sent to fieldworkers 2 months before the end of the study by the SCIN trial research team. The fieldworkers gave the ICU/SCBU nurses the questionnaires at recall for hand photographs. Completed questionnaires were returned to the fieldworker or directly (by post) to the SCIN trial research team in London. If it was not possible for an ICU/SCBU nurse to complete the questionnaires at the time, then they were asked to return the completed questionnaire to the SCIN trial team directly using a business reply envelope. At the follow-up time point (i.e. t = 12 months), participants who were unable to attend to have hand photographs taken by the fieldworker could send the SCIN trial research team self-taken photographs of their hands.

All participants who decided to withdraw from the study were asked to complete a shortened version of questionnaire C and were invited to have follow-up hand photographs taken after giving further consent. The collection of these data enabled the researchers to report further on the study’s primary objective measure (i.e. changes in the point prevalence of visible hand dermatitis from baseline).

Extension to the follow-up period

To optimise the response rate at follow-up, the research team felt that it was important to adopt a pragmatic approach that encouraged fieldworkers to try and follow up as many participants as practicable beyond the 12- to 15-month time point. As described in Chapters 3 and 4, an ancillary analysis (sensitivity analysis II) was conducted for the purpose of excluding the sample of participants who were followed up between 12 and 15 months.

Primary outcome

• Difference between intervention plus and intervention light sites in the change in point prevalence of visible hand dermatitis from baseline to the end of follow-up, as assessed by two dermatologists.

Secondary outcomes

• The difference between intervention and control sites in the change in point prevalence of severity of visible hand dermatitis from baseline to the end of follow-up (as ascertained by the dermatologists).

• Days lost from sickness absence and days of modified duties because of hand dermatitis per 100 days of nurse time during the 12 months of follow-up.

• The change from baseline to after completion of the BCP, and to the end of the 12-month follow-up period, in beliefs about dermatitis prevention behaviours.

• The change from baseline to the end of follow-up in the reported frequency of:

  • use of hand rubs for hand cleansing

  • hand-washing with water

  • use of moisturising creams (for student nurses, who had not started clinical attachments at the beginning of the study, this was reduced to differences between the intervention and control sites at the end of the follow-up).

• The change from baseline to the end of follow-up in quality-of-life score.

• The use of moisturiser provided for the intervention (in terms of requests for further supplies by student nurses and orders for supplies of moisturisers by ICUs).

Original sample size calculations

The aim was to recruit at least 40 student nurses and 40 nurses from the ICUs at each trust. To give an indication of power, it was assumed that:

• the expected baseline prevalence of hand dermatitis, overall, was 5% in student nurses and 25% in ICU nurses

• the expected rates, overall, at the control trusts at the end of follow-up were 25% for student nurses and 23% for ICU nurses; these estimated rates allow for limited impact in control trusts

• the expected prevalence rates for individual trusts varied by a multiplying factor, which was normally distributed with a mean of 1.0 and a SD of 0.2

• after allowance for other variables, hand dermatitis at baseline in an individual carried a relative risk of 2.5 for the presence of hand dermatitis at follow-up.

With these assumptions and a 5% level of statistical significance (two-sided), the study would have approximately 89% power to detect a prevalence at follow-up in the intervention trusts of 10% in student nurses and 95% power to detect a prevalence of 10% at follow-up in ICU nurses. For final prevalence rates of 12%, the corresponding power values would be 73% for student nurses and 82% for ICU nurses.

Revised sample size calculations

Fieldworkers were encouraged to recruit as many eligible student nurses and ICU/SCBU nurses as possible, with the expectation that they would recruit at least 40 student nurses and 40 critical care nurses at each site. From the outset it was acknowledged that, as a result of the low workforce populations, a number of smaller sites would not reach this target. To give an indication of power, it was assumed that, at the end of follow-up, the expected prevalence rates overall at the intervention light sites would be 24% in both student and ICU/SCBU nurses, the intraclass correlation coefficient (ICC) would be 0.05 and 20% of participants would be lost to follow-up. With a 5% level of statistical significance (two-sided), the study would have approximately 83% power to detect a reduction in prevalence of dermatitis at follow-up in the intervention plus sites to 10% in student nurses and a 91% power to detect a reduction in prevalence to 10% at follow-up in ICU nurses. For final prevalence rates of 12%, the power values would be 68% for student nurses and 78% for ICU/SCBU nurses, whereas for final prevalence rates of 14%, the corresponding power values would be 51% and 61%, respectively. The power would be higher if the ICC was < 0.05. [These calculations were carried out using the ‘clustersampsi’ command in Stata^® (version 12.1; StataCorp LP, College Station, TX, USA) for difference in proportions.]

Per-protocol analysis

As the primary research question was whether or not the change in the prevalence of hand dermatitis was different between those nurses who were in the intervention plus arm and those nurses in the intervention light arm, in the analysis it was important to consider the possibility that participants in the intervention plus group did not access the intervention at all (intervention access rates described above). If the number of participants who did not access intervention was large, it is likely that the intervention effect was underestimated. For this reason, the main analysis for the primary outcome, and that for health beliefs, was repeated, after restricting the intervention plus group to those nurses who accessed the intervention.

Process evaluation

The intervention process was evaluated by describing the:

• uptake of the intervention, as measured by the proportion of eligible nurses who accessed the BCP (online or paper based), determined by responses in the study questionnaires

• acceptability of the intervention, as measured by the perceived interest, relevance to role and likelihood of recommending it to colleagues.

These data were collected from study questionnaires B and C at intervention plus sites only.

Health economics

Method

Study procedures were followed as for the first stages of the main study at an intervention plus site. In addition, the hands of participants were swabbed to look for the carriage of pathogenic bacteria. As the web version of the BCP was not developed at the time of the feasibility study, a paper version of the BCP was tested. As in the main study, all student nurses who were due to start their first clinical placement, and who had a history of atopic disease or hand dermatitis (as indicated on the responses provided on a standard health questionnaire), were identified by the occupational health department in the course of its routine occupational health assessment and invited to take part. In addition, all full-time ICU nurses (i.e. those working ≥ 30 hours per week) on the ICU unit at the local health board were invited to participate. Those nurses who agreed to participate were asked to complete questionnaire A (with modifications for ICU nurses, as in the main study), and their hands were photographed and swabbed for organisms using the proposed methodology for the main study.

Once the student nurses started clinical work, they were offered the written leaflet Dermatitis: Occupational Aspects of Management. Evidence-based Guidance for Employees and a paper version of the BCP. The nurses were asked to develop implementation plans and were provided with personal supplies of emollients for use during the study period. Once ICU nurses were recruited to the study, optimal equipment for hand cleansing and drying was placed on their ward, along with dispensers of moisturising cream. The written leaflet and paper version of the BCP were then offered to all ICU staff of the ward concerned (i.e. not only those individuals who had consented to take part in the study), and the uptake was documented.

Two weeks after the paper version of the BCP was offered, the nurses were asked to complete questionnaire B about beliefs regarding prevention of dermatitis and participation in, and views on, the BCP.

At the end of the study period, a number of participants recruited to the study were invited to take part in an in-depth telephone semistructured interview designed to seek qualitative feedback on the acceptability and user-friendliness of the methods of data collection and the BCP. The lead occupational health clinician from the trust was also asked for feedback on the ease of implementation of the study procedures, data collection tools and study protocols.

Results of the feasibility study

Thirty-two student nurses were identified as being eligible for the study, 22 (69%) of whom were recruited into the study. Twenty-two (100%) of those nurses recruited returned a completed questionnaire A and 10 out of 22 (45%) nurses returned a completed questionnaire B. Two (20%) of those nurses who returned questionnaire B did not participate in the BCP. One hundred and forty-eight ICU nurses were identified as being eligible for the study (including those not on shift). Twenty-six eligible ICU nurses were on shift on the day of recruitment and all 26 (100%) were recruited into the study. Twenty-two (77%) of those nurses recruited returned a completed questionnaire A and 13 out of 22 (60%) returned a completed questionnaire B. One (8%) of those nurses who returned questionnaire B did not participate in the BCP. The baseline prevalence of hand dermatitis, as diagnosed by the dermatologists, was 5 out of 21 (23.8%) student nurses (one had missing data) and 10 out of 26 (38.5%) ICU nurses. The photographic method for assessing the presence and severity of hand dermatitis was found to be easy to implement by the fieldworker and the dermatologists. However, as one of the dermatologists was found to have a lower threshold for diagnosing hand dermatitis, the dermatologists considered it necessary to develop a set of mini consensus rules that they would follow during the main trial. (One dermatologist tended to diagnose erythema in isolation as evidence of dermatitis and the other diagnosed dermatitis only when erythema was present in association with dryness of the skin.) The development of these mini rules ensured that a consistent and uniform approach was followed by the dermatologists when assessing the hand photographs. Without prior agreement of the rules of diagnosing dermatitis between the dermatologists, the interobserver agreement in the assessment of photographs was 0.5%. Importantly, there was poor correlation between the participants’ self-assessment of the presence of hand dermatitis and between the fieldworkers and the dermatologists. All participants agreed to have their hands swabbed, but only one swab (from a student nurse) was positive for a mild growth of Streptococcus pyogenes (Lancefield group A); the other swabs were negative for the organisms looked for. The paper version of the BCP that was piloted was acceptable in format and appeared to be clearly understood by the study participants.

None of the study participants agreed to be interviewed after the feasibility study, but the fieldworker was interviewed by the trial manager and gave valuable feedback on the practicalities of running the study at a grass-roots level. The fieldworker’s feedback allowed the research team to address ambiguities with respect to several of the data collection tools, the procedure for setting up study material prior to commencement of participant recruitment, the practicalities associated with the taking of hand photographs and for conducting physical inspections of hands.

Changes to the main study protocol as a result of the feasibility study

The feasibility study provided a valuable opportunity to test the study protocol/procedures and to identify and address issues that were likely to arise in the main trial. As a consequence, minor but important amendments, as outlined below, were made to the study protocol to enhance trial processes and procedures, and the study team is confident that these amendments contributed to the overall success of the main trial.

As there was poor agreement in the diagnosis of dermatitis between the fieldworkers, the study participants and the dermatologists, the requirement for fieldworkers to assess whether or not hand dermatitis was present was removed. As a result, the dermatologists’ assessment to report on the primary outcome measure was relied on solely. As a consequence, all participants’ hands were photographed at baseline (when questionnaire A was administered) and 12–15 months after the BCP was delivered in intervention plus sites (and at an equivalent time at the intervention light sites). The trial dermatologists agreed a set of rules by which they would diagnose hand dermatitis from the photographs and assessed a further 100 hand photographs to ensure that they achieved a good agreement between themselves (kappa > 0.6). In addition, an experienced dermatology research nurse was employed to conduct an initial screening assessment of all hand photographs collected during the trial. The swabbing of the participants’ hands yielded little useful information. As a consequence, it was a recommendation from the Trial Steering Committee that the collection of swab samples be removed from the main study. This recommendation was upheld by the National Institute for Health Research Health Technology Assessment programme.

Severity of hand dermatitis

At follow-up, 2 out of 36 (5.6%) student nurses and 1 out of 93 (1.1%) ICU nurses with dermatitis at follow-up had severe dermatitis. Owing to the very low number of cases of dermatitis, it was not possible to conduct further analysis of this secondary outcome.

Days lost

The basic descriptive statistical analysis found that of the 324 out of 337 (96.1%) student nurses who answered the days lost question, only four in the intervention plus arm and four in the intervention light arm reported having taken sick leave because of dermatitis. Likewise, of the 459 out of 483 (95.0%) of ICU nurses who answered the days lost question, only three in the intervention plus arm and four in the intervention light arm reported having taken sick leave because of dermatitis. The sparsity of the data meant that this aspect of the data would not be analysed in regression modelling.

Health beliefs

Table 8 shows descriptive statistics for each of the 25 health belief scores separately for the two arms of the trial and the two types of nurses among those nurses who returned the baseline and intermediate questionnaires (n = 1090). Similar to Table 8, descriptive statistics of health beliefs scores among those nurses who returned the baseline and 12-month follow-up questionnaires are presented in Table 9. Figure 7 shows a graphical illustration of the change in health beliefs from baseline (x-axis) to the 12-month follow-up (y-axis) separately for those in the intervention light arm (green dots) and intervention plus arm (blue dots), and for student (Figure 7a) and ICU/SCBU nurses (Figure 7b). Among student nurses, those in the intervention light arm had health beliefs scores that were very similar at both baseline and follow-up (green dots lying on the line of equality). In contrast, student nurses in the intervention plus arm had higher health belief scores at follow-up than at baseline (blue dots lying on the left side of the line of equality), indicating a shift towards more positive health beliefs. For ICU/SCBU nurses, the change in health beliefs from baseline to follow-up was smaller and parallel for both nurses in the intervention light and those in the intervention plus arm (both green and blue dots lying to the left side of, but very close to, the line of equality).

FIGURE 7.

Scatterplot of health belief scores assessed at follow-up against health belief scores assessed at baseline separately for (a) student and (b) ICU/SCBU nurses and for the two arms of the trial. The dashed line represents the line of equality.

The change in health beliefs was assessed further through regression modelling before and after adjusting for age, sex and follow-up time. In regression modelling, both changes from baseline to after accessing the online BCP and from baseline to the 12-month follow-up were explored. As for the primary outcome (i.e. objectively assessed dermatitis), the change in each health belief score was modelled by using the score at follow-up as the outcome variable, adjusting for the corresponding score at baseline. Effect estimates summarised by regression coefficients and their corresponding 95% CIs are presented in Table 10. A regression coefficient of value β indicates that those nurses in the intervention plus arm changed their average score in the given health belief variable by β units from baseline to follow-up compared with those nurses in the intervention light arm. A graphical display of the effect of intervention on change in health belief scores from baseline to the 12-month follow-up for each type of nurse is shown in Figure 8. In Figure 8, for each of the two graphs (for student nurses and ICU/SCBU nurses), regression coefficients to the left side of the vertical line (regression coefficient 0 = 0) indicate higher scores in the intervention light arm than in the intervention plus arm by a value of β. Regression coefficients to the right side of the vertical line indicate higher scores in the intervention plus arm than in the intervention light arm by a value of β. For student nurses, change was more positive for those in the intervention plus arm than for those in the intervention light arm for almost all health beliefs, except for action plans about consulting occupational health services if symptoms occur. However, almost none of those changes in health belief scores was significantly different between the two intervention arms. An exception to this was change in PBC for use of hand rubs and in intentions to wash hands only when appropriate; both changes were significantly more positive for those nurses in the intervention plus arm than for those in the intervention light arm. Similar to student nurses, ICU/SCBU nurses in the intervention plus arm had a more positive change than ICU/SCBU nurses in the intervention light arm in most of the health beliefs assessed. However, almost none of the changes was statistically different between the two trial arms. The only exception was the change in action plans for use of hand cream, for which the change from baseline to the 12-month follow-up was more positive for nurses in the intervention plus arm than for those in the intervention light arm.

FIGURE 8.

Associations between change in health beliefs scores from baseline to the 12-month follow-up and intervention group after adjusting for age, sex and follow-up time. For each of the two graphs (for student nurses and ICU/SCBU nurses), regression coefficients to the left side of the vertical line (regression coefficient = 0) indicate higher scores in the intervention light arm, and regression coefficients to the right side of the vertical line indicate higher scores in the intervention plus arm.

Health behaviours

Five health behaviours were assessed in the trial (frequency of hand-washing with soap and water, use of hand rubs, use of moisturising creams before shifts, use of moisturising creams during shifts and use of moisturising creams after shifts). These were assessed at both baseline and the 12-month follow-up for ICU/SCBU nurses and only at the 12-month follow-up for student nurses, as they had not started their placement yet.

Considering the ordinal responses to these questions as a continuous measure, the average scores for each of the five measures assessed at follow-up (and their corresponding 95% CIs) are illustrated in Figure 9. For most of the protective measures, frequency of use of hand rubs and hand creams at the 12-month follow-up was higher for nurses in the intervention plus trial arm than for those in the intervention light arm. Differences were more marked for frequency of use of moisturising creams before, during and after shifts among ICU/SCBU nurses. Estimates from the regression analysis (Table 11) confirmed this observation (the proportionality of odds assumption of the model was tested). However, the differences in frequency of use of hand-protective measures between those nurses in the intervention plus and light arms were mostly non-significant in the adjusted models. The only exception was the difference between the two trial arms in the frequency of use of moisturising creams during shifts among ICU/SCBU nurses, which was statistically higher in the intervention plus arm. The composite score of frequency of use of hand cream (before, during and after shifts) was statistically higher in the intervention plus arm.

FIGURE 9.

Mean scores and 95% CIs for use of dermatitis preventative measures at the 12-month follow-up for (a) student nurses and (b) ICU nurses.

Quality of life

See Health economics.

Use of moisturising (in terms of requests for further supplies by student nurses and orders for supplies of moisturisers by intensive care units)

Because of logistical difficulties, it was not possible to reliably collect data on orders for supplies of moisturisers by ICUs, that is supplies of moisturisers were replenished from the main hospital stock rather than being ordered separately by ICUs. As an alternative, monthly inspections were conducted of all wall dispensers available on the ICUs and crude measures were recorded of the quantity of moisturising cream, as seen through the clear display field on each dispenser. However, at the end of the study it was agreed that this was not a robust method for taking reliable measures because it was not possible to ascertain how many times each dispenser may have been refilled between inspections and the dispensers were used by all ward staff and not just by those nurses taking part in the study.

Among student nurses in the intervention plus arm, only 10 nurses from four sites returned completed forms with information regarding requests for an extra supply of moisturising creams. The overall quantity requested was 33 tubes in total. As the reason for the additional supply request, the nurses reported that the existing supply had been used.

Per-protocol analysis

The main analysis presented above is the ITT analysis. However, a large proportion of the nurses in the intervention plus arm reported that they did not access the BCP intervention, although they may have accessed other components of the intervention (e.g. provision of moisturising creams). The per-protocol analysis is based on those nurses who reported that they did not access the BCP. The number of participants who did not access the BCP intervention are shown by categories of baseline characteristics in Table 12. Among student nurses, fewer of the females accessed the BCP intervention than those who did not (38% vs. 61%). Nurses who did not access the BCP were slightly younger and were more likely to have atopic tendency and self-reported atopic dermatitis at baseline. Among ICU/SCBU nurses, all baseline characteristics were well balanced between those who accessed the BCP intervention and those who did not. For both student and ICU/SCBU nurses, the outcome variables, including the primary outcome and health beliefs about action plans, were very similar between those nurses who accessed the BCP intervention and those who did not.

The trial participants reported the main reasons for not accessing the BCP intervention. The predominant reason for student nurses was that they forgot (55%), whereas for ICU/SCBU nurses it was a lack of time (40%). Many (student and ICU/SCBU) nurses also reported that they wanted to access the BCP intervention but did not get around to it (23% for both student and ICU nurses), whereas a few nurses also reported that they were not interested (10% among student and 13% among ICU nurses).

Primary outcome

When the analysis was repeated after excluding those nurses who reported that they did not access the BCP intervention, the change observed in objectively assessed dermatitis from baseline to the 12-month follow-up was more pronounced. As shown in Figure 10, dermatitis among student nurses increased from 8.1% at baseline to 11.4% at follow-up in the intervention light arm, but it decreased from 17.1% at baseline to 3.9% at follow-up in the intervention plus arm. Among ICU/SCBU nurses, dermatitis prevalence decreased from 16.1% at baseline to 13.6% at follow-up in the intervention light arm, whereas in the intervention plus arm it decreased from 16.4% at baseline to 10.2% at follow-up. The greater difference between the two intervention arms, especially for student nurses, after excluding participants who reported not having accessed the BCP intervention, is also depicted in the effect estimates from regression analysis shown in Table 13.

FIGURE 10.

Prevalence of objectively assessed dermatitis by intervention arm and time point when dermatitis was assessed (baseline and follow-up), for (a) student and (b) ICU/SCBU nurses after excluding participants who reported not having accessed the BCP intervention.

Sensitivity analysis I (after excluding self-taken photographs)

Among student nurses, the prevalence of dermatitis increased from baseline to follow-up among those in the intervention light arm, but it decreased among those nurses in the intervention plus arm (Figure 11). Among ICU/SCBU nurses, prevalence of dermatitis decreased from baseline to follow-up for both arms. However, the decrease was somewhat larger among those nurses in the intervention plus arm. Regression analysis (Table 14) confirmed this observation, with effect estimates being stronger for student nurses and closer to the null value of 1 (OR = 1) for ICU/SCBU nurses. None of the calculations reached the threshold of statistical significance.

FIGURE 11.

Prevalence of objectively assessed dermatitis by intervention arm and time point when dermatitis was assessed (baseline and follow-up), for (a) student and (b) ICU/SCBU nurses, after excluding self-taken photographs.

Sensitivity analysis II (after restricting the sample to those who were followed up between 12 and 15 months)

Among student nurses, the prevalence of dermatitis increased from baseline to follow-up for both the intervention light and intervention plus arms (Figure 12). However, the increase in prevalence of dermatitis was smaller in the intervention plus group. Among ICU/SCBU nurses, prevalence of dermatitis decreased from baseline to follow-up for both trial arms, but the decrease was larger in the intervention plus arm. In the fully adjusted model, effect estimates (Table 15) were stronger for ICU/SCBU nurses than for student nurses. None of the calculations reached the threshold of statistical significance.

FIGURE 12.

Prevalence of objectively assessed dermatitis by intervention arm and time point when dermatitis was assessed (baseline and follow-up), for (a) student and (b) ICU/SCBU nurses, after restricting the sample to those who were followed up between 12 and 15 months.

Sensitivity analysis III (after excluding University Hospital Southampton NHS Foundation Trust and Guy’s and St Thomas’ NHS Foundation Trust)

The University Hospital Southampton NHS Foundation Trust and Guy’s and St Thomas’ NHS Foundation Trust achieved a very high rate of recruitment among the student nurse population. Both these sites were in the intervention plus arm of the trial and this resulted in unequal numbers of student nurse participants in the intervention plus and intervention light sites.

In student nurses, the prevalence of dermatitis increased from baseline to follow-up in those in the intervention light arm, and it decreased considerably in those in the intervention plus arm (Figure 13). In ICU/SCBU nurses, prevalence of dermatitis decreased from baseline to follow-up, but decrease was larger in those in the intervention plus arm. Fully adjusted effect estimates for both student and ICU/SCBU nurses were 1.2 (Table 16). None of the associations reached statistical significance.

FIGURE 13.

Prevalence of objectively assessed dermatitis by intervention arm and time point when dermatitis was assessed (baseline and follow-up), for (a) student and (b) ICU/SCBU nurses, after excluding University Hospital Southampton NHS Foundation Trust and Guy’s and St Thomas’ NHS Foundation Trust.

As anticipated, the number of recruited nurses was considerably higher in some trusts than in other trusts. To explore the possibility that clusters of sites with higher numbers of recruited nurses were different from those with lower numbers of recruited nurses in terms of health beliefs and behaviours, correlation coefficients between size of the cluster and mean scores of health beliefs and behaviours were calculated (Table 17). Correlation coefficients between size of the cluster and health beliefs and behaviours were predominantly negative, suggesting that the larger the size of the cluster, the more negative the health beliefs and the lower the frequency of use of dermatitis prevention measures. In most cases, the correlation coefficients were weak and not statistically significant. The exception was the association between size of the cluster and attitude towards appropriate hand-washing in student nurses, cluster size and subjective norm for use of hand rub and subjective norm for appropriate glove use in ICU/SCBU nurses.

Analysis of total health-care costs

At baseline, the total mean health-care costs for student nurses were £184 in the intervention plus arm and £157 in the intervention light arm. The difference was not statistically significant (bootstrapped 95% CI –£4 to £58). For ICU/SCBU nurses, the costs were £187 and £179 for the intervention plus and intervention light arms, respectively. This difference was not statistically significant (bootstrapped 95% CI –£35 to £55).

At follow-up, the mean total health-care costs for the student nurses were £162 for the intervention plus arm and £159 for the intervention light arm. After adjusting for baseline, it was found that the intervention plus arm had mean costs that were £2 less than for the intervention light arm, a difference that was not statistically significant (bootstrapped 95% CI –£44 to £45). The mean costs for the ICU/SCBU nurses were £142 for the intervention plus arm and £128 for the intervention light arm. The intervention plus arm had costs that were, on average, £3 higher after adjusting for baseline, which again was not statistically significant (bootstrapped 95% CI –£31 to £34).

When only cases for which there were QALY data were included, it was found that the student nurses in the intervention plus arm still had costs that were £2 less than for the intervention light arm (bootstrapped 95% CI –£44 to £49) and the ICU/SCBU nurses in the intervention plus arm had costs that were £4 higher than for the intervention light arm (bootstrapped 95% CI –£23 to £38).

Analysis of total societal costs

At baseline, the student nurses in the intervention plus arm had mean societal costs of £700, whereas the intervention light arm had costs of £918. The difference of £218 was not statistically significant (bootstrapped 95% CI –£160 to £649). The ICU/SCBU nurses in the intervention plus arm had mean societal costs of £2312 compared with £2013 in the intervention light arm, a difference of £298 (bootstrapped 95% CI –£637 to £1173).

During the follow-up period, the societal costs in the student nurse group were, on average, £477 for the intervention plus arm and £494 for the intervention light arm. After adjustment for baseline, the intervention plus arm was shown to have costs £32 higher than for the intervention light arm (bootstrapped 95% CI –£159 to £234). The ICU/SCBU nurses in the intervention plus arm had mean societal costs at follow-up of £1723, compared with £1489 for the intervention light arm. The intervention plus arm costs were £250 higher after adjustment (bootstrapped 95% CI –£484 to £943).

When only cases for which QALY data were included, it was found that the student nurses in the intervention plus arm still had costs that were £33 more than for the intervention light group (bootstrapped 95% CI –£156 to £231) and the ICU/SCBU nurses in the intervention plus arm had costs that were £233 higher than for the intervention light arm (bootstrapped 95% CI –£417 to £922).

Quality-adjusted life-years

The baseline and follow-up EQ-5D-5L scores are shown in Table 21. There was virtually no difference in the EQ-5D-5L scores between trial arms at baseline or at follow-up or for QALYs over the follow-up period. Adjusting for baseline EQ-5D-5L scores among student nurses in the intervention plus arm resulted in just 0.0002 more QALYs (bootstrapped 95% CI –0.0054 to 0.0062 QALYs). Among ICU/SCBU nurses, the intervention plus arm had 0.0015 fewer QALYs (bootstrapped 95% CI –0.0018 to 0.0049 QALYs).

For student nurses for whom baseline and follow-up costs were available, the intervention plus arm had 0.00002 more QALYs (bootstrapped 95% CI –0.0051 to 0.0055 QALYs). For ICU/SCBU nurses, the intervention plus arm had 0.0016 fewer QALYs (bootstrapped 95% CI –0.0018 to 0.0046 QALYs).

Cost-effectiveness analysis

Using complete cases (baseline and follow-up costs and QALYs available), the above results show that, from a health-care perspective, the intervention plus arm for student nurses dominated the intervention light arm in that it resulted in lower mean costs (by £2) and more QALYs (0.00002). For ICU/SCBU nurses, the intervention plus arm was dominated – it had higher costs (by £4) and fewer QALYs (0.0016) than the intervention light arm. The ICERs are, consequently, both negative. However, the CIs around the cost and QALY differences are very wide. Figures 14 and 17 show the uncertainty around these estimates. Each point on the scatterplots represents a cost and QALY difference obtained from 1 of 1000 bootstrapped resamples. The spread of replications in Figure 15 means that there is a 31% likelihood that intervention plus is cost-saving and outcome-improving for student nurses, 23% likelihood of lower costs and worse outcomes, 23% likelihood of worse outcomes and higher costs and 20% likelihood of higher costs and better outcomes. The corresponding cost-effectiveness acceptability curve shows that at a threshold of £20,000 per QALY there is a 53% likelihood that intervention plus was cost-effective (Figure 15).

FIGURE 14.

Cost-effectiveness plane for student nurses (health-care perspective).

FIGURE 15.

Cost-effectiveness acceptability curve for student nurses.

Figure 16 shows that, for ICU/SCBU nurses, intervention plus has a probability of 10% of resulting in lower costs and better outcomes, a 29% probability of resulting in lower costs and worse outcomes, a 53% probability of resulting in worse outcomes and higher costs and an 8% probability of resulting in higher costs and better outcomes. At a threshold of £20,000 per QALY, there is a 19% likelihood that intervention plus was cost-effective (Figure 17).

FIGURE 16.

Cost-effectiveness plane for ICU/SCBU nurses.

FIGURE 17.

Cost-effectiveness acceptability curve for ICU/SCBU nurses.

As stated earlier, societal cost differences were higher for intervention plus by £33 for student nurses and £233 for ICU/SCBU nurses. For student nurses, intervention plus has an ICER of £1.65M (£33 divided by 0.00002), whereas for ICU/SCBU nurses, intervention plus is dominated by intervention light.

Elsewhere it has been observed that intervention plus, relative to intervention light, resulted in a smaller proportion of nurses having dermatitis. The incremental percentage point difference between baseline and follow-up was 8.1 for student nurses and 3.4 for ICU/SCBU nurses. The intervention plus arm was therefore ‘dominant’ for student nurses (i.e. lower costs and a higher proportion of nurses free of dermatitis than with intervention light). For the ICU/SCBU nurses, the average extra cost for one more nurse to avoid dermatitis was £3 (the incremental cost) divided by 0.034 (the incremental outcome), resulting in a cost of £88 per case averted.

Sensitivity analyses

When intervention plus arm participants from the student nurse group who declared that they had not accessed the intervention were assigned a zero time cost, the mean health-care cost savings increased from £2 to £5. This makes intervention plus more likely to be cost-effective. For ICU/SCBU nurses, intervention plus resulted in costs that were £2 higher than for intervention light (compared with £3 for the base-case analysis).

Publications (up to the time of publication of the final report)

Parsons V, Madan I. From vision to action part 1: building an evidence-base in occupational health. Occup Health Work 2015;12(4).

Parsons V, Madan I. Developing a successful national occupational health research proposal. Part 2. Occup Health Work 2015;12(5).

Madan I, Parsons V, Cookson B, English J, Lavender T, McCrone P, et al. A behavioural change package to prevent hand dermatitis in nurses working in the National Health Service. The SCIN trial: study protocol for a cluster randomised controlled trial. Trials 2016;17:145.

Parsons V, Madan I. Implementing a national OH clinical trial. Part 3: the national SCIN trial – roll out and lessons learned. Occup Health Work 2016;13(5).

Weston D, Parsons V, Ntani G, Rushton L, Madan I. Mixed contact methods to improve response to a postal questionnaire. Occup Med 2017;67:305–7.

Parsons V, Williams H, English J, Llewellyn J, Ntani G, Madan I. The development of a protocol for diagnosing hand dermatitis from photographic images. Contact Derm 2018;79:270–5.

Conference proceedings (up to the time of publication of the final report)

Parsons V, Coggon D, Cookson B, English J, Lavender T, McCrone P, et al. The SCIN (Skin Care Intervention in Nurses) Trial. A Cross-sectional Feasibility Study in Wales (Stage 1). Poster presentation at the 8th Georg Rajka International Symposium on Atopic Dermatitis, Nottingham, UK, 22 May 2014.

Parsons V, Din S, Wright A, Coggon D, Ntani G, Williams H, et al. The SCIN (Skin Care Intervention in Nurses) Trial. A Cross-sectional Feasibility Study in Wales. Poster presentation at the Society of Occupational Medicine Scientific Meeting, London, UK, 28 May 2014.

Madan I, Parsons V, Cookson B, English J, Lavender T, McCrone P, et al. The SCIN (Skin Care Intervention in Nurses) Trial. A Cluster Randomised Trial. Oral presentation at the Epidemiology in Occupational Health (EPICOH) Conference, Barcelona, Spain, 5 September 2016.

Weston D, Parsons V, Madan I. The Use of Mixed Methods of Contact to Improve Response rate to a Postal Questionnaire. Poster presentation at the Epidemiology in Occupational Health (EPICOH) Conference, Barcelona, Spain, 5 September 2016.

Madan I, Parsons V, Cookson B, English J, Lavender T, McCrone P, et al. The SCIN (Skin Care Intervention in Nurses) Trial. A Cluster Randomised Trial. Poster presentation at the 13th Congress European Society of Contact Dermatitis (ESCD) Conference, Manchester, UK, 15 September 2016.

Madan I, Parsons V. Update on the Conduct of the SCIN Trial. Oral presentation at the Eczema and Contact Dermatitis: An Evidence-based Update Meeting, Nottingham, UK, 17 May 2017.

Madan I, Parsons V. Development of the Hand Photography Protocol. Oral and poster presentation at the StanDerm workshop, Berlin, Germany, 18 May 2017.

Madan V, Parsons V, Ntani G, Harling C, McCrone P, Brough C, et al. Presentation of the SCIN Trial Results. Paper presented at the SCIN trial dissemination conference, London, UK, 12 July 2017.

Madan I, Parsons V, Cookson B, English J, Lavender T, McCrone P, et al. SCIN Trial (Skin Care Intervention in Nurses). Dissemination of results. Oral presentation at the 32nd Triennial Congress of the International Commission on Occupational Health (ICOH), Dublin, Ireland, 29 April–4 May 2018.