1 Recommendations

Haemophilia A is caused by a gene mutation that results in the inability or reduced ability to produce factor VIII, which is vital in stable blood clot formation. This leads to prolonged bleeding after injury and, when severe, bleeding into joints and muscles without any injury. Haemophilia A is an inherited condition that mostly occurs in men and boys. Women and girls who carry the haemophilia gene may have mild or, rarely, moderate to severe symptoms of bleeding. For this evaluation, the company only presented clinical- and cost-effectiveness evidence for efanesoctocog alfa in severe haemophilia A (see section 3.2). The clinical experts explained that severe haemophilia A usually presents in the first few years of life with joint or muscle bleeds. Occasionally, it may cause spontaneous and potentially fatal bleeds in any tissue. The clinical experts explained that subclinical bleeds are also associated with the condition. These bleeds can cause chronic pain and joint damage, potentially affecting mobility and, over time, needing surgery. The patient experts highlighted that the risk of bleeding can limit jobs, sports and other activities. It also has a substantial psychological effect on people with the condition, and affects the quality of life of carers of children with the condition. Also, because haemophilia A is inherited, there may be several siblings with the condition in the same family, increasing its impact on carers. The committee recognised that severe haemophilia A is a chronic condition that significantly affects the lives of people affected by it.

The NICE scope and anticipated licence for efanesoctocog alfa includes everyone with haemophilia A. But the decision problem in the company submission only included people with severe haemophilia A. The severity of haemophilia A is classed according to the amount of clotting factor remaining compared with expected levels. Mild haemophilia is defined as over 5% of normal clotting factor, moderate as between 1% and 5%, and severe as less than 1%. The company explained that it had excluded people with mild or moderate haemophilia A from its decision problem because there was no evidence for efanesoctocog alfa in these groups. Also, it did not expect efanesoctocog alfa to be routinely used in these groups. The clinical experts explained that, generally, treatment for severe haemophilia A differs from that for mild and moderate forms (see section 3.3). But some people with moderate haemophilia A and factor VIII activity levels between 1% and 2% would be offered the same treatments as people with the severe form. They added that healthcare professionals would be keen to use efanesoctocog alfa in these people. The committee considered this but concluded that it had not been presented with clinical- and cost-effectiveness evidence for people with mild to moderate haemophilia A. Also, it considered that differences in the treatment pathway meant that it was likely that the clinical- and cost-effectiveness outcomes would differ between people with severe and mild to moderate haemophilia A. So, it was unable to make recommendations for using efanesoctocog alfa in mild to moderate haemophilia A.

The clinical experts explained that the main aim of treatment for severe haemophilia A is to prevent bleeding and resulting long-term damage, especially to joints. This is through prophylaxis to prevent bleeds, and on-demand treatment for bleeding episodes when needed. The available treatment options for long-term prophylaxis are:

• Factor VIII replacement therapy to replenish missing clotting factor in the blood through an intravenous injection: standard and extended half-life factor VIII replacement therapies are available. The standard half-life (SHL) factor VIII replacement therapies licensed for prophylaxis in the NHS (all every 2 to 3 days) are:

  • octocog alfa

  • simoctogog alfa

  • morcotogog alfa

The clinical and patient experts highlighted that current treatment options do not always prevent bleeding episodes and are associated with administration challenges. Frequent factor VIII replacement injections can damage veins, resulting in pain on administration and increasing the chance of 'vein collapse'. The frequency of injections is especially challenging in older people and young children, who often have poor venous access. It can reduce adherence to and eventually prevent the use of prophylactic factor VIII replacement therapies, leading to poor bleed control. Young children often need a central venous access device, which needs placing surgically and has an infection risk. The patient experts highlighted the potential stigma associated with visible bruising from frequent intravenous injections. They also explained that the volume and frequency of factor VIII replacement injections needed can make travelling a challenge. It can also be hard to plan injections around daily life. Also, about 5% to 7% of people with haemophilia A develop antibodies to factor VIII (called inhibitors). This makes treatment with clotting factor replacement less effective. In NHS practice, most people with haemophilia A have emicizumab, which is given subcutaneously, rather than intravenous factor VIII replacement therapies. But the patient and clinical experts highlighted that some people choose not to have emicizumab for reasons including:

• There is uncertainty about the level of bleed coverage with emicizumab compared with factor VIII replacement therapies.

• It is not a factor VIII replacement therapy, so factor VIII activity levels are not monitored via a blood test. This means there is no clinical marker of protection from bleeds.

• It cannot be used as an on-demand treatment, so people need further factor VIII replacement injections for individual bleeding episodes. These can be hard to manage, especially in young children who may need hospitalisation if they are not used to intravenous injections.

• People who contracted hepatitis C from contaminated factor VIII blood products were treated subcutaneously for the hepatitis C, so may find this administration route traumatic.

The NICE scope considered the relevant comparators for efanesoctocog alfa to be established clinical management, including:

• factor VIII replacement therapy (prophylaxis and on-demand)

• emicizumab.

In light of the evidence provided, the committee first considered the relevant comparators for PUPs. The clinical experts explained that subcutaneous administration is the only option for children until their veins are sufficiently robust to have intravenous injections. In some treatment centres, people start having emicizumab from diagnosis and then are offered SHL factor VIII replacement therapies or efmoroctocog alfa. The clinical experts explained that efmoroctocog alfa is the only EHL licensed for people under 12 years and that other EHLs are not used off-licence in young people. In some treatment centres, people start by having SHL factor VIII replacement therapies or efmoroctocog alfa, especially for children who are diagnosed with haemophilia A after the first few months of life. So, the committee concluded that emicizumab, SHL factor VIII replacement therapies (octocog alfa, simoctocog alfa or morcotocog alfa) and efmoroctocog alfa were all relevant comparators in PUPs.

The committee next considered the relevant comparators for efanesoctocog alfa in PTPs. It noted that the company had not included EHL factor VIII replacement therapy as a comparator (see section 3.5). This was because the company considered that efanesoctocog alfa would be used after EHL factor VIII replacement therapies in clinical practice. The committee noted data from the National Haemophilia Database about factor VIII therapy use in the NHS for severe haemophilia in people 12 years and over. This suggested that both SHL (octocog alfa, simoctocog alfa or morcotocog alfa) and EHL (efmoroctocog alfa, turoctocog alfa pegol or rurioctocog alfa pegol) factor VIII replacement therapies are used. The committee acknowledged that some PTPs were likely to be under 12 years, so would not be able to have turoctocog alfa pegol or rurioctocog alfa pegol. It recalled that the decision to use emicizumab or factor VIII replacement therapies is individual and based on many different factors (see section 3.2). So, efanesoctocog alfa will likely be considered for people who would otherwise have emicizumab or factor VIII replacement therapies. So, it agreed that the company's choice to exclude factor VIII replacement therapy as a comparator in PTPs was not justified. It concluded that the relevant comparators in PTPs were emicizumab, and SHL (octocog alfa, simoctocog alfa or morcotocog alfa) and EHL (efmoroctocog alfa, turoctocog alfa pegol or rurioctocog alfa pegol) factor VIII replacement therapies.

The clinical evidence for efanesoctocog alfa came from XTEND‑1, a phase 3 open-label non-randomised trial. XTEND‑1 enrolled PTPs 12 years and over with severe haemophilia A and no inhibitors to factor VIII. It had 2 arms:

• Arm A enrolled 133 people who had had a prophylaxis regimen with factor VIII replacement therapy or emicizumab for at least 6 months in the last year. People could not have had emicizumab within 20 weeks of screening. People in arm A had 50 IU/kg efanesoctocog alfa weekly for 52 weeks.

• Arm B enrolled 26 people who had had on-demand SHL or EHL factor VIII replacement therapies and had a history of 1 or more bleeds per month over the past 6 or 12 months. People in arm B had efanesoctocog alfa 50 IU/kg on-demand for the first 26 weeks, then switched to weekly efanesoctocog alfa prophylaxis for another 26 weeks.

The results of XTEND‑1 suggested that

• People having prophylaxis with efanesoctocog alfa had a reduction in ABR from baseline (prior prophylaxis). For people in arm A, the ABR for treated bleeds reduced from 3.20 at baseline to 0.71 (95%confidence interval [CI] 0.52 to 0.97) after 52 weeks. The upper limit of the one-sided 97.5% confidence interval was less than the companies prespecified value, denoting a clinically meaningful treatment effect.

• People having on-demand treatment with efanesoctocog alfa had a reduction in ABR from baseline (prior on-demand treatment). For people in arm B, the ABR for treated bleeds reduced from 35.70 at baseline to 21.42 after 26 weeks (standard deviation [SD] 7.41). After people switched to efanesoctocog alfa weekly prophylaxis for the last 26 weeks of XTEND‑1 the ABR for treated bleeds was 0.69 (SD 1.35).

• Similar improvements in bleeding rate were seen when considering any bleeds, regardless of whether or not the bleed was treated (exact results are confidential and cannot be reported here).

• Weekly prophylaxis with efanesoctocog alfa reduced the risk of bleeding compared with prestudy SHL and EHL factor VIII replacement therapy prophylaxis in an intrapatient comparison in people who participated in both arm A of XTEND‑1 and the prospective observational study (difference in mean ABR for treated bleeds -2.27, 95% CI -3.44 to -1.10; p<0.0001).

• While having weekly efanesoctocog alfa prophylaxis, 65% of arm A had no bleeds after 52 weeks of prophylaxis and 77% of arm B had no bleeds after 26 weeks of prophylaxis. Everyone in arm B had at least 1 bleed during the 26 weeks they had on-demand treatment.

• Factor VIII activity levels after weekly injections were maintained at week 26 in people having prophylaxis, suggesting a maintained response to treatment. Similar postinjection factor VIII activity levels were seen in the on-demand and prophylaxis arms.

• Improvements in baseline were seen for Haem‑A‑QoL Physical Health and EQ‑5D scores.

The NICE scope and anticipated licence for efanesoctocog alfa included all people with haemophilia A. The EAG highlighted that the population in XTEND‑1 was narrower than the scope and anticipated licence for efanesoctocog alfa because it excluded:

• people with mild and moderate haemophilia A

• people under 12 years

• PUPs

• people with inhibitors to factor VIII.

There were no trials directly comparing efanesoctocog alfa with emicizumab, so the company did an ITC to establish the relative efficacy. The clinical-effectiveness data for emicizumab came from HAVEN‑3. This was an open-label study in 152 PTPs 12 years and over with severe haemophilia A and no inhibitors. It had 4 arms:

• People who had had on-demand regimens were randomised to have prophylaxis with 1.5 mg/kg emicizumab weekly (arm A), 3 mg/kg every 2 weeks (arm B) or no prophylaxis (arm C).

• People who had had prophylaxis regimens had 1.5 mg/kg emicizumab weekly (arm D).

There were no trials directly comparing efanesoctocog alfa with efmoroctocog alfa. So, the company did an ITC using a propensity score matching (PSM) approach to establish the relative efficacy. The clinical-effectiveness data for efmoroctocog alfa came from the A‑LONG trial. This was an open-label study with 3 arms in 165 PTPs 12 years and over with severe haemophilia A and no inhibitors:

• People who had had prophylaxis entered arm 1, in which the dose of efmoroctocog alfa was increased over time from 25 to 65 IU/kg.

• People who had had on-demand therapy could enter arm 1 or be randomised to arm 2 (weekly 65 IU/kg efmoroctocog alfa) or arm 3 (on-demand therapy with 10 to 50 IU/kg efmoroctocog alfa).

The results of the MAIC using the company's preferred arms suggested that efanesoctocog alfa reduced the bleeding rate when compared with emicizumab. The incidence rate ratio (IRR) ABR for any bleed was 0.28 (95% CI 0.10 to 0.81) and for any treated bleed was 0.47 (95% CI 0.15 to 1.44). The committee noted that only the comparison of ABR for any bleed was statistically significant. Using the EAG's preferred arms showed similar results. The IRR ABR for any bleed was 0.32 (95% CI 0.19 to 0.56) and for any treated bleed was 0.50 (95% CI 0.29 to 0.86). Compared with efmoroctocog alfa, efanesoctocog alfa reduced the bleeding rate for all outcomes (IRR ABR for any treated bleed 0.29, 95% CI 0.17 to 0.51). The committee recalled that the methodology in the unanchored MAIC with emicizumab and, to a lesser extent, the PSM with efmoroctocog alfa were uncertain (see section 3.11 and section 3.12). It questioned the face validity of the company's results when incorporating the company's preferred IRRs from the MAIC and PSM into the model. This was because the ABR for any bleed with emicizumab was higher than that with efmoroctocog alfa. This contrasted with the results for any treated bleed, in which the ABR was lower for emicizumab than for efmoroctocog alfa. The committee considered that this raised concerns about the ITCs and the evidence they were based on because:

• The clinical experts considered that the difference in effectiveness between emicizumab and factor VIII replacement therapies such as efmoroctocog alfa was uncertain (see section 3.4)

• In clinical practice, most people have emicizumab (see section 3.4), and people are most likely to choose the treatment that they consider to be most effective.

The committee concluded that the results of the company's ITCs were not appropriate for decision making. This was because of concerns with the methodology of the ITC (see section 3.11, section 3.12 and section 3.13), and the lack of face validity of the results. It agreed that multiple alternative approaches should be explored to estimate the effectiveness of efanesoctocog alfa compared with the relevant comparators, either directly or through an indirect comparison. These should include but not necessarily be limited to:

• A MAIC adjusting both the A‑LONG and XTEND‑1 populations to the aggregate data from HAVEN‑3: the committee acknowledged that a PSM approach is normally preferred when individual patient data is available (see section 3.11 and section 3.12). But the committee noted the differences in the trial populations. It thought that adjusting both A‑LONG and XTEND‑1 to the HAVEN‑3 population should be explored to compare the relative effects of each comparator with efanesoctocog alfa after adjustment to the same population (that of HAVEN‑3).

• Exploring methods to anchor an ITC by:

  • Using the on-demand arms in each trial: the EAG highlighted that the treatments considered as on-demand differed by trial (efanesoctocog alfa in XTEND‑1, efmoroctocog alfa in A‑LONG and SHL factor VIII replacement therapies in HAVEN‑3). There was also no randomisation between the on-demand and prophylaxis arms in XTEND‑1 and A‑LONG, implying the need for an unanchored comparison. But the committee noted that both the Institute of Clinical and Economic Review and Canada's Drug and Health Technology Agency had done a network meta-analysis for emicizumab compared with factor VIII replacement therapies. The committee considered that 1 possible approach to include XTEND‑1 in a network would be to do a PSM between arms A and B. This would provide a relative effect adjusted for confounders. After this, the on-demand arm could be used to anchor an indirect comparison, if all on-demand treatments are assumed to be equally effective.

  • The committee considered whether the intrapatient comparisons in XTEND‑1 and HAVEN‑3 could be used to anchor an ITC using prior SHL and EHL factor VIII therapy as the common comparator. The company highlighted that the standard care with factor VIII replacement therapies was different in the 3 trials. It also noted that people with haemophilia A switch treatments on a regular basis.

• The committee recalled that both SHL and EHL factor VIII replacement therapies were relevant comparators for efanesoctocog alfa (see section 3.6 and section 3.7). It considered that there was direct evidence available for efanesoctocog alfa compared with SHL and EHL factor VIII replacement therapies from the XTEND‑1 intrapatient comparison. It considered that this direct evidence could be used to inform comparative clinical effectiveness in the economic model (see section 3.9), instead of doing an ITC.

The company developed a 3‑state Markov model to determine the cost effectiveness of efanesoctocog alfa. The health states were 'no bleeds', 'any bleeds' and 'death'. All people entered the model in the 'no bleeds' health state, after which a proportion were assumed to have a bleed each cycle. Some of these were treated with extra, on-demand factor VIII replacement injections, and others were untreated. All bleeds were associated with a short-term (7‑day) and long-term (6‑month) utility decrement, and treated bleeds accrued an extra cost. The company also modelled a utility decrement for the proportion of people assumed to have factor VIII activity levels below 20%. The cycle length was 6‑months with a half-cycle correction and a lifetime time horizon. A proportion of people transitioned to death each cycle, aligned with general population mortality. That is, no mortality benefit was assumed for efanesoctocog alfa, and people with haemophilia A were assumed to have same mortality as general population. The EAG commented that the model may have missed the granularity in bleeding severities and locations, but it expected this to have a limited impact on the results. The committee concluded that the company's general model structure was simplistic but may be acceptable for decision making.

The company's model estimated the cost effectiveness of efanesoctocog alfa compared with comparators using the following evidence sources:

• The quality-adjusted life years (QALYs) were determined by the number of treated and untreated bleeds. These were calculated using the proportion of people with a bleed each cycle and, to determine the bleeding rate in people with bleeds, the ABRs for treated bleeds and any bleeds.

• The costs were estimated using the proportion of bleeds treated (based on the ABR for treated bleeds).

The company assumed that people without bleeds and factor VIII levels above 50% would have the same quality of life as the age-adjusted general public. The company applied 2 disutilities for people who had a bleed:

• a short-term disutility applied for 7 days to reflect the pain and discomfort of bleeding, and the burden of further factor VIII injections

• a long-term disutility applied for the full 6‑month cycle, to capture anxiety related to the risk of a further bleed and limits to daily activities.

The company also applied a disutility for people whose factor VIII activity levels were under 20%. This was based on clinical expert opinion to the company that the higher risk of bleeds in people with lower factor VIII activity levels can cause anxiety and limit daily activities. The EAG highlighted that, although low factor VIII activity levels were associated with reduced quality of life in the XTEND‑1 trial, these people were monitored more frequently than they would be in clinical practice. If people were unaware of low factor VIII activity levels, they would be unlikely to limit activities or have anxiety over the risk of bleeds. So, the company's approach may have overestimated the disutility in people with low factor VIII activity levels. The patient experts stated that they were likely to be more cautious and adjust their daily activities if they knew their factor VIII activity levels would be low. The patient experts explained that they would be aware that their factor VIII activity levels would be low shortly before their next dose. This would be the case even if they had not measured their factor VIII activity levels. One clinical expert estimated that people were unlikely to have spontaneous bleeds with factor VIII activity levels of over 10% or bleeds with minor trauma with levels of over 15%. The committee also noted that people with a factor VIII activity level of 20% were classed as having mild haemophilia A, so would have a relatively low risk of bleeding. So, the committee agreed that applying a disutility for people with a factor VIII activity level of 20% was inappropriate. The committee also noted that the company had modelled everyone having emicizumab as having factor VIII levels of between 5% to 20%. This was based on a study by Shima et al. (2016) in non-human primates. So, everyone having emicizumab accrued a disutility for having low factor VIII activity levels. This was a much higher percentage than for people having efanesoctocog alfa and efmoroctocog alfa. The committee also considered whether it was appropriate to apply a disutility based on factor VIII activity at all for people having emicizumab. It noted that emicizumab does not work by replacing factor VIII so factor VIII activity levels cannot be used to measure bleeding protection. The committee recalled that the relative clinical effectiveness between treatments was unknown (see section 3.13). So, it was concerned that the company's approach to modelling a disutility for low factor VIII activity levels may have introduced a bias against emicizumab. The committee concluded that it was plausible that having low factor VIII activity levels reduced quality of life in people having factor VIII replacement therapies. But it considered that it was unclear whether having low factor VIII activity levels would affect quality of life in people having emicizumab. Also, the committee recalled that there were substantial limitations to the company's approach to capturing utilities using Tobit models that affected the model (see section 3.17). It agreed that it would like to see:

• justification that factor VIII levels affect quality of life in people with haemophilia A and how they do this, including people who are having emicizumab

• modelling of quality of life that is consistent and coherent in how it reflects differences in quality of life between treatments (see section 3.17) and the effects of factor VIII levels

• a scenario in which disutility is applied to people with factor VIII activity levels of under 15%.

The company's model assumed that a proportion of people with bleeds each cycle would need further treatment with on-demand factor VIII therapies (see section 3.15). People having prophylactic efanesoctocog alfa had on-demand treatment with efanesoctocog alfa and people on efmoroctocog alfa had treatment with efmoroctocog alfa. People having prophylactic emicizumab had on-demand octocog alfa. The EAG was concerned that the company had modelled a 50 IU/kg on-demand dose for efmoroctocog alfa and octocog alfa, but only 25 IU/kg for efanesoctocog alfa. The company based this on clinical expert opinion that the sustained pharmacokinetic profile of efanesoctocog alfa would mean a lower dose would be effective at controlling bleeds. The EAG noted that XTEND‑1 used an on-demand dose of 50 IU/kg to treat bleeds that occurred on efanesoctocog alfa prophylaxis, followed by a further 30 IU/kg if needed. It noted that, in XTEND‑1, 77% of people with a bleed on efanesoctocog alfa prophylaxis in arm A had around 50 IU/kg efanesoctocog alfa to stop bleeding. So, there was no clinical-effectiveness data using the company's preferred dose of 25 IU/kg. For this reason, the EAG used a dose of 50 IU/kg for bleeds for all modelled treatments in its base case. The patient experts confirmed that, in their experience of treating bleeds with on-demand efanesoctocog alfa in XTEND‑1, 1 dose of 50 IU/kg was usually sufficient in controlling bleeds to an extent to which people could work and carry out daily activities. They explained that many people prefer to avoid further retreatment after the initial on-demand dose unless their symptoms worsen. So, it was unlikely that an extra 30 IU/kg would be used in clinical practice. The committee noted that XTEND‑1 showed that factor VIII levels were maintained to around 10% a week after injection. One clinical expert stated that a 50 IU/kg dose of efanesoctocog alfa would usually not be needed to increase the factor VIII activity levels to such an extent that the bleeding stops. The expert explained that the dose used would be based on bleed location and time since last injection. But the committee was concerned that the pharmacokinetics of on-demand and prophylactic therapy may be different. This is because stopping a bleed uses factor VIII faster than preventing a bleed, meaning that a higher dose of efanesoctocog alfa would be needed. The clinical experts confirmed that this was plausible. They highlighted that the half-life of factor VIII replacement therapies are reduced after surgery. The committee recalled that people would only have spontaneous bleeds if their factor VIII activity levels were under 15% (see section 3.18). So, the committee was concerned that a 25 IU/kg dose of efanesoctocog alfa may not be enough to stop the bleeding. The patient experts also highlighted that people having efanesoctocog alfa prophylactically would be likely to use the same dose for treating bleeds. This is because they would have this dose readily available and it would avoid wastage of unused vials at a lower dose. The committee concluded that a dose of 50 IU/kg of efanesoctocog alfa for treating bleeds occurring on prophylaxis should be used to align with data from XTEND‑1.

The company included costs for treatment acquisition and bleed management in its model. It did not model any treatment administration costs because treatments are self-administered. It assumed wastage costs only for octocog alfa, the on-demand treatment for people having emicizumab (see section 3.19), using the proportion of people in HAVEN‑3 who remained bleed free on emicizumab. The company stated that wastage costs would not be relevant for efanesoctocog alfa, efmoroctocog alfa and emicizumab. This was because the number of doses used for prophylaxis was rounded up or down to a full vial by healthcare professionals. The patient experts highlighted that, for people on emicizumab, additional vials of factor VIII replacement therapy are needed at home for on-demand treatments. These can be wasted if no bleed occurs during the lifetime of the product (around 2 years in the fridge). The EAG highlighted that it considered the company's approach to modelling wastage of octocog alfa inappropriate. This was because it implied that the same people had bleeds each cycle, and the company had not justified the dose of octocog alfa assumed to be wasted. So, the EAG considered that the company should provide further justification, including consultation with clinical experts, for its approach. It also thought the company should consider revising its approach to modelling the expected wastage cost for octocog alfa. The committee recalled its preferred doses for treating bleeds with efanesoctocog alfa and efmoroctocog alfa aligned with those used for prophylaxis, which would minimise waste (see section 3.19). But the clinical experts highlighted that haemophilia A treatments are weight based. This means that people may 'round up' and use more drug than needed because of poor correlation between vial size and weight-based dose. So, although there would be no wastage (that is, no excess vials would be thrown away), the NHS cost of people using a higher dose than needed should be reflected in the model. The company's model did not account for this cost. The committee noted that the company had provided a scenario assuming wastage costs for emicizumab but not for efanesoctocog alfa or efmoroctocog alfa. So, the effect on the incremental cost-effectiveness ratio (ICER) was unknown. The committee concluded that the treatment costs for the amount of efanesoctocog alfa, efmoroctocog alfa, emicizumab and octocog alfa expected to be used in clinical practice, including 'rounding up' of doses, had not been appropriately considered in the company's model. So, the committee thought that the company should amend its model to account for the expected use of each treatment in clinical practice.

The company assumed that each bleed incurred a cost for management, including emergency, specialist and nurse visits. The number of emergency and specialist visits was based on Shrestha et al. (2017). Because this paper reported the need for multiple specialist visits per bleed, the company assumed that no additional nurse visits would be needed. The company's clinical experts also confirmed that people would be keen to avoid emergency visits if possible, and that bleeds are often managed by a combination of in-person and telephone consultations with specialists. The EAG noted that most bleeds in XTEND‑1 were joint and muscle bleeds that resolved with 1 injection, so were likely mild to moderate. It considered that mild to moderate bleeds would likely be managed over the phone and often by specialist nurses instead of doctors. It also noted that the company had not used the most recent NHS reference costs and the costs for specialist nurse visits had increased substantially. So, the EAG was concerned that the company's assumption may not represent the costs of treating bleeds in the NHS. To account for this uncertainty, the EAG provided scenarios in which the current number of specialist visits were spread among specialists and nurse visits (to account for bleeds resolved by phone), and in which all resource use was because the management of bleeding episodes were omitted. The committee considered that, given the uncertainties highlighted by the EAG, the costs of managing bleeds was uncertain. It agreed that the company should provide further justification for its proposed bleed management costs and amend its approach to modelling these if necessary.

The committee noted that there were uncertainties in the evidence base and modelling assumptions, specifically:

• the comparators for efanesoctocog alfa (see section 3.6 and section 3.7)

• the effectiveness of efanesoctocog alfa compared with comparators from the ITC (see section 3.11, section 3.12 and section 3.13)

• the utility values used by the company, including any disutility for low factor VIII activity levels (see section 3.17 and section 3.18)

• wastage costs for efanesoctocog alfa and efmoroctocog alfa (see section 3.20)

• resource use and costs for managing bleeds (see section 3.21).

Because of confidential commercial arrangements for efanesoctocog alfa, the comparators and other treatments in the model, the exact cost-effectiveness estimates are confidential and cannot be reported here. The ICERs for the comparisons against emicizumab (both in PUPs and PTPs) were within the range normally considered an acceptable use of NHS resources in both the company's and EAG's base cases. In the company's base-case analysis for the comparison against efmoroctocog alfa in PUPs, the ICER was within the range normally considered an acceptable use of NHS resources. In the EAG's base-case analysis for the comparison against efmoroctocog alfa in PUPs, the ICER was above the level normally considered an acceptable use of NHS resources. The committee noted that the driver for the ICER being higher in the EAG's base case was using a 50 IU/kg dose of efanesoctocog alfa for treating bleeds.

The committee preferred the model to:

• include emicizumab, SHL (octocog alfa, simoctogog alfa and morcotogog alfa) and EHL (efmoroctocog alfa and, in PTPs only, turoctocog alfa pegol and rurioctocog alfa pegol) factor VIII replacement therapies as relevant comparators for both PTPs and PUPs

• explore alternative approaches to estimating the relative clinical effectiveness of efanesoctocog alfa compared with comparators, including SHL (octocog alfa, simoctogog alfa and morcotogog alfa) and EHL (efmoroctocog alfa and, in PTPs only, turoctocog alfa pegol and rurioctocog alfa pegol) factor VIII replacement therapies

• explore alternative methods of modelling utilities

• explore alternative approaches to modelling wastage

• using an on-demand dose of 50 IU/kg of efanesoctocog alfa for treating bleeds while having prophylaxis.

The committee noted that people who carry the haemophilia gene may have mild or, rarely, moderate to severe symptoms of bleeding. It noted that all carriers of haemophilia A have XX chromosomes, so carrier status is affected by biological sex. But it recalled that it had not been presented with any evidence for the mild or moderate haemophilia A populations. It also noted that a recommendation in severe haemophilia A would not be restricted by biological sex. It recalled that there were differences in the treatment pathway and potential treatment effect. These differences meant clinical- and cost-effectiveness outcomes would likely be different between people with severe and mild to moderate haemophilia A (see section 3.2). So, it could not make a recommendation for this population. Stakeholders also highlighted that some of the treatments for haemophilia A, including efanesoctocog alfa, are derived from human blood or human or animal cells. This may not be considered acceptable by people with some religious beliefs. The committee was aware that there are several treatment options from different sources that people may choose. These include emicizumab, which is not derived from human blood products. The committee did not identify this as an equalities issue that would affect its recommendations. The committee concluded that all equalities issues for efanesoctocog alfa had been considered in its decision making.

The committee noted that some potential benefits of efanesoctocog alfa may not have been included in company's model. The company, and the patient and clinical experts described the uncaptured benefits of weekly dosing of efanesoctocog alfa, compared with more frequent dosing of factor VIII replacement therapies, including:

• a reduced treatment burden for people with the condition and their carers (especially considering that severe haemophilia A may affect several siblings in the same family)

• improved vein health, especially in older people who have been using factor VIII replacement therapies for a long time; the committee considered it unlikely that the need for a venous access device for children would decrease as weekly injections would still be needed

• improved treatment adherence

• freedom to travel and participate in sports more easily, which can reduce obesity levels and related comorbidities in later life.

The committee noted the important uncertainties in the clinical-effectiveness evidence, and agreed that the inputs in the economic model made it unsuitable for decision making. It also noted that evidence comparing efanesoctocog alfa with all relevant comparators was not available. This meant it was not possible to reliably estimate the cost effectiveness of efanesoctocog alfa. So, it is not recommended. The committee concluded that the company should provide additional information for consideration at the next evaluation committee meeting (see section 3.25).