Apremilast for treating moderate to severe plaque psoriasis

Clinical management

4.1 The committee heard about the experience of people with psoriasis. It heard that the disease results in itchy, dry, scaly and thickened skin, which can be physically and psychologically debilitating, particularly if located on the hands, feet, face and genitals. Severe psoriasis is also associated with a shortened life expectancy. The committee heard that, because psoriasis is visible to others, it can make people feel isolated and lonely, which could lead to them losing self-confidence and avoiding social situations, and could also affect career opportunities and influence intimate relationships. It heard from clinical experts that people with severe psoriasis are about 6 times more likely to have suicidal thoughts or commit suicide than the general population. The committee agreed that severe psoriasis has a significant psychosocial impact and substantially decreases quality of life.

4.2 The committee discussed how clinicians assess the severity of disease in people with psoriasis. It understood that several indices are used, and heard that clinicians routinely use both the Psoriasis Area Severity Index (PASI) and the Dermatology Life Quality Index (DLQI) to monitor the disease. The committee was aware that the UK marketing authorisation for apremilast includes people with moderate and severe disease. It understood that, while the marketing authorisation did not specify the criteria for moderate or severe disease, the regulatory decision was based on trials that included people with a baseline PASI score of 12 or more (mean score 19), a 10% or greater of body surface area affected, and a static Physician Global Assessment score of at least 3. The committee noted that the trials did not stipulate that participants have a particular DLQI score at baseline. The committee was aware that previous NICE technology appraisals defined severe psoriasis as a PASI score of 10 or more and a DLQI of more than 10. It understood that there is no universally accepted definition of moderate psoriasis. However, the committee heard from clinical experts that, in practice, moderate disease would be characterised by a lower PASI score (between 5 and 9). It noted that the company, in its submission, presented analyses in people with a PASI score of 10 or more and a DLQI score of 10 or less to represent people with disease not severe enough to be treated with biologicals in the UK. Clinical experts explained that the disadvantages of the DLQI are that it is not specific to psoriasis and does not capture all of the impacts of the disease (such as anxiety and depression). Clinical and patient experts suggested that some people with chronic psoriasis can develop coping mechanisms and so adjust to the impact of the disease, resulting in lower DLQI scores. The committee heard that clinicians use the DLQI for treatment decisions with biologicals, but do not generally use it to define different levels of severity. The committee acknowledged that PASI and DLQI, which reflect the outcomes used in the trials, are relevant measures used in clinical practice in the NHS. The committee concluded that the evidence base for apremilast reflected people with severe disease, as defined in UK clinical practice.

4.3 The committee considered the treatment pathway for people with psoriasis. It was aware that people have topical treatments as first-line treatment, followed by conventional (non‑biological) systemic therapies (such as methotrexate or ciclosporin), and phototherapy. If these treatments do not adequately control the psoriasis, people may have biological therapies, which they continue to have as long as the drugs work. The committee understood that if the disease no longer responds to a biological therapy, people will be offered another biological therapy. This pattern is likely to be repeated over their lifetime; clinical experts noted that people with psoriasis will often try many alternative biological agents in a short timeframe. The committee heard that, for people whose disease does not respond to multiple biological agents, the only remaining treatment option is best supportive care, which usually consists of topical agents and bandaging. The committee questioned whether best supportive care was effective in treating psoriasis (that is, whether it improved PASI score or other measures of disease). It heard from clinical experts that best supportive care can be effective in the short term. The committee was aware that best supportive care can be accompanied by disutility because of the intensive time-consuming, inconvenient and unpleasant treatments, and that the psoriasis may worsen sooner than with biological therapies. The committee concluded that best supportive care is associated with limited, short-term efficacy and recognised the value of having a range of treatments with different mechanisms of action available.

4.4 The committee considered the potential positioning of apremilast in the treatment pathway in NHS clinical practice. It noted that the marketing authorisation for apremilast allows it to be positioned:

4.5 The committee considered whether apremilast would be used before biological therapies were considered, that is, after 1 but not all other systemic treatments had failed. It noted that the company had not presented analyses for apremilast in this position, and that it did not hear from clinical experts that they would offer the drug at this point. The committee therefore did not consider this position further.

4.6 The committee discussed whether apremilast would be used if biological therapies were not tolerated or after all biological therapies had failed. It heard from a clinical expert that apremilast would generally be prescribed for people who had already tried biological therapies, or for those who are unable to take them. The committee understood from the clinical experts that, in general, apremilast would not displace a biological therapy in the treatment pathway. It concluded that the most likely position for apremilast in the treatment pathway was if biological therapies were not tolerated or after all biological therapies had failed.

4.7 The committee discussed whether apremilast would be used at the same point in the treatment pathway as biological therapies (that is, once all other systemic therapies had failed). It heard from clinical experts that the positioning of apremilast (either before, or instead of, a biological therapy) would be driven largely by patient choice and intolerance or contraindications to biological therapy such as serious infections. The committee heard from:

Clinical effectiveness

4.8 The committee appreciated that the clinical evidence for apremilast came from 4 multicentre placebo-controlled double-blind randomised controlled trials (RCTs) in people with moderate to severe chronic plaque psoriasis: PSOR‑005, -008, -009 and -010. The primary endpoint was the same in all 4 trials: a 75% reduction in the PASI score at week 16 (known as a PASI 75 response). The randomised period of all trials was only until week 16. The committee considered the baseline characteristics of the patients in the apremilast trials and discussed whether the trials reflected the UK population with psoriasis who would be eligible to have apremilast. It noted the evidence review group (ERG's) comment that only 13% of the PSOR‑008 trial population would be eligible for apremilast. The committee heard from the company that, in its view, the ERG had misinterpreted the intended positioning of apremilast. The company clarified that, according to the UK marketing authorisation, apremilast could be used after only 1 conventional systemic treatment, which could be phototherapy. The company stated that about 65% of the PSOR‑008 and PSOR‑009 trial populations had any prior systemic therapy (including conventional therapies but also biological agents). The committee noted the ERG's comments that the PSOR‑008 and PSOR‑009 trials may have overestimated the benefit of apremilast; some patients in the trial had no previous systemic treatment, so their disease would have been more likely to respond to apremilast because it was less severe than in people who had more treatments. The committee heard from clinical experts that there is no robust evidence to suggest that previous treatment with conventional therapy affects response to subsequent treatment. The committee heard from the company that other criteria for inclusion and exclusion used in apremilast trials were similar to trials of biological therapies. The committee concluded that the apremilast trials provided an appropriate basis for its decision-making.

4.9 The committee discussed the results of the placebo-controlled apremilast clinical trials, and considered the company's network meta-analysis, which indirectly compared apremilast with other biological agents specified in the scope. It noted that clinical trial evidence showed that apremilast was more effective than placebo for key outcomes at 16 weeks, and that this benefit was consistent across subgroups studied. The committee heard from clinical experts that biological therapies are more effective than apremilast for treating psoriasis, and was aware that the company's network meta-analysis showed that the probability of response to treatment is highest with biological therapies and lowest with apremilast. The committee heard from the ERG that the company's network meta-analysis was technically robust, but any bias from trials would carry through to the results from this analysis. The committee heard that the ERG compared the meta-analysis results for the apremilast: etanercept comparison with odds ratios calculated using results of the PSOR‑010 trial, confirming that psoriasis is more likely to respond to etanercept than apremilast. The committee was aware of the drawbacks of the PSOR‑010 study in that it was powered to compare apremilast with placebo and etanercept with placebo (but not apremilast with etanercept), but still concluded that apremilast is more effective than placebo, but not as effective as biological therapies.

4.10 The committee discussed the long-term effectiveness of apremilast, noting that a substantial proportion of people who had a PASI 75 response after 16 weeks of apremilast treatment subsequently lost their PASI 75 response during the treatment withdrawal phase (marked as 'academic in confidence' by the company). The committee heard from the company that some of the people in the apremilast trials who lost response, later regained it. The company stated that everyone who had a PASI 75 response at week 16 had at least a PASI 50 response by week 52. The clinical experts stated that fluctuating PASI scores characterise the natural history of psoriasis and any decline in efficacy with apremilast is expected to be similar to a decline with biological therapies. The experts also stated that the same level of response would not be seen if patients were to stop and restart treatment with apremilast. The committee concluded that some response to treatment with apremilast is maintained up to 1 year, but noted uncertainty about longer-term effectiveness beyond the clinical trial data.

4.11 When considering the tolerability of apremilast, the committee heard from clinical experts that apremilast is associated with a number of adverse events early in treatment. The committee heard that the most common adverse events with apremilast are related to the gastrointestinal tract (including diarrhoea and nausea), and that people are willing to tolerate gastrointestinal adverse effects if they are benefiting from the drug. The committee concluded that apremilast is associated with gastrointestinal adverse events, but these would not preclude the use of apremilast.

4.12 The committee discussed the probability of people stopping apremilast treatment, and heard from clinical experts that the rate of withdrawal from apremilast is similar to (or potentially lower than) the rate of withdrawal from biological therapies. It heard from the company that the ERG's reference to a higher withdrawal rate with apremilast based on PSOR‑008 data was factually inaccurate (the company marked this withdrawal rate as 'academic in confidence'). The company stated that 2 year data from PSOR‑008 indicate a withdrawal rate of 19.5% per year for apremilast, which is similar to the company's assumption of a 20% withdrawal rate for biological therapies, which it chose based on the rate used for biological therapies in previous NICE appraisals (adalimumab, etanercept, infliximab, secukinumab and ustekinumab). The committee concluded that the probability of people stopping treatment with apremilast is likely to be similar to the probability of people stopping treatment with biological therapies.

Cost effectiveness

4.13 The committee considered whether the company's health economic model included relevant treatment sequences and positions. The committee noted that the treatment sequences modelled by the company reflected the likely positioning of apremilast according to the clinical experts (see sections 4.4 to 4.7). The committee heard from clinical experts that etanercept is the least effective of the existing biological therapies for treating severe psoriasis. When queried, the company stated that it had selected etanercept for the model because it is the most used biological therapy for psoriasis in Europe. The ERG stated that this would not be a key driver of the results because the model includes biological therapies in both arms in the model and would not be displaced in the sequence including apremilast. Based on the likely positioning of apremilast in the treatment pathway (see sections 4.4 to 4.7), the committee concluded that, although the company did not compare the sequence in which apremilast came after biological therapies with a sequence without apremilast, the positions and comparisons modelled by the company were generally sufficient for decision-making.

4.14 The committee discussed the sources used by the company to estimate resource use and costs associated with best supportive care. It noted that the incremental cost-effectiveness ratios (ICERs) were highly sensitive to these inputs, and specifically whether the model included hospitalisation rates and costs from Fonia et al. (2010; the ERG's preferred assumption of best supportive care costs of £348 per cycle based on 6.49 days of hospitalisation per year) or NICE's psoriasis guideline (the company's base case assuming best supportive care costs £888 per cycle based on 26.6 days of hospitalisation per year). The committee heard from the clinical experts that both sources were likely to overestimate the actual number of hospital days and resource use associated with best supportive care. This is in part because the populations described in Fonia et al. and NICE's guideline differed from the population covered in this appraisal; NICE's psoriasis guideline was for a specific, high-need subpopulation with very severe psoriasis, and Fonia et al. described care in a tertiary care centre known for treating the most severely affected people. The committee heard from the company that the Fonia et al. study reflected a site that offered day-care and therefore admitted fewer people to hospital than would normally be admitted in clinical practice. The clinical expert stated that the Fonia et al. study describes a day unit that offered on‑site hotel accommodation to people. The committee noted that this option is much less costly than a hospital stay because it would not incur nursing and other hospital costs and, increasingly, is the model of care for people with psoriasis in the NHS. The committee also heard from the clinical experts that, in recent years, the number of people hospitalised for severe psoriasis has fallen, and that clinicians give best supportive care to people during their outpatient visits; therefore, hospitalisation costs associated with psoriasis have fallen, and are continuing to fall. The committee noted that after consultation, the company provided NHS hospital episode statistics data that showed that the average length of hospital stay associated with best supportive care was 3.5 days. It heard from the company that in its view, these values underestimate actual length of NHS hospitalisation because they include people with different disease severities as well as people receiving concomitant medication and that, in patients who had received inpatient care, the average length of stay was 10.74 days. The clinical experts agreed that the hospital episode statistics data underestimated length of hospitalisation. The committee agreed with this, but considered that the most plausible estimate would be lower than the ERG and company assumptions of 6.49 and 26.6 days per year. With respect to the proportion of people admitted to hospital, the clinical experts noted that the actual proportion is much lower than the 30% annual probability assumed by Fonia et al. (the lower of the 2 estimates). The committee recognised the significant uncertainty in this model input, and agreed that the best supportive care costs are likely to be lower than in Fonia, and also noted that assuming a lower cost would increase the ICER. Additionally, the committee noted that costs associated with hospitalisation were consistently applied to all days in hospital. The committee, however, considered that it was reasonable to expect that the first few days in hospital would generate more expenses than later days; therefore the company may have overestimated the overall costs of best supportive care associated with hospitalisation in the model and accounting for this was likely to increase the ICERs presented. The committee recognised the considerable uncertainty and concluded that shortcomings exist among all sources of data for resource use, but that resource use for best supportive care is closer to Fonia et al. than to the estimates from NICE's guideline on psoriasis.

4.15 The committee went on to discuss the cost estimates for people whose disease does not respond to one systemic treatment and who then go on to get another one, during which time they need elements of best supportive care. The committee recognised this as a significant driver of cost-effectiveness results. The committee heard that the company had originally assumed that this resource would be lower than best supportive care costs and included an estimate of £460 per cycle, based on their preferred best supportive care costs from the NICE guideline on psoriasis (see section 4.11). The ERG instead based its preferred estimate on the study by Fonia et al. (2010) and also assumed that this additional resource is the same as for best supportive care, therefore including an estimate of £348 per cycle. After consultation, the company explored the effect on the cost‑effectiveness results of using the costs from Fonia et al. for the 12 month period after patients start biological therapies. However the company noted that because it had already included costs for outpatients and systemic treatments in its model (also from Fonia et al.), it deducted these costs to avoid double counting, resulting in a cost for people whose condition does not respond of £45 per cycle. The committee agreed that avoiding double counting was appropriate, but also agreed with the ERG that using costs from the period after starting biological therapies includes costs for people whose condition responds, as well as costs for people whose condition does not respond to biological therapies; therefore underestimating the true costs. Instead, the committee preferred the ERG's inclusion of costs from Fonia et al. that reflect the 12 month period before a person starts a biological therapy while at the same time reducing the costs of outpatient and systemic treatments (£225 per cycle). The company agreed during the second committee meeting that this estimate was plausible, but highlighted that the ERG applied this cost to all people starting a new biological therapy for the 16 week 'trial' period, whereas a proportion of people on new treatments benefit immediately or at least quickly. The company acknowledged that there was uncertainty around the proportion of patients to which this 'non-responder' cost should apply. The clinical experts agreed that response varies among people who try new active treatments, and that it is unlikely that all people will incur the same 'non-responder' costs. The committee heard from the ERG that the costs from Fonia et al. include patients at different levels of response and therefore the model already accounted for variable non-responder costs during the period in which modelled patients try new drugs. The committee concluded that there was considerable uncertainty about the actual costs associated with starting a new therapy if a person's disease has previously not responded to another therapy in the model, but concluded that the ERG's estimate of £225 per cycle was the most plausible of those presented.

4.16 The committee considered the quality-of-life and utility values used in the company's original model. It was aware that the company did not use the baseline EQ‑5D data collected in the apremilast clinical trials. The committee heard that the company sourced the baseline utility value in the model (0.7) from previous technology appraisals, and noted that people in the apremilast clinical trials had a baseline utility value of 0.8. The company stated that the baseline EQ‑5D value in the apremilast trials differed from the baseline value in trials of biological therapies (although other key baseline characteristics such as PASI score were consistent across the trials). The ERG confirmed that changing the baseline utility value used in the model would not significantly affect the cost-effectiveness results because the model is driven by the incremental changes in utility score from baseline. The committee then discussed the company's approach to incorporating utility increments associated with a response to treatment in the model. The committee was concerned that the company had included trial-based EQ‑5D data in its model reflecting a DLQI score of 10 or less but that, in the model reflecting a DLQI score of more than 10, the company had used a mapping algorithm instead of clinical trial data. The committee was also concerned that the company used etanercept data in the mapping algorithm instead of apremilast data. The company justified this on the basis that EQ‑5D data in the apremilast trials were subject to a ceiling effect. However, following the first committee meeting the company clarified that it had, in error, derived EQ‑5D data from US instead of UK tariffs. The company agreed that the updated trial-based EQ‑5D data was appropriate and presented revised results for the change in utility from baseline associated with the different PASI response categories in the model. The committee noted that the company's models did not take into account the disutility values associated with adverse events, but the ERG was unable to comment on how including these values would have affected the ICER. The committee concluded that the utility gains estimated from the company's revised model (for people with a DLQI score of more than 10) were plausible.

4.17 The committee discussed the assumption in the company model that apremilast is associated with fewer visits to a clinician than biological therapies (1 annual visit for apremilast, compared with 4 visits per year with biological therapies). The committee heard from the clinical experts that because apremilast is a new drug dermatologists would be cautious in allowing fewer visits to a clinician and therefore the number of clinician visits (including GP visits in 'shared care' arrangements) for apremilast and biological therapies would be the same initially, but, in the long term, the clinical experts expected that the number of monitoring visits would be lower for apremilast. The committee did not consider it realistic that people receiving apremilast would visit their clinician only once each year, and concluded that 4 visits per year (including visits to GPs) is more appropriate (based on the company's assumption of 4 visits for biological therapies). The clinical experts agreed with this, but noted that monitoring costs for apremilast were likely to be lower than for biological therapies. The clinical experts also confirmed that for treatment with apremilast GPs could monitor patients under shared care arrangements. However, the committee noted that the cost of monitoring people on apremilast in primary care should reflect the assumption of additional payments to GPs, similar to arrangements for monitoring methotrexate. The committee noted that if lower monitoring costs for apremilast were incorporated in the model, this would lower the ICERs presented, but acknowledged the lack of robust estimates available. Therefore, the committee concluded that the assumption of equal monitoring visits was appropriate. The committee discussed the potential for waste with apremilast. The clinical experts acknowledged that when people did not adhere to or withdrew from treatment, some tablets would be wasted. The committee considered that it would be reasonable to account for some treatment waste with apremilast. It noted the company's revised model, which included an assumption of 14 days' wasted treatment; it heard from the ERG and experts that this was plausible. The company additionally stated that a waste assumption should be applied equally to biological therapies and apremilast. The ERG, however, considered that the waste for biological therapies in each arm of the model would cancel each other out but the company stated that would be true only if a lifetime horizon was assumed. The clinical experts also stated that while there will be waste with biological therapies, this is low because people get rigorous training before being prescribed biological therapies, in order to boost adherence. The committee concluded that the assumption of 14 days' apremilast waste in the revised model was appropriate.

4.18 The committee discussed the company's assumption that the probability of response for each treatment was the same regardless of its position in the sequence. It heard from the company that clinical trials show that the efficacy of apremilast appears consistent across the positions in the treatment sequence. The company noted that the response was slightly lower if apremilast were positioned after biologicals, and that its model accounted for this reduced efficacy. The committee was satisfied that the company had included the efficacy of apremilast appropriately in its model.

4.19 The committee considered the other assumptions in the company model in light of its clinical discussion. It concluded that the model should include the possibility that psoriasis can improve with best supportive care (in contrast to the company's assumption of no effectiveness, see section 4.3). The committee agreed with the company's assumptions that withdrawal rates are similar for apremilast and biologicals (see section 4.9), and that response rates remain relatively constant over time (see section 4.7).

4.20 The committee discussed the ICERs for apremilast positioned before biological therapies in a population with a PASI score of 10 or more and a DLQI score of more than 10.The committee considered the company's revised base-case results and the ERG's exploratory analyses. The committee concluded that the most plausible ICER available for decision-making was about £30,300 per quality-adjusted life year (QALY) gained and noted that this was above the range normally considered cost effective. However it noted that there was considerable uncertainty about key factors driving this ICER, such as monitoring costs (see section 4.13), amount of drug waste (see section 4.14), the likely costs associated with best supportive care (see section 4.11) and the costs associated with 'non-responders' (see section 4.12). The committee considered that these uncertainties could drive the ICERs in different directions and the magnitude of impact was uncertain. However, the committee recalled its consideration in section 4.11 that the costs associated with best supportive care are likely to be even lower than those estimated by the ERG from Fonia et al. (2010) and accounting for this would increase the ICER. Moreover, being mindful that apremilast was not as effective as biological therapies, the committee noted comments from consultation and from the patient expert that apremilast in a sequence before biological therapies could delay access to more effective treatment, and may therefore not be preferred. The committee heard from the patient expert that achieving clear skin in the shortest possible time is important to people and that a PASI 75 response shown with apremilast means that people are still affected by psoriasis. Together with the uncertainties in the economic modelling, the committee concluded that apremilast could not be recommended for severe psoriasis after the failure of conventional systemic therapy but before biological therapy.

4.21 The committee discussed the ICERs for apremilast positioned before biological therapies in a population with a PASI score of 10 or more and a DLQI score of 10 or less (moderate disease), and where best supportive care was the only comparator because patients with moderate disease are not offered biological therapies. At its first meeting, the committee concluded that the most plausible ICER for the apremilast sequence lay somewhere between £97,500 and £125,300 per QALY gained, taking into account its preferred assumptions. The committee noted that the company had not updated this analysis with the UK tariff-based utility values, and estimated, based on the original modelling, that the ICER in the less severely affected population could be twice that seen for the population with a PASI and DLQI of 10 or more, that is, about £60,000 per QALY gained. The committee noted that the evidence base for apremilast did not include people with moderate disease as defined in UK clinical practice (a PASI score of 5 to 9). Given that the company's model indicated that apremilast had a higher ICER in a less affected population (that is, people with a DLQI score of 10 or less), the committee concluded that the ICER for apremilast for treating moderate psoriasis would not be within the range considered to be a cost-effective use of NHS resources.

4.22 The committee considered the company's cost-effectiveness results for apremilast positioned after biological therapies and before best supportive care. It noted that the sequence in which apremilast was positioned after biological therapy was dominated (provided fewer QALYs at a higher cost) by the sequence in which apremilast came before biological agents. Having already concluded that apremilast, as a treatment in a sequence before biological therapy, is not a cost-effective use of NHS resources (see section 4.17), the committee concluded that a treatment sequence that provides fewer QALYs but costs more could not be considered a cost-effective use of NHS resources.

4.23 The committee considered the company's cost-effectiveness results for apremilast as a replacement treatment for 1 of the biological therapies in the sequence, even though the clinical experts stated that apremilast was unlikely to displace a biological agent in the treatment pathway. It noted that the sequences containing apremilast were cost saving – but less effective – than the comparator sequences, resulting in ICERs that reflected 'savings per QALY lost' (ranging from £21,100 to £39,100 per QALY). The committee considered that the ICERs were based on uncertain assumptions and noted that that ICERs based on its preferred assumptions were not available. The committee noted that, in situations in which an ICER is derived from a technology that is less effective and less costly than its comparator, the commonly assumed decision rule of accepting ICERs below a given threshold is reversed, and so the higher the ICER, the more cost effective a treatment becomes. The committee concluded that the ICERs for apremilast as a replacement for biological therapies for severe psoriasis were not within the range considered to be a cost-effective use of NHS resources.

4.24 The committee heard differing views about whether apremilast was innovative in its potential to make a significant and substantial impact on health-related benefits. It agreed that apremilast appeared to be innovative in its novel use of an existing mechanism of action, and agreed that it demonstrated innovation by providing an additional novel oral therapy. However, the committee recalled that some people with psoriasis prefer less frequent injectable treatments to more frequent oral ones (see section 4.4) and therefore the committee concluded that, in this respect, there were no additional gains in health-related quality of life over those already included in the QALY calculations.

4.25 The committee considered when appraising apremilast whether it should take into account the consequences of the Pharmaceutical Price Regulation Scheme (PPRS) 2014, and in particular the PPRS payment mechanism. The committee noted NICE's position statement in this regard, and accepted the conclusion 'that the 2014 PPRS payment mechanism should not, as a matter of course, be regarded as an applicable consideration in its assessment of the cost effectiveness of branded medicines'. The committee heard nothing to suggest that there is any basis for taking a different view on the relevance of the PPRS to this appraisal. It therefore concluded that the PPRS payment mechanism was irrelevant for the consideration of cost effectiveness of apremilast.

4.26 The committee discussed whether the recommendations could be considered unfair because a small group of people are unable to take any biological therapies because of contraindications that could be associated with protected characteristics according to the Equality Act. The committee concluded that these patients would otherwise have best supportive care, and although the company did not present a revised ICER comparing apremilast with best supportive care for this group, the committee expected the ICER to be around £60,000 per QALY gained. Considering that this was much higher than what is normally accepted as good use of NHS resources, the committee concluded that to recommend apremilast for this group would not achieve the legitimate aim of providing advice on cost-effective treatments.

Rapid review

Summary of appraisal committee's key conclusions