Pressure garment to prevent abnormal scarring after burn injury in adults and children: the PEGASUS feasibility RCT and mixed-methods study

Phase 1

Phase 1 consisted of a series of surveys and interviews with patients and health-care providers. The following objectives were assessed:

1. The current UK practice of PGT, including the number of patients receiving treatment, how garments are manufactured, costs of the garments to each NHS burns service and an estimate of the total costs to the NHS – data were collected by a web-based survey that was sent to all UK NHS burns units and data of the duration of PGT across the burn care providers were also collected. (Details and results are reported in Chapter 3.)

2. The number of children and adults eligible to be included in a PGT clinical trial in UK NHS burns units – data were collected by the same web-based survey used in objective 1. (Details and results are reported in Chapter 3.)

3. The views of HCPs [including consultants, nurses, physiotherapists and occupational therapists (OTs)] on conducting a RCT to compare PGT and no PGT, including attitudes to participation, perceived equipoise and barriers to, and facilitators of, undertaking and participating in a full-scale RCT of PGT – data were collected through a second web-based questionnaire, which was sent to individual HCPs, and in-depth interviews. (Results are reported in Chapter 4.)

4. The experience of patients and carers of PGT and their willingness to be randomised between PGT and no PGT – the impact of PGT on psychological and social functioning, body image and self-esteem and the psychological protection that some patients feel that PGT provides were explored. PGT treatment burden, perceived benefits and compliance (adherence to the treatment) were also assessed. Qualitative interviews were undertaken with patients who previously received PGT and/or their carers, either as face-to-face interviews or telephone interviews. (Details and results are reported in Chapter 5.)

5. Patient-centred outcomes that specifically measure post-burn scarring and appropriate methods of evaluating these outcomes – specific domains that are important to the patients, carers and HCPs were established in order to select the most suitable primary outcome for the full-scale trial. Data from the staff attitudes survey (objective 3) and from qualitative interviews with patients and carers (objective 4) and clinicians and HCPs were used. (Details and results are reported in Chapter 5.)

6. Appropriate methods for evaluating outcomes for the health economic analysis and resource use. (Details and results are reported in Chapter 6.)

Phase 2

Phase 2 was a multicentre pilot RCT that was conducted in four paediatric and six adult recruiting centres across England and Wales. This trial was an open two-arm trial with 12 months’ follow-up. Participants were randomised to PGT or no PGT, with both groups receiving other modalities of scar management, such as massage, moisturisation or silicone treatment.

The following objectives were assessed in the pilot RCT:

1. Recruitment rates and willingness to randomise between PGT and no PGT in multiple sites within the pilot RCT. (Details and results are reported in Chapter 7.)

2. Adherence with randomised allocation and PGT therapy. (Details and results are reported in Chapter 7.)

3. Possible processes for achieving blinded assessment of outcomes. (Details and results are reported in Chapter 7.)

4. Outcome distributions to inform a sample size calculation for a future trial. (Details and results are reported in Chapter 7.)

5. The perspectives, and experiences, of the patients, carers and HCPs who participated in the trial on trial participation and processes by qualitative process evaluation. (Details and results are reported in Chapter 8.)

6. The feasibility and success of the economic assessment to inform the assessment of cost-effectiveness in a future trial. (Details and results are reported in Chapter 9.)

The original study plan involved running P1 and P2 sequentially to allow findings from the qualitative work and surveys to influence the design of the pilot trial. However, the study duration was reduced on the request of the funding board and the two phases were run concurrently, which reduced the opportunity of findings from P1 to influence the design of the pilot trial in P2.

Responses

A total of 26 NHS burns services were identified that provided care to adults and 25 that provided care to children. Responses were received from 15 of the adult services (58%) and 15 (60%) of the child services (Table 1).

Numbers of patients receiving pressure garment therapy

Currently, within the UK, patients are treated for their burns injury within one of the three following NHS burns services:

1. burns facility – treatment of burns with a TBSA of < 10%

2. burns unit – treatment of burns with a TBSA of < 30% for adults and < 20% for children

3. burns centre – treatment of any size burn and all complex burns.

The number of eligible patients for PGT varied across each of the NHS burns services. From the 30 NHS burns services that responded to the survey, five (three for adults and two for children) are burn facilities (for a TBSA of < 10%), twelve (five for adults and seven for children) are burns units (for a TBSA of < 30% for adults and < 20% for children) and thirteen (seven for adults and six for children) are burns centres (for all percentages of TBSA).

Survey responses indicated that the number of eligible patients from the burn facilities ranged from 0 to 1 patients per month, burns units had 1–15 patients per month and burn centres had 2–18 patients per month. The services that responded to the survey estimated that, in total, 879 adults and 1104 children are treated each year, 55% in burns centres and 43% in burns units. Assuming similar patient numbers in facilities, units and centres that did not respond, we estimate that there are 1369 adults and 1476 children treated across the UK in burns services each year.

Source and cost of pressure garments

Information on garment source was available for all 26 services for adults and 25 of the services for children. Thirty-seven (72%) NHS burns services, the majority, choose to externally source pressure garments; of these services, 35 use the manufacturer Jobskin Ltd (Nottingham, UK), one DM Orthotics Ltd (Redruth, UK) and one Gottfried Medical, Inc. (Toledo, OH, USA). The remaining 14 (28%) NHS burns services provide an in-house service, which is mostly provided by the OT department, although one NHS burns service used their orthotics department.

Of the adult services, 31% produce pressure garments in-house; of the child services, 26% produce them in-house. The 30 services that completed the web-based survey consisted of eight with in-house pressure garment services and 22 with externally sourced pressure garment services, which were equally split between paediatric and adult services.

The approximate cost for the in-house garments per NHS burns service was £3237.50 per month, giving a yearly spend of £38,850 per service. The approximate cost for the external garments per NHS burns service was £4329.58 per month, giving a yearly spend of £51,995. We noted variation in the monthly spends per burns service, with a range of £300–5420 and a mean of £2986.

Extrapolating to all UK units based on the predicted patient numbers and standardised costs and garment numbers, and accounting for the split between internal and external provision, we estimate that the total NHS annual spend on pressure garments is £765,053 for adult services and £1,305,216 for paediatric services.

Funding for the provision of pressure garments varies depending on the burns unit and the hospital trust. The majority of the burns units obtain funding for pressure garments from the hospital budget, which is allocated to the following departments:

• occupational therapy department

• therapy department

• burns and plastics

• emergency budget.

Only three of the respondents received the pressure garment funding through their general practice.

Duration of pressure garment therapy

It is advised by the manufacturer and by the therapist that the patient wears the garment for 23 hours a day until the scars are fully mature. ‘Fully mature’ is a term that is used when the scars are no longer responding to pressure, appear pale in appearance when looking at vascularity, and do not alter in height when the garments are removed. When asking the question in regard to wear time, it was found that the time to scar maturity varied across the units and between patients. The length of time that patients are expected to wear pressure garments ranged across the burns centres from 3 months to 2 years, with the majority of burns centres recommending patients wear pressure garments for 12–18 months and with no difference noted between adult and paediatric recommendations.

Pressure garment selection

Pressure garment selection is routinely done by an experienced burns therapist who will consider many external factors (including physical and psychological factors) when selecting whether or not the patient is appropriate for PGT and the design of the garment. A wide range of designs are available both externally and internally, with bespoke garments often being designed and altered by the in-house technician or external manufacturer to ensure that the scar area is covered and the appropriate pressure is applied.

Within the survey we asked the NHS burns services ‘What guides you to select a particular garment?’. In 16 of the responses, the following criteria were listed as being important when selecting a particular garment:

• burn location

• burn size

• patient’s age

• functional ability

• dexterity.

In addition to these factors, one site mentioned that all patients who have suffered a burn injury are prescribed PGT regardless of factors such as scar size and location. A blanket referral into pressure garments is unusual, as the therapist will usually complete a subjective assessment as part of their initial assessment that will determine the scar therapy requirements of the individual patient.

Factors for selecting a particular garment did vary between adults and paediatrics; one NHS burns service stated that it would consider the patient’s type of work and whether or not certain pressure garments would be suitable in some environments.

Two of the paediatric sites considered factors relating to children and their age, such as the need for current or future toilet training, when selecting particular pressure garments.

Local protocols for scar management using PGT vary across the service but are transferable within networks and the wider scar management service throughout the UK. The British Burn Association’s Burn Therapists’ Special Interest Group allows scar management therapists to meet annually to discuss current practice and benchmark their service to ensure that a gold standard of practice is available and delivered. Scar management guidelines are used by each of the networks for therapists to read and follow as their service allows; many different scar modalities are available but finance and staffing has an impact on whether or not these are available in each of the UK burns services. It is important to note that, as a result of staffing levels, three of the burns services that responded to the web-based survey only partially completed the survey.

Online survey and analysis

A staff attitude survey was developed by the PEGASUS team in-house, given the lack of appropriate validated questionnaires in the literature (see Appendix 1). We sought input from our patient and public involvement (PPI) group during development, and pilot tested the survey with local burns staff, making revisions to language and ordering when necessary. The questionnaire, via fixed (e.g. yes/no, Likert scaling) and free-text responses, covered the following areas: (1) respondent demographics, (2) experience with PGT, (3) views on PGT for children and young people, and adults, (4) views on a full-scale RCT of PGT, (5) participation in a full-scale RCT of PGT, (6) patients’ (children and young people, parents, and adults) willingness to participate in a RCT and (7) clinical and non-clinical outcomes. Fixed-response data were analysed using simple descriptive statistics. Free-text responses were analysed using a simple content analysis (thematic) approach.

Survey sampling

Between October 2014 and March 2015, 30 burns services across England, Wales and Scotland were approached to participate in the survey (Figure 1). The survey link was sent out to all staff involved in PGT or scar management by the lead therapist in each service. Reminder e-mails were sent out up to three times, with all staff actively encouraged to participate within the data collection window.

FIGURE 1.

Map of PEGASUS P1 and P2 burns services in England, Scotland and Wales.

Qualitative telephone interviews and analysis

At the end of the online survey, participants were invited to provide their details to give permission for the PEGASUS research team to contact them to explore their survey responses in more detail. Respondents who left their details were then purposively sampled based on the following variables: (1) profession, (2) yes versus no response to a RCT for PGT and (3) agreement versus disagreement with attitudinal survey questions. A semistructured discussion guide was developed (see Appendix 2) based on the survey questions, the literature, discussions with our PPI group and the wider PEGASUS research team. Interviews were undertaken by telephone, audio-recorded, transcribed clean verbatim and analysed using a thematic approach. 45 Quotations in Results are identified using the participant’s unique identifier code and indicate whether the quotation was from an adult patient, a parent of a paediatric or adolescent patient or a member of NHS burns service staff. Staff quotations indicate profession and, when required, whether or not the quotation was from a survey free-text response.

Sample characteristics and response rates of survey and interview participants

Responses were received from 245 burns service staff from 28 out of 30 NHS burns services in England, Scotland and Wales that were approached to participate. Following cleaning of the data, 223 surveys were usable for analysis. It was not possible to accurately calculate the number of staff who were approached to participate within each service to calculate the response rate for the survey, but the mean number of respondents from each service was eight (range 1–24). Fifteen survey respondents from 13 burns services also participated in a semistructured telephone interview (average interview length was 32 minutes, range 19–41 minutes). Table 2 provides a summary of respondent sample characteristics for the survey and interview participants. Participants were predominantly female, nurses or therapists and caring for both adults and children.

Perspectives on pressure garments as part of scar management therapy

One hundred and sixty-seven of the survey respondents reported that they provided care either to adults only (n = 67) or to adults and children (n = 100). Table 3 provides a summary of staff perspectives on the use of PGT for scar management in adult patients. Staff reported that they agreed with the statements that PGT is one of the most important treatments (n = 123/167; 74%) and that it is beneficial to patients (n = 150/167; 90%), and just over half (n = 93/167; 56%) thought that it should remain the standard scar management treatment. However, 58% of staff (n = 97/167) were unsure about whether or not there was research evidence on the clinical effectiveness of PGT and some (n = 43/167; 26%) were uncertain of the benefits of PGT.

One hundred and fifty-six of the survey respondents reported that they provided care either to children only (n = 56) or to adults and children (n = 100). Table 3 provides a summary of staff perspectives of the use of PGT for scar management in paediatric patients. In a similar pattern to responses to the adult data, staff reported that they agreed with the statements that PGT is one of the most important treatments (n = 115/156; 74%) and that it is beneficial to patients (n = 145/156; 93%) and half (n = 77/156; 51%) thought that it should remain the standard scar management treatment. However, 60% of the staff (n = 92/156) were unsure about whether or not there was research evidence on the clinical effectiveness of PGT and some (n = 45/156; 29%) were uncertain of the benefits of PGT.

The majority of staff reported that they perceived pressure garments to be important and of benefit to patients, even though they were unsure whether or not there was research evidence of clinical effectiveness. Staff, often based on their own clinical experience, reported that pressure garments were important because of the benefits to scar management and the fact that PGT was better than other treatments, such as the use of silicone and massage. Reported benefits included (1) improving scar appearance, (2) helping to control symptoms such as itch, (3) helping functionality, (4) reducing the time it takes for the scar to mature and (5) providing physical (e.g. protection) and psychological (e.g. confidence and security) benefits:

It makes the scar paler, it makes it softer, it can help minimise itching, it can . . . where there are bands forming that make contractures it can hold those tightly against the body and minimise those, softens, flattens, pales, less lumpy, more pliable. So it’s about the colour, the texture and the height.

TLR05 OT

There were, however, some interviewees who articulated that they did not perceive any benefit of using pressure garments. Rather, they believed that the patients perceived a benefit as a result of the ‘scenario’ in clinical practice, in which patients are influenced by the treatment plan specified by the clinician and have an expectation of pressure garments:

Well, I don’t know whether they are benefits actually, I don’t think they are benefits. I think we have created a scenario where patients feel bad without them. So in other words if I gave them joss sticks and said if you use these three times a day and massage with some lavender and it’s really important you do that, and if you do that for 3 months non-stop they might get the same benefit, we don’t know. It’s that whole scenario about it. Even before the patient’s fully healed people start talking about we’ll measure you up for the pressure garments, and so it’s almost a done deal.

TEG15 consultant

A small number of interviewees expressed some uncertainty about the benefits of PGT. For example, benefits were reported as being dependent on the location of the scar, the type of scar and the characteristics of the patient. Some interviewees stated that the benefits were only apparent if the pressure garments were well fitted and the patients were compliant with treatment but for some patients, even if they were compliant, scar outcomes could be poor. There was discussion of the facts that scars ‘settle’ over time anyway, irrespective of the treatment, but that pressure garments, although perhaps not better than other treatments, were better than no treatment at all:

I think in some cases some patients do comply really well, and they still get some quite bad scarring and they are wearing a pressure garment. But you have no comparison to say well would that scarring have been a lot worse if they hadn’t worn the garment. There are some individual things I think does influence it, and I do think there are some quirks, so I think some children that absolutely comply to their pressure garment can still get a bad scar.

TSC06 OT

There was a split decision, to some extent, within the survey respondents about whether or not PGT should remain the standard treatment for scar management. Interviewees highlighted that they perceived there to be a role for pressure garments in scar management but that the decision to use pressure garments was dependent on clinical judgement and experience. This ‘judgement’ was also linked to the perceived benefits of PGT based on characteristics of the scar and the patient. The role of PGT in scar management and perceived benefits were particularly apparent for those working with children:

I think it’s a clinical judgement on what that scar is doing, how it’s behaving. But I definitely feel there’s a clear role for the use of pressure garments, and that for some, particularly I think the tricky more aggressive scars, I think pressure garments work better than some of the other options out there.

TSC06 OT, children only

Some interviewees also reflected that PGT should be part of a ‘toolkit’ of treatment options and that the pressure garments should only be used when required, based on clinical experience, rather than as standard care in which all patients are expected and expect to receive them:

I think it is [PGT] definitely a very important treatment modality to have in my toolkit, and I do use it.

TSA11 OT

When explored in further detail during the interviews, there was a range of knowledge about the research evidence on the clinical effectiveness of PGT. Some participants were fully aware that there is currently little evidence to show that PGT is clinically effective; however, others stated that the survey had prompted them to look at the evidence and led to the realisation that they had made an assumption about the existence of good evidence of clinical effectiveness. In addition, for some, they found in their own anecdotal experience that PGT was effective:

. . . this is proper research evidence as in . . . so this is the group I think in Canada that did a very good comprehensive review of pressure garment use a long time ago.

TEG15 consultant

This is very interesting, and I own this, this was a complete assumption that there was evidence, because everybody does it, and it was only after I had filled out your survey and then was having a conversation with one of my therapy colleagues who said actually that isn’t necessarily the case. So I was erroneously confident in my knowledge [laughs].

TCW12 clinical psychologist

Attitudes to a full-scale randomised controlled trial for pressure garment therapy

In response to the question ‘Do we need a full-scale RCT of PGT?’, 67% of staff participants (n = 149/223) answered yes, 8% (n = 17/223) answered no, and 26% (n = 57/223) were undecided. Table 4 shows the breakdown of responses to this question by professional group and pilot versus non-pilot site. Nearly all doctors (n = 26/29; 90%) and the majority of therapists (OT, n = 32/43, 75%; and physiotherapists, n = 24/32, 75%) identified a need for a trial; however, there was less certainty for nurses (n = 52/89; 58%) and other professional groups (n = 15/30; 50%). There was a similar proportion of pilot site (n = 61/88; 69%) and non-pilot site (n = 83/127; 65%) respondents reporting the need for a full-scale RCT of PGT. Although only 8% (n = 17/223) of survey respondents overall said that there was no need for a full-scale RCT, further exploration of these responses as part of the free-text survey and the interview data highlighted the complexity of the underpinning rationale. The rationale for responding yes to this question was centred, in the main, around providing evidence to support current PGT practice:

Yes, and what I really would like to see is a good trial that supports it, because my feeling is that if we run a good trial with enough numbers then we’ll give ourselves the evidence. So if we can get a NICE [National Institute for Health and Care Excellence] recommendation to use pressure garments on burns then we won’t face a situation in the future where the people holding the purse strings say well actually that’s not a proven technique to improve scar maturation so we’ll not pay for it. I am worried about the implications on our care in the future if we don’t get the evidence for doing it.

TCW01 consultant

Staff did, however, acknowledge that there was a need to challenge embedded practice that was not evidence based, particularly given the cost and treatment burden of PGT:

I think that it would be very helpful for everyone if we had objective evidence that pressure garments worked, because as time goes on, funding, etc., is going to be more and more dependent on evidence-based treatment, and I think that having strong evidence to support the pressure garments and the use of pressure garments, and therapy time to make garments or to measure, etc., and carry out the treatments, it’s always easier to justify if we’ve got some hard evidence.

TSC07 consultant

One consultant reflected that, even though they did not see the benefits of PGT, they were engaged in the study because they were concerned about the potential impact of changes to burns scar management on staffing and morale within the team:

. . . we wouldn’t be doing this project if I wasn’t, but generally to some extent it’s a problem [lack of PGT evidence] because I could take a very authoritarian view in my department and just stop using pressure garments. That puts one person out of a job, possibly two, and it would in one single act turn all of the burns therapists against me, because they have been taught, fed, aligned that these are effective, and they have categorical anecdotal proof. So I tread a careful line, and I’ve taken a view that I will stop people from wearing them once I review them in my clinic. But it’s an awful expense as well, because if you think about the number of clinical attendances, and the number of times the patient is seen by highly specialised staff, the garments that are worn, then actually it’s a serious cost and we could use that on something else.

TEG15 consultant

It was perceived that having ‘hard’ evidence might help to placate burn service management, give a definitive answer whether or not PGT should be used and would also to help to boost staff confidence in talking to patients about treatment options:

I am confident in the way I explain the things [pressure garments] to patients, and to advise them as to how these things work, and what I’ve seen from my experience about the other members of the team. But to actually have that hard evidence would just I suppose give us another backup to say if that is the case that we can show that to patients.

TWY10 physiotherapist

There were two main reasons why participants said that there was not a need for a RCT of PGT: first, they had limited awareness of any lack of clinical effectiveness evidence; and, second, they were happy with PGT as a treatment and were concerned, from an ethical point of view, that there would be a no-pressure-garment arm in a trial:

I guess it was interesting because I hadn’t even thought about people not being offered pressure therapy, and perhaps that was my ignorance, is that I had assumed that it was something that was a treatment that had been researched and proved to be something that it was clinically effective and that we should be using. So I was very much, I’ll own it, I was coming from a position of ignorance that hadn’t actually realised that wasn’t necessarily the case . . . Yes, because I was ignorant before [when responded no to a RCT in survey], I thought we had one, why would you do it again?

TCW12 clinical psychologist

Those staff who were undecided about a RCT reflected that their reasoning centred around staffing and ethical concerns. The latter may highlight a lack of equipoise. Participants raised concerns about not having enough staff to participate in a RCT and around the specific design of the RCT (see Chapter 8 for further discussion). The ethical concerns were similar to those who said no to a RCT for PGT. Staff, in particular nurses, struggled to reconcile the fact that by undertaking a RCT there was a chance that patients, who they believed should be in pressure garments, might not receive them, that is, ‘withholding’ the best treatment. This was particularly true for patients with certain scar characteristics, such as very large or ‘severe’ scars. Furthermore, the staff were concerned that patients would want the ‘best treatment’, which in their view is PGT:

Yes, mine was mainly from an ethical background of if it is seen, and I don’t know if it is or not, as the gold standard treatment now, how the study would take place, some people not having pressure garments, some people having pressure garments, I don’t really . . . because obviously with a study you need to provide the treatment that’s at least as good as the treatment you . . . or think is as good as the treatment you’ve got now, so I couldn’t get my head around, I couldn’t understand what could be as good as having pressure garments, because otherwise you’re disadvantaging perhaps some patients who wouldn’t have pressure garments to measure against.

TSW14 nurse

Attitudes to undertaking and participating in a full-scale randomised controlled trial for pressure garment therapy

Table 4 shows the proportion of staff and perceived proportion of burns services who would or would not be willing to participate in a full-scale RCT of PGT, split by profession and pilot versus non-pilot site. Overall, two-thirds of respondents (n = 143/223, 64%) reported that they and their NHS burns service (n = 144/223, 65%) would be willing to participate in a full-scale of RCT for PGT; however, nearly one-third did not know (staff, n = 65/223, 29%; NHS burns services, n = 75/223, 33%). When split by professional group, doctors (n = 23/29, 79%) and nurses (n = 60/89, 67%) were more likely to say yes to participating than other professional groups, for example physiotherapists (n = 18/32, 56%). Respondents from pilot sites were also more likely to say that they would participate than those from non-pilot sites (pilot site, n = 73/88, 83%; non-pilot site, n = 68/127, 54%). When asked whether or not their NHS burns services would be willing to participate, doctors (n = 21/29, 72%), nurses (n = 61/89, 69%) and physiotherapists (n = 21/32, 66%) were more likely to say yes than other groups. Similar to staff participation, respondents from pilot sites were more likely to state that their NHS burns service would take part than those from non-pilot sites (pilot site, n = 73/88, 83%; non-pilot site, n = 68/127, 55%).

As part of the interviews, the barriers to, and facilitators of, staff and service participation in a full-scale RCT of PGT were explored (Figure 2). Many of the factors identified, for example appropriate resourcing for staff, worked in both directions; if they were present they acted as a facilitator of participation and if they were absent they acted as a barrier. Linking with the concerns given in Attitudes to a full-scale randomised controlled trial for pressure garment therapy concerning the need for a full-scale RCT, staff reported their general willingness to be part of a future RCT, although there were some reservations, particularly around what involvement would actually entail, as well as the feasibility and practicability of undertaking this type of clinical effectiveness study for an embedded clinical treatment:

FIGURE 2.

Barriers to, and facilitators of, NHS burns staff and service participation in a full-scale RCT of PGT.

I was a bit unsure in what my role would . . . say we were to participate exactly what my role would be. Would I be the therapist that then does the assessment and then sends that off somewhere for it all to be compiled together? Because there would need to be training to make sure that we’re all rating the scars in the same way to show that they are improving or not improving, and I think at the moment people use different scales, people do things slightly differently. So I think it would be training and then support throughout the process.

TSC06 OT, children only

Yes, I think that’s reasonable, because I think at the moment the staffing levels that we have are very . . . we are struggling to meet the need. Obviously by the time the trial comes round it would depend on staffing levels and the situation of the service at the time, the changes that we’re looking to making we’re looking to . . . there may be an investment needed there as well from what we’re just trying to do within the department let alone adding to that. But I know that there is a lot of support for research and development, and if that was the case perhaps that would be something that the trust would consider.

TSA11 OT

I think it’s a very tricky thing to do [run a RCT of PGT], I don’t know how. Again just with it being such an emotive thing with children, to see . . . every parent wants the best for their child. I would be . . . I think we need to do something like this.

TRH04 physiotherapist, children only

Key barriers included ethical concerns over the ‘usual-care’ treatment group, patient characteristics and burn type, whether or not patients would agree to participate and, for those who did consent, would they comply with treatment and trial processes:

I would have a problem with that, because if they are consented . . . if they know that if they do this they will get this benefit, and if they don’t do it they won’t get that benefit I don’t think people who are fully understanding are going to consent to be in it. I think you will probably only get people who don’t quite understand what they are signing up to . . . I would much rather the study design assessed patients who did not comply with the treatment in the first place and look retrospectively at their outcome, and the assessors not knowing whether they have pressure garments or not.

TLR05 OT

I think, I thought I can see where it’s coming from, my brain . . . I can’t see how you can get across all the practicalities of or the barriers that might come up to actually prove it, because I would have thought across different centres you’re going to have to make sure that how you’re going to rate the scar. There’s so many questions about how scarring is assessed anyway and what scales to use, I just don’t . . . Well we’d have to be very clear that didn’t, because otherwise it would bias the study. That would just be one of my questions, is whether it would be better just to do this study with adults rather than including children.

TSC06 OT, children only

Data sources

Data for this analysis of willingness to participate in a trial, patient-centred outcomes and subsequent outcome domains were drawn from three different sources: (1) the free-text responses in the staff attitude survey (n = 223; 28/30 NHS burns services), (2) semistructured telephone interviews with a sample of the staff survey respondents (n = 15; 13 NHS burns services) and (3) semistructured interviews with adult patients and parents (n = 40, 24 adults and 16 parents; four NHS burns services). Please see Chapter 3 for a detailed description of the methods used for data sources (1) and (2).

Sampling and recruitment for adult and parent interviews

To be eligible to participate in an interview, patients or parents of paediatric (aged 0–8 years) and adolescent (aged 9–15 years) patients had to have ≥ 6 months’ experience of PGT and have finished PGT no more than 2 years prior to the interview. We attempted to include a range of patients according to sex, age, ethnicity and type and severity of burn to facilitate a maximum variation sample. Participants were recruited by OTs and/or research nurses (RNs) working at four of the PEGASUS pilot trial sites in England: Queen Elizabeth Hospital, Birmingham (adults only); Birmingham Children’s Hospital (BCH) (parents only); St Andrews Centre for Plastic Surgery and Burns, Broomfield Hospital, Essex (adults only); and Queen Victoria Hospital, East Grinstead (adults and parents). Clinical staff provided information sheets to potential interviewees and took written consent to pass contact details on to the PEGASUS qualitative research team. A member of the qualitative research team then contacted potential interviewees, provided further information and answered questions as necessary, and then arranged a suitable time, date and venue for the interview. Interviews were mainly conducted in the patient’s home, which was the preferred venue; however, a small number took place on university and hospital premises or via telephone. Written informed consent was taken prior to the start of interview data collection.

Originally, we had planned to also conduct two discussion groups with older children aged 9–16 years. After initial delays with ethics and governance approvals and site initiation, and considering the logistics and time needed to arrange these discussion groups, it was decided to conduct a small number of further interviews with parents of older children instead. This decision was taken with the input of the Trial Management Group (TMG) and Trial Steering Committee.

Interview format and content

A semistructured discussion guide was developed (see Appendix 3) based on the literature, discussions with our PPI group and the wider PEGASUS research team. Interviews were conducted in a participant-focused manner, allowing issues and perspectives that were important to participants to emerge naturally. The topic guide and interview process was refined after reflection on a small sample of early interviews. Topics discussed included:

• accounts of the accident and injury (when interviewees were happy to talk about these in order to provide context for the remainder of the data)

• accounts of subsequent treatment

• the experience of PGT and other scar management techniques

• hopes and expectations for treatment, recovery and scar management

• perspectives on a trial of PGT (e.g. would the interviewee have considered participating in a trial?)

• patient-centred outcomes.

Analysis

Interviews were undertaken either face to face or via telephone, audio-recorded, transcribed clean verbatim and analysed using a thematic approach. 45 Early analytic findings were discussed among the members of the TMG and shared for discussion and feedback at PEGASUS investigator meetings, at which clinical staff delivering the pilot trial and at least one patient representative were present. Following agreement of the final themes, example quotations were identified from each data source. Quotations in Results are identified using the participant’s unique identifier code, and indicate whether the quotation was from an adult patient, a parent of a paediatric or adolescent patient or a member of NHS burns service staff. Staff quotations indicate profession and, when required, whether or not the quotation was from a survey free-text response.

Sample characteristics

Table 5 describes and summarises respondent sample characteristics for the P1 adult patient and parents of paediatric or adolescent patient interviews. A total of 40 interviews with adults (n = 24) and parents (n = 16) were undertaken across four NHS burns services. Interviews lasted, on average, 51 minutes (range 19–108 minutes), with 33 conducted face to face and seven via telephone. Table 5 is supplemented with information about the burn injury, sex and age of the adult patients and the paediatric and adolescent patients whose parents were interviewed. The sample was approximately evenly split between males (52.5%) and females, across a range of ages. Most participants were white (77.5%), with the type of burn predominantly reported as flame (45.0% of total; 54.2% for adults and 31.3% for paediatric and adolescent patients) or scald (37.5% of total; 29.2% for adults and 50.0% for paediatric or adolescent patients).

Perspectives on a trial of pressure garment therapy

During interviews with adult patients and parents of paediatric or adolescent patients who had already experienced PGT, we discussed whether or not, in principle, they would have agreed to take part in a RCT had they been approached to do so following their initial acute treatment. Many of the participants stated that they believed that they would have agreed to take part, citing reasons that were similar to those later observed during the qualitative process evaluation of our pilot trial (see Chapter 8), for example altruism towards clinical staff and a desire to help generate knowledge that would benefit future patients. This was the case for most of the adult interviewees and approximately half of the parents that we spoke to:

I would have said yes, I would take part in it. I think because being in such a strange situation I think if it helps somebody else, and it’s going to do good whatever happens, so basically I see myself as lucky, because my scars are covered virtually. Seeing me without looking just about there you can’t really tell, so I’m a lucky person, that’s how I put myself down as. But if you have got on your face, or you have got somewhere that really stands out you need as much help as you can possibly get. If by doing a trial helps somebody with something facial or whatever well brilliant.

How would you have felt having this 50/50 chance?

Yes, fine, I would have taken that, I would yes, I would have said yes whatever.

CA04, adult patient

However, when considering this during interviews, several parents felt that they would have been reluctant to take part or would have definitely refused to do so. The reasons given reflect our observations later during the qualitative process evaluation of our pilot trial (see Chapter 8), and centre upon the need for parents to feel that their children were receiving the best treatment available. Interestingly, this parent reflected on the potential for differences in attitudes towards the trial concerning the participation of adults and children:

I would have said no.

That’s fine, I need to know.

I don’t like 50/50 chances [laughs].

So you’d have said no and you would have taken . . .

I think that idea is more better for adults, and then you would say well we’ve tried it on adults, we would like to try it on children, because for a child to do a 50/50 . . . I wouldn’t risk my child like that, I would be like you do what’s best for my child. I wouldn’t be thinking if that could have worked, and I didn’t use it I feel like I would regret it. So I wouldn’t want to make that decision for my child. I’d do it [take part in the trial] for probably me . . .

BC02, parent of paediatric patient

A small number of participants were unsure as to how exactly they would have reacted to a request to consider taking part in a trial.

General outcome measure and domain identification

A large number of burn scar-specific outcomes were reported by, and perceived to be important to, patients, parents and clinicians:

I feel all outcomes are important and valid and should include clinical observations and patient-reported outcomes. I think it is very important to look at scar appearance, pain and itch but also include function, body image and impact on life as using pressure garments affects all these aspects.

ID25, survey free-text response, physiotherapist

I think from the doctors themselves they measure the size of the scar, how deep it is, how firm it is, the colouring, and they go from that, and how he is, is he scratching it, is he pulling at the vest, is he doing . . . they ask all sorts of questions, so probably everything that they have asked is going to be asked in the trial. There’s nothing that I can think of on top of that.

EC03, parent of paediatric patient

Several of the NHS burns service staff were able to identify and articulate the importance of outcome measures that focus on how scar management affects the patient rather than ones that they perceived as being easy to measure objectively. One consultant, for example, highlighted the significance of identifying outcomes and using outcome measures that are important to the patient and their ability to return to ‘normal life’:

The main focus is really how does it affect the patient? So the things that I think I picked [in the survey: itch, function, ability to work, psychological state, social interaction with family and peers and cost] were related to the impact that it will have on the patient. It’s sometimes very easy to focus on . . . very easy to measure objective data, but if it actually makes no difference to the patient in terms of their ability to work, the symptoms that they have, the difficulties or the number of visits that they have to make, well then ‘so what?’ is what I would say. So if they score seven on a scar score but they’re working and they haven’t had to visit the hospital at all because they’re getting back on with their lives compared to scoring five, which is a better score, but they’ve had to make 17 visits and haven’t gone back to work, is that a success? I don’t think so. I think it’s about how we can pick outcomes that are related to returning to normal life, not getting good objective scores. So yes, I think what will actually . . . as with any study you need to balance what are things that are easy to measure with things that matter to the patient . . .

TEG15, consultant

Another member of staff, who responded to the free-text questions about outcomes in the survey, again highlighted the importance of focusing on patient-reported, rather than clinically reported, outcomes:

As with most outcomes, I believe patient-reported ones that focus more on well-being, function and psychosocial status, etc. should be more influential than pure[ly] ‘clinical’ ones.

ID28, survey free-text response, consultant

Identification of outcome domains

Five core outcome domains, covering the list of burn scar-specific outcomes, were identified within the data:

1. scar characteristics and appearance

2. movement and function

3. itch and pain

4. psychological and social functioning

5. treatment burden.

Health-related quality of life

The wide age range of patients recruited for the study was recognised as a complication from the outset, necessitating the use of multiple instruments and piloting of HRQoL instruments that are not yet widely used in health economics. Furthermore, it was acknowledged that the small sample of patients completing each instrument (i.e. falling into the age range for which each instrument had been validated) would limit full evaluation of these instruments.

In the National Institute for Health and Care Excellence (NICE) reference case, the EuroQol-5 Dimensions (EQ-5D) is specified as the preferred measure of HRQoL in adults. 46 The EuroQol-5 Dimensions, five-level version (EQ-5D-5L), was therefore chosen as the sole health economics instrument in patients aged ≥ 17 years. The timing of the data collection would coincide with the release of the value set for the new EQ-5D-5L instrument and it was therefore planned to use corresponding preference values from that publication. 47

In the case of assessing HRQoL in children, NICE acknowledges that ‘when necessary, consideration should be given to alternative standardised and validated preference-based measures of HRQoL that have been designed specifically for use in children’. 46

A 2014 systematic review48 of UK economic evaluations of paediatric interventions was critical of published economic evaluations for failing to make use of instruments specifically developed for use with children, and identified the Child Health Utility Index 9D (CHU-9D) as a potentially useful measure for future research. Although CHU-9D had not been used in the economic evaluation studies in this review, empirical work from the West Midlands (UK) demonstrated the superiority of CHU-9D over the EuroQol-5 Dimensions, Youth (EQ-5D-Y) for children aged 6–7 years, and CHU-9D has been validated for use with children aged 11–17 years. 49,50 We therefore selected the CHU-9D for use in this study as the sole health economics instrument for patients aged 5–6 years with proxy completion by the parent, for both child and parent/proxy completion for ages 7–10 years and for child completion only in patients aged 11–15 years. Young people aged 16 years also completed the CHU-9D alongside the EQ-5D-5L. Dual completion by parents and children in the 7–10 years age range presented us with the opportunity to assess whether or not child completion would be feasible as part of a full study.

Although there were initially no plans to include an instrument to assess HRQoL in patients aged < 5 years (the question on CHU-9D about ‘school work/homework’ suggests a natural cut-off point at 5 years as a minimum age in the UK), feedback from clinical colleagues stressed the significant prevalence of burn injuries within younger age groups (the original lower age range for the trial was removed as part of the first amendment, which was submitted to the Research Ethics Committee in September 2014). A decision was therefore taken to introduce the Pediatric Quality of Life Inventory (PedsQL), after the start date of the study, for proxy completion by parents on behalf of children aged 2–4 years. It should be noted that no HRQoL or health economics instrument was identified for ages 0–1 year.

Although there is currently no set of tariff or preference values for PedsQL, we were aware of plans to predict CHU-9D utility values from PedsQL scores using regression mapping algorithms, as part of the National Institute for Health Research (NIHR) Health Technology Assessment (HTA)-funded PREDnisolone in NephrOtic Syndrome (PREDNOS) study (reference number 08/53/31). 51 It was therefore expected that a mapping algorithm would be in use in time for any full trial of PGT.

Costing/resource use

At the time of preparing the trial protocol, a search of the Database of Instruments for Resource Use Measurement (DIRUM) database (www.dirum.org; accessed 30 January 2014) revealed no previous costing/resource use questionnaires for burn treatments within the disease subcategories of ‘wounds’, ‘skin’ or ‘child health’. A new resource use questionnaire was therefore designed, but drew to some extent on previous work relating to leg ulcers and eczema (see Appendix 4). 52,53

A decision was taken to explore the feasibility of collecting cost data from a societal and a NHS perspective. It was expected that patients in both treatment arms may potentially need to travel long distances to receive care at central locations (particularly those travelling in from more rural areas) and that, outside of the trial environment, if PGT is not received, there may be fewer visits to the clinic. As well as direct travel costs, long journey times may necessitate time off work for both adult patients and parents. It was also expected that those patients not receiving PGT may purchase alternative over-the-counter dressings, bandages, tightly fitting clothing or ointments, hence shifting the cost, to some extent, from the NHS to patient or parent. Based upon feedback from our patient and public representatives, we considered the possibility that if patients deem a pressure garment to offer a degree of protection (versus no pressure garment), they may return to work or have increased productivity at work earlier, particularly in those in more manual professions. A decision was therefore taken to include questions relating to employment and presenteeism (reduced productivity in the workplace as a result of health problems).

A decision was taken not to ask patients or parents to self-report income, as there was concern that they may be reluctant to provide such personal information; we did, however, obtain data on occupation.

The wide range of ages necessitated the development of two versions of the resource use questionnaire, one for adults (defined in this context as aged ≥ 16 years to coincide with the adoption of the EQ-5D-5L), and one for parent completion in the case of children (aged ≤ 15 years).

The number of visits to the burns clinic was specified in the trial protocol, as was the management by the site of massage, moisturisation and the application of silicone gel. The location of the pressure garment and the date on which pressure garments were fitted was recorded by staff at the treatment centres via the case report form (CRF). It was planned that cost data relating to different types of pressure garment would be obtained from a commercial price list used by the majority of the sites.

Willingness to pay

The inclusion of the WTP questionnaire was intended both to (1) value the impact of the treatment using a single, monetary metric, across all age ranges, given that – ultimately – two different sets of utility values would be elicited (from EQ-5D-5L and from CHU-9D – directly and indirectly) and (2) allow patients/parents to value outcomes that they have reason to consider relevant and important. This was especially important as the HRQoL instruments could only crudely incorporate outcomes, such as the patient’s own perceptions of changes in scar appearance, or aspects of treatment such as itch; information and compassionate care provided by staff; and inconvenience. The value of such aspects of treatment (Mooney54 also suggests that these may include dignity and autonomy) can be thought of in terms of ‘process utility’. 54 This is not a limitation of the measures as such, but instead stems from a normative stance within cost–utility analysis that the sole objective is to maximise health outcomes.

Two versions of the WTP questionnaire were developed, for PGT and no-PGT patients (see Appendices 5 and 6), both to be administered at the 6-month follow-up. Both groups were asked to value the impact of the treatment that they had actually received (an ex post perspective), although the PGT patients were also provided with (or rather were reminded of, given that they had received an information leaflet at the time of recruitment) a short description of the difference between the two treatment arms, and asked the additional question: ‘If you could pay a smaller amount per month to receive the same treatment but without the tight fitting Lycra garment, would you choose to do so?’, with the follow-up question, if yes, of ‘I would pay £__ per month for this treatment’.

The 6-month follow-up was chosen to give sufficient time for patients or parents to experience treatment because not all patients were followed up at 12 months. Owing to resource constraints, it was planned that the questionnaire would be given out at the centre (by the usual centre staff) with the ‘standard’ questionnaires.

Trial design

The second phase of the PEGASUS study, P2, was an open, pilot, two-arm RCT with a maximum follow-up period of 12 months. Participants were randomised to PGT or no PGT and both groups were allowed standard massage, moisturisation and silicone treatments. The trial aimed to recruit in seven specialist burns units in the UK.

Participants

The study recruited participants of any age who were at risk of abnormal scarring following burn injury and had been referred to one of the specialist burns services. Following standard practice, consideration of scar management started before the wound had healed. When 90% wound healing had occurred (following visual assessment by a therapist), patients attended an initial appointment during which they were screened for eligibility by the local research team.

When the eligible patient’s burn was assessed as having healed, the clinical team introduced the concept of the PEGASUS study, explaining the treatment options available and the level of evidence behind each option. The clinical team provided the eligible patient or their parent or carer with the Research Ethics Committee-approved patient information leaflet (PIL) on NHS trust headed paper. Potential participants then had 24 hours to read the PIL, discuss it with their family, ask questions and decide whether or not they would like to consent to participate in the trial. Once consented, patients were randomised to either PGT or no PGT. All patients were allowed to receive other modalities of scar management such as massage, moisturisation or silicone treatment.

Randomisation

After informed consent had been obtained, participants were randomised into the trial via a telephone randomisation service provided by the Birmingham Clinical Trials Unit (BCTU). The randomisation schedule was generated by computer using permuted random blocks of variable length to allocate eligible patients in a 1 : 1 ratio to either the intervention (PGT) or the control (no-PGT) arms. No stratification was employed in the allocation.

A trial number was allocated only after eligibility had been confirmed via the inclusion and exclusion criteria on the randomisation form and patient consent had been given. Once a trial number was allocated, a confirmatory e-mail was sent to the local principal investigator (PI) and the named RN or scar management therapist. With the participant’s permission, the GP was notified using the standard letter provided for this purpose.

Health technologies being assessed

Participants in both arms of the trial were allocated to receive standard education concerning scar management techniques from their clinical team. This typically involves demonstration of massage techniques, recommendation to undertake massage three or four times per day with moisturiser, stretching and other exercises, sun cream advice and a discussion concerning returning to normal functioning. To minimise treatment difference between the trial arms, participants were allowed to receive silicone products, including gels, sprays or sheets, from month 1 in the PGT arm and week 1 in the no-PGT arm, if deemed appropriate.

Participants in the PGT arm received pressure garments. Each centre in the trial used its standard supply route, either purchasing garments externally or manufacturing its own. Participants were measured for their garments after randomisation and were scheduled to return to the burns clinic for garment fitting 1 week later, at which point any alterations could be made. To avoid interruption of the application of the pressure garment for the recommended 23 hours a day, a second garment was supplied if needed. The fit and wear of the garments was checked throughout the trial at the routine clinic visits and garments were replaced as required.

All participants were required to attend routine follow-up scheduled appointments, with additional appointments allowed (for issues such as vulnerable adults and children), in line with standard management practice in each centre and not as part of the trial appointments.

Study setting

To evaluate the ability to recruit across multiple sites, seven specialist NHS burns centres were invited to participate: two centres that specialised in treating children, three sites that treated only adults and two that treated both. The five sites treating only adults or only children were:

1. BCH NHS Trust (children)

2. Central Manchester University Hospitals NHS Foundation Trust (Manchester) (children)

3. North Bristol NHS Trust (Bristol) (adults)

4. University Hospitals Birmingham (UHB) NHS Foundation Trust (adults)

5. The Welsh Centre for Burns and Plastic Surgery, Abertawe Bro Morgannwg University Health Board (Swansea) (adults).

The two sites treating both adults and children were:

1. Queen Victoria Hospital NHS Foundation Trust (East Grinstead)

2. Mid Essex Hospital Services NHS Trust (Essex).

Study procedures and schedule of assessments

The schedule of assessments for the trial is detailed below and summarised in Appendix 7, Table 27. In the original study protocol, it was intended that, as information on valid outcome measures emerged from P1 of this project, it would be desirable to refine the outcome measures collected in P2. However, with the delayed start to recruitment and the reduced follow-up time of the trial, there was insufficient separation between P1 and P2 to allow new outcome measures to be identified and included. The assessments made in the RCT were as follows.

Blinding

Owing to the nature of the intervention, it was not possible to blind either the participants or their parents/carers to the allocated treatment. The clinician/therapist scales (VSS and POSAS observer-reported scale) were completed by the therapist at the site, who is likely to have been aware of treatment allocation. The feasibility of potential methods for introducing blinding in a future trial was discussed at investigator meetings during the trial.

Sample size

As this is a feasibility study, it was not powered to detect any clinically meaningful difference between the PGT and no-PGT arms in any of the recorded outcomes. The principal objective of the feasibility component was to obtain reliable estimates of the rates of recruitment and retention ahead of a full-scale RCT. This study aimed to recruit at least 44 participants in each group, a total of 88 participants over an estimated 9-month recruitment period. This would allow the recruitment rate to be estimated with a 95% confidence interval (CI) and with a maximum width of 20%.

Statistical methods

Outcomes of interest are described using appropriate summary statistics. For continuous outcomes, means and standard deviations (SDs) are used if the distributions are approximately normal; medians and interquartile ranges (IQRs) are used if they are not. For categorical outcomes, summary statistics consist of numbers and percentages. The changes from baseline to the follow-up assessments at week 1 and at months 1, 3, 6 and 12 were calculated for each of the treatment arms, along with the IQRs. Estimates of the variability of any outcome measures identified as being suitable for a main trial are also calculated to inform sample size calculations for a future full-scale RCT. Recruitment rates are given as an average number of patients per centre per month. Dropout rates are expressed as a percentage separately for each arm, along with the reasons for dropout.

All analyses were performed on the intention-to-treat principle.

The data analyses were generated using SAS^® software (version 9.4; SAS Institute Inc., Cary, NC, USA). SAS and all other SAS Institute Inc. product or service names are registered trademarks or trademarks of SAS Institute Inc. in the USA and other countries. ^® indicates US registration.

Ethics approval and trial registration

The study was reviewed by the National Research Ethics Service Committee West Midlands – Coventry and Warwickshire on 28 May 2014. Following clarification of the content of the PIL, refinement of the adult consent form and further details of the composition of the monitoring committee, these amendments were submitted on 27 June 2014 and subsequently approved. The PEGASUS trial was given a favourable opinion by the committee on 8 July 2014.

Participant screening and recruitment

The study was scheduled to open in September 2014 and to recruit for 9 months. Owing to delays in contract negotiations at sponsor level and gaining local approval at sites, the study opened to recruitment 4 months late, in December 2015. Support was obtained from the HTA programme manager in identifying strategies to resolve these protracted delays and engaging with the sponsor to help expedite the approval process. Recruitment opened on 1 December 2014 and ended, 9 months later, on 1 September 2015.

However, the original end date of the trial was not changed and follow-up ended for all participants by 31 May 2016. This meant that, although all participants were followed up for ≥ 6 months, only 56 out of the 88 participants (those recruited prior to 31 May 2015) were followed up as planned for 12 months.

The first participant was recruited to the trial on 6 January 2015, and six out of seven sites recruited their first participant within 3 weeks of opening. Recruitment of the 88 participants was completed on 19 August 2015, within the allotted 9 months, with 50% of the target attained in 4 months. During this 9-month period of recruitment, a total of 211 patients were screened, with 186 (88%) deemed eligible and 88 (47% of eligible patients) recruited into the trial (Figure 3). The eligibility rate of 88% (95% CI 83% to 92%) suggests that > 80% of screened patients would be eligible in a large-scale RCT. Of the 25 patients screened but deemed ineligible, nine (36%) were considered not suitable for PGT, seven (28%) were excluded because of their medical history (e.g. diabetes, memory loss, unstable personality), four (16%) were not suitable for screening, two (8%) required PGT, two (8%) were unlikely to comply with PGT and one (4%) was still undergoing treatment as the wound had not fully healed.

FIGURE 3.

Recruitment to the trial.

The recruitment of adult and paediatric patients is summarised in Table 7.

Consent to participation in the trial

The consent rate was 47% (95% CI 40% to 55%) among eligible patients and 42% (95% CI 35% to 49%) among all screened patients. Reasons for non-participation were offered by 19 out of the 98 eligible patients who did not give consent: six (6%) patients did not want PGT, six (6%) wanted PGT, one (1%) did not want to be randomised and six (6%) offered other individual reasons. The study team judged 10 (10%) further patients as unable to give informed consent, 12 (12%) were unable to speak/understand English and 10 (10%) were unlikely to complete follow-up. Thirty (31%) were recorded as simply declining participation and 12 (12%) as not stating any reason. A further five (5%) were not recruited because of staffing issues.

The main reason for failure at three of the sites recruiting paediatric patients (Central Manchester University Hospitals NHS Foundation Trust, Mid Essex Hospital Services NHS Trust and BCH NHS Trust) was because of the patient or parent declining to participate. This accounted for 41–59% of total screening failures at these centres compared with 13–24% at two of the sites recruiting adults only (North Bristol NHS Trust and UHB NHS Foundation Trust). The main reason for screening failure at these sites was lack of suitability for the trial. The other adult site, The Welsh Centre for Burns and Plastic Surgery, Abertawe Bro Morgannwg University Health Board, had 33% of screening failures because of staffing resources and 40% because patients declined to enter the trial. The Queen Victoria Hospital NHS Foundation Trust did not complete screening logs fully, so no information is available. Adequate staff provision and training is a key element to delivering any future full-scale trial of this type. Reasons for patients choosing to consent or decline their participation in the trial are analysed in detail through the qualitative work in Chapter 8.

Randomisation

Participants were equally allocated between study arms. Baseline demographic and clinical characteristics by allocated treatment are provided in Table 8. The majority of recorded burns were of either thermal flame or thermal scald aetiology (77% and 70% in the no-PGT and PGT arms, respectively). Few participants recorded any other injuries.

Withdrawals

Five out of 88 (6%) participants were withdrawn: four who were initially allocated to the no-PGT treatment arm and one to PGT treatment. The reasons for withdrawal for the four participants who were randomised to the no-PGT treatment arm were as follows:

1. The participant asked to be treated with pressure garments, as they felt that their scar was becoming lumpy and looking worse.

2. Study investigators at the site decided to withdraw the participant from the no-PGT treatment arm and started PGT for all three monitored scars.

3. The participant was elderly and relied on friends to travel to appointments. The participant did not wish to travel for any follow-up appointments and wished to choose the treatment that could be monitored at long intervals between follow-up appointments. The participant withdrew from the study before completion of baseline assessments.

4. The participant attended the initial baseline assessment but cancelled the first follow-up visit and then withdrew from the study and any further follow-up. The participant met the inclusion criteria but only had minimal evidence of scarring.

The one participant who withdrew from the PGT treatment arm withdrew from the study prior to the week 1 visit but gave no reason.

There was evidence of site confusion regarding the correct trial procedure for withdrawals in the event of withdrawal from treatment, as participants should have continued in the study (this applied to the first and second withdrawals listed above). Inadequate screening may also have led to the third participant inappropriately being recruited and randomised.

Lost to follow-up

Fourteen participants (16%) were lost to follow-up (LTFU): seven in each treatment arm. A further PGT participant did not attend follow-up at 6 months and was not seen at 12 months because the study had terminated before this follow-up point had been reached. Counting this participant as LTFU gives a LTFU rate of 17%. Reasons for being LTFU were known for only three of these participants: one was inappropriately included and two moved away from the area. Loss to follow-up occurred after month 1 in 11 out of the 15 participants (73%). Loss to follow-up showed variation between centres (see Table 7), although results are based on small numbers. The North Bristol NHS Trust and Queen Victoria Hospital NHS Foundation Trust had the highest losses to follow-up, of 36% and 38%, respectively, whereas both specialist children’s hospitals (BCH NHS Trust and Central Manchester University Hospitals NHS Foundation Trust) had no participants LTFU. The nature of the participant population at the centres (e.g. adult vs. paediatric participants) may relate to the levels of loss to follow-up, although at both the Mid Essex Hospital Services NHS Trust and The Welsh Centre for Burns and Plastic Surgery, Abertawe Bro Morgannwg University Health Board, only a single participant was LTFU. The possible impact of clinical equipoise on participant retention is discussed in detail in Chapter 8.

Owing to the reduced study timeline, all participants were able to be followed up for 6 months but not all could be followed up for the full 12-month period. Sixty-seven of the 88 participants (76%, 95% CI 66% to 85%) completed their 6-month visit: 32 (73%, 95% CI 57% to 85%) participants in the no-PGT arm and 35 (80%, 95% CI 65% to 90%) in the PGT arm. Of the 21 participants who were not assessed at 6 months, 14 were defined as LTFU, five were withdrawals, one had been discharged and one did not attend at 6 months and the study terminated before they reached 12 months. The estimated rate of loss to follow-up by 6 months is 16% (95% CI 9% to 25%). Aggregating losses to follow-up and withdrawals up to month 6, a total of 19 participants dropped out of the study; hence, the estimated drop-out rate is 22% (95% CI 14% to 32%).

The participant flow through the trial is shown in the Consolidated Standards of Reporting Trials (CONSORT) flow diagram (Figure 4).

FIGURE 4.

Consolidated Standards of Reporting Trials flow diagram.

Adherence and crossover

Pressure garment therapy assessments

For participants randomised to the PGT treatment arm of the trial, details regarding the manufacture and fitting/refitting of the pressure garments were collected at each visit. Table 11 summarises the pressure garment provision at baseline. The two participants who did not respond to this question were using version 1.0 of the baseline form, which did not include this particular question. Of the 42 respondents, all were measured for pressure garments, with most pressure garments being produced externally [24/42 (57%)] compared with being produced in-house [17/42 (40%)], with information not being recorded for one participant. Nearly half of the PGT treatment arm participants, 20 out of 42 (48%), had not been fitted with their pressure garments for their main scar (scar site 1) at baseline.

Pressure garments were measured, fitted or refitted for 17 out of 39 (44%) participants at month 1, 28 out of 36 (78%) at month 3 and 23 out of 35 (66%) at month 6. The number of participants on the PGT treatment arm having a new pressure garment fitted to at least one of their scars was 17 out of 42 (40%) at baseline, 31 out of 41 (76%) at week 1, 15 out of 39 (38%) at month 1, 24 out of 36 (67%) at month 3, 20 out of 35 (57%) at month 6 and 7 out of 21 (33%) at month 12.

The question wording used in the pilot trial did not clearly distinguish between remeasurement, refitting and replacing garments, which caused some confusion in the responses. Assessment in a main trial requires separately recording these three different processes to fully assess the resource used. It is also important to distinguish between garments for the multiple scars that may be treated at the same time.

Blinding

Assessment of the feasibility of achieving blinded assessment of outcomes was one of the study objectives. The possibilities of undertaking each of the outcome assessments while being blinded to participant allocation was discussed with the therapists at the investigator meetings during the trial.

During the pilot trial, the physical presence of a pressure garment prohibited the blinding of either the participants or their parents/carers to the allocated treatment arm, and the study researcher who treated the participants at their follow-up assessments could not be blinded to treatment allocation because of the small number of adequately trained staff at each site. For most participants, a single researcher would measure and refit the pressure garment, if necessary, and conduct all assessments at follow-up appointments.

To perform a blinded assessment of scar appearance, a second study researcher, unaware of the treatment allocation, would need to be present at each site to complete the scar assessment forms. However, the therapists indicated that this would not guarantee blinded outcome assessment, as pressure garments are, by design, tight-fitting garments that leave indentations or markings on the skin. They would need to be removed for a considerable length of time before any blinded assessment could be made. This was viewed as being unnecessarily burdensome to the participants.

Although clinical photography was considered as a potential method of allowing blinded assessment of scar appearance, obtaining photographs was logistically difficult, as the variability in available facilities at different study sites (e.g. medical imaging vs. a handheld digital camera) meant that standardisation of the photographs was complex and lay outside the standard duties and experience of the study researcher performing the assessments. The requirement for use of medical imaging at some sites also proved prohibitively expensive for this small feasibility study. Moreover, clinical photography would still be subject to the problem of indentations or marks being left on the skin of participants who used pressure garments.

Outcomes

The PEGASUS study was not powered to provide definitive data to assess the clinical effectiveness of PGT compared with no PGT. However, we report the outcome data collected for incorporation in future meta-analyses, as well as to assess the feasibility of collecting outcome data on the chosen scales.

Global scar appearance

For all summaries of global scar appearance, the data are based on the scar that was identified as scar 1 on the baseline assessment. Results for other monitored scars were similar and are not presented here. Table 12 summarises global scar appearance at baseline by treatment arm, as measured using the clinician-reported VSS.

TABLE 12 - Vancouver Scar Scale scores at baseline by treatment arm

VSS domain	Treatment arm
	No PGT (n = 42)	PGT (n = 43)
VSS, median score (IQR)
Pigmentation (scored 0–3)	2 (0–2)	2 (0–3)
Pliability (scored 0–5)	2 (1–2)	2 (2–3)
Height (scored 0–3)	1 (0–1)	1 (1–1)
Vascularity (scored 0–3)	2 (1–2)	2 (1–2)
Total (scored 0–14)	6 (4–7)	7 (5–9)
VSS categories, n (%)
Pigmentation
Normal	12 (28)	11 (26)
Hypopigmented	7 (17)	6 (14)
Mixed	13 (31)	15 (35)
Hyperpigmented	10 (24)	11 (25)
Missing	0 (0)	0 (0)
Pliability
Normal	4 (10)	0 (0)
Supple	8 (19)	10 (23)
Yielding	21 (50)	18 (42)
Firm	9 (21)	15 (35)
Ropes	0 (0)	0 (0)
Contracture	0 (0)	0 (0)
Missing	0 (0)	0 (0)
Height
Flat	13 (31)	8 (19)
< 2 mm	26 (62)	29 (67)
2–5 mm	3 (7)	5 (12)
> 5 mm	0 (0)	1 (2)
Missing	0 (0)	0 (0)
Vascularity
Normal	1 (2)	0 (0)
Pink	17 (40)	13 (30)
Red	20 (48)	22 (51)
Purple	4 (10)	19 (19)
Missing	0 (0)	0 (0)

A summary of VSS domains on follow-up and for each domain individually is provided in Figure 5. Numerical values, as well as the proportion of participants for whom that domain improve from baseline, are presented in Appendix 7, Table 28. The VSS domains were similar at baseline. There were more participants with flat scars in the no-PGT treatment arm than in the PGT arm: 13 out of 42 (31%) versus 8 out of 43 (19%), but there was little difference in overall VSS score between the two treatment arms at any time point.

FIGURE 5.

Summary of baseline and follow-up scores on the VSS by visit and treatment arm for each VSS domain. Plots contain box plots (the central box depicting the IQR and median, whiskers depicting range with outliers indicated individually) and means (open circles) with their 95% CIs. (a) Pigmentation; (b) pliability; (c) height; (d) vascularity; and (e) total.

Tables 13–16 summarise the baseline POSAS domains on both the observer-reported and patient-reported scar scales. Results for surface area assessment demonstrated clumping as a result of the recording behaviour of a single therapist who argued that this domain was irrelevant as they would not expect most burns scars to contract or expand in surface area.

TABLE 13 - Patient and Observer Scar Assessment Scale overall observer-reported scores at baseline by treatment arm

POSAS observer scale (scored 1–10)	Treatment arm, median score (IQR)
	No PGT (n = 42)	PGT (n = 43)
Vascularity	6 (4–7)	6 (4–7)
Pigmentation	3 (2–6)	4 (2–7)
Thickness	3 (2–4)	3 (2–5)
Relief	3 (2–4) (n = 41)	4 (2–5)
Pliability	3 (3–4)	4 (2–5)
Surface area	1 (1–3) (n = 37)	1 (1–2) (n = 39)
Overall opinion	4 (3–5)	5 (4–6)

TABLE 14 - Patient and Observer Scar Assessment Scale observer-reported scores for domains at baseline by treatment arm

POSAS observer scale domains	Treatment arm, n (%)
	No PGT (n = 42)	PGT (n = 43)
Vascularity
Pale	0 (0)	0 (0)
Pink	14 (33)	7 (16)
Red	19 (45)	18 (42)
Purple	2 (5)	5 (12)
Mixed	6 (14)	13 (30)
Missing	1 (2)	0 (0)
Pigmentation
Hypopigmented	6 (14)	9 (21)
Hyperpigmented	9 (22)	9 (21)
Mixed	26 (62)	25 (58)
Missing	1 (2)	0 (0)
Thickness
Thicker	36 (86)	41 (95)
Thinner	2 (5)	0 (0)
Missing	4 (9)	2 (5)
Relief
More	28 (67)	25 (58)
Less	1 (2)	1 (2)
Mixed	12 (29)	16 (37)
Missing	1 (2)	1 (2)
Pliability
Supple	10 (24)	15 (35)
Stiff	20 (48)	21 (49)
Mixed	11 (26)	5 (11)
Missing	1 (2)	2 (5)
Surface area
Expansion	5 (12)	4 (9)
Contraction	9 (21)	4 (9)
Mixed	19 (45)	26 (61)
Missing	9 (21)	9 (21)

TABLE 15 - Patient and Observer Scar Assessment Scale patient-reported pain-scale domains (adults) at baseline by treatment arm

Adult patient scale (scored 1–10)	Treatment arm, median score (IQR)
	No PGT (N = 28)	PGT (N = 27)
Has the scar been painful in the past few weeks?	3 (1–6) (n = 27)	3 (2–5) (n = 27)
Has the scar been itching in the past few weeks?	4 (2–7) (n = 27)	4 (2–7) (n = 27)
Is the scar colour different from the colour of your normal skin at present?	8 (5–10) (n = 27)	8 (6–10) (n = 27)
Is the stiffness of the scar different from your normal skin at present?	8 (5–8) (n = 27)	7 (5–8) (n = 27)
Is the thickness of the scar different from your normal skin at present?	4 (3–6) (n = 26)	5 (4–8) (n = 27)
Is the scar more irregular than your normal skin at present?	5 (4–8) (n = 27)	7 (4–9) (n = 27)
What is your overall opinion of the scar compared with normal skin?	7 (5–8) (n = 27)	9 (6–10) (n = 27)

TABLE 16 - Patient and Observer Scar Assessment Scale patient-reported pain-scale domains (paediatric) at baseline by treatment arm

Paediatric patient scale (scored 1–10)	Treatment arm, median score (IQR)
	No PGT (n = 14)	PGT (n = 16)
Itch scale (younger patients) 1 (scored 0–4)	1 (0–3) (n = 3)	1.5 (1–2) (n = 8)
Itch scale (older patients) 2 (scored 0–10)	No data	3 (n = 1)
Pain scale (younger patients) 3 (scored 0–10)	0 (0–4) (n = 3)	2 (0–2) (n = 9)
Pain scale (older patients) 2 (scored 0–10)	No data	No data

The POSAS baseline scores are similar between treatment arms on both the patient-reported and observer-reported forms. The paediatric patient-reported scale pain and itch domains were poorly completed, with only 3 out of 14 (21%) and 9 out of 16 (56%) participants, in the no-PGT and PGT arms, respectively, giving any assessment of pain and itch.

A summary of POSAS observer-reported domain scores on follow-up and for each domain is provided in Figure 6. Numerical values, as well as the proportion of participants for whom that domain improved from baseline, are presented in Appendix 7, Table 29. The scores and trends are similar across most domains; scars are rated as getting slightly worse by month 1 before improving by months 3 and 6.

FIGURE 6.

Summary of baseline and follow-up scores on the POSAS observer-reported scale by visit and treatment arm for each POSAS domain. Plots contain box plots (the central box depicting the IQR and median, whiskers depicting range, with outliers indicated individually) and means (open circles) with 95% CIs. (a) Vascularity; (b) pigmentation; (c) thickness; (d) relief; (e) pliability; (f) surface area; and (g) overall opinion.

A summary of POSAS patient-reported domian scores on follow-up is provided in Figure 7. A summary of each domain, the numerical values and the proportion of participants for whom that domain improved is presented in Appendix 7, Table 30. Similar patterns are observed in the patient- and observer-reported scales. The response rate for the POSAS observer-reported form was higher than that for the patient-reported form, with 64 observer-reported forms returned at month 6, but only 36 patient-reported forms returned. The POSAS paediatric follow-up was completed very poorly, with very limited data supplied.

FIGURE 7.

Summary of baseline and follow-up scores on the POSAS patient-reported scale by visit and treatment arm for each POSAS domain. Plots contain box plots (the central box depicting the IQR and median, whiskers depicting range, with outliers indicated individually) and means (open circles) with 95% CIs. (a) Pain; (b) itch; (c) colour; (d) stiffness; (e) thickness; (f) irregular; and (g) overall.

Overall scores for POSAS and VSS are presented in Appendix 7, Table 31.

Sampling and recruitment: adult and parent interviews

We interviewed adult patients and parents of paediatric patients recruited to the pilot trial and allocated to the PGT and no-PGT treatment arms at two time points: time point 1 (T1) and time point 2 (T2). T1 interviews were undertaken soon after treatment allocation. T2 interviews were conducted approximately 12 months after recruitment to the pilot. We also attempted to recruit patients who had declined to participate in the pilot in order to explore that decision. The recruitment process was identical to that described in Chapter 5. T1 interviews were mainly conducted face to face in patient’s home, although a small number took place on hospital premises or via telephone. Owing to project staffing issues and timelines, the follow-up interviews for the non-Birmingham recruiting sites (two of the four sites) were conducted by telephone.

Interview format and content: adult and parent interviews

Interviews were semistructured according to a predevised topic guide and conducted in a participant-focused manner. The topic guide and interview process were refined after reflection on a small sample of early interviews. Topics included the following:

• T1

  • accounts of the accident and injury

  • accounts of the initial approach regarding the pilot trial

  • reasons for agreeing to participate in the pilot trial (or declining participation for decliners)

  • understanding of the pilot trial

  • thoughts on treatment allocation received

  • hopes and expectations for treatment, recovery and scar management

  • early experience of scar management techniques

• T2

  • update: progress with scar management and other treatments when applicable

  • experience of scar management techniques

  • adherence to scar management techniques

  • experience of participation in the pilot trial, including perspectives on the frequency and convenience of trial visits, the nature and content of scar assessment and other outcome measures

  • perceptions of scar outcomes achieved to date and broader impacts of burn injury and treatment.

Recruitment: staff interviews

We interviewed eight staff from six of the participating PEGASUS pilot sites. Of these, two were RNs and the remainder were front-line clinical staff who were delivering the pilot trial on a day-to-day basis (OTs/physiotherapists). Six interviews were conducted by telephone and two were carried out face to face in Birmingham. The quotations from staff interviews are identified using codes S1–S8.

Analysis

Interviews were audio-recorded, transcribed clean verbatim and analysed using a thematic approach. 45 Early analytic findings were discussed among the TMG and at PEGASUS investigator meetings, at which clinical staff delivering the pilot trial and at least one patient representative were present.

Sample characteristics

Sample characteristics for T1 interviews are given in Table 21.

Recruitment to pilot trial

Barriers to recruitment: staff perspectives

Patient views on recruitment to the pilot trial

Interviews with patients and parents of paediatric patients illuminate their experience of recruitment to the pilot trial and the factors that have influenced their decision-making regarding participation.

Scar assessments and other outcome measures

Scar assessments and other outcome measures: patients’ views

Patient interviewees did not express any significant concerns regarding the nature of scar assessments undertaken as part of the pilot trial. Similarly, most patients were happy with the QoL questionnaires that they were asked to complete.

However, some interviewees did question aspects of the generic QoL measures and the validity of some of the self-report content of the paediatric scar assessments. One parent was keen to point out that some measures, including the Burn Man Itch Scale, were difficult for children to understand and complete accurately:

Yes, there were a couple of answer as well on that form which will have gone back into the data mix, which again that was [de-identified]’s answer and that was obviously what was recorded, but at no point to us apart from the first couple of days did he ever say that his feet were itchy, and yet one day we went again on the chart about the itchy man, I think he said it was 9 or 10, it was extraordinarily itchy, and I said, ‘Are you sure [de-identified]?’ He said yes, ‘Are you absolutely sure, because you never said your feet were so itchy before?’ ‘Yes, I think it’s that one,’ and he went for that, and it was in the car back where he said, ‘Daddy my feet don’t really itch,’ I thought, well why did you say that? ‘I like the look of the man’, or whatever it was, he just decided on a whim to go and pick that as his answer.

BCP07, parent of paediatric patient, PGT

One interviewee noted that the process of filling out the questionnaires could get repetitive, especially if they did not perceive any significant change between visits:

It would be all sorts of silly questions, and then it would be like on a scale of 1 to 10 how do you feel, happy or not? I’m like well 10 happy, and then [de-identified] would do the on a scale of 1 to 10 where would you compare your thickness between your scar and your normal skin, and you would be like well that’s soft, that’s fairly hard, so 5, stuff like that, all the time, every bloody time. I said to her, ‘Can’t you just photocopy it, photocopy my answers? Because it’s the same on a different day’.

Willingness-to-pay instrument

Volume of assessments and associated case report forms: staff perspectives

Although staff interviewees expressed some concerns about the content of some of the assessments, as described in Staff concerns, and that initially the volume of paperwork and time taken was considerable, most felt that over time, with familiarity, they became accustomed to it:

Actually it wasn’t too bad. To start with it seemed quite lengthy but once you had done it a couple of times it was quite easy.

Interviewee S3

Staff also reported that patients seemed to cope with the demands of completing the required paperwork and questionnaires:

. . . once they were on it [the trial] they were quite happy. They filled in their forms, went through the questionnaires, other than the ones who then just didn’t just come back. The ones who stayed on and come back have all been absolutely fine, no trouble at all.

Interviewee S2

PEGASUS assessment appointments: arrangements and impact on staff capacity

We discussed the frequency of PEGASUS study-related appointments with interviewees and how these were managed alongside routine care. Although there appear to be slight differences in routine scar management follow-up practice across the pilot sites, on the whole, interviewees did not report a need to arrange specific PEGASUS study appointments/clinics or that the number and frequency of follow-up assessments were difficult to manage. Longer appointment slots were allocated to the PEGASUS study patients in routine follow-up clinics:

Yes, they were fitted in, booked into our [mentions specific clinic] which is run once a week, and we just allotted an hour slot rather than the normal half-an-hour slot. So we didn’t see them outside of the clinic, we just put them into the clinic. I think initially we were going to be seeing them outside of the clinic up in the consultants’ clinic, but that didn’t work, because it was too disjointed, because sometimes we don’t attend that clinic, so it was a bit . . . it was easier just to slot them into the normal clinic and just make sure that you blocked out an hour for them, because then you knew you were going to see them and you were able to give the proper time to them.

Interviewee S8

We also discussed the impact of the PEGASUS study assessment appointments on staff time and clinical capacity. Views were mixed on this: some staff reported significant time requirements, whereas others reported minimal disruption and additional time requirements:

So it is time consuming. I don’t want to sound really negative about it because I think it’s been very interesting, and I’m all for studies that look at what we do and examine it. But I think that to get it to . . . in order to be able to do it properly and not just focus on the research and therefore not do what you would normally do in a treatment session, which certainly isn’t the objective is it? It’s supposed to be in addition, then you’ve got to appreciate that it takes time, and I would suggest it takes a lot more time when you’re dealing with children than it does when you’re dealing with adults.

Interviewee S1

One member of staff noted that, because of the relatively small number of patients who were recruited to the PEGASUS pilot at their site, the overall impact on time and capacity was minimal.

Patient perspectives on frequency of pilot trial appointments and assessments

On the whole, patients did not express any concern about the frequency of pilot trial follow-up appointments, and most understood this to be in line with routine follow-up practice. However, several patients discussed the practicalities of attending frequent follow-up appointments, for example the need for flexibility with working arrangements for those in employment or for parents of paediatric patients. Parents, in particular, noted how this could be an emotional burden, for example if only one parent was able to attend follow-up and trial appointments as a result of work commitments. Although the actual time taken at appointments was prolonged because of trial assessments, on the whole this was valued by patients because of the attention that was given to them and also as a consequence of very well-developed relationships with core clinical staff.

Loss to follow-up

The staff interviewed noted that they had been unable to follow up some patients throughout the whole PEGASUS pilot trial assessment period. Two main explanations for this were given. First, some patients with minor burns, who at lesser risk of poor scar outcomes, may not attend long-term follow-up. Second, interviewees noted that the social or socioeconomic circumstances of a proportion of their patient population could influence the likelihood of attendance for follow-up appointments:

Yes, so one of them was definitely . . . she was at so low risk for scarring issues, also didn’t want to come back to the hospital anyway whatever, because she got persuaded really, and probably wasn’t ever going to come back and follow through . . . Another one he probably would have benefited from scar management but equally because of his social setup, etc., you had an inkling that maybe he was not going to come back for follow-up appointments . . .

Interviewee S2

Withdrawals or crossover from the pressure garment therapy group

Some of the staff interviewees described situations in which participants who had originally been allocated to the no-PGT treatment arm were withdrawn from the trial. Predominantly, this was described as a result of clinical judgements that patients should be receiving PGT and may, therefore, be related to strong views in favour of PGT. Staff at both of the sites treating paediatric patients related instances in which this had occurred:

. . . one parent withdrew because she didn’t get pressure garments and she was really worried about her child’s scar and said that she didn’t want to be in the trial anymore, and there was another child who wasn’t in pressure garments whose scars got very hypertrophic and started restricting movement, so I got him to see the consultant in clinic and between the two of us we agreed that they needed to go into garments, and all the others stayed in the arms that they were assigned to.

Interviewee S1

Interestingly, in a further interview, one member of staff recalled that they had been encouraged to withdraw a patient from the PEGASUS pilot study because of concerns expressed by a more senior colleague. In this situation, the interviewee stressed that the patient did not report any functional issues or concerns with the appearance of his scars, or treatment received and, therefore, was continued in the trial:

. . . there was one patient who went into the non-pressure group who he clearly wouldn’t have used pressure garments anyway, so from that point it was good, and his scarring wasn’t great but he was doing alright and he was happy with it. But at one point I was being pushed by the [senior member of staff] to withdraw him, ‘Oh no he needs to come off, he needs to be put in pressure’, etc., but actually then he talked through with the patient and actually he wasn’t bothered, so again it helped perhaps challenge the [senior member of staff] back a little bit more when they were saying, ‘No we need this treatment’. ‘Well actually he’s on this trial, I’ve talked to him, it’s not causing him any functional problems, he’s not bothered by the look of his scars, the treatment he’s having he feels is working, leave it at that’.

Interviewee S2

Suggested improvements for a definitive trial

It is worth noting that none of the interviewees at any of the pilot sites stated that a further definitive trial of PGT is not feasible or that they or their patients would not participate in a further trial. Interviewees did, however, make some suggestions for improvements: first, in terms of capacity, on-site trial co-ordination and site initiation and, second, regarding some streamlining of the paperwork and assessments required.

‘Standard’ treatment options for scar management and associated clinical decision-making

Staff noted that there is a range of treatment options normally available for scar management, including moisturisers, silicone cream and sheets, massage, splints and PGT. On the whole, interviewees indicated that patients will be advised to use moisturisers and massage. When PGT is used in scar management, this will often be in combination with other treatments and often involves the addition of silicone:

You might get silicone stitched into the garment, you get pads stitched into the garments, you can get panels stitched into the garment, you can increase the pressure in the garment, so whole load of different things you can add in depending again on the problem that is presented in front of you, or something is not responding as well as you would hope then you might tackle it with extra things added in. I generally start simple and then add in if needed, unless again someone is perhaps presenting with a severe scar at that point.

Interviewee S2

Interviewees discussed factors that would influence clinical decision-making regarding treatment options for scar management and the use of PGT outside of the context of the pilot trial. Therapists described a complex holistic assessment of patients made on an individual basis. Components of this assessment included clinical and patient characteristics, as well as patient preferences. The main components that influenced decisions to use PGT were:

• length of time to healing

• location of scar

• size of scar

• skin type/tone

• age of patient

• social and family circumstances of patient

• mobility.

Staff interviewees talked about time taken to heal and the location of the scar:

It’s experience, so signs of it raising at an early stage, if it’s had delayed healing, any history of scarring before or family history, sometimes the area where it is on the body as well, there might be some areas that are more likely to put a garment on than others. I would probably say the main thing would be how long it’s taken to heal.

Interviewee S3

Location of scarring could mean that function and movement were of particular concern and, therefore, patients were deemed suitable for PGT, for example, when a burn injury and scar involved joints and could affect mobility. Similarly, hand burns and potential complications, such as web creep, were mentioned specifically by interviewees as indications for PGT:

I think it is the length of time it’s taken to heal, and we’re talking well over 4 weeks or so, well over that, also whether it’s over a joint, the position of the . . . so hands I’m more prone to give pressure garments to, and really the depth the burn, the type of burn that it is. So that’s when I’ll lean towards pressure garments or think that might be a good idea.

Interviewee S8

But because we found that pressure garments really help to prevent web creep and maintain your webs . . .

Interviewee S6

As described in Strong views on pressure garment therapy among staff, when clinical staff may struggle emotionally to recruit patients whom they consider on the edge of eligibility, clinical characteristics, such as potential to restrict movement and function via involvement of joints, may influence recruitment behaviour.

Some interviewees indicated that the size of the scar would influence clinical decisions, with PGT deemed to be less suitable for smaller scars:

Smaller scars, which is why we had a bit of an issue with the 1% burns, they’re such tiny scars that patients don’t necessarily feel, and neither do I, that it warrants a pressure garment.

Interviewee S5

Associations between skin tone/type and scarring also influenced decisions, with non-white patients being seen as more prone to poor scar outcomes. Because of the potential difficulties of removing and applying a pressure garment, factors such as the patient’s age, the perceived importance of appearance and their mobility were also taken into account. Some staff suggested that garments were more likely to be used for children:

But even if they have taken 4 weeks to heal and it’s not a particularly deep burn, it’s not over a joint, I will . . . if it’s an adult . . . children it’s separate completely, I tend to be more . . . what’s the word? . . . giving pressure garments more than I would to an adult. With an adult I would be more watch and wait and let them have a look and take responsibility, and see. But with a child it’s slightly different. I’ll put more children into pressure garments maybe than I would an adult.

Interviewee S8

Finally, judgements may be made around the suitability of a treatment regime because of expected adherence based on family circumstances.

Patients’ understanding of the pilot trial

Hardly any of the patient interviewees were aware that the pilot trial was part of a feasibility study rather than a definitive trial to test the effectiveness of PGT. Perhaps this is unsurprising, as, despite the information materials being clear about the nature of the study, most knowledge and understanding will be consequent to discussions and interactions with HCPs, who were more likely to talk in terms of a trial of PGT, rather than a feasibility study or pilot trial. There were a very small number of exceptions to this. Generally, there was a good basic understanding of the treatment options in both treatment groups. For example, when asked about different scar management treatment options and how they relate to the trial, one parent gave the following response:

So my understanding is that there are a couple of different treatments available for the ongoing management of burn injuries, which is the silicone treatment and also pressure garments, or a combination of the two, or none of them.

BCP07, parent of paediatric patient, PGT

With regard to PGT specifically, again there was a good appreciation that garments were intended to be worn for up to 23 hours each day. However, there was some ambiguity about the length of time that patients would be in garments. This perhaps reflects variation in clinical practice and the variable length of time that patients may be asked to wear garments for.

Patients’ reaction to group allocation

We discussed patients’ and parents’ feelings about the group that they had been allocated to. Reactions were varied and there did not seem to be a clear pattern to reactions that could be associated with allocation to the PGT or no-PGT treatment arm. Many interviewees seemed happy with the treatment allocation they had received, although this may be related to any preconceived preferences for PGT or no PGT. For example, this parent reflected on the point at which they were told which treatment (no PGT) their child had been allocated to:

How did you find the randomisation process? . . .

Well we were busy talking to [name of staff member] and [name of staff member] came in and said he’s been randomised this way, and we said great, so that was fine. I’m sure there was someone with a random number generator at the end of a phone or a big red button that lights up a particular light somewhere and tells you which way it’s going to go, but that’s fine.

BCP05, parent of paediatric patient, no PGT

This interviewee went on to state that they (the mother and father) were pleased that their child had been allocated to the no-PGT treatment arm, as they felt it would allow the child to get back to normal as soon as possible:

I think we both liked the idea of a non-pressure garment treatment because he would feel more normal sooner. If he went with a pressure garment it would be extending that bandage feeling, and that there’s still it feels like there’s still something special going on here, there’s something out of the ordinary happening with the pressure garment. [. . .] So from that perspective we didn’t mind, but we liked the idea of a return to normality for him and his psychological adjustment to what he had been through. Kids are great, they bounce back really well anyway, and [name of child] is testament to that. He’s been completely unfazed by the whole thing, but we like the idea of him being able to get back to normal as quickly as possible. So the idea of a pressure garment was not ideal for us, we would have preferred an alternative, and we were lucky that we were randomised that way.

BCP05, parent of paediatric patient, no PGT

However, several interviewees expressed some disappointment with the treatment allocation, in both the PGT and no-PGT arms. For example, one adult patient with severe injuries said that they were disappointed when they had learned that they had been allocated to receive PGT and described those in the no-PGT treatment arm as ‘lucky’. This patient was concerned about the burden of wearing garments. Some parents expressed similar views with reasoning along the lines of that expressed by interviewee BCP05 in the previous quotation.

Some of the interviewees allocated to no PGT were disappointed by this. One adult interviewee stated that they had expected to receive the garments as part of the trial, even though they understood the nature of the treatment allocation. Allocation to no PGT therefore seemed to come as a surprise:

I was, I don’t know why, I suppose I think just in my head I’d thought well I’m in this trial, I reckon they’re going to use it on me to see whether it works. I hadn’t thought about the fact that I might not be wearing one. I hadn’t really give it any consideration, so that’s probably why it came as a surprise to me, that was not the fault of anybody, that was just the way my mind had gone, that I would be wearing one.

CAP04, adult, no PGT

One parent decided to withdraw her child from the trial when they were told that they had been allocated to no PGT:

How did you feel at that time of being allocated in the no pressure garment group?

Quite upset really, because I’d got my hopes up and I thought that I’d been led to believe that she would get them. I’d gone all the way there especially from here thinking that we were going to get them for her, and then to be told that we wouldn’t I thought well there’s no point doing the trial if I’ve got to go all that way, and all we’re going to do is just use cream. So at that point I started to look myself as to how much it would be to get the garments and pay for them myself.

CPPD03, parent of paediatric patient withdrawer

Patient reports of adherence to allocated treatment

Interviewees in both arms described variable adherence to their allocated treatment. Some interviewees who were allocated to receive PGT, described being able to maintain regimes that accommodated complete adherence to the PGT regime, for example wearing garments for 23 hours a day, and appropriate washing and care of the garments. Some interviewees allocated to the no-PGT treatment arm also seemed happy with the process and routine of massage and creaming, although this could involve some experimentation with the types of cream used. However, other patients, and parents of paediatric patients, described variable adherence and non-adherence to prescribed PGT regimes, but also to massage and creaming as part of the no-PGT treatment arm. Factors that appear to influence adherence include:

• the need for, and availability of, social support (e.g. spouse, family) to enable scar management regimes

• the fit and design of pressure garments

• discomfort or pain associated with allocated treatment

• emotional burden of allocated treatment

• feelings of physical and emotional protection offered by pressure garments.

For some adult participants, especially those with more severe burn injuries, the physical and emotional support from the family was paramount in treatment regimes, influencing adherence:

All day, I wouldn’t take it off, it would just be on for a couple of days and then I would change it. I would have a shower and then my wife would cream on my back and my arm and up the side, and then it would all go back on again, and it would stay on for another 2 days and then I would have a shower again. The only thing is with the garments and the wounds, obviously showering is a nightmare, because I’ve got to wait for somebody to be here, my wife, to cream me afterwards. That was the major thing, and to help me put my garments back on.

QAP16, adult, PGT

This patient, among others, also described how the original fit and design of one of the pressure garments had influenced adherence, as it had not stayed in position and, therefore, it had had to be altered over time. Both adult patients and parents of paediatric patients frequently mentioned this issue, and some interviewees had been through several iterations of garment design and fit.

Other interviewees admitted to not having worn allocated garments or to having stopped wearing a garment after a short period of time. One adult patient, again with more severe injuries, stopped wearing one foot pressure garment soon after allocation to treatment as it was causing significant pain and discomfort and also discontinued wearing another arm garment after 2–3 months:

. . . I just started getting fed up of wearing, because like my feet hypersensitive, so of course you’re trying to get this bloody tight pressure garment over the top of it, and it was just a nightmare . . . My foot it’s like hypersensitive it is, it’s just you touch it and it buzzes like mad, it’s alright in the boot, and I’m alright standing up on my own, but if you touch the bottom of it or the side of it, the toes or whatever, it just buzzes like pins and needles, but it’s like a whole new level. So I give up with that. . . . So I just decided not to wear it on my leg. I wore my arm one, and then eventually I just thought well it’s looking quite tidy now, so I just didn’t bother wearing that one either.

QAP13, adult, PGT

This interviewee also described priorities, such as the need to regain movement and function, that meant that garments and appearance were less of an immediate priority. They said that this influenced their attitude towards adherence to PGT.

One parent, whose child had crossed over to wearing garments from the no-PGT arm, described how the perceived emotional burden of the PGT on their child influenced adherence to treatment, with a designated ‘break’ during 1 day each week:

I give him a day’s grace, so his day as I said is Saturdays.

So he gets Saturdays as a break?

He gets Saturdays and I just think he goes swimming, so wake up in the morning, he will put on his swimming things, he goes swimming, he comes out, has a shower at the swimming baths, by the time we come home I usually give him a bath when we get home, so I wash his hair properly and wash him properly, and then I just let him have a day’s grace without wearing it, and then on a Sunday he puts it on usually Sunday afternoon, and that’s because usually in the mornings we don’t really do anything, and then he has another bath in the afternoon, and then I put it back on. But I have never told anybody at the hospital that, but I feel that, and I am . . . like I say he’s not going to have it on today. It’s now got into a bit of a routine that Saturdays he doesn’t wear it.

So just a bit of a break?

Yes, just a bit of a break for him, because sometimes you take it off and you can see him doing that, and then you just think of my gosh the sweat [. . .] He just curls up, you know a release, you released me.

BCP01, parent of paediatric patient, no PGT

However, some interviewees, particularly adults, described how the garments could provide a feeling of both physical protection, as a barrier to knocks and injury, and also emotional protection in hiding some of the visible injuries and scarring. These were a positive influence on adherence to treatment.

As already mentioned, non-adherence was also evident in the no-PGT arm, with interviewees describing similar variability in adherence to advised treatment regimes and also similar influences on adherence. For some of the parents we have interviewed, the process of massage and creaming, particularly with young children, was described as an emotional burden:

. . . I found it difficult to do as well, because it meant that I had to have contact with quite a sensitive area emotionally and physically. So I found that daily routine of doing that time consuming, and hard emotionally and I felt physically on [de-identified]. So I found that difficult as well, and it meant that I had to look at the scar, and I had troubles at the beginning looking at the scar every day twice a day, really looking at it, because I had to then massage it, plus he was a 1 year old, trying to keep him still to do something on an area that’s [. . .] I was probably a bit more reluctant [to look at the scar] so I probably wasn’t massaging it as much as I should have done, just because I found it hard.

BCP01, parent of paediatric patient, no PGT

Subjective patient perceptions of effectiveness

When we discussed progress with scarring and impressions of the effect of treatments, most of the interviewees held positive views about their scar management treatments, regardless of whether it was PGT or no PGT (massage and creaming). As per the P1 interview work with patients concerning key outcome domains, interviewees would often discuss aspects of scar characteristics, including thickness, texture and colour, as illustrated in the following quotations:

The colour has gone down. All these parts here that have . . . veiny looking things, you can see it’s all red like this, all this was like that and really red, so it’s gone a lot paler now, more towards my skin colour.

QAP16, adult, PGT

So you’re happy with the massage and the creaming for your hands and how it’s gone?

Yes, couldn’t want for any better, but I think that’s just fantastic [. . .]

It sounds like you were really keen on the massage and . . .

Yes, I am really keen and it does work. I didn’t think it would ever get as good as this, what stage you’re at now, no way, not this time last year when I’m looking at my hands just thinking Jesus, they’re a right mess they are, they ain’t going to get much better than that, but they have, so surprised, so it’s well worth doing. Get stuck in and get the massaging done.

QAP12, adult, no PGT

However, some patients’ perspectives on progress and outcomes and the relationship with scar management treatment varied over time, and for those patients with extensive burns, comparisons would be made between different areas and scars:

Are you pleased yourself with how things are progressing?

I am now, but I wasn’t at the beginning. I just didn’t think the garments were doing anything for me. But I found at times when I have taken them off and I’ve really missed them, and I can feel my lumps and bumps, and then I put my garments back on for a while and only took them off when I should have done, and it felt smoother. So it’s mainly when late at night or when I get home from work and sit down I only have to have a little tingling sensation down on the scar and I start scratching, and I just scratch myself to pieces.

Is that all the scars or is there one in particular which is . . .

It’s the two big scars that I’ve got on my left leg that are playing me up. On the other leg which we’re monitoring as well, that’s doing really well, and that doesn’t scratch, itch, nothing, and it’s very smooth at the moment, and it’s fading in colour, so I’m really pleased with that side. But my left leg was very bad, and I’ve got one scar which is really bad, quite deep. The other one above it, which was bad as well, seems to be that is progressing nicely now, and it’s not so raised and hard and thick.

EAP02, adult, PGT

The views of clinicians and their feedback to patients are obviously key in the judgements of progression and outcome that patients are making:

The lady I saw last time I’d not seen for I think maybe I can’t remember if it was 3 months or 6 months, but a big chunk of time, and she was really impressed, because she had seen it right from the beginning when I first was in hospital with it, and she couldn’t quite believe how good it was looking. So definitely really pleased.

EAP05, adult, PGT

There are a small number of examples in which interviewees were wearing, or had worn, garments on one burn area and had not worn them on another. In the following example quotation, the patient compares outcomes between these, questioning the difference and, therefore, the effect of PGT on scarring:

It makes them softer, I can . . . but literally you have the pressure garment on, you take it off, and then it’s real smooth, and it’s real soft, give it about 4 hours and it will be back to being rough and bumpy again. So to me while you’re wearing it it’s doing it good, but the second you take it off it’s reverting back to its usual form if you like. So it would almost be like you’d have to wear them for a lifetime, and I ain’t doing that, no way mate. No literally you would wear it, but to be fair that’s no different. When you take it off that is smooth and soft, so it used to be just like that when I took off, and I ain’t wore this garment now for probably nearly 5/6 months, and it’s exactly like just the second I took it off, it’s no different, and these used to be . . . these here, this line used to be real thick, again I haven’t worn the pressure garment now for 5 or 6 months and it’s flat. It used to be real ridgy, like a real thick old ridge it would be, gone.

QAP13, adult, PGT

Pediatric Quality of Life Inventory

Parents of 7 out of the 10 patients aged between 2 and 4 years completed at least part of the PedsQL at one or more visits. Accounting for the unavailability of PedsQL at visits prior to June 2015, PedsQL completion rates were two out of three (67%) at baseline, two out of three (67%) at visit 1 (week 1), three out of three (100%) at visit 2 (month 1), six out of seven (86%) at visit 3 (month 3) and six out of eight (75%) at visit 4 (month 6). There was unnecessary proxy completion of the PedsQL in a further seven cases for patients aged < 2 years at the point of randomisation (i.e. completion was not in accordance with the study protocol).

Table 23 reports the completion rate, by item, of the PedsQL for all time points. Completion rates for items relating to the physical and emotional functioning were high (range 95–100%), but decreased for the social functioning domain (range 77–86%). Completion of the final item on PedsQL is optional, with the instruction, ‘Please complete this section if your child attends nursery or day care’, which is likely to explain the fact that the completion rate for this final item on the PedsQL was 77%.

Child Health Utility Index 9D: parent version

The CHU-9D was at least partially completed at one or more visits on behalf of six out of a total of seven eligible patients who were aged between 5 and 10 years at randomisation [baseline, n = 6/7, 86%; visit 1 (week 1), n = 4/7, 57%; visit 2 (month 1), n = 5/7, 71%; visit 3 (month 3), n = 5/7, 71%; visit 4 (month 6), n = 4/7, 57%]. In the case of proxy completion of the CHU-9D, there was one single missing response (to one item) by one single parent (see Table 23) at one single time point. The CHU-9D was unnecessarily completed by two parents of patients aged < 5 years at the point of randomisation (i.e. completion was not in accordance with the study protocol).

Child Health Utility Index 9D: child version

The CHU-9D was fully completed (no item non-response) at one or more visits by six young patients out of a total of eight eligible patients aged between 7 and 16 years at randomisation [baseline, n = 5/8, 63%; visit 1 (week 1), n = 6/8, 75%; visit 2 (month 1), n = 5/8, 63%; visit 3 (month 3), n = 4/8, 50%; visit 4 (month 6), n = 5/8, 63%]. The CHU-9D was completed unnecessarily by two patients who were aged < 7 years at the time of randomisation (i.e. the child completed the instrument in addition to the parent).

In the case of two patients aged between 7 and 10 years, both patient and parent (proxy) completed the CHU-9D in accordance with the study protocol. The UK value set was used to convert the CHU-9D profiles into utility values. 59 Scores for the proxy and patient matched exactly in the case of one child across all five time points (0.963 at baseline, and at visits 1 and 2; and then 1.0 at visits 3 and 4). Scores differed for the second child across observations, with no consistency in terms of the magnitude or direction of the discrepancies. No values are available for this second patient at baseline; at visit 1 values were 0.699 (child) and 0.752 (parent); at visit 2 values were 0.915 (child) and 0.595 (parent); at visit 3 values were 0.882 (child) and 0.915 (parent); and at visit 4 values were 0.652 (child) and 0.925 (parent). Several of these differences were substantial – and raise some concerns about the validity of the responses.

EuroQol-5 Dimensions, five-level version

The EQ-5D-5L was completed by all 28 PGT-arm participants at one or more visits [baseline 25/28, 89%; visit 1 (week 1) 25/28, 89%; visit 2 (month 1) 23/28, 82%; visit 3 (month 3) 22/28, 79%; and visit 4 (month 6), 18/28, 64%]. Out of a total of 29 eligible no-PGT participants, the EQ-5D-5L was completed by 27 participants at one or more visits [baseline 23/27, 85%; visit 1, 22/27, 81%; visit 2, 17/27, 63%; visit 3, 16/27, 59%; and visits 4, 18/27, 67%]. There was an appropriate completion rate of at least 98% on every item of EQ-5D-5L (see Table 23). ‘Appropriate completion’ indicates that a tick was placed against a single level/box for a specific item. If more than one level was selected for a single item, this was classed as an error. One participant ticked all of the boxes on the form.

Mean utility scores by treatment arm and by visit are presented in Table 24. Aggregated (mean) utility in the PGT treatment arm was greater at baseline, but the utility values within both treatment arms remain fairly stable across all time points (albeit with slightly more fluctuation in the no-PGT arm). Looking at the different dimensions of the EQ-5D-5L at baseline for both treatment arms, participants rated their health as worse in relation to the pain/discomfort and usual activities items compared with other items.

Resource use questionnaire

Forty-seven out of a total of 57 patients (82%) aged ≥ 16 years filled out the adult version of the resource use questionnaire. The breakdown by occupation at week 1 was as follows: 14 out of 47 (30%) adult patients reported being employed full-time, 5 out of 47 (11%) reported being employed part-time, 7 out of 47 (15%) were self-employed, 6 out of 47 (13%) were retired, 1 out of 47 (2%) patients reported being a full-time parent/guardian, 1 out of 47 (2%) reported being a full-time carer, 9 out of 47 (19%) were unemployed, 3 out of 47 (6%) were studying in full-time education and 1 out of 47 (2%) patients recorded their occupation type as ‘other’. In the case of this last patient, free-text comments revealed a mixture of self-employment and time as an employee.

Eleven of these 47 adult patients (23%) recorded a change in their employment status over the duration of the study. One (no-PGT arm) self-employed patient transitioned into full-time employment, and one (PGT arm) patient did the opposite. Two patients who were self-employed (one patient in the PGT arm and one in the no-PGT arm) and one employed full time (PGT arm) reported being unemployed at later time points. Two (no-PGT arm) patients transitioned from self-employment to ‘not actively working’, then back to self-employment; one (no-PGT arm) patient transitioned from part-time employment to full-time student. Two (no-PGT arm) patients who undertook a mixture of self-employment and part-time employment changed to full-time employment. One (PGT arm) patient temporarily switched from working full time to self-employment.

Qualitative work (see Chapter 5) confirmed that return to work was important to at least some trial participants, as evidenced by the following quotation:

The range of movement was the most important for me . . . because at the end of the day I am right-handed, all of the jobs I’ve ever had have been manual work jobs which involve your hands, so to me being able to use it again was the most important thing.

CA06, adult

Some patients responding to the question ‘Since your last appointment at this clinic, how many days of work have you missed because of your burn injury?’ appear to have recorded responses that are implausible, in the sense that days off work reported exceed the number of days between visits. This potentially indicates that patients misinterpreted the question as being the days off work since the start of the burn injury. This issue could be addressed in a full trial with careful amendment, and possible piloting, of the wording of this particular question.

Thirty-four patients recorded that they were prescribed medicines or creams by their GP for their burn injury at some point during the trial. The most prescribed product was moisturiser (29 reported prescriptions), followed by prescriptions relating to pain relief (12 reported prescriptions). Sun cream was only prescribed twice, and three prescriptions for antibiotics were reported. Thirty patients reported that they purchased over-the-counter medicines or creams at some time during the trial period; again, this was mostly moisturiser (22 purchases). Sun cream was purchased 17 times and painkillers nine times. Eleven patients (23%) purchased clothing specifically as a result of their burn injury (six in the no-PGT arm and five in the PGT arm). Evidence from the CRFs (in this specific case, notes relating to non-adherence), appear to confirm the fact that at least some patients randomised to the no-PGT arm personally purchased items intended to act in the same way as the tight-fitting Lycra garment:

Patient has started using Tubigrip on scar. He feels scar looks better and is protected by using Tubigrip.

Notes from a CRF on non-adherence

Most patients reported driving to the clinic. The mean distance travelled (as a return journey) varied by centre, with, on average, patients travelling the farthest (113.4 miles) distance to attend St Andrew’s Centre for Plastic Surgery and Burns, Broomfield Hospital, Essex, and patients attending Manchester’s Children’s Hospital travelling the shortest (25.1 miles) distance. Mean distances at other centres were: Bristol’s Southmead Hospital, 32.1 miles; Swansea’s Morriston Hospital, 50.7 miles; BCH, 55.6 miles; UHB, 60.3 miles; and East Grinstead’s Queen Victoria Hospital, 76.3 miles.

Pressure garment costs

The costs for the initial fitting of pressure garments ranged from £36 to £141 for outsourced garments (mean cost £63, median cost £45) and from £34 to £114 for garments manufactured in-house (mean cost £55, median cost £40). The cost of each specific garment type (e.g. glove, sock, etc.) is not reproduced here, as the costing information was taken from a commercial price list and hence is likely to be commercially sensitive. As it was unclear from some of the CRFs (as completed by staff at the treatment centres) whether or not a refitting had occurred at a specific visit, the average number of refittings for pressure garments (across the PGT sample) was not calculated. It was necessary to retrospectively obtain data from sites as to the type of garment fitted, as this was not recorded on the CRF and there was some ambiguity as to which type of garment would best fit certain burn sites on the body.

A minority of sites manufactured and supplied garments in-house, although cost differences were small. In a full trial, the impact of manufacturing garments in-house versus purchasing from a commercial supplier could be explored through sensitivity analysis.

Willingness to pay

A total of 47 patients (70% of patients still participating in the trial at the 6-month follow-up) completed the WTP questionnaire, 26 in the PGT arm (out of 35 PGT patients followed up at 6 months) and 21 in the no-PGT arm (out of 32 no-PGT patients followed up at 6 months). The amount patients were willing to pay ranged from £0 to £500 per month in the PGT arm (mean £110, median £100) and from £20 to £1000 per month in the no-PGT arm (mean £174, median £100). There was one zero amount and there was evidence that it was a true value and not a ‘protest vote’, as the participant recorded the following free-text comment:

Had I been a young man I may have been willing to pay, however, at my age and with the size of the scar I wouldn’t have bothered paying for any scar treatment.

The £1000 value reported by a single participant in the no-PGT arm increased the mean value (from £133 to £174) and may not be an affordable monthly amount.

Additional free-text comments indicated that one patient/parent was responding in terms of actual monetary cost, and not expressing a value:

I had to work out the prices of the amount of each box of [silicone adhesive tape] that I use on my daughter’s scars and the cream.

Parents in particular found the WTP question challenging:

Very difficult to answer – how do you place a monetary value on the care of your child?

These free-text comments support the qualitative findings, reported in Chapter 8.

As supported by the qualitative work, the free-text comments also gave some indication that the experience of receiving care from staff was of high value to patients/parents:

It’s not specific treatments so much as the care of the hospital staff that really helps.

Patients in the PGT arm were asked two additional, follow-up questions: ‘If you could pay a smaller amount per month to receive the same treatment but without the tight-fitting Lycra garment, would you choose to do so?’ and ‘What would be the lower amount that you would be willing to pay in this case?’. Five of the 26 PGT patients/parents indicated that they would be willing to pay less for treatment that did not include the pressure garment, although two of those five failed to record the amount that they would be willing to pay in that case. Three did report a lower amount: £50 rather than £100, £5 rather than £20 and £140 rather than £180.

Sites and recruitment

Interviews with staff, parents and patient participants in the pilot trial further demonstrate that established views regarding the role of PGT in burns scar management have the potential to influence trial conduct. Staff may influence patients’ and parents’ perspectives on the trial if expectations are set in favour of the need for PGT, either by the staff who are recruiting to the trial or by other burns service staff whom they interact with earlier in the care pathway. There is evidence from interviews with patients and parents that staff who were not involved in the pilot trial have discussed PGT with them, thereby influencing expectations and attitudes towards the trial. This is unsurprising, given the embedded nature of PGT in clinical practice. Some interviewees did also report that they felt that the recruiting clinician favoured the PGT option, and there is an isolated example of an interviewee reporting that they were told that their child would have been withdrawn from the pilot had they not been allocated to PGT. Although the latter was not a common feature in the patient and parent interviews, when discussing recruitment to the pilot trial some staff interviewees did also indicate that they had been selective on occasion about which patients to approach regarding the trial. We believe that these data reflect the findings from the staff attitudes survey and that, although recruitment to the pilot has been successful, these observations need to be taken into account during the design and delivery of a definitive trial.

With regard to recruitment to the pilot study, our other main finding from the qualitative research was that several of the adult patients reported approaches regarding entry into the trial relatively early in the patient care pathway, during acute treatment and at a time when many were in pain and/or under the influence of painkillers. Although some parents were similarly approached early on, their recollection of the process and knowledge and understanding of the trial generally seemed to be more complete. Although we cannot say definitively that this is related to the point at which recruitment takes place in the care pathway, some consideration of this is needed for a definitive trial. In terms of feasibility, this is not straightforward, as approaches later in the pathway, prior to scar management, may mean that patient expectations for treatment at that stage, and particularly regarding the role of PGT, could be different.

We would recommend that the whole burns service is involved in recruitment and delivery of a trial of PGT. Staff who are not investigators or recruiting to the trial should also be involved in site initiation activities to raise awareness of the trial and the rationale for it. Site initiation visits need to be face to face and multiple site initiation visits may be required, for example when the burns and therapy units are separate. This should be resourced appropriately. Any definitive study should be designed with an integrated mixed-methods process evaluation that has a targeted recruitment investigation. This should include clear reporting and feedback to staff recruiting to the trial, the study team, the clinical trials unit and study centres and investigators.

A component of site initiation should be clear training, and information given to recruiting staff and clinicians should include the evidence and rationale for the trial, relevant findings from the feasibility study, including observations regarding recruitment, and strategies for the discussion of the trial and its relationship with established routine scar management practice. Consideration should be given to calling this a ‘scar management trial’, which would enable more neutral conversations with patients about treatment options included in the trial.

Multiple burns

Patients with multiple burns raised challenging issues in the trial. We indicated that all burns suitable for PGT should be assessed in the trial and pressure garments used on these if the participant was allocated to the PGT arm, with a maximum of three scars being selected for inclusion in the outcome assessment. Difficulties were experienced in keeping track of the selected scars, which may have been a source of data errors.

Given the proposed emphasis in a future trial on holistic assessment, it would be preferable to consider treatment of a patient rather than of a single scar site, with outcome assessments based across all scars. Evaluation of the suitability of the chosen patient-reported outcome assessment tool to achieve this is required.

Crossovers, drop-out and adherence to therapy

We identified confusion at sites in the definitions of crossover and withdrawal from the trial. Patients who had a successful early response to therapy (with or without PGT) were misclassified as crossovers and/or study withdrawals. Clear instruction and training on how individuals should be classified and followed up once therapy is complete is required for a future trial to ensure that final outcome assessments are completed and data available for analysis.

Data collection

The patients and clinic staff experienced some confusion in completion of the outcome assessments. For example, there were a number of times when the wrong QoL questionnaire was filled out by the patient or parent, which was most likely caused by the inclusion of three separate and age-specific HRQoL questionnaires in the study CRF. Particular problems were experienced for children who crossed age boundaries during the duration of the trial. The wide range of ages for the feasibility study undoubtedly increased the complexity and need for different validated QoL instruments. A definitive trial would benefit from using an enhanced electronic CRF system that ensured that only the single appropriate tool is completed for each patient at each visit.

Questions relating to resource use appear broadly relevant but some proved difficult to analyse or interpret, and it is unclear if missing responses are a true indication of no resource use or if respondents simply avoided questions that were complex or time-consuming. A future trial needs to ensure that participants provide a response (zero or a quantity) for each item of resource use.

Implications for health care

It is not possible to propose any modifications to current health care based on the findings of this feasibility study.

Recommendations for research

The feasibility study has demonstrated that it is possible to complete a definitive randomised evaluation of the value of PGT in the management of burn scars within the UK NHS context. Key aspects that need to be defined in such a study include the recruitment of infants and children in addition to adults, specificity of the suitability for inclusion in a trial based on the degree of burn severity (i.e. TBSA burn percentage), whether or not the clinical effectiveness of PGT should be assessed according to burn severity, the selection of an appropriate outcome measure and the choice of tools for the assessment of cost-effectiveness in infants and children.

Although the BBSIP tool appears to be a suitable outcome measure, evaluations of the tool in a UK context are necessary to ensure that it has validity and to obtain estimates of measurement distributions and the clinically important differences required to compute the sample size for a trial.

Trial development and management group

Naiem Moiemen (Chief Investigator/Applicant), University Hospitals Birmingham NHS Foundation Trust, Queen Elizabeth Hospital, Birmingham; e-mail: naiem.moiemen@uhb.nhs.uk.

Professor Jonathan Deeks (Co-investigator/Applicant, Biostatistics and Trial Design), BCTU, University of Birmingham, Birmingham.

Dr Emma Frew [Co-Investigator/Applicant, Health Economics (design phase)], Institute of Applied Health Research, University of Birmingham, Birmingham.

Melanie Calvert, Jonathan Mathers and Laura Jones (Co-investigators/Applicants, PROMs work), Institute of Applied Health Research, University of Birmingham, Birmingham.

Dr Philip Kinghorn (Co-Investigator/Applicant, Health Economics), School of Health and Population Sciences, University of Birmingham, Birmingham.

Amy Bamford (Current Practice Research, RN), University Hospitals Birmingham NHS Foundation Trust, Queen Elizabeth Hospital, Birmingham.

Fay Gardiner (Current Practice Research, OT), University Hospitals Birmingham NHS Foundation Trust, Queen Elizabeth Hospital, Birmingham.

Liz Coombes (Counselling Psychologist), University Hospitals Birmingham NHS Foundation Trust, Birmingham.

Karen Turner (née Doyle) (PPI involvement), Cancer Research UK Clinical Trials Unit, University of Birmingham, Birmingham.

Jonathan Bishop (Trial Statistician), BCTU, University of Birmingham, Birmingham.

Trial co-ordination and trial management (day to day)

Margaret Grant (Unit Co-ordination), Gemma Slinn (Senior Trial Co-ordinator), Dawn Brant (Trial Co-ordinator), Dana Bridges [Trial Co-ordinator (until May 2015)] and Hannah Watson (née Lack) (Data Manager), BCTU, University of Birmingham, Birmingham.

Patient-reported outcome measures research (not mentioned above)

Sue Wright, Ian Lichfield and Nicole Andrews (qualitative research), Institute of Applied Health Research, University of Birmingham, Birmingham.

Health economics research (not mentioned above)

Mark Monahan (health economic evaluation), School of Health and Population Sciences, University of Birmingham, Birmingham.

Central Manchester University Hospitals NHS Foundation Trust, Royal Manchester Children’s Hospital

Mamta Shah (PI), Katrina Keating (RN), Julie Walmsley (charge nurse), Lorna Stanyer (OT) and Anna Nixon (physiotherapist).

Abertawe Bro Morgannwg University Health Board, Morriston Hospital

William Dickson (previous PI), Tom Potokar (PI), Janine Evans (OT), Natalie Blytt Jordens (RN) and Jane Griffiths (RN).

University Hospitals Birmingham NHS Foundation Trust, Queen Elizabeth Hospital

Naiem Moiemen (chief investigator), Amy Bamford (senior RN), Fay Gardiner (OT), Elizabeth Chipp (consultant), Kwang Chear Lee (trial management) and Laura Mee (administration).

Mid Essex Hospital Services NHS Trust, St Andrews Centre for Plastic Surgery and Burns, Broomfield Hospital

Peter Dziewulski (PI), Helen Gerrish (RN), Natalie Whybro (RN), Victoria Dudman (OT) and Karen Cranmer (Research Co-ordinator).

Queen Victoria Hospital NHS Trust, Queen Victoria Hospital

Bajlit Dheansea (PI), Simon Booth (RN), Louise Rodgers (OT) and Poppy Cole (physiotherapist).

North Bristol NHS Trust, Southmead Hospital

Ian Mackie (PI), Anthony Sack (RN), Alison Guy (OT) and Sankhya Sen (Surgeon).

Birmingham Children’s Hospital NHS Trust, Birmingham Children’s Hospital

Yvonne Wilson (PI), Kate Whiting (OT), Sarah Payne (RN) and Federica D’Asta (Burns Fellow).

Trial Steering Committee

Stuart Watson (Independent Chairperson, Consultant Burns and Plastic Surgeon), Glasgow Royal Infirmary, North Glasgow University Hospitals NHS Trust, Glasgow.

Jamie Kirkham [Independent (Biostatistics), Lecturer in Medical Statistics], Institute of Translational Medicine, University of Liverpool, Liverpool.

Annemarie Nelson [Independent (PROMs), Department Director Marie Curie Palliative Care Research], Marie Curie Palliative Care Research Centre, Cardiff.

Sukhbir Rayatt [Independent (Consultant), Consultant Burns and Plastic Surgeon] University Hospitals of North Midlands NHS Trust, Stoke-on-Trent.

David Udale [Independent Member (PPI representative)].

Naiem Moiemen (on behalf of the TMG, Chief Investigator).

Jonathan Deeks (Professor of Biostatistics).

Data Monitoring and Ethics Committee

Alan Phipps (Chairperson, Consultant Plastic Surgeon), Department of Plastic Surgery, Mid Yorkshire Hospitals NHS Trust, Wakefield.

Vikram Sharma (Independent member, Consultant Plastic Surgeon), Department of Plastic and Reconstructive Surgery, Cambridge University Hospitals NHS Foundation Trust, Cambridge.

Sarah Brown (Independent member, Senior Medical Statistician), Clinical Trials Research Unit, University of Leeds, Leeds.

For the pilot randomised controlled trial

Outcome data will be made available, on request to the BCTU, for incorporation in future meta-analyses.

For the qualitative research studies

This is a qualitative study and, therefore, the data generated are not suitable for sharing beyond that contained within the report to protect the identities of the participants.