Home-based narrowband UVB, topical corticosteroid or combination for children and adults with vitiligo: HI-Light Vitiligo three-arm RCT

Importance of the topic to patients and health-care practitioners

A James Lind Alliance Priority Setting Partnership identified priority topics for future vitiligo research, which were important to patients and health-care practitioners. 21 The Home Interventions and Light therapy for the treatment of vitiligo (HI-Light Vitiligo) trial has been designed to address two of the priority topics:

1. Which treatment is more effective for vitiligo: steroid creams/ointments or light therapy?

2. How effective is ultraviolet B light (UVB) therapy when combined with creams or ointments in treating vitiligo?

The Priority Setting Partnership also highlighted the importance of testing vitiligo treatments in children; so the HI-Light Vitiligo trial recruited both children and adults.

Informed consent

Written informed consent was obtained from each participant (or parent/carer in the case of children) prior to any trial procedures being carried out. Children provided assent as well if they wished to. Separate written consent was obtained for participation in the process evaluation, supported by a separate age-appropriate information sheet.

Topical therapy

Light therapy

Primary outcome

The primary outcome was participant-reported treatment success at 9 months.

The primary outcome was assessed for each participant at 9 months (at the end of the treatment phase) at the target patch. Treatment success was defined as the participant reporting that their vitiligo was either ‘a lot less noticeable’ or ‘no longer noticeable’ in response to the question ‘Compared with the start of the study, how noticeable is the vitiligo now?’, using the previously validated VNS. 42

Secondary outcomes

1. VNS treatment success by blinded review of digital images at 9 months.

   Assessed at 9 months at the target patch by three independent patient reviewers using digital images of the trial participants and using the same question for the primary outcome. Treatment success was derived from the score given by the majority of the three blinded reviewers.

2. Participant-reported treatment success by body region.

   Assessed at 9 months, measured using the VNS and analysed by body region (A, B and C). Each participant assessed up to three patches from three different body regions, including the one chosen as the target patch. During the no-treatment follow-up phase, the same question was used at 12, 15, 18 and 21 months to assess long-term patient-reported noticeability for each body region.

3. Onset of treatment response.

   Investigator-assessed onset of treatment response (including cessation of spread) for the target patch. To be assessed at 3, 6 and 9 months using the following question – ‘Compared with the start of the study, has there been a change in the vitiligo patch?’

   • Stayed the same (not worsened).

   • Improved.

   • Got worse.

   A treatment response was considered to have occurred if the response given was ‘stayed the same’ or ‘improved’. Analyses for this secondary outcome used investigator-assessed responses because they were more likely to remain unblinded than the participants.

1. Maintenance of treatment response.

   Participant-assessed maintenance of treatment response (including cessation of spread) for the target patch. This was assessed at 12, 15, 18 and 21 months to assess long-term patient-reported noticeability using the following question – ‘Compared to since you stopped using the study treatments, has there been a change in the vitiligo patch?’

   • Improved.

   • Stayed the same.

   • Got worse.

   Loss of maintenance of treatment response was defined as ‘got worse’.

2. Percentage repigmentation at 9 months.

   Percentage repigmentation was assessed at 9 months by a blinded independent dermatologist using digital images taken at baseline and at 9 months for the target patch. Investigator assessment of percentage repigmentation was also conducted at 3, 6 and 9 months.

3. Quality of life at the end of treatment (9 months) and at the end of follow-up (21 months). Assessments included:

   • Vitiligo-Specific Health-Related Quality-of-Life Instrument (VitiQOL)43 for adults (aged ≥ 18 years).

   • Skindex-2944 for adults (aged ≥ 18 years).

   • EuroQol 5-Dimensions, five-level version (EQ-5D-5L) for those aged ≥ 11 years. 45,46

   • Child Health Utility 9D (CHU-9D)47 for children aged ≤ 17 years.

4. Time burden of treatment: time per session for active light treatment and participant-reported treatment burden for TCS and light treatments during the treatment phase.

Safety outcomes

The safety end points are the number of adverse reactions during the treatment phase.

Participants were asked to record any AEs in their treatment diary and were also asked at the 3-, 6- and 9-month clinic visits about any AEs that they had experienced. Any AEs that were deemed related to trial treatments (adverse reactions) were reported in the CRF. Erythema (redness) of grade 1 or 2 was not considered an AE, as this is an expected treatment response from use of NB-UVB. All serious adverse events (SAEs) were reported directly to the trial co-ordinating centre and assessed for seriousness, expectedness and causality by the chief investigator, or delegated medical monitor. SAEs were recorded and reported to the Medicines Health Regulatory Authority and Research Ethics Committee as part of the annual reports.

Cost-effectiveness analysis

The within-trial economic evaluation estimates the incremental cost-effectiveness from an NHS perspective of:

• NB-UVB light therapy (plus placebo ointment) compared with TCS (plus dummy light).

• Combination of NB-UVB light therapy and TCS compared with TCS (plus dummy light).

The economic analysis uses individual participant-level data from the trial. The base-case analysis undertakes a cost-effectiveness analysis from an NHS perspective for all participants. Secondary analyses consider the cost–utility of the comparators of interest for those with EQ-5D-5L data available (participants aged ≥ 11 years) and separately for those with CHU-9D data available (participants aged 5 to ≤ 17 years). Full details of the methods can be found in Chapter 4, Methods.

Data collection

Trial data were entered into a web-based electronic case report form (eCRF) (MACRO 4.2.1 version 3800, Elsevier, London, UK). Staff at research sites had access to the data from their site only, with access controlled through person-specific login credentials. Access to the trial database and database maintenance was managed by NCTU.

To facilitate the data collection process, site staff members were provided with CRF workbooks that mirrored the data required for the eCRF. Investigators were asked to transcribe the data into the eCRF within 7 days of the data being collected, when possible.

Participants used a trial handbook, which included a detailed treatment diary, AE record, the use of any health-care resources and any prescribed medicines. Site staff reviewed these handbooks at the 3-, 6- and 9-month clinic visits and entered summary data into the eCRF.

The primary outcome was collected at the 9-month clinic visit. For those who did not attend this visit and who had not withdrawn from the trial, primary outcome data were obtained via telephone, post or text message, when possible.

After the treatment period (9 months), follow-up continued for a further 12 months, with participant-completed questionnaires at 12, 15, 18 and 21 months. These questionnaires were sent either by post with the data entered and returned on paper, or via e-mail using electronic questionnaires designed by staff at NCTU. Reminders were sent (via e-mail or post) if the questionnaire remained uncompleted after 2 weeks, and again after 3 weeks. Members of NCTU staff chased up outstanding questionnaires after 3 weeks by telephone.

Primary outcome

The number and percentage of participants achieving ‘treatment success’ (defined as a response of either ‘a lot less noticeable’ or ‘no longer noticeable’ in response to the question ‘Compared with the start of the study, how noticeable is the vitiligo now?’) is reported for each treatment group at 9 months post randomisation.

The primary analysis was performed on the intention-to-treat (ITT) analysis set, where multiple imputation was used to account for missing primary outcome data at 9 months. Prior to primary analysis, baseline characteristics were summarised further by treatment arms and the availability of primary outcome at 9 months to check the missing at random assumption of multiple imputation.

Randomised groups were compared using a mixed-effects model for binary outcome adjusted by recruitment centre, body region of the target patch and age at randomisation (continuous). The primary effectiveness parameter comparing NB-UVB light with TCS alone, and NB-UVB light plus TCS with TCS alone, was the risk difference (risk ratio will also be included) in the percentage of participants achieving treatment success at 9 months along with 95% confidence interval (CI) and exact p-value. By default, risk differences are reported because these estimates are more clinically intuitive for binary outcomes. However, where models estimating risk difference do not converge, odds ratios will be reported instead of risk differences.

Sensitivity analyses were conducted to (1) adjust for any variables with imbalance at baseline, (2) repeat primary analysis based on participants whose primary outcome was available at 9 months and (3) investigate the effects of treatment adherence.

Planned subgroup analyses were (1) children versus adults, (2) body region of the target vitiligo patch, (3) hypomelanotic patch (an indicator of disease activity): definitely or maybe versus no and (4) ≥ 4 years duration of vitiligo versus < 4 years. These analyses were conducted by inclusion of appropriate interaction terms in the regression model and were considered as exploratory. An additional post hoc subgroup analysis explored the impact that skin type (types I–III vs. types IV–VI) had on the results.

Secondary outcomes

1. VNS treatment success by blinded review of digital images at 9 months.

   Between-group comparisons were performed using a mixed-effect regression model for binary outcome, adjusting by recruitment centre, body region of target patch and age (continuous). The analysis was performed on a modified ITT set, where no imputation of missing data was required.

2. Participant-reported treatment success by body region (at 9 months).

   VNS treatment success at 9 months for all assessed patches (up to three) was analysed using a multilevel mixed-effects model, accounting for potential correlation between treatment effects at different body regions in the same person. This analysis was conducted with multiple imputation of missing treatment success data. Patient-reported treatment success by body region at 3 and 6 months is presented descriptively.

3. Onset of treatment response (during treatment phase).

   Summary data of all three categories (stayed the same, improved, got worse) is presented by treatment group and by timeline (3, 6 and 9 months). The cumulative percentage of participants who achieved a treatment response (i.e. stayed the same or improved) at target patch is presented. Analysis of treatment response at 9 months was analysed using a mixed-effect regression model for binary outcome, adjusting by recruitment centre, body region of target patch and age (continuous).

   Participant-reported onset of treatment response is summarised as for investigator-assessed treatment response.

4. Maintenance of treatment response (during follow-up phase).

   Maintenance of treatment response is presented separately for those who achieved and those who did not achieve treatment response at the end of the treatment phase. The cumulative percentage of participants with loss of maintenance of treatment response is presented by treatment arm. Data are reported for the target patch only.

5. Percentage repigmentation at 9 months (by blinded dermatologist and investigator).

   Analysis of blinded dermatologist-assessed percentage repigmentation at 9 months was analysed using a mixed-effect regression model for binary outcome, adjusting by recruitment centre, body region of target patch and age (continuous). Where available, data from investigator assessments at 9 months were used for missing data based on blinded clinician assessment of digital images.

   Treatment success based on investigator-assessed percentage repigmentation at 9 months is reported descriptively.

   Assessments carried out by investigators at 3 and 6 and 9 months are presented descriptively.

6. Quality of life at the end of treatment (9 months) and at the end of follow-up (21 months).

   Total scores for VitiQOL, Skindex-29, CHU-9D and EQ-5D questionnaires at 9 months and 21 months are summarised by treatment arm using appropriate summary statistics.

7. Time burden of treatment.

   For active light therapy, the average time per treatment session was estimated using data collected at 3, 6 and 9 months. Time burden of TCS application was assumed to be minimal. The percentage of those who reported difficulties with the interventions are summarised, along with a description of the difficulties experienced.

Vitiligo Noticeability Scale treatment success from blinded patient and public involvement reviewers

Treatment success from blinded image assessment by patient reviewers were broadly consistent with the primary analysis but were more likely to suggest that there was a benefit from NB-UVB, with evidence of significant differences in treatment success for both the NB-UVB and the combination groups, compared with TCS (Table 11).

Participant-reported Vitiligo Noticeability Scale treatment success by region of the body (including all assessed patches)

Patches on the hands and feet were less likely to respond to treatment than patches on other parts of the body, regardless of the treatments being used. However, the between-group comparisons given in Table 12 indicate that there was no evidence of a differential treatment effect according to the location of assessed patches (see Table 12 and Figure 5). Participant-reported VNS at 3 and 6 months by body region is summarised in Appendix 3.

FIGURE 5.

Treatment success at all assessed patches at 9 months.

Overall, 94% of participants had achieved onset of treatment response by 3 months for all groups [defined as the active target patch having improved or stayed the same (i.e. not worsened) as assessed by investigators] (Figure 6): TCS (40% improved, 57% stayed the same), NB-UVB (61% improved, 35% stayed the same) and combination (60% improved, 38% stayed the same).

FIGURE 6.

Investigator assessed onset of treatment response.

Participant-reported onset of treatment response is summarised in Appendix 4.

Treatment success: percentage repigmentation

Percentage repigmentation was assessed by a dermatologist using digital images taken at baseline and at 9 months. The results were supportive of the primary outcome, although the rates of treatment success were lower: 3% (4/115) for the TCS group, 8% (9/116) for the NB-UVB group and 15% (18/120) for the combination group. The adjusted odds ratio was 2.22 (95% CI 0.66 to 7.51) for NB-UVB compared with TCS, and was 4.62 (95% CI 1.50 to 14.24) for combination compared with TCS (Table 13). Review by blinded investigators during clinic visits were also supportive of the primary outcome (Table 13). Full details of repigmentation rates at all time points are summarised in Appendix 5.

Long-term follow-up (post intervention)

Long-term follow-up rates at 12, 15, 18 and 21 months were 56%, 52%, 47% and 43%, respectively, and so the results are presented descriptively.

By 21 months (12 months after stopping treatment), just over 40% (149/338) of participants reported that repigmentation had been lost (Table 14). These percentages were similar for those who achieved ‘treatment success’ at 9 months (Table 15).

Participant-reported Vitiligo Noticeability Scale throughout the trial (treatment and follow-up)

The VNS scores throughout the study treatment period (0 to 9 months) and follow-up period (12 to 21 months) are shown in Figure 7. The number included at each time point varies according to follow-up completion rates, but shows treatment success to be achieved by 6 months in the combination group and maintained for approximately 3 months before loss of gained pigmentation in the longer term.

FIGURE 7.

Percentage of participants reporting treatment success for target patch during the trial (VNS treatment success for target patch over time).

Quality of life

There was no difference between the groups in any of the generic or vitiligo-specific quality-of-life instruments at any time point (Table 16).

Adverse and serious adverse events

Identification of resources

In keeping with the chosen perspective, the base case captured the intervention costs (including any side-effect costs) to the NHS and the participant’s wider use of the NHS (including health-care visits and prescriptions) as a result of vitiligo. Participants’ personal out-of-pocket expenses incurred as a result of their vitiligo were also captured in a separate analysis taking a broader perspective. The time spent by patients administering the interventions is presented descriptively in Chapter 3, Adherence to trial treatment and treatment burden, but participant time burden administering treatment was not costed.

Measurement of resource use data

Resource use for the intervention phase was collected at 3, 6 and 9 months, using information recorded by participants in daily diaries and in CRFs collected at follow-up visits. In the follow-up period, resource use was collected via online participant questionnaires at 12, 15, 18 and 21 months (or via paper copies if preferred).

Valuation of resource use data

The cost of the intervention was estimated at the individual level as follows.

Unit costs

All resource use relevant to the NHS perspective, including wider NHS usage due to vitiligo, was valued using UK unit costs [in Great British pounds (GBP)] for the 2017 price year (the most recent price year available at the start of the analysis). Unit costs were identified from published sources, such as Unit Costs of Health and Social Care,61 Prescription Cost Analysis58 and NHS Reference Costs 2017. 60 A table of unit costs, together with their sources, is presented in Results.

Personal costs incurred by participants as out-of-pockets costs due to their vitiligo were valued using patient-reported estimates. These were not adjusted to reflect the year in which they were incurred because timing is likely to have had a negligible effect on price for the types of items reported [for instance, the majority of items were sun creams, emollients or camouflage products that are (or are similar to products) also available on prescription and the net ingredient cost (NIC) per item in the prescription cost analysis barely changed between 2016 and 2017 (from £8.34 in 2016 to £8.29 in 201757,62)].

Total costs

The cost of all reported resource use (relevant to an NHS perspective) was calculated for each participant. These figures were then summed for each participant, giving a total cost over the 9-month treatment period in the primary analysis. For each of the different intervention arms, a mean cost per participant was estimated.

Vitiligo Noticeability Scale

The primary clinical outcome measure in the HI-Light Vitiligo trial is participant-reported treatment success, measured at 9 months, using the VNS. 42 Treatment success, a binary outcome, is defined by whether the participant responds that their target vitiligo patch is ‘a lot less noticeable’ or ‘no longer noticeable’ in response to the question: ‘Compared with the start of the study, how noticeable is the vitiligo now?’. To the best of our knowledge, no previous studies have compared the treatments being compared in this study, hence the use of single study-based estimates of effectiveness.

Quality of life

Quality-adjusted life-years (QALYs) were estimated in secondary analyses using the utility scores obtained from the EQ-5D-5L instrument for participants aged ≥ 11 years and the CHU-9D in the analysis focused on children aged < 18 years. 63,64 For participants aged 5–6 years, the CHU-9D was completed by parental proxy, but for all other ages these instruments were self-completed.

The decision to use the EQ-5D-5L was based on the EuroQoL EQ-5D-Y user guide (https://euroqol.org/publications/user-guides/; accessed November 2020) available at the time of study design, which stated that although the EQ-5D-Y ‘is generally recommended’ the adult version might be possible. We chose to use just the one version of the EQ-5D in the study for consistency and because the EQ-5D-Y does not currently have a UK valuation set. The CHU-9D was chosen over the EQ-5D-Y because a UK valuation set exists for it.

Neither generic utility instrument had been used in this disease area before. Therefore, their inclusion was somewhat experimental, seeking to start to build up some evidence for their potential for use in vitiligo.

Quality assurance

The process of setting up and checking the quality of devices before they were issued to patients was estimated using expert opinion from staff at the Nottingham University Hospitals medical physics department. The quality assurance process involved device in and device out processes. Before devices were issued to participants they were tested for electrical safety and output, spectral characterisation was carried out, and some data administration was involved. When devices were returned, they again had their output tested and some data administration was involved. Table 19 shows the time and cost for each aspect. Staff time was assumed to be a mid-point band 5 on Agenda for Change and the batch size was assumed to be 10 devices at once. Quality assurance costs were also multiplied by the annuity factor to gain the cost over the study period. In reality, quality assurance might be carried out more frequently than every 5 years or may be provided using a different service model (e.g. specialist vs. local sites undertaking the activity) that may affect cost but the impact of this assumption is tested in the sensitivity analysis section, where price is varied to see the impact on cost per treatment success.

It was assumed that devices would be given to patients at an appointment with the dermatologist. It was assumed they would have four visits with a dermatologist over the 9-month treatment period, whichever treatment group they were in. Those receiving the NB-UVB would also have had an appointment with a nurse in the dermatology department and a training session with the nurse. Tables 20 and 21 show that the training time had a mean of 73.08 minutes in the NB-UVB-only group and 69.17 minutes in the combination treatment group. In addition to routine visits to the dermatologist and nurse at set intervals, unscheduled contacts were also recorded. Such visits could either be face to face or over the telephone and occurred because of side effects or concerns over the use of treatments. The number of such contacts was small in all groups, although the combination treatment group had the most (see Tables 20 and 21).

Those participants receiving active TCS received two 90-g tubes of mometasone furoate 0.1% ointment at the outset of the study and any requests for further tubes were recorded and costed accordingly and similar amounts were requested in the TCS only and combination treatment groups (see Tables 20–22).

Cost-effectiveness analysis of narrowband UVB only compared with topical corticosteroid only

The unadjusted risk difference for NB-UVB compared to TCS was 3.64% (adjusted risk difference 5.20%), this equates to a NNT of 27 (19 adjusted); in other words, 27 (19) participants would need to be treated for one of them gain treatment success.

The incremental difference in cost was £174.65 (95% CI £152.75 to £96.55) unadjusted or £173.44 (95% CI £150.55 to £196.32) adjusted for age and body region of the target patch. The unadjusted incremental cost was £4801.92 (£3335.74 adjusted) per additional successful treatment. Figure 8 shows the probability that NB-UVB only is cost-effective at different possible levels of willingness to pay for an additional treatment success; probability increases as willingness to pay increases. It can be seen that there is a lot of uncertainty surrounding the decision of whether or not NB-UVB alone, compared with TCS alone, represents value for money as there is always at least 40% probability of making the wrong decision if choosing to fund NB-UVB alone below a threshold value of willingness to pay of £10,000 per additional treatment success.

FIGURE 8.

Cost-effectiveness acceptability curve for NB-UVB only vs. TCS only.

Cost-effectiveness analysis of combination treatment compared with topical corticosteroid only

The unadjusted risk difference for combination treatment compared to TCS was 9.16% (adjusted 10.94%). This equates to a NNT of 10 (9 adjusted) [i.e. 10 (9 adjusted) participants would need to be treated for one of them to gain a treatment success].

The incremental difference in cost was £213.40 (95% CI £190.02 to £236.78) unadjusted or £211.46 (95% CI £188.10 to £234.81) adjusted for age and location of the target patch. The unadjusted incremental cost was £2328.56 (£1932.35 adjusted) per additional successful treatment.

Figure 9 shows the probability that combination treatment is cost-effective at different possible levels of willingness to pay for an additional treatment success. It shows that combination treatment is likely to be cost-effective if the decision-maker is willing to pay > £3000 per additional treatment success. There is, however, currently no evidence to indicate how much a decision-maker would be willing to pay for an additional treatment success as defined in this study. Should the decision-maker’s willingness to pay per additional treatment success be low (i.e. < £2500) then it can be seen that uncertainty surrounding the decision to fund combination treatment is high.

FIGURE 9.

Cost-effectiveness acceptability curve for combination treatment vs. TCS only.

Sensitivity analysis

A number of sensitivity analyses were undertaken to explore key uncertainties around important parameters in the economic evaluation. The results of this are summarised in Table 24 with greater detail below for each analysis.

Complete-case analysis

The base case assumed data to be missing at random and undertook imputation to allow for this. 70 Table 24 presents the results for a complete-case analysis that includes participants with complete resource use and outcome data only to see if this changes the conclusions reached in the base-case analysis. In total, 348 participants had complete data on both cost and outcome (success of treatment): 113 in TCS only, 115 in NB-UVB only and 120 in combination treatment.

The cost of the narrowband UVB device

The cost-effectiveness of the interventions is likely to be driven significantly by the cost of the NB-UVB device. There is uncertainty about how the device would be prescribed and used in the NHS. If adopted as an effective treatment, patients may have to pay for the device themselves (with training, support and quality assurance paid for by the NHS), or the device might be adopted and provided free at point of use by the NHS for NHS patients. The base-case analysis annuitised the device cost, assuming that the device would be used for a period of 5 years, but there is uncertainty surrounding this period of use and, in practice, it may be that the devices are not returned by patients at the end of treatment.

We re-estimated the incremental cost per successful treatment assuming that patients paid for the device, quality assurance, glasses and goggles as one extreme and at the other we doubled the price of the device, quality assurance, goggles and glasses to provide an upper estimate.

As expected (see Table 23), reducing the cost of devices to zero reduced the incremental cost per treatment success, thereby lowering the amount that an NHS decision-maker would have to be willing to pay for this treatment to be implemented in the NHS. Conversely, doubling the cost of the device increased the incremental cost per treatment success and meant that the NHS would have to value a treatment success more highly than the base case to be willing to adopt the treatments.

Changes in the price of the device had less of an impact on the combination treatment versus TCS only comparison, due to the greater treatment success observed in the combination group. As noted in the primary base-case analysis, it is not clear how much a decision-maker would be willing to pay to achieve one more additional treatment success as measured on the VNS. Therefore, these figures just provide a range around the likely cost per treatment success.

Wider cost perspective

As part of the trial, participants were asked about the out-of-pocket costs (if any) incurred by themselves or their families as a result of their vitiligo. These costs were added to the base-case results (NHS perspective only) to see how they would have an impact on the incremental cost per treatment success. A total of 47 (11.1%) participants reported incurring out-of-pocket costs during the 9-month treatment period: 17 in the TCS-only group, 17 in the NB-UVB-only group and 13 in the combination group. The mean number of items and mean cost per participant by group can be seen in Tables 20–22. The type of items included (from most to least purchased) camouflage/makeup, sun cream and sun care, clothes/scarves, face creams/moisturisers/emollients, fake tan/tanning products, travel for appointments, private appointment including multivitamins, and herbal remedies.

Taking into account the participant out-of-pocket costs in relation to vitiligo reduced the incremental cost per treatment success, as these costs were higher in the standard care group (TCS only) (see Table 23 for results).

Impact of adherence

Because significant clinical effectiveness was found and a little under half of the participants used the treatment for > 75% of the expected duration, the primary economic analysis was repeated including the adherent sample only, where adherence was estimated as total sessions used divided by total expected sessions. In total, 227 participants adhered to treatments > 75% of the time; this sample was used as the adherent sample, minus three participants (one of whom had the primary outcome missing and two who had cost data missing).

The intervention was more cost-effective for patients who adhered to treatment, as they were the ones most likely to achieve a successful outcome (see Table 23 for estimates).

Longer-term analysis (12 to 21 months)

In the health economic analysis plan we intended to explore the longer-term cost-effectiveness of the comparators of interest beyond the 9-month treatment period (if either were found effective), to see if value for money was sustained. In the trial, only 30.4% of participants had complete data on NHS resource use in months 10–21, 44.5% of participants aged ≥ 11 years completed the EQ-5D-5L at 21 months and 43.3% of participants aged < 18 years at the beginning of the study had completed the CHU-9D at 21 months. Given the sparsity of data, we have not performed an economic evaluation over the longer-term follow-up as it would be too speculative. However, we report mean estimates of the participants’ (all ages, n = 517) wider NHS use over months 10 to 21 (the follow-up period) and utility at 21 months. Only 157 participants had complete resource use data for the whole 12-month follow-up period, which may have been for zero use, 64 had 9 months’ worth of data available, 56 had 6 months’ worth of data available, 59 had 3 months’ worth of data available and 181 had no resource use data recorded for the follow-up period. The mean quarterly NHS cost per participant over the 12-month follow-up period was £21.26 (SD £46.32) for combination treatment (n = 114), £25.89 (SD £52.82) for NB-UVB alone (n = 117) and £21.74 (SD £42.33) for TCS alone (n = 105). The mean prescription cost per participant over the 12-month follow-up period was £14.82 (SD £45.22) for combination treatment (n = 114), £13.78 (SD £45.63) for NB-UVB alone (n = 117) and £13.20 (SD £51.44) for TCS alone (n = 107). The mean out-of-pocket cost per participant over the 12-month follow-up period was £42.85 (SD £398.74) for combination treatment (n = 114), £3.62 (SD £16.93) for NB-UVB alone (n = 117) and £8.48 (SD £39.41) for TCS alone (n = 107).

The mean utility (EQ-5D-5L) per participant aged ≥ 11 years at 21 months was 0.856 (SD 0.230) for combination treatment (n = 73), 0.865 (SD 0.231) for NB-UVB alone (n = 61) and 0.833 (SD 0.274) for TCS alone (n = 69). The mean utility (CHU-9D) per participant (aged < 18 years at the outset of the study) at 21 months was 0.938 (SD 0.054) for combination treatment (n = 20), 0.941 (SD 0.056) for NB-UVB alone (n = 16) and 0.937 (SD 0.118) for TCS alone (n = 16).

Cost–utility analysis for those aged ≥ 11 years

Of the 517 participants in the trial, 456 (88%) participants were aged 11 years, 155 were randomised to TCS only, 148 were randomised to NB-UVB only and 153 were randomised to combination treatment.

The cost–utility analysis was planned as a secondary analysis due to the fact that, to our knowledge, no prior study had utilised the EQ-5D-5L, or CHU-9D for children, in patients with vitiligo. There were some concerns that such generic quality-of-life instruments might not be appropriate for this condition, as much of the effect may be visual or psychological rather than on physical quality of life. Such concerns seem to have been borne out in the study, Table 25 shows the domains on the EQ-5D-5L selected by participants at baseline. In total, 55% of participants reported having no problems on any of the five domains of the EQ-5D-5L at baseline, suggesting that over half of the sample started the study in perfect health as defined by this instrument. This is a large ceiling effect that was also observed at subsequent follow-up (Table 25). No floor effect was observed at any time point.

Cost–utility analysis for participants aged ≥ 11 years for narrowband UVB only compared with topical corticosteroid only

The unadjusted mean cost per participant in the NB-UVB-only treatment group (n = 131) was £774.64 (SD £83.71, 95% CI £760.17 to £789.11) compared to £599.99 (SD £96.18, 95% CI £583.43 to £616.55) for the TCS-only group (n = 132) giving an unadjusted mean incremental cost per participant of £174.65 (95% CI £152.75 to £196.55). The imputed, adjusted and bootstrapped mean incremental cost per participant was £169.58 (95% CI £165.50 to £173.65) more for the NB-UVB-only treatment group than the TCS-only group.

The imputed, adjusted and bootstrapped mean incremental QALYs gained were 0.0204 (95% CI 0.0180 to 0.0229) in favour of the NB-UVB only compared with TCS only (Table 26). The adjusted incremental cost per QALY was £8293.88.

Cost–utility analysis for participants aged ≥ 11 years for combination treatment compared with topical corticosteroid only

The unadjusted mean cost per participant in the combination treatment group (n = 136) was £813.38 (SD £111.39, 95% CI £794.49 to £832.27) compared with £599.99 (SD £96.18, 95% CI £583.43 to £616.55) for the TCS-only group (n = 132) giving an unadjusted mean incremental cost per participant of £213.40 (95% CI £188.33 to £238.46). The imputed, adjusted and bootstrapped mean incremental cost per participant was £203.93 (95% CI £199.39 to £208.47) more for the combination treatment group than the TCS-only group.

The imputed, adjusted and bootstrapped mean incremental QALYs gained were 0.0145 (95% CI 0.0123 to 0.0167) in favour of the combination treatment compared with TCS only (see Table 26). The adjusted incremental cost per QALY was £14,081.

Alternative utility value sets were tested to see if they had any impact on the results but because the size of QALY gains were so small in this study (see Appendix 7) the choice of value set did not affect the results presented above in the cost–utility analysis.

Cost–utility analysis for participants aged < 18 years

A total of 119 participants were aged < 18 years, 40 received TCS only, 39 NB-UVB only and 40 the combination treatment. Complete cost and outcome data were available for only 91 (75.8%) of these participants. The results presented here are based on a complete-case analysis only, as an imputed, adjusted analysis was not possible due to the small sample size. Table 27 shows the utility, as measured on the CHU-9D at baseline and 9 months, and QALYs for the 9-month treatment period. The incremental QALYs were non-significantly different from zero.

The ceiling effect on this instrument was better than the EQ-5D-5L but still high; 30% of participants had no problems according to any of the nine dimensions on the CHU-9D (Table 28). The domains of worry, tiredness and sleeping were those in which problems were reported most often.

Narrowband UVB compared with topical corticosteroid: cost–utility analysis for participants aged < 18 years (Table 27)

For those participants with complete cost and utility data, the unadjusted mean cost per participant in the NB-UVB-only treatment group (n = 28) was £818.47 (SD £91.20, 95% CI £787.16 to £849.81) compared with £597.51 (SD £49.31, 95% CI £579.43 to £615.60) for the TCS-only group (n = 31). The unadjusted mean incremental cost per participant was £171.50 (95% CI £137.35 to £205.65). The unadjusted incremental cost per QALY was £92,381.98. This figure is significantly higher than accepted threshold values and thus would not be considered cost-effective.

Combination treatment compared with topical corticosteroids: cost–utility analysis for participants aged < 18 years

For those participants with complete cost and utility data, the unadjusted mean cost per participant in the combination treatment group (n = 35) was £818.47 (SD £91.20, 95% CI £787.16 to £849.81) compared with £597.51 (SD £49.31, 95% CI £579.43 to £615.60) for the TCS-only group (n = 31). The unadjusted mean incremental cost per participant of £220.97 (95% CI £184.23 to £257.69). As mean cost was higher for the combination treatment group and QALYs less (albeit by a very small amount), it is possible to say that for this group of participants standard care (TCS only) dominates (i.e. it is both cheaper and more effective than combination treatment). However, one should note the small sample sizes.

Study design

This is a mixed-methods process evaluation incorporating stakeholder interviews, interviews with NHS commissioners, an online survey of those who delivered home-based treatment in the trial and focus groups with those who delivered the trial.

Ethics approval was obtained for the process evaluation on 10 April 2017 (ethics reference 14/EM/1173, SA04) from the National Research Ethics Service (NRES) Committee East Midlands – Derby.

Participants

1. Trial participants were recruited from the main HI-Light Vitiligo trial.

   Sampling was purposive, focused initially on age, treatment allocation, recruiting site and treatment success (based on the primary outcome). Other factors, such as treatment adherence, early stopping of treatment, ethnicity/skin type, gender, extent of vitiligo, number of patches being treated, and whether or not the participant experienced problems with treatment, guided later-stage recruitment of interviewees.

   Participants were approached at the 9-month time point to minimise the impact that this had on treatment adherence.

2. Commissioners were identified via online directories of Clinical Commissioning Groups and via personal contact with members of the study team.

3. On completion of the trial, site investigators (principal investigators and research nurses) at all recruiting centres were invited to take part in an online survey and/or a focus group to review the delivery of NB-UVB.

4. To avoid having any impact on recruitment to the trial, all activities with recruiting site staff were conducted after recruitment had finished.

Data overview

1. In total, 25 interviews with trial participants (or parents) were conducted between 13 July 2017 and 20 July 2018 (Table 29).

   Twelve out of the 16 recruiting sites were represented in the sample.

2. Nine commissioner interviews, involving 10 individuals, were conducted between 5 June 2017 and 10 October 2017.

   Participants included strategic and operational roles in the commissioning process, and represented a geographic spread across England. Most participants were medically trained, and included general practitioners (GPs) with a special interest in dermatology.

3. Twenty-four recruiting site staff completed the online survey: seven doctors, 16 nurses and one other.

   Ten of these had prior experience of phototherapy services, for others the HI-Light Vitiligo trial had been their introduction to this treatment. To support anonymity we did not collect data about which site they represented.

   Thirteen site staff participated in the focus groups representing 10 recruiting sites. Eleven nurses were split into two groups; two doctors formed the final focus group.

Is home-based narrowband UVB treatment and topical corticosteroid for vitiligo manageable for participants?

All bar two of the (health-care professional) survey respondents agreed that home-based treatments are ‘easy’ for participants; discussion group data (with site investigators) reinforced this, with health-care professionals reporting that the phototherapy device is (superficially) simple to operate and most trial participants seemed to understand the instructions offered about light therapy and TCS ointment. Trial participants offered a similar assessment indicating that they generally understood how to use the individual treatments; training and support offered by the research nurses and demonstration video had helped in this.

Some practical difficulties with individual treatments were described (by both site investigators and trial participants) but nothing of a magnitude to prevent participants effectively managing their treatments. Practical difficulties for light therapy were timers that failed, guard teeth that broke off, difficulties reaching parts of the body and difficulties using (a flat device) on curved parts of the body. Practical difficulties for the use of TCS ointment were an unpleasant smell, a greasy feeling and poor absorption. The time commitment required for light therapy was a common cause for comment:

It felt like an awful amount of time, I am pretty busy and to eventually be spending in excess of three-quarters of an hour per 2 days just felt like an inordinate amount of time.

Adult participant 3

Treating multiple patches could leave participants feeling overburdened:

I started with more than that because I was quite positive, I was doing different parts of my body like six or seven or something . . . Then I just did three, the three patches they were interested in so I was just treating them, no more.

Adult participant 4

Time seemed to be more of an issue for parents treating a young child (i.e. ‘keeping them still!’), or if there were other children in the household that required attention. Participants described linking treatment to ‘treat time’ for children (e.g. watching television); others described building routines that facilitated their time commitment:

Yes it was always around 6 o’clock after my tea and after I’d washed up and whatever you know, then I didn’t have time to just sit and do it then.

Adult participant 12

Yeah, so I used to sort of do each week in advance so that I knew what I’d got to do the next week and how long each treatment was going to be, and I’d put my notes next to it so I knew I’d done it, so no, I was quite comfortable with that.

Parent of child participant 3

Although time-consuming, home-based phototherapy was considered less disruptive than regular hospital visits for phototherapy, and the potential for treatments at home was important to the majority of trial participants.

Despite a generally positive assessment of each treatment, health-care professionals and some participants flagged the complexity of treatments in combination. Site investigators reported that trial participants who they thought had a good understanding of treatments made errors with TCS and phototherapy dose, and suggested that some individuals had disregarded instructions and used the ointment/device excessively on multiple body sites. One trial participant admitted as much:

I just ramped it up pretty much straight away back to what it was, but again no redness whatsoever which only really served to confirm it’s a dummy.

Adult participant 11

Some trial participants said that they found the combination treatment protocol complicated, particularly early on, and expressed caution when initially using the treatments:

Yeah I found it confusing for the first few weeks, it was like 1 week on 1 week off [for TCS], and every other day for the light and stuff.

Parent of child participant 5

Some trial participants acknowledged that they had made mistakes with treatment:

I was completely knackered and was [. . .] at the end of the day, had done the light treatment. So, I sat and did my chest which was one the areas being treated and part of, one of my, part of my left hand which is the other bit of the treatment and then started to do the second bit on the left hand and fell asleep so I ended up burning myself.

Adult participant 3

Stepping-up or down NB-UVB dose (as part of the treatment protocol or in response to erythema) was recognised to confuse and cause difficulties, with site investigators concerned that some participants never appeared to fully understand the process of incremental dose change. Trial participants indicated that their treatment diary was essential in guiding them:

Yes, without that it would be nowhere, without the form that you fill in with boxes I mean and writing down the time you would be absolutely nowhere, there’s no chance in a million that you would actually keep to anything like the protocol.

Adult participant 3

A further area of complexity identified by trial participants was in assessing whether or not treatment was making any difference and the importance of the photographs in determining any change:

Well I thought [things had improved] but when we got the photographs, you know we got the photographs on the computer; it didn’t seem to be any different to be honest.

Adult participant 12

Only when I went back and saw the difference in the photographs, so as I was treating it I couldn’t really see any difference, when you saw the photographs versus my face in the mirror actually, there was a difference.

Adult participant 5

Seasonal variation in skin tone might add to this complexity:

. . . because usually I do find I do get quite tanned and therefore, between the summer and winter there is a contrast, so without looking at the photographs I couldn’t tell whether actually it was making any difference or not.

Adult participant 5

Trial participants described nurses as having an important role in supporting them in assessing whether or not treatment was leading to improvement; nurses were also considered important in supporting the management of erythema (especially when it occurred for the first time). Overall, participants viewed nurse support positively; they responded quickly, were helpful and friendly, they offered reassurance, and generally ensured that participants were doing the right thing.

In contrast, participants were sceptical of the potential for GP-led support:

In all honesty if I rang my GP up and I had an issue, they’re useless anyway, you have to wait God knows how long to get an appointment and whatever else.

Parent of child participant 6

Despite the acknowledged complexity of the combination treatment, trial participants reported that they felt able to adhere to treatment regimen without any fundamental difficulty. Where they had not adhered to protocol this would more often be with light therapy and participants would point to legitimate (practical) reasons for this, such as other health conditions or holidays:

. . . like I say it was literally just when we went away on holiday, I just, I probably wouldn’t have done it if I did take it to be honest when I’m on holiday.

Adult participant 9

Others had not adhered to the light treatment protocol when they recognised no benefit from treatment, many associated this with receiving a dummy device:

I think I only really found it onerous because I was just convinced it was a dummy, and I just felt as if I was, about 20 minutes I was really just wasting basically because I thought this was not going to be any good at all.

Adult participant 11

As soon as I realised that it wasn’t even tanning my skin I just, it was really hard to continue because it was really time-consuming.

Adolescent participant 2

These comments suggest the importance of expectations in shaping how treatments are managed; some individuals ceased treatment because expected improvement had not occurred. That noted, it should be made explicit that most trial participants demonstrated generally realistic expectations, often borne out of previous treatment experience:

I mean it’s not like sunburn or a suntan is it? Where the skin sort of changes overnight practically, if you’ve got something happening in the cell structure maybe that takes a much longer time because you’re waiting for the cells to regenerate.

Adult participant 10

. . . 6 months I would have expected to have seen something.

Adult participant 2

So, it’s just I didn’t have great expectations, possibly because of previous treatments and stuff.

Adult participant 13

Expectations were, however, often tempered by an intuitive, emotional response to the offer of a new treatment:

I was hoping for them to shrink or bring some of the pigmentation, like away, get her back to her normal colour.

Parent of child participant 5

Actually, I was very pessimistic about the whole thing but I was, I didn’t really think that I was going to get any benefit.

Adult participant 13

Should home-based treatment be made more widely available?

Although potentially complex (and confusing for some), there was a general sense that this type of combination treatment should be made available to more vitiligo patients; 18 out of 24 health-care professional survey respondents agreed or strongly agreed with this. Focus groups with site investigators supported this by highlighting a clinical population that have few treatment options, and for whom the impact of vitiligo can be very distressing:

We have always said that it is the best of a bad bunch of treatments, and it probably still is. There is no fantastic treatment out there for vitiligo, there doesn’t seem to be, and the trial doesn’t show that it’s fantastic. It’s shown that for patients it’s worthwhile doing because the quality of life is impaired for a lot of patients. They are pinning hopes on it.

Site investigator 9 – research nurse

This is reflected in those reasons offered by trial participants for taking part in the HI-Light Vitiligo trial. Some hoped that participation would bring them access to new treatments for themselves or their children, some subsequently hoped for complete remission, whereas others hoped that their disease would stop spreading. For a minority of participants there was a sense of having ‘nothing to lose’:

. . . had hoped it would totally recover the 9 months or earlier you know, the sort of blemishes would disappear.

Adult participant 5

I decided to take part because why not, it would be working on my skin or not but I just decided to take part to see what happened.

Adult participant 4

I don’t know, probably half and half of me was hoping that yes, something would work and it would help her, but if it didn’t then we wasn’t really going to lose anything.

Parent of child participant 7

Others hoped that their involvement in the HI-Light Vitiligo trial would benefit the broader vitiligo population by contributing to the development of a new treatment pathway. Given these assessments, it is understandable that site investigators recognised the importance of new treatments for their patient population, even suggesting that (irrespective of the clinical impact) new treatments offer vitiligo patients hope and the potential to engage with their condition.

As this desire for new treatments might suggest, commissioners confirmed that treatment pathways for vitiligo are often lacking, suggesting that dermatology, let alone vitiligo, is unlikely to be a priority in commissioning discussions. They also indicated that some commissioners perceive vitiligo to be a ‘cosmetic’ condition, which adds a further barrier to it being considered a priority area. Variation in cosmetic impact (and associated concerns) was also manifest in comments made by trial participants, which suggests that willingness to pursue (complex) treatment might vary according to site or visibility of vitiligo:

. . . if I had, say if it was like more in a cosmetic important place I’m pretty sure I would be prepared to have a go at it long term, when saying long term, I mean over a period of years or whatever is required.

Adult participant 3

I never felt that for a chap of my age, I mean I’m 74, it’s a bit irrelevant because they’re worse when you’ve got a bit of a sun tan obviously, but it’s not like being perhaps a lady who cosmetically can look a bit odd.

Adult participant 7

These comments illustrate and support the assessment of site investigators that home-based therapy might not be appropriate for all vitiligo patients. In the online survey, 12 site investigators indicated that home-based therapy would be appropriate for most people, 11 indicated that it would be appropriate for some, but none indicated that it would be appropriate for all people with vitiligo. Considerations for providing home-based treatment might be (1) clinical, (2) practical, or related to (3) personal circumstances:

1. Comments from consultant dermatologists (as part of the survey and site investigator discussions) speculated that combination treatment might be particularly beneficial for new patients as an early intervention.

   Trial findings point to variation in outcome according to body site of vitiligo patch (see Chapter 3).

   Target patches were chosen by trial participants, and had to have been active in the preceding 12 months, as reported by participants. Discussions with site investigators suggested that some participants may not have been able to judge this very well, with some investigators feeling that potential participants may have exaggerated the activity of their vitiligo to obtain access to the treatments offered in the trial.

2. The practical challenge of managing combination treatment (long-term use, dose fluctuation, potential side effects, etc.) was recognised by both participants and site investigators. Participants suggested that not all people are sufficiently organised for these treatments; they thought that individuals need to plan ahead, be committed to the treatments and to be willing to incorporate them into their routine:Adult participant 1

   If you’re an organised type of person it becomes part of second nature after a while.

   In some cases, hospital-based light therapy was considered more appropriate:Adult participant 4

   Yes, to go to the cabinet and spend 5, 10 minutes and that is all because you will have treated all your body.

   The duration of home-based treatment was also a factor to be considered when considering individuals:Parent of child participant 2Adult participant 3

   I think doing it for any longer [than 9 months] probably would have been a bit challenging probably, because it became, it did start to become a bit more of a hassle to do it so regularly.

   Towards the end, I mean I really did find, it just felt like I was spending a lot of time.

   Health-care professionals felt that mental health issues, other health complaints, or significant caring responsibilities (e.g. multiple children) might all challenge an individual’s ability to maintain a complex treatment regimen over a long period of time.

3. Beyond practical challenges, there was concern from health-care professionals that an individual’s level of understanding about vitiligo and their expectations of treatment might also be important

4. Those with potential difficulties adhering to a long-term treatment programme may be unlikely to benefit from home-based treatment and those with unrealistic expectations might find it difficult to adhere. Trial participants recognised that some individuals might exaggerate, or provide inaccurate information about patches, to gain access to a new treatment.

   Establishing which individuals might benefit from home-based treatment was considered difficult by health-care providers; fully sharing information about effectiveness, treatment burden and treatment duration may help support shared decision-making. Some site investigators even suggested that some kind of test, formal or informal, of whether or not a patient understands the treatment regimen might also be appropriate.

Could home-based treatment be made more widely available outside the trial?

A small number of site investigators indicated that they were already reusing devices from the trial and incorporating them into clinical practice for vitiligo. Indeed, some sites in the UK (e.g. Ninewells Hospital, Dundee) are already offering such devices for use at home, often to treat other skin conditions such as scalp psoriasis.

Other site investigators pointed to the importance of situating new pathways within existing phototherapy provision for other skin conditions, with appropriate support from medical physics to monitor device output. However, they recognised that this would not be possible in all locations and some suggested that home-based treatment should be managed regionally by specialist centres.

These perspectives mirror commissioners’ concerns that any new treatment would need to sit within (or at least not disrupt) existing service pathways. In contrast to site investigators, commissioners were, however, less aware of a need for new treatment pathways and perceived no explicit demand from patients, clinicians or health-care providers:

I’m not getting any complaints for example about the services that we provide. Like GPs aren’t coming to me saying, we’re not happy with this. As far as our GPs are concerned, they’re getting a good service because their patients aren’t complaining to them. It’s not coming up on our monitoring in terms of performance.

MC: commissioner

Should a new treatment pathway be considered, commissioners stressed that clinical and cost-effectiveness would be key to any decision, they warned that significant changes to services would need to be supported by considerable evidence of clinical or cost improvement:

I think the only issue would be time . . . a change in service is something that can be time-consuming. So the benefit has to be significant. So we’re making an assessment how significant the change is.

JB: commissioner

Investigators recognised that commissioners might be reluctant to commission new services and speculated on different mechanisms for the provision of phototherapy devices.

All investigators recognised that these devices can be easily purchased online and some reported that participants had indicated that they might buy one independently of NHS support. Similarly, several participants described considering purchasing a device, although for some the thought of ‘going it alone’ deterred them:

I think they’re about £100 aren’t they? They’re not fantastically expensive but I didn’t then think I might go and buy one of those, largely because I wasn’t sure how I would use it you know, it’s very secure and comforting isn’t it to have that kind of regime and do this, that and the other every day, and then you think ‘right OK so I know where I’m up to’ and so on, so to suddenly be cut loose from that would be a little bit more you know, anxiety provoking, when you know that it’s potentially dangerous.

Adult participant 6

Investigators recognised that the publication of (positive) trial findings might accelerate this type of independent use among a patient population desperate for treatment options. Most investigators were apprehensive about this, in the online survey only 2 out of 24 felt that NHS involvement was not important and 13 felt that this was essential or very important. Concerns for safety led some investigators to suggest that devices should be automatic (i.e. patients cannot adjust) or managed by a dermatology nurse. Some patients also suggested that devices could have safety features, such as automatic cut-outs, to prevent overuse. Investigators also suggested that patient monitoring should be frequent and sooner than 3 months (as in the trial). Some participants recognised the value of an earlier monitoring visit, whereas others felt that this was not necessary:

. . . personally think it needs an interim visit, if only to compare the photograph, because I do think that you forget what it was like and you do think ‘oh it’s not making any difference’, but then when you see the photograph and you see the shape changing.

Parent of child participant 3

The potential for some form of ‘mixed economy’, where patients lease or purchase a phototherapy device within an NHS service, was considered by site investigators to be the most likely way that effective provision could be offered. However, this is not without its difficulties. Both trial participants and site investigators were concerned about unequal access for those that cannot afford to purchase or lease a device. Some health-care professionals suggested that ‘purchasing health care’ may lead to unreasonable expectations and/or incorrect use (‘if I’m paying for it will work!’). In addition, the failure to return leased devices might make a service economically not viable (127 light therapy devices were not returned during this trial).

Trial fidelity

The process evaluation offers some insight to support the interpretation of the clinical effectiveness and cost-effectiveness data.

Participant interviews demonstrated that adherence to the treatments was not hampered by a lack of knowledge about how to use the treatments, nor by a lack of support when using them. Intermittent non-adherence to treatment was acknowledged (and is perhaps unsurprising given the treatment burden associated with light therapy) but this was most often a pragmatic response to life circumstances and events. Trial adherence data (see Table 7) show that around two-thirds of participants used the treatments as specified in the treatment protocol while they were still using them.

All of this suggests that trial procedures were adequate in supporting normal practice, and that clinical findings were neither inflated nor diminished by unrealistic or disastrous levels of treatment adherence.

Where trial procedures were, however, less successful was in blinding participants to treatment allocation. The trial data demonstrates that a relatively high percentage of participants stopped treatment early, with only around half using treatments for at least three-quarters of the expected duration (see Table 7). Although some participants stopped treatment because they had achieved complete remission, more commonly participants indicated that a lack of effect led them to cease treatment with some suggesting that they knew that their light device was a dummy. A lack of any redness in the skin or other indications such as warmth meant that around half correctly guessed that they had the dummy device (89% of those using active light guessed correctly).

With regard to trial outcomes, participants were largely able to judge the primary outcome (VNS). The use of photographs was crucial in this, owing to the duration between visits, and/or because of difficulties in establishing whether or not minor changes had taken place. It is perhaps worth noting that participants indicated that seasonal variation in the noticeability of their vitiligo (i.e. it was more noticeable in summer) may have had an impact on their assessments. The consistency of other trial outcomes, however, suggests that this was not a major issue.

Population in need

Site investigators recognised a clinical population with few treatment options, commissioners acknowledged that vitiligo is not a priority area, and trial participants expressed a desire to try new treatments that might work where normal clinical practice had failed. Health economic assessment suggests that many individuals with vitiligo manage with little or no input from the NHS. The culmination of these insights demonstrate that there is an unmet need for effective treatments for vitiligo.

Participants were willing to go to great lengths to accommodate the time-consuming and complex treatment regimen; many were willing to continue treatment in the absence of any effect. Expectations were realistic, participants hoped for partial improvement or halting the spread of vitiligo (rather than complete repigmentation) and few expected immediate results. These characteristics may suggest that the levels of treatment success observed in the trial, and ability of the interventions to stop the spread of vitiligo, offer sufficient potential for individuals with vitiligo to be willing to try (and persist with) home-based treatment in the future.

Assuming that vitiligo does, however, have at least some degree of psychological impact in many people with the condition, it seems important that future recipients of vitiligo treatments should be fully informed about the likely success rate of home-based treatment, to ensure realistic expectations of treatment.

Easy to do but complex to use

Site investigators and trial participants recognised that treatments were relatively straightforward to use, with appropriate instruction and support.

However, economic assessment points to the cost-effectiveness of combination treatment (more so than treatments in isolation); a complex treatment protocol coupled with a considerable time burden (both each day and over a period of months) creates potential for incorrect use that can result in either increased side effects or reduced effectiveness. Site investigators were particularly concerned about the potential for participants to harm themselves. Treatment diary planners and site staff were considered essential by trial participants in helping them to navigate this complexity. It is notable that some participants had considered purchasing a light therapy device, but had decided against this because of a lack of ongoing NHS support. Site investigators stressed the importance of ongoing support and monitoring of patients in any future clinical service.

The clinical data indicate a need for some form of intermittent maintenance therapy, as effectiveness diminished once treatments had been stopped. Many trial participants were relieved to stop light therapy after 9 months, suggesting that maintenance therapy with TCS is more likely to be a preferred approach over maintenance light therapy.

Treatments may not be suitable for all

Treatment burden means that home-based therapy is less appropriate for those wishing to treat a high number of patches. Lifestyle and personal circumstances may make adhering to a complex treatment regimen over an extended period of time difficult for some. It is also pertinent to remember that not everyone with vitiligo wants or is seeking treatment.

Site investigators stressed the difficulty of predicting which participants were most likely to benefit from home-based treatment. Fully discussing the advantages and disadvantages, treatment burden and time scale of home-based treatments are essential in helping people with vitiligo reach informed decisions about treatment.

Integrating within the NHS

Site investigators were positive about the trial results (‘any improvement is worthwhile’ was a common sentiment) and participants were keen to see effective treatments for vitiligo become available. However, it is unclear whether or not sufficient improvement is manifest here to convince commissioners of the value of home-based phototherapy for the management of vitiligo. Home phototherapy services that support treatment of a broad range of skin conditions, as seen in existing specialist phototherapy units offering home phototherapy (e.g. Ninewells Hospital in Dundee), are likely to be more attractive.

Both site investigators and trial participants recognised that light therapy devices might be privately purchased. Some site investigators were concerned about this, and some trial participants indicated interest in privately purchasing, but at the same time being concerned about a lack of clinical support if they did so. It is pertinent to stress that training, treatment diaries, technical support and staff support were all considered essential by site investigators (and many trial participants) to the success and safety of home-based treatment. Medical physics is required to ensure that devices are appropriately calibrated and to ensure that bulb output is consistent; phototherapy services are required to support a complex treatment regimen, monitor effects and support/temper patient expectations.

The culmination of these two strands points to the potential for some form of mixed economy provision where light therapy devices are leased or privately purchased (independently or via the NHS) with treatment defined, supported and monitored by NHS services. The number of devices (127) not returned after the end of the treatment period in this trial might make NHS leasing of devices a less appealing prospect to commissioners.

Site investigators recognised that not all settings are well placed to provide medical physics support and commissioners suggested that new pathways are more attractive if they sit within existing provision (rather than requiring infrastructure development). This might point to a hub and spoke model of regional delivery whereby specialist sites with existing medical physics expertise could provide access to home phototherapy devices across a number of NHS trusts, but clinical provision including training and monitoring of side effects could be delivered locally.

It is perhaps appropriate to conclude by recognising that the challenge of any future home-based treatment for vitiligo will be navigating the needs of patients and their enthusiasm for new treatment alongside the concerns of health-care professionals about the potential side effects associated with light therapy and long-term TCS use.

Contextual factors relating to stakeholder involvement

Stakeholders involved in the HI-Light Vitiligo trial

Stakeholders (including people with vitiligo and health-care professionals) were involved across all areas of trial design and delivery. They were involved in prioritising the initial research question; completing surveys to inform trial design; developing and testing the primary outcome; assisting in trial conduct, recruitment and oversight; and contributed to the analysis and interpretation of the trial results. The number and types of stakeholder involved at each stage in the trial life cycle are shown in Figure 10.

FIGURE 10.

Stakeholder contribution to the design, development and execution of the HI-Light Vitiligo trial.

Stages of research and opportunities for stakeholder impact

The impact that stakeholder contribution had in the development and delivery of the HI-Light Vitiligo trial is summarised in Table 31.

FIGURE 11.

Stills from the HI-Light Vitiligo trial training videos. Figures reproduced with permission (2019, personal communication). (a) A patient researcher and (b) a young person with vitiligo and her mother demonstrate the use of the NB-UVB hand-held light device.

Use of narrowband UVB devices in the HI-Light Vitiligo trial

Before commencing trial recruitment, we undertook a photometric characterisation set of measurements as described in Study 1: photometric characterisation of the narrowband UVB devices prior to their use in the trial to achieve objectives 1, 2 and 4.

The HI-Light Vitiligo trial recruited 517 participants (children aged ≥ 5 years and adults) who were randomised into one of three parallel groups. In total, 425 live devices and 175 placebo devices were tested. This was more than the number of participants because of some tested devices not being suitable for participants, and anticipation of some devices requiring replacement during treatment. Participants (and their carers) were recruited at 16 secondary care sites around the UK and trained in how to use the NB-UVB device by watching a training video, and receipt of a written manual during a thorough face-to-face training with a research nurse. If a participant felt that the NB-UVB device was not working properly, they reported this to the co-ordinating clinical trials unit and a replacement unit was sent out from the central trial pharmacy. Faulty devices were returned by the participant and replaced. The faulty devices were sent back to the manufacturer.

Once the trial started recruitment, we performed the tests described in Study 2: quality assurance of devices prior to distribution to trial participants prior to issuing the devices, to achieve objective 3. All study tests were performed by a single team of scientists and technologists experienced in ultraviolet measurements.

Devices and test equipment

The hand-held NB-UVB device used in the HI-Light Vitiligo trial was the Dermfix 1000MX unit (Androv Medical, Leatherhead, UK). This unit is provided with a suggested dosing schedule, to be used after consultation with a supervising medical professional. We sought to develop a simplified dosing schedule that could be used in the trial. Initially the manufacturer supplied 10 Dermfix 1000MX units and two fluorescent tubes for characterisation [LightTech LTC 9W/G23 (LightTech Lamp Technology Ltd., Dunakeszi, Hungary) and Philips PL-S 9W/01/2P tubes (Koninklijke Philips N.V., Amsterdam, The Netherlands)]. Although there were small differences in spectral emissions (differing relative intensities at equivalent wavelengths) and a reduced output from the LightTech tube, the cost differential between the tubes was felt to outweigh these emission differences and so for all further characterisation and trial utilisation LightTech tubes were used in the hand-held devices. This is the standard configuration for this unit.

A Bentham DMc150 spectroradiometer (Bentham Instruments, Reading, UK), comprising a radiometer and double monochromator, was used to verify spectral outputs and an ILT 1700 radiometer (International Light Technologies, Peabody, MA, USA) was used for instantaneous irradiance measurements and integrated dose measurements. The spectroradiometer’s double monochromator and radiometer were calibrated against a mercury lamp with traceable spectral emissions. The ILT 1700 radiometer was field calibrated against the spectroradiometer.

To ensure consistency of output measurements, a jig to hold the Bentham and ILT sensors was designed and built by the Clinical Engineering Department at Nottingham University Hospitals NHS Trust (Figures 12 and 13). This ensured that the sensors were positioned over the centre of the lamp at the comb tip to consistently simulate desired clinical use.

FIGURE 12.

Elevation drawing of the measurement jig.

FIGURE 13.

Photographs showing three differing views of the test jig.

Study 1: characterisation of devices

Ten tubes from the same manufacturing batch were tested for spectral irradiance. The results of these measurements are shown in Figure 14. The results show almost exact coincidence of the spectral irradiance for all 10 tubes. They also show good coincidence with the manufacturer’s specified spectral irradiance. The main irradiance was at 313 nm, with subsidiary peaks at 365, 405 and 435 nm. There was a very small ultraviolet A light (UVA) component (365 nm).

FIGURE 14.

Spectral irradiance for the first batch of 10 tubes (spectral irradiance for LightTech tubes).

Table 32 shows the irradiance results from the sample of 10 tubes. These results are to be compared with the manufacturer’s specification (‘typical irradiance at 120 seconds’) at the tube’s mid-point at the comb tip of 7 mW per cm². These results showed that most tubes would fall within our acceptance criteria of an output ± 20% from the value used to inform the clinical treatment protocol.

The mean device output was only 53% of the manufacturer’s specified output. For clinical protocol derivation, a mean tube irradiance of 4 mW per cm² was used.

The results of simulating various participant protocols on the irradiance are shown in Table 33.

The table shows the average irradiance and percentage drop from the initial maximum average irradiance at various time points during the treatment protocol for skin type VI for three lesions (the maximum number of lesions to be assessed in the trial). It allows for the repeated turning on and off of the device on different days and, therefore, includes multiple warm-up times, as would be the case for a real participant treatment regimen.

Study 2: results of devices prior to issue to participants

Although the spectral irradiance test in study 2 was designed to ensure that no devices with gross set-up errors (e.g. incorrect tube wavelength) were issued to participants, it also allowed us to track changes in tube manufacturing. The peak irradiance wavelength was noted to change from 313 to 314 nm after about one-fifth of the batch processing. This slight change in the wavelength of the peak irradiance was not due to any calibration drift of our monochromator and was probably due to changes in the phosphor composition of the fluorescent tubes. This change did not produce any significant change in the amount of UVA radiation.

Twenty-one batches were tested and 9 out of 21 (43%) batches had devices where one or more devices showed an irradiance outside the range 3.22–4.82 mW per cm² (mean value ± 20%). A total of 54 live devices were rejected out of a total tested of 425 (13%). One batch (24 live devices) was tested where all devices lay outside the required range. After confirming that this was not due to faulty bulbs, combs or voltage supply, it was assumed that a power supply fault was responsible and the devices were returned to the manufacturer. The batch sizes ranged from 15 to 25 units.

Relevance to the wider literature

These results show that combination treatment with NB-UVB and potent TCS is more effective than a single intervention (in this case, TCS). This is consistent with previous research, which has shown that combination treatments are generally more effective than monotherapies in treating vitiligo, although overall response rates, both in our study and in previous research, are generally modest. 22,25,26

Although there have not been any studies assessing the same interventions as those used in this study, the response rates are comparable with other studies. A meta-analysis of studies assessing phototherapy for vitiligo,82 including 29 prospective studies of NB-UVB, reported a ‘marked response’ (> 75% repigmentation) in around 19% of participants after 6 months of NB-UVB monotherapy. This is similar to the rates of treatment success in our study, measured using the VNS (18% for NB-UVB only and 28% for combination at 6 months), although we observed lower success rates based on ≥ 75% repigmentation (5% for NB-UVB only and 11% for combination at 6 months). The same meta-analysis reported better response rates for vitiligo on the head and neck, which is consistent with our study. 82

No other studies have compared the specific combination of NB-UVB and mometasone furoate with mometasone alone, so direct comparison is difficult. One study comparing the combination of NB-UVB and clobetasol propionate (a more potent TCS) with NB-UVB alone83 was identified in the Cochrane systematic review of interventions for vitiligo. 22 This study suggested that combination treatment may be more effective than NB-UVB monotherapy, but the study was small and so lacked power to demonstrate any statistically significant difference between the intervention groups; the relative risk ratio for achieving > 75% repigmentation was 1.38 (95% CI 0.71 to 2.68). 83

No significant safety issues have been identified in previous small studies of home-based hand-held phototherapy devices for vitiligo, used instead of hospital NB-UVB therapy,25,26 and this is confirmed by the findings of our study. Long-term NB-UVB treatment (mean number of treatments = 211) in a study of patients with darker skin types conferred no increase in skin cancer risk, suggesting that NB-UVB can safely be continued for longer periods of time than in our study, although most patients in that study were skin types IV–VI. 84 Large cohort studies of patients having long-term treatment with NB-UVB have also shown no significant increased risk in skin cancer risk from this treatment. 38,39

Although combining NB-UVB and calcineurin inhibitors (e.g. tacrolimus) was discouraged in the past, because of concerns over a possible increased risk of skin cancer, a number of studies assessing this combination of treatment have been published over the last few years. A systematic review by Arora et al. 85 identified three studies comparing a combination of NB-UVB and tacrolimus with NB-UVB monotherapy. Meta-analysis of two of these studies showed combination treatment to be more effective than NB-UVB monotherapy in achieving > 75% repigmentation, although only just (risk ratio 1.34, 95% CI 1.05 to 1.71). It is possible that further studies comparing these interventions will provide sufficient data to make the confidence estimates stronger, but this remains to be seen.

Strengths and limitations

This was a large, pragmatic trial that was designed and managed in collaboration with an accredited clinical trials unit. Using a patient-reported primary outcome meant that treatment success reflected the views of people with vitiligo and was supported by blinded outcome assessment using digital images for both VNS and percentage repigmentation; both of which have been recommended for inclusion in vitiligo clinical trials by people with vitiligo. 35

As found in other vitiligo trials,22 retention throughout the trial was challenging. Just over 70% of participants provided primary outcome data at 9 months, and < 50% provided data by 21 months. This limited interpretation of some of the results, especially during the long-term follow-up phase.

Because loss to follow-up was higher than originally anticipated, the trial lacked power to provide a high level of precision around the point estimates.

Adherence to treatment regimens was quite low (see Table 6). This was probably due to the time burden of treatments, particularly the active or dummy NB-UVB devices. This is a limitation of the study but this was a pragmatic trial and treatments were delivered by the participant and/or their carers at home (with nursing support). It is possible that participants adhered more to trial interventions as a result of being in a trial, and this may have led to an overestimate of treatment effects. However, we think that, overall, the level of use of the treatments is reflective of how they would be used in real life.

We used a single, standardised treatment schedule, which we asked all participants to follow. This started at a very low dose and then built up to higher doses in small increments. This will have meant that for participants with darker skin types, the first few doses will have been lower than those used in conventional hospital-based phototherapy, where starting doses are determined by measuring the MED prior to starting treatment. However, participants would all move up the schedule to longer treatment times and over the course of treatment (up to 9 months) these smaller initial doses are not likely to have had a significant impact on the total NB-UVB dose received by those participants using active devices.

Participants were encouraged to choose a target patch that was genuinely the one in which they most wanted to see a difference. If they had two patches on different parts of the body that they were equally keen to see an improvement in, and if one of the patches was on their hands or feet, they were advised that the response may not be as good on the hands and feet, but they could still choose for the hand/foot patch to be their target patch. Many participants still chose a hand/foot patch as their target patch, but it is possible that some participants may have decided to change the target patch to one on the head and neck or rest of the body. This could have introduced bias into the study findings. However, this is similar to the situation in clinical practice, where a patient may be advised that treatment of vitiligo in certain anatomical locations may be less effective and that it may make more sense to concentrate on treating areas of vitiligo that are more likely to respond to treatment.

Generalisability

This trial has good external validity as it was a large, pragmatic trial with few exclusions, although all participants were required to have active vitiligo that affected < 10% of their body surface area. People with more extensive vitiligo are unlikely to find these interventions helpful as the treatments would become overly burdensome.

The trial included both children and adults and treated different body sites. Planned subgroup analyses explored the impact of these characteristics but found no evidence of differential treatment response by age or body site, other than the overall poorer response rates on the hands and feet. People with all skin types and ethnicities were included in the trial as this reflected the types of patients typically presenting for vitiligo treatment within the UK NHS. We did not exclude participants with lighter skin types (types I and II), as vitiligo can cause considerable distress in such people as well as those with darker skin types. 16,86 A post hoc analysis by skin type found no differential treatment response in people with paler skin types (types I–III) or those with darker skin types (types IV–VI), although we would emphasise that this was an exploratory, post hoc analysis and the study was not specifically powered for this analysis.

The trial was designed to reflect normal clinical practice as far as possible. Hand-held NB-UVB devices such as those tested in this study are not widely available within the UK NHS at present, although a few sites do offer treatment with similar devices and they can also be purchased online and used at the user’s own risk. In the trial, nurses in secondary care dermatology departments delivered training on use of the treatments, and participants were reviewed every 3 months during clinic visits. Additional support was provided by telephone as required, if participants had queries about use of the interventions or experienced side effects.

The process evaluation conducted alongside this study identified the importance of the support provided to participants, to enable them to use the treatments safely. Participants and investigators agreed that the complexity of the treatments meant that support and close monitoring were essential. Some participants had considered purchasing a light therapy device, but had decided against this because of a lack of the necessary support infrastructure. If the treatments were introduced into the NHS, the cost of providing this support infrastructure would need to be taken into account and health-care decision-makers would have to decide how much they are willing to pay to achieve a successful treatment. The relative lack of other treatment options, and the likely high cost of newer drug treatments currently being developed, would be important to consider when making such decisions.

Implications for health care

The HI-Light Vitiligo trial demonstrates that combination treatment with NB-UVB and potent TCS is superior to potent TCS alone, although the benefits are likely to be modest. Combination treatment was safe, well tolerated and was potentially cost-effective for people with limited vitiligo, but there is uncertainty over how much a decision-maker would be willing to pay to achieve an additional treatment success.

Patients starting vitiligo treatments should be made aware of the considerable time commitment required, and the likely duration of treatment over many months. Clinical review at 3 months appears to be an appropriate time point at which to judge whether or not further treatment is likely to be beneficial.

Our study confirmed that vitiligo on the hands and feet responds less well to treatment, so treating these anatomical areas may be difficult to justify when resources are limited.

Hand-held, home NB-UVB therapy appears to be a useful treatment option for people with vitiligo and provides considerable advantages over hospital NB-UVB therapy (which requires hospital visits two or three times per week). Home NB-UVB requires training and support from health-care professionals with experience of delivering phototherapy services and is time intensive for patients.

Use of mometasone furoate 0.1% ointment (a potent corticosteroid) as first-line treatment for vitiligo is supported, as it achieved treatment success in one in six individuals and was effective in stopping the spread of active vitiligo patches. Stopping the spread of vitiligo is an important treatment outcome to people with the condition. 21,33 These trial results suggest that potent TCS is safe in both adults and children when used 1 week on, 1 week off for 9 months.

Treatment effects were lost once interventions were stopped, suggesting that maintenance therapy is likely to be needed to prevent further loss of pigment.

Compared with potent TCS, combination treatment had a lower incremental cost-effectiveness ratio than NB-UVB monotherapy (meaning that an additional treatment success can be attained for a lower cost), although the mechanism for widespread implementation of a home-based NB-UVB service for skin disorders within the NHS has yet to be established.

Qualitative findings from our mixed-methods process evaluation study suggested that people with vitiligo and health-care professionals who treat them would value the provision of home NB-UVB as a useful treatment option for the management of vitiligo, despite the relatively modest treatment effects. Both trial participants and health-care professionals suggested that some form of ‘mixed economy’ may be the most effective way of providing home-based light therapy. This could potentially involve patients leasing or purchasing a phototherapy device, and the NHS providing the necessary training, quality assurance and support for patients. This would reduce the likely cost to the NHS (see Chapter 4, Sensitivity analysis) but would have equity implications in that treatment would only be accessible to those patients able to afford it.

These findings need to be disseminated to a wide audience. People seeking treatment for vitiligo are unlikely to receive any treatment if they do not receive appropriate advice from health-care professionals. In the UK, people with vitiligo are likely to consult a GP initially, and research among members of The Vitiligo Society suggest people view their GP as their primary source of information, although GPs appear to have low awareness of vitiligo. 87 The NICE Clinical Knowledge Summaries guideline suggests that people seeking treatment for vitiligo may be prescribed TCSs and/or referred to dermatology. 88 However, anecdotally, such management does not always seem to be followed. The safety data from this trial suggest that GPs can be reassured that adverse effects are rare if potent TCSs are used long-term once daily on alternate weeks (i.e. 1 week on, 1 week off).

Implications for research

Participants in the HI-Light Vitiligo trial reported relatively high quality-of-life scores at baseline using both generic and vitiligo-specific quality-of-life instruments. Despite having good quality of life, all participants were keen to access vitiligo treatments and were wiling to use them over many months, suggesting that something other than quality of life was motivating treatment choices. It is not clear whether this was because the trial focused on people with limited vitiligo (which had a limited impact on their quality of life), or the quality-of-life instruments themselves were insufficiently sensitive to detect the impact of vitiligo, particularly in relation to the psychological impact of the condition.

Because home-based phototherapy services for the management of skin disorders are currently available in only a small number of specialist centres, further research is required to establish the best ways of implementing a home-based light therapy service across the UK. This might usefully involve a hub and spoke model whereby specialist medical physics units perform the testing and maintenance of devices for a number of departments.

We used participant-reported treatment success as the primary outcome, based on the noticeability of the vitiligo (VNS), to ensure that vitiligo treatments were judged against criteria that are meaningful to people with vitiligo. Further work is required to establish the validity, responsiveness and interpretability of the VNS. In particular, it would be helpful to establish how patients value a ‘partial treatment’ response as measured by the VNS.

The HI-Light Vitiligo trial was designed to address two of the questions prioritised by health-care professionals and people with vitiligo in the James Lind Alliance Vitiligo Priority Setting Partnership. 21 Many of the top 10 priorities remain unanswered (Box 2).

About you

Are you a doctor/a nurse/a specialist dermatology nurse/other?

Prior to the HI-Light Vitiligo trial had you been involved in any form of phototherapy service? Yes/No.

What was your role in the HI-Light Vitiligo trial?

Are you already aware of the HI-Light Vitiligo trial results? Yes/No.

Question 1

Do you agree that home-based phototherapy should be made more widely available for vitiligo patients?

Strongly agree/agree/neutral/disagree/strongly disagree.

We appreciate that the HI-Light Vitiligo trial results will ultimately inform this decision, but at this point we would welcome your intuitive response.

Can you explain your response? What is it about home-based phototherapy (and your experiences as part of the HI-Light Vitiligo trial) that encourages, or discourages, you about its use?

Free-text response box . . .

Question 2

Do you think that home-based phototherapy is appropriate for all vitiligo patients?

All patients/most patients/some patients/few patients/no patients.

Could you explain your answer? What factors might influence whether a patient is appropriate for home-based phototherapy?

We would be interested to hear if you think that there are types of vitiligo presentation which are more, or less, appropriate for home-based phototherapy.

We would also be interested to hear if you think that lifestyle/personality/personal circumstance are important in this decision.

Free-text response box . . .

Question 3

Do you agree that delivering a home-based phototherapy service is feasible in the NHS?

Strongly agree/agree/neutral/disagree/strongly disagree.

We appreciate that ultimately this is a decision that commissioners will make, but we would invite your comment about the practical challenges that this might involve.

What were the difficulties and challenges that you found in delivering the HI-Light Vitiligo trial? Do you have any suggestions that would make a home-based phototherapy service easier to deliver or manage?

Free-text response box . . .

Question 4

Do you think that participants (and their families) found hand-held phototherapy easy to do at home?

Very easy/easy/neutral/difficult/very difficult.

We would be interested to hear about any difficulties or challenges that participants experienced with the hand-held phototherapy (and/or steroid cream).

We would be interested to hear about any strategies or techniques that participants used to manage, and about the nature of support that you offered them in this.

Free-text response box . . .

Question 5

How important do you think it is for any hand-held phototherapy devices to be provided and maintained by an NHS provider?

Can you explain why you think this?

Do you have any thoughts about patients purchasing their own hand-held phototherapy unit? Via the NHS? Via a commercial provider?

Free-text response box . . .

Question 6

Do you have any other comments, or recommendations, that would help others to establish and run a home-based phototherapy service for vitiligo?

Do you have any top-tips that you would like to share?

Free-text response box . . .