The ReCell Spray‑On Skin system for treating skin loss, scarring and depigmentation after burn injury

The External Assessment Centre excluded all the studies identified by the sponsor because they were outside the scope of the evaluation, although it did recognise that the studies contained useful cost information for standard care. The External Assessment Centre initially considered that the study by Wood et al. (2012) (not presented by the sponsor as part of its economic evidence) may provide some relevant evidence on costs. However, quality assessment showed several limitations (small number of patients, non‑UK care pathway, heterogeneous population), so the External Assessment Centre concluded that the evidence could not be generalised to support the sponsor's economic case.

The sponsor submitted a de novo cost analysis comparing the ReCell Spray‑On Skin system ('ReCell') plus conventional dressings, ReCell plus Biobrane, Biobrane alone, and conventional dressings alone, for treatment of a partial thickness 640 cm² burn. Full details of all cost evidence and modelling considered by the Committee are available in the assessment report overview in the supporting documentation.

The sponsor submitted a base case analysis modelling patients treated for partial thickness burns including scalds, for which meshed grafting was not needed. The model covered a 21‑day period. The extent of healing was observed after the first 10 days, assessing whether wounds had 100% epithelialisation (completely healed) or incomplete healing. This determined whether a patient's wound was likely to heal without intervention within the 21‑day timescale or whether skin grafting would be needed. Other clinical parameters in the model included length of time as an inpatient and need for a skin graft.

In the base case, the sponsor assumed a time to healing of 15 days (based on the time to 100% epithelialisation) for the conventional dressing treatment arm. This was based on the median from 3 studies observing conventional topical burn treatments (Caruso et al. 2006; Cuttle et al. 2007; Silverstein et al. 2011). The sponsor then used a percentage reduction in healing time for the conventional dressings to calculate healing times for the other 3 interventions. The estimated reductions were based on results from Wood et al. (2012) and Echlin et al. (2012b) and were 30% for Biobrane or ReCell alone and 40% for the combination. The estimates of the number of people who were treated as inpatients were based on clinical opinion obtained by the sponsor. The proportion of patients having a standard skin graft was based on clinical studies (Caruso et al. 2006; Cuttle et al. 2007; Ostlie et al. 2012; Silverstein et al. 2011) which may have overestimated the proportion because of the included populations. Clinical opinion also suggested this figure may be lower (5% to 10%). The sponsor used clinical opinion to inform this parameter for the other interventions.

The sponsor's base case included several key assumptions:

• A burn size of 640 cm² (5% to 10% total body surface area depending on the age and size of the patient).

• The burn was considered to be partial thickness with no definite areas of deep involvement.

• Burns were considered sufficiently severe to warrant initial debridement in theatre. Patients having conventional and Biobrane treatment were assumed to need 20 minutes theatre time at the start of treatment. Those treated with ReCell would need 30 minutes theatre time (based on clinical opinion).

• All patients would remain as inpatients until day 2. Those who were discharged at this point would have re‑dressing either as outpatients or as day visitors to the ward (based on clinical opinion).

• All patients were assumed to have their burn managed on a general burns ward (or in an outpatient clinic if discharged). The sponsor excluded intensive care unit costs because it considered these costs would obscure other treatment cost differences.

The sponsor's base case analysis showed Biobrane to be the lowest cost treatment (£6,398), followed by the ReCell Spray‑On Skin system plus Biobrane (£7,787), the ReCell Spray‑On Skin system (£7,892) alone and then conventional dressings (£9,543). Biobrane was the lowest cost treatment because of its lower acquisition cost compared with the alternative treatments, fewer dressing changes needed and reduced healing time compared with conventional dressings. Cost savings for the ReCell Spray‑On Skin system were driven by a reduced proportion of patients needing a standard skin graft and shorter healing times compared with conventional dressings. The sponsor explored the uncertainty around the model parameters and the effect this had on the incremental cost of ReCell using a one‑way sensitivity analysis. The results of the sensitivity analysis showed that in all of the scenarios presented, ReCell was cost saving compared with conventional dressings, except for the smaller wound size of 320 cm², when ReCell and ReCell plus Biobrane were more costly than conventional dressings. In all of the other scenarios Biobrane was the lowest cost option followed by ReCell plus Biobrane, ReCell alone and conventional dressings.

After reviewing the available economic evidence, the Committee asked for additional information on the cost consequences of using ReCell to treat full or deep partial thickness burns in conjunction with grafting, in the form of a revised economic model (see section 5.18). The Committee also requested further information on 3 parameters used in the sponsor's existing model for partial thickness burns not needing grafting that it considered to be particularly uncertain, which were length of hospital stay, time to epithelialisation and need for a skin graft. The External Assessment Centre gathered additional information from clinical experts working in NHS burns units and centres to further inform the revisions to the sponsor's model.

The External Assessment Centre was asked to consider full or deep partial thickness burns in 2 subgroups:

• Patients with full or deep partial thickness burns judged to need skin grafting and likely to need 1 operation with inpatient stay on a burns ward. The interventions for this group were skin grafting in conjunction with ReCell and ReCell alone. The comparator was skin grafting alone.

• Patients with large area full or deep partial thickness burns judged to need wide meshed skin grafting and likely to need multiple operations and grafting and time in an intensive care or high dependency unit. The intervention for this group was meshed grafting with ReCell and ReCell alone at the donor site. The comparator was meshed skin grafting alone and standard donor site treatment.
  
  The External Assessment Centre further classified these 2 groups as patients with burns covering 10% and 40% of the total body surface area respectively, based on referral thresholds to burns services. The External Assessment Centre intended to use these parameters as the base case for the model and to vary them in sensitivity analyses. It also classified the partial thickness burns not needing grafting as those covering 10% of the total body surface area for the purposes of gathering information about the uncertainties in the sponsor's model. Adults and children were considered separately within the burn groups.

In order to find further information to inform the analysis of full or deep partial thickness burns in the 2 subgroups, the External Assessment Centre carried out an initial literature search around burn care in the UK. Results from this search, the sponsor's model and previously identified literature were used to create a list of parameters for which information was needed. The External Assessment Centre devised a questionnaire to gather data on these parameters, with the help of the lead clinical experts from the evaluation. The questionnaire was aimed at capturing quantitative information for the revised modelling and also for the 3 uncertain parameters from the sponsor's model. It also included questions to gather qualitative data on the cosmetic outcomes of using ReCell in burns and vitiligo. The questionnaire was administered as a semi‑structured interview with optional email follow‑up because of the volume of questions included and the opportunity to gain additional insight from the clinicians.

The External Assessment Centre tried to obtain data from the International and National Burn Injury Databases (iBID and NBID) but was unable to gain access within the timeframe for the additional work.

The External Assessment Centre contacted lead clinicians from all specialist burns units and centres identified in England and Wales (adult and paediatric; 18 in total). Interviews were conducted with 10 clinicians, 9 consultant burns surgeons and 1 specialist burns nurse. Another burns surgeon provided comments on a summary of the data. Of those interviewed, 3 had not used ReCell but all respondents were familiar with it.

The data collected for each of the 3 groups are summarised in tables 5 to 7 of the External Assessment Centre's additional report (pages 14 to 17). For the 10% total body surface area partial thickness burns not needing grafting, the External Assessment Centre found that many adults would be treated with conventional dressings rather than Biobrane and would not be taken to an operating theatre. Children would be more likely to be treated with Biobrane, which would need a general anaesthetic in an operating theatre. The External Assessment Centre found some variation between sites in terms of the care pathway and little quantitative data. Most respondents indicated that this type of burn would usually heal within 2 weeks without the need for alternative treatments such as ReCell. Those using ReCell in this group indicated that it would need additional theatre time and could not provide any quantifiable data about reduced healing time, subsequent need for grafting or scar outcome.

For the 10% total body surface area full or partial thickness burns needing grafting, the External Assessment Centre found that unmeshed sheet rather than meshed skin grafting was more likely to be used. Most respondents indicated that ReCell would not be used alone or with Biobrane in patients who were treated with sheet skin grafts.

For the 40% total body surface area full or partial thickness burns likely to need meshed grafting, the External Assessment Centre found considerable variation in practice between sites. Available data from the International Burn Injury Database indicated that there would be around 8 or 9 patients treated in each burns centre each year making generalisable estimates of the treatment pathway difficult. Many respondents described using autologous cultured cells in conjunction with meshed skin grafts to improve the speed of healing and to improve the appearance of the healed burn. Autologous cultured cells are available from 1 NHS laboratory and 1 commercial provider and take around 2 weeks to produce. ReCell was generally considered for use when cultured cells were not available. The respondents expected healing time to decrease and cosmetic outcome to improve with the addition of autologous cells but were unable to provide quantitative data.

The External Assessment Centre asked all interviewees what they considered the role of ReCell in treating acute burns to be. All interviewees thought that ReCell might have some benefit, particularly in large burns needing skin grafting. The immediate availability of cells produced using ReCell was considered to be an advantage over the use of cultured cells needing a 2‑week wait. However, some clinicians expressed a preference for cultured cells as a result of available volume and cell viability. Another advantage of ReCell identified was the potential for its use with left‑over pieces of donor skin from grafting, which would otherwise be discarded. However, the External Assessment Centre noted that this type of use may extend theatre time. The respondents identified 2 other scenarios for the use of ReCell. In mid‑dermal, mixed depth, intermediate or indeterminate burns ReCell could be used with the aim of reducing healing time and the need for later skin grafting. The other option identified was its use in deep facial burns as an additional treatment when the burn had not healed well after 2 weeks.

The External Assessment Centre summarised the available evidence from the assessment report, the additional literature search and the findings from the interviews with experts for 9 main parameters in section 3.3 of its additional report (pages 18 to 23). The External Assessment Centre stated that the lack of quantitative data for the clinical benefits or resource impact of using ReCell meant that it was unable to develop economic modelling for burns needing skin grafting.

The External Assessment Centre concluded that the uncertainties in the parameters used in the sponsor's model (length of hospital stay, time to healing and proportion of patients needing skin grafts) could not be verified. The findings from the interviews with experts indicated that ReCell was unlikely to be used in the NHS to treat the population included in the sponsor's model, because these burns are likely to heal without the need for skin substitutes and so there would be no additional benefit from its use.

The Committee noted that the sponsor's economic model did not include the treatment of large area full or deep partial thickness burns needing skin grafting, as identified in the decision problem for the evaluation. The Committee heard expert clinical advice that the use of ReCell in this patient group might have substantial clinical benefits. Therefore, the Committee asked the External Assessment Centre to produce economic analysis for the treatment of these burns to aid its decision‑making in developing recommendations for ReCell. The Committee was aware of the lack of published data for this patient group but it considered that parameters for the economic analysis could reasonably be obtained by expert opinion from a broad range of clinicians working across NHS burns units and centres. The Committee also suggested that hospital‑based audit data might be available to inform the model.

The Committee was unconvinced about the validity of some of the assumptions in the sponsor's model for partial thickness burns, in particular the length of hospital stay, time to healing and the proportion of patients needing skin grafts. The Committee accepted the conclusions of the External Assessment Centre's additional work that several parameters in the sponsor's model for smaller partial thickness burns not needing grafting could not be validated (see section 5.17). It also accepted the advice that ReCell is unlikely to be used by clinicians in the NHS for this group of patients because their burns would usually be treated with conventional or biosynthetic dressings without skin substitutes.

The Committee considered the External Assessment Centre's conclusion that there were insufficient data to inform the cost modelling to include the use of ReCell in full or deep partial thickness burns needing grafting. The Committee queried the effect of potential reductions in healing time on length of hospital stay for these burns in the NHS. In the light of expert clinical advice that treating these burn injuries has a high cost to the NHS and that a reduced stay could have a relatively large effect on resource use, the Committee considered that further research on use of ReCell should include evaluation of its cost and resource impact.