Strimvelis for treating adenosine deaminase deficiency–severe combined immunodeficiency

The committee heard from patient experts that all aspects of life for both the patient and their family are affected by the condition. The experts highlighted that diagnosis can be delayed because the rarity of the condition can lead to it not immediately being recognised. This can cause profound anxiety to the family as they watch their child suffer recurrent infections without knowing why. The patient experts highlighted that anxiety remains after diagnosis because of the strain of having to think about what lies ahead, including life changes such as isolation to manage the condition before treatment, the possibility of needing to stop work to be a full-time carer and the possibility of embarking on a treatment that carries a mortality risk. The committee concluded that ADA–SCID is a rare, serious, life-threatening and debilitating condition that also severely affects the lives of families and carers.

The committee heard from clinical experts that, when ADA–SCID is suspected, people will be referred to 1 of the 2 expert centres in England, where a diagnosis will be confirmed by gene testing. The experts explained that the current first-line treatment option for people with ADA–SCID is a haematopoietic stem cell transplant (HSCT) from a human leukocyte antigen (HLA)-matched related donor (MRD) but that for most people, there would not be one available. If an MRD is not available, the next option is an HSCT from an HLA‑matched unrelated donor (MUD). To establish whether a MUD is available, a database search for a donor would be started. This is quick, and gives an indication if one might potentially be available. Then, further high-resolution tissue typing and donor availability are confirmed to establish if a transplant is possible. The committee noted the evidence submissions indicated that current clinical practice when a MUD is not available is to enrol people in clinical trials. It noted that, if these trials were unavailable, the next option would be an HSCT from a haploidentical donor. It heard from clinical experts that this option has not been used in clinical practice for a long time but is used for other related conditions, and would be used if clinical trials were unavailable.

The committee heard that people diagnosed with ADA–SCID at the 2 expert centres are immediately started on enzyme replacement therapy with polyethylene glycol-modified adenosine deaminase (PEG-ADA). The committee was aware that, in England, PEG‑ADA is used as a 'bridge' to stabilise the patient's condition before an HSCT. The clinical experts stated that the duration of PEG‑ADA as a bridging therapy is highly variable, depending on the patient's condition and the availability of an HSCT.

The committee heard from clinical experts that the success of current treatment and the severity of lifelong impairment due to the systemic metabolic defect is probably improved by early diagnosis and treatment of the condition. The clinical experts and patient groups noted that, in other countries such as the US, a newborn screening programme identifies people with ADA–SCID at birth. NHS England highlighted that, in England, a consultation is currently in place to determine whether a pilot newborn screening programme for ADA–SCID should be started. The clinical experts considered that starting such a programme could potentially improve the success of treatment and reduce the proportion of people with lifelong impairments.

The committee noted that no deaths have occurred in people with ADA–SCID who have had Strimvelis in the clinical trials. It also heard from the clinical experts that there have been no deaths in people who have had Strimvelis in clinical practice. The committee considered the overall survival to be extremely positive, although it acknowledged that the small patient numbers made the results uncertain. In comparison, overall survival was 67% (10/15) for patients who had an HSCT from a MUD between 1995 and 2009), and was 71% (5/7) for patients who had an HSCT from a haploidentical donor. Results for HSCTs from haploidentical donors were based only on data from 2000–2009 because the company noted substantial improvements in effectiveness over time. The committee heard from the clinical experts that an HSCT from a haploidentical donor is generally expected to be less effective than an HSCT from a MUD, and considered that the higher survival reported in Hassan et al. (2012) for HSCTs from haploidentical donors was unlikely to be borne out in the long term. The committee noted that the figures for overall survival with an HSCT from a MUD have accrued over a long time period. It heard from the clinical experts that they would expect the current overall survival to be higher than reported because of advances in clinical practice over time. The clinical experts noted that they do not have any data on the current overall survival for ADA–SCID treated with an HSCT from a MUD but estimated it could currently be up to 70% to 75%. The committee concluded that Strimvelis improves overall survival compared with HSCT.

The committee noted that intervention-free survival after Strimvelis (defined in the Strimvelis clinical trials as survival without post-gene therapy PEG‑ADA used continuously for 3 months or more, a further HSCT or death) was reported as 82% (14/17). However, it heard from the company that a patient excluded from the analysis because of missing data had since been confirmed as meeting the criteria for intervention-free survival; including these data gives an intervention-free survival of 83% (15/18). It also heard from the clinical experts that they would expect the intervention-free survival of Strimvelis to be greater in clinical practice than reported in the clinical trials because of:

• the restriction of the use of Strimvelis to only people who are expected to produce enough CD34+ cells

• the expertise gained in administering Strimvelis during the 15‑year timeframe of the trials.
  
  Looking at the evidence for intervention-free survival after an HSCT, the committee noted that reintroduction of PEG‑ADA was not reported systematically for the whole population in Hassan et al. (2012). However, there was evidence that 1 person of the 15 who had had an HSCT from a MUD then had a second HSCT. Of the 7 people who had had an HSCT from a haploidentical donor, 2 did not engraft; 1 of them subsequently had gene therapy and 1 of them had 2 rescue HSCTs and then died. The committee recalled that overall survival with an HSCT is likely to have improved since Hassan et al. (2012; see section 4.7). However, the company argued that, if overall survival increases, it is possible that intervention-free survival would decrease because people who historically would have died may survive but need more interventions to manage their condition. The committee concluded that intervention-free survival was an important measure of Strimvelis's efficacy but, because of differences in reporting, it was difficult to make a comparison with HSCTs.

The committee was aware that the company had collected several measures of immune function in the clinical trials. The committee heard from clinical experts that these measures correlate strongly with clinical outcomes, and are used in clinical practice to inform treatment decisions. It noted that treatment with Strimvelis was broadly shown to lead to immune reconstitution, and considered this was likely to translate into long-term clinical improvement. The committee recalled that recurrent infections because of the lack of lymphocytes are one of the main features of ADA–SCID (see section 2.1). In the Strimvelis clinical trials, severe infections were defined as infections that led to or prolonged hospitalisation, and the estimated rates of infection were 1.17 pre-treatment, 0.26 at 4 months to 3 years post-treatment, and 0.07 at 4 years to 8 years post-treatment. Severe infections had not been clearly reported in the available literature for HSCTs, so rates of severe infection could not be calculated nor compared with those for Strimvelis. However, the committee heard from clinical experts that they would not expect the rate of severe infection to be different between successful treatment with Strimvelis or an HSCT. The committee concluded that the severe infection rate is likely to be similar for Strimvelis and HSCT, and that the rate estimated in the Strimvelis clinical trials represented the best available evidence for decision-making.

The committee noted that non-immunological events were reported by all but 1 patient after Strimvelis treatment, with the most common being hearing impairment. The committee heard from clinical experts that Strimvelis treatment is also unlikely to improve these aspects of ADA–SCID. However, there is hope that early diagnosis and treatment might limit these symptoms. The committee recalled that treatment with an HSCT does not alleviate the non-immunological aspects of ADA–SCID (see section 4.1). The reporting of these aspects in trials in people who have had an HSCT has been limited. However, the committee noted evidence highlighting that people treated with an HSCT have a mean IQ of more than 2 standard deviations below the general population mean and had a greater risk of behavioural problems. The committee heard from the clinical experts that, in clinical practice, they would expect some people to present with no learning difficulties and a few to present with moderate learning difficulties. It heard from patient experts that they consider the potential neurological abnormalities to be relatively minor in comparison with the importance of a treatment being life-saving. However, the committee concluded that neither Strimvelis nor HSCTs improve the non-immunological aspects of ADA–SCID, and that a substantial proportion of people who have successful treatment will have lifelong impairments.

The committee was aware that busulfan conditioning is needed to suppress a person's immune system before either an HSCT or Strimvelis, and that this may cause adverse effects. It heard that reduced-dose busulfan protocols are now used when doing an HSCT from a MUD, and that the busulfan dose used before Strimvelis is even lower. The committee accepted that adverse events due to busulfan conditioning are dose dependent, and so conditioning before Strimvelis treatment would be likely to cause fewer adverse events than conditioning before an HSCT. The experts also highlighted that, in infants, the conditioning is known to cause fertility problems in later life and that reduced doses of busulfan are expected to lead to a lower incidence of these; the committee understood that there is no evidence for this reduction yet because people who have had Strimvelis in clinical trials have yet to reach child-bearing age. The committee concluded that Strimvelis treatment is expected to cause fewer adverse events during treatment because of the lower busulfan conditioning needed.

The committee was aware that graft-versus-host disease (GvHD), in which donor immune cells reject the new host, is an adverse event that can lead to ongoing morbidity or mortality after an HSCT. It noted that no one who has had Strimvelis has developed GvHD, which is to be expected because it is a treatment that uses the patient's own cells. The committee heard from the company, patient experts and clinical experts that they consider the lack of GvHD to be a key benefit of using Strimvelis compared with HSCTs. The clinical experts noted that GvHD is a substantial factor in post-treatment mortality and morbidity after HSCT, and that there is an inverse correlation between how closely matched the HSCT donor is and the risk of GvHD. However, they noted that, even with an HSCT from an MRD, there remains a risk of GvHD. The committee concluded that the lack of GvHD is a valuable benefit of Strimvelis.

The committee was aware that the use of gene therapy treatment for conditions other than ADA–SCID has been associated with cancer, and that the European Medicines Agency concluded that the long-term carcinogenic potential of Strimvelis could not be determined at the time of assessment. The committee heard from the company that, since 2000, 40 people with ADA–SCID have had treatment with gamma retroviral vectors, including Strimvelis, and that there have been no reports of leukaemia. It heard from the clinical experts that there are theoretical reasons why gene therapy for ADA–SCID is less likely to cause cancer than in other conditions, but that the risk cannot be entirely ruled out. The clinical experts noted that, in other conditions, people who develop cancer present with signs and symptoms relatively soon after treatment, and the company highlighted the median follow-up for Strimvelis in the clinical trials was 6.9 years. The company stated that, as part of its regulatory commitments, it is running a long-term patient registry that will identify and investigate any cases of cancer. NHS England also stated that, as part of the commissioning process, it would produce a patient pathway that would detail the long-term follow-up needed for people in England who have Strimvelis. The committee acknowledged that the risk of cancer, although probably small, could not be excluded. It was reassured that the company and NHS England have measures in place to identify the risks of cancer associated with the treatment in general, and to follow individual patients over time to provide care and treatment if it occurs.

The committee discussed the generalisability of the clinical evidence to people who would have Strimvelis in clinical practice in England. It noted that the median age of people in the clinical trials was 1.7 years, but that the oldest were up to 6 years. The committee recalled that, for most people, signs and symptoms of ADA–SCID begin within the first year of life (see section 2.2). It heard from the clinical experts that they would expect people who have Strimvelis in clinical practice to be younger than those in the clinical trial. The committee heard that it is plausible that, because of this, Strimvelis would be more effective in clinical practice than in the clinical trials. This is because a younger population would be expected to produce a greater harvest of CD34+ cells needed for Strimvelis manufacture, and may have fewer non-immunological aspects of the condition. The committee concluded that the age of the population who would have Strimvelis in clinical practice may be lower than that in the clinical trial, and that this could lead to greater clinical benefit.

The committee heard from patient experts that they considered having the option of Strimvelis would be life changing for patients with ADA–SCID. They highlighted that, because treatment with Strimvelis carried less risk of post-treatment mortality and GvHD than an HSCT from a MUD or haploidentical donor, anxiety for family and carers would be reduced before treatment. They also highlighted that the busulfan conditioning needed for Strimvelis treatment was less harsh than for HSCT, and that this would make a substantial difference to the patients' and carers' quality of life during treatment. The patient experts discussed the upheaval of travelling to Milan for treatment and noted that, although this may be difficult for some people, evidence from a survey of parents suggested that all respondents would be willing to travel for the benefits offered by treatments like this. The committee also heard from 1 patient expert that they expected the financial impact, and impact on family and work, might be similar to currently available treatments because of the need for pre-treatment isolation and travel to 1 of the 2 expert centres in England for HSCTs. The committee recognised that treatment abroad with Strimvelis is potentially a major upheaval for some people. However, it concluded that, on balance, people considered Strimvelis to have a substantial benefit over HSCTs because of the expectation of lower risk of mortality and adverse events.

The committee was aware that the company and NHS England expected 3 people to be diagnosed with ADA–SCID in England per year, but noted that the submission from the patient group stated that it expected 6 to 10 people to be diagnosed per year. The committee heard from the clinical experts and NHS England that this higher estimate relates to the number of people who had treatment in the expert centres each year, which includes people who travel from other European countries. The company estimated that, of the 3 people diagnosed each year, 1 will have an MRD and 1 will choose not to have Strimvelis because of the need to travel to Italy. The committee heard evidence from the clinical experts that about 20% to 25% of people have an MRD. It recalled that the patient experts expected travelling to Italy would be an upheaval, but that families had reported that they would be willing to travel for treatment if necessary (see section 4.15). The committee concluded that the company's estimated patient numbers were reasonable but noted the apparent greater willingness to travel for treatment expressed by the patient experts. It accepted that, because ADA–SCID is a very rare condition, the number of people who would have Strimvelis each year would vary.

The committee was aware that the company assumed 1 person a year would have Strimvelis treatment. It noted that the total net 5‑year budget impact was £2.35 million and £0.93 million for Strimvelis compared with an HSCT from a MUD and a haploidentical donor respectively. The committee noted that the results were highly sensitive to the number of patients having treatment. It concluded that the assumptions made by the company in the budget impact analysis were reasonable.

The committee was aware that the model was based on a decision tree to establish the proportion of patients whose initial treatment was successful, with long-term survival extrapolated over a lifetime time horizon using a Markov modelling approach. The company noted that people entered the model aged 1 year, which is younger than reflected in the clinical evidence. It was aware that a younger age affects drug costs because of weight-dependent doses, especially for costs associated with PEG-ADA. The committee recalled that the age of people in the clinical trials was higher than in clinical practice (see section 4.14), and heard from the clinical experts that the starting age of people in the model was more reflective of people with newly diagnosed ADA–SCID who would be identified in clinical practice. The committee concluded that the model structure was suitable for decision-making.

The committee was aware that the ERG highlighted alternative treatment pathways to initial and rescue transplant after treatment failure. The ERG also highlighted that people who choose Strimvelis may do so after searching for a MUD. The committee recalled that the time-limiting factor when searching for an HSCT donor is contacting and testing potential donors (see section 4.3). It heard from the clinical experts that, if Strimvelis were to be recommended, they would discuss the choice of treatment after doing a database search, but before contacting and testing potential donors. The ERG suggested that, after treatment failure, people who have not had Strimvelis may choose to have it as a rescue therapy. However, the committee heard from the clinical experts that this is not possible because people who have already had busulfan conditioning would be unable to donate sufficient CD34+ cells for the treatment. The committee concluded that the treatment pathway used in the company model was appropriate for decision-making.

The committee noted that the company's model was based on evidence from the 'integrated population' of the Strimvelis clinical trials, but excluded the evidence from the NPP. The ERG considered that the NPP evidence should have been included because of low patient numbers in the evidence. The results of the NPP are confidential and cannot be reported here. However, the committee was aware that the ERG had also identified issues with the generalisability of the NPP evidence. The committee heard from the clinical experts that the patient characteristics of the NPP population were distinctly different from those seen in clinical practice. It recalled that, based on the marketing authorisation of Strimvelis, treatment is only suitable if enough CD34+ cells can be harvested for Strimvelis production, and this can be detected by a bone marrow biopsy (see section 3.1). The committee considered that the population of the NPP was not similar to what would be expected in clinical practice. It also considered that people similar to the NPP population would likely be ineligible for Strimvelis because the treatment is restricted to people in whom sufficient CD34+ cells can be harvested. The committee therefore concluded that the NPP evidence could be excluded from the model.

The committee recognised that survival after an HSCT or Strimvelis was one of the most influential factors affecting the model results. It was aware that values in the model were based on survival as reported in the Strimvelis clinical trials and Hassan et al. (2012), but that these estimates were highly uncertain because of the limited evidence. The committee recalled that it had considered that the overall survival values reported for an HSCT from a MUD in Hassan et al. were an underestimate because of improvements in outcomes with HSCTs over time (see section 4.7). The clinical experts, when pressed, suggested that current overall survival mortality would be around 70% to 75%. The committee considered that the exact survival rate after an HSCT from MUD was uncertain but considered that a more conservative assumption than used by the company or ERG would be appropriate. It concluded that a value of 72.5% would be reasonable to inform decision-making.

The company assumed that people having Strimvelis would have 10 weeks less PEG-ADA therapy than people having an HSCT because of the need to search the registry for a donor. The ERG highlighted that the duration of PEG‑ADA in practice is uncertain, and that the lengths of time it was given in the Strimvelis trial were substantially longer than those estimated in the model. The ERG stated that it preferred to use equal pre-treatment durations for Strimvelis and HSCTs. The committee heard from the clinical experts that the duration of PEG‑ADA in the clinical trial was extended because of the trial inclusion criteria and protocol. It accepted that this was not representative of clinical practice in England. However, it recalled that, in clinical practice, people would not proceed to an HSCT or Strimvelis therapy if their condition was not stable, and would remain on PEG‑ADA (see section 2.4). The committee recognised that the duration of pre-treatment PEG-ADA in practice was uncertain, but considered that the ERG's preferred assumption that the durations were equal between HSCTs and Strimvelis was plausible.

The committee was aware the company assumed that if treatment failure occurred the person would have a rescue HSCT from a newly born matched sibling donor, and that this subsequent treatment would always be successful and would carry no risk of post-treatment adverse events. The committee heard from the clinical experts that, in clinical practice, most people who have a subsequent HSCT would have it from a MUD, and that the success rate of this subsequent transplant would be similar or perhaps slightly lower than for people who were having a first transplant. After consultation, the committee acknowledged that a matched sibling donor would be the first choice if available, although it recalled that a MUD transplant would be more common. The company used the intervention-free survival clinical evidence to inform the rates of rescue transplant in the model, whereas the committee heard from the clinical experts that they expect rescue rates to be equal across the different treatments. The ERG and company highlighted inaccuracies in the modelling of rescue transplants because:

• the rates were not conditional on survival of the initial transplant

• rescue was assumed to occur a year before chronic GvHD had resolved

• a patient previously excluded from analysis because of incomplete data had now been confirmed to have not had a rescue transplant.
  
  The committee was also aware that the model assumed that all rescue transplants occur 2 years after the initial transplant. However, it noted that, in clinical practice, rescue therapy could be done sooner for people treated with Strimvelis compared with HSCTs because Strimvelis does not cause GvHD, and that this would reduce the duration of PEG‑ADA needed as a bridge to transplant. The committee concluded that it was more appropriate to assume rescue transplant from a MUD and that the identified inaccuracies should be corrected. However, it considered that the rates were very uncertain because of limitations in the intervention-free survival data. The committee concluded that it was reasonable to consider the scenario in which the rescue rates were equal across the treatment arms in its decision-making.

The committee was aware that the Strimvelis clinical trials collected some quality-of-life data but these had not been included in the model because they were limited; the company instead used utilities from the literature. The ERG highlighted that, in the company model, no disutility was considered for people having intravenous immunoglobulin (IVIG) or for those who had severe infections; after year 8, people were assumed to have regained full health and to have utilities equal to the general population. The committee heard from the company that the assumptions used in the model were for simplicity because no specific long-term utilities for ADA–SCID were identified. The impact of these assumptions was explored in sensitivity and scenario analyses. The ERG preferred to incorporate the company's scenario analysis in which a utility weight of 0.75 was applied to people who had IVIG. The committee heard from clinical experts that people who need IVIG are likely to be sicker than people who do not need it. So, it considered that applying a disutility for these people was reasonable. Moreover, the committee considered it implausible that people would return to full health because it recalled that a substantial proportion of people who have successful treatment will have lifelong impairments (see section 4.10). The committee was aware that the ERG preferred to include costs and utilities associated with bilateral deafness, which has been identified as a common aspect of the condition that would be unaffected by treatment (see section 4.10). The committee heard from the company that these costs and utility values were derived from people with congenital hearing loss and may not reflect people with ADA–SCID whose hearing loss is acquired during infancy. The company preferred to reflect uncertainty over specific utility values by exploring sensitivity analyses that reduced the utilities by up to an extreme value of 20%, rather than including specific utility values. The committee considered it appropriate that restoration to a lower utility than that seen in the general population should be used. However, it was unclear whether specific costs associated with lifelong impairments should be included because these costs are not associated with the treatment. It agreed that the specific utility values were highly uncertain, but noted that quality-of-life gains were similar between the ERG's preferred approach and the company's sensitivity analysis. The committee concluded that, because the ERG's preferred assumptions were based on available evidence, they were preferable for decision-making.

The committee was aware that the company included a scenario analysis in which an additional disutility was associated with bereavement after the death of a child. The committee considered that this would not fully reflect the quality-of-life benefit to carers after successful treatment. The committee recalled that the patient experts stated that caring for someone who has ADA–SCID affects all aspects of life for a carer (see section 4.2). It heard from the patient experts that improvements to the quality of life of the carer occurred immediately after a successful treatment. The committee heard from the company and ERG that there were no specific carer-related utilities that could be identified to reflect the benefit in quality of life after successful treatment. The committee considered that improvements in carer-related quality of life were an important aspect of successful treatment, but acknowledged that a specific value could not be identified. It concluded that improvements to carer-related quality of life should be qualitatively taken into consideration in the committee's decision-making.

The committee was aware that, at clarification, the company submitted an alternative 'secondary' analysis that included updated unit costs for PEG‑ADA and IVIG, costs to the NHS for providing travel to Milan, and ambulance costs to and from the airports. NHS England stated that a policy for NHS funding of travel and accommodation will be developed, building on its experience of commissioning proton beam therapy. It also stated that the cost of travel to Milan for patients and families would be one of the key additional costs to the NHS associated with Strimvelis treatment. The committee therefore considered that these costs should have been included in the model. The ERG also highlighted further alternative unit costs in its critique, which were included in its preferred analysis. These included alternative costs for an HSCT, admission to hospital for GvHD and costs incurred for testing people ineligible for Strimvelis treatment. The committee considered the alternative costs identified by the ERG to be reasonable. It concluded that it was reasonable to include the updated costs in the company's secondary analysis and the ERG's analysis.

The committee was aware that NICE's guide to the methods of technology appraisal (2013) and NICE's interim process and methods of the highly specialised technologies programme (2017) specify that the reference case discounting rate is 3.5%. However, it also states that a non-reference-case rate of 1.5% may be used when treatment restores people to full or near-full health when they would otherwise die or have severely impaired lives, if it is highly likely that there will be long-term benefits, and if the treatment does not commit the NHS to significant irrecoverable costs. The committee heard from the company that it considered a discount rate of 1.5% to be most appropriate because Strimvelis restores people who would otherwise die or have a very severely impaired life to full or near-full health, and sustains gains over a very long period. The committee acknowledged that Strimvelis has a high one-off cost, whereas the benefits are accrued over the lifetime of the patient. It considered that it was likely that the alternative 1.5% discounting rate was intended to cover situations similar to this – that is, when costs are incurred up-front but benefits are accrued over a longer period. The committee acknowledged that the technology was transformative for people who, without treatment, would otherwise die. However, it recalled that people who have successful treatment often have lifelong impairments (see section 4.10). The committee was highly uncertain about whether people treated with Strimvelis would be considered to have 'normal or near-normal health'. The committee was reassured by the clinical experts that the immune reconstitution seen with Strimvelis was expected to translate to long-term immune competence. However, it also recognised that there were uncertainties in whether the long-term benefits of treatment would be achieved because of the limited evidence. The committee concluded that it was uncertain about whether Strimvelis fully met the criteria to use a discounting rate of 1.5%, and that both discount rates should be considered by the committee during its decision-making.

The committee noted that, in the company's deterministic base case, Strimvelis was associated with incremental cost-effectiveness ratios (ICERs) of £36,360 and £14,645 per quality-adjusted life year (QALY) gained compared with an HSCT from a MUD and a haploidentical donor respectively. The committee highlighted that the costs and QALYs accrued over the lifetime of the model were high; Strimvelis was associated with 13.6 and 11.7 incremental QALYs and £494,255 and £170,668 incremental costs when compared with an HSCT from a MUD and a haploidentical donor respectively.

The committee recalled that the ERG made several changes to the company's base case. These were:

• inclusion of NPP data (see section 4.21)

• assuming equal wait time and pre-treatment PEG‑ADA use across treatment arms (see section 4.23)

• assuming rescue therapy has cost and health outcomes equal to an initial HSCT from a MUD (see section 4.24)

• including ongoing healthcare costs and morbidity associated with bilateral deafness (see section 4.25)

• using alternative assumptions from the company's secondary analysis and inclusion of alternative unit costs (see section 4.27).
  
  The committee noted that applying all of the ERG's changes increased the ICERs, at a discount rate of 1.5% for Strimvelis, to £86,815 and £16,704 per QALY gained compared with an HSCT from a MUD and a haploidentical donor respectively. The committee recalled that it had considered including the NPP data or costs associated with hearing loss to be inappropriate. Furthermore, it considered that the overall survival after an HSCT from a MUD was underestimated and it was reasonable to use a more conservative estimate of 72.5% (see section 4.22), and that there was also uncertainty over the most appropriate discount rate that should be used (see section 4.28). The results of the cost–utility analysis incorporating the committee's preferred assumptions at a 3.5% discounting rate were £120,506 and £12,106 per QALY gained compared with an HSCT from a MUD and a haploidentical donor respectively. At a 1.5% discounting rate, the results of the model were £74,430 per QALY gained compared with an HSCT from a MUD, and Strimvelis was dominant (that is more effective and less costly) compared with an HSCT from a haploidentical donor.

The committee recalled that there was uncertainty in the rates of rescue treatment used in the model, and that it was plausible that the rates were equal across treatment arms (see section 4.24). The ERG explored the ICERs when the rescue rate was equal across all the treatment arms. In this scenario analysis, incorporating the committee's preferred assumptions and at a 3.5% discounting rate, the ICERs were £91,910 and £84,172 per QALY gained compared with an HSCT from a MUD and a haploidentical donor respectively. At a 1.5% discounting rate, the ICERs were £54,072 and £49,429 per QALY gained compared with an HSCT from a MUD and a haploidentical donor respectively. The committee acknowledged that the rate of rescue was a key driver of the ICER, but was reassured that it was plausible that the ICER for Strimvelis compared with an HSCT from a MUD may be lower than that estimated using the committee's preferred assumptions.

The committee understood that NICE's interim process and methods of the highly specialised technologies programme (2017) specifies that a most plausible ICER of below £100,000 per QALY gained for a highly specialised technology is normally considered an effective use of NHS resources. For a most plausible ICER above £100,000 per QALY gained, judgements about the acceptability of the highly specialised technology as an effective use of NHS resources must take account of the magnitude of the incremental therapeutic improvement, as revealed through the number of additional QALYs gained. The committee noted that, based on the most plausible assumptions, the undiscounted QALY gains were 14.0 and 19.6 when Strimvelis was compared with an HSCT from a MUD and a haploidentical donor respectively. The committee understood that a weight between 1 and 3 can be applied when the QALY gain is between 10 and 30 QALYs, and that this would result in a QALY weighting of 1.40 and 1.96 for Strimvelis compared with an HSCT from a MUD and a haploidentical donor respectively. The committee recalled that rates of rescue treatment were uncertain (see section 4.31), and was reassured to note that the potential QALY weighting would be higher if rates of rescue transplant were equal across all the treatment arms. The committee was satisfied that there was sufficient evidence that Strimvelis offers significant QALY gains, and therefore applied this weighting in its consideration of the value for money of Strimvelis.