Fingolimod for the treatment of highly active relapsing–remitting multiple sclerosis

Clinical effectiveness

4.6 The Committee noted that only part of the population covered by the marketing authorisation for fingolimod was considered in the manufacturer's submission, that is, people with highly active relapsing–remitting multiple sclerosis who have an unchanged or increased relapse rate or ongoing severe relapses compared with the previous year despite treatment with beta interferon (population 1b). The Committee noted that the manufacturer's reason for this was that this population represented the largest subgroup in the clinical trials of fingolimod. The Committee concluded that it could only make a recommendation on the use of fingolimod for the population presented in the manufacturer's submission (that is population 1b). It would be unable to make any recommendations for the use of fingolimod in any other populations covered by the marketing authorisation without evidence on the cost effectiveness of fingolimod for these populations from the manufacturer. The Committee also noted that the manufacturer compared fingolimod with only one formulation of beta interferon (Avonex) in its original base-case analysis, and that it included no other comparators from the decision problem. The Committee heard from the manufacturer that Avonex and Rebif are the most commonly prescribed forms of beta interferon in the NHS. The Committee acknowledged the concerns of the ERG that a beta interferon should not be the only comparator in an analysis for a patient group who have had a suboptimal response to prior beta interferon therapy, and that best supportive care should also be considered. The Committee concluded that limiting the analyses to comparisons with Avonex only was not appropriate, and instead comparisons with other beta interferons and best supportive care need to be included as a basis for any recommendations in this appraisal.

4.7 The Committee considered the evidence presented by the manufacturer on the clinical effectiveness of fingolimod. It noted that the manufacturer derived data from two clinical trials that assessed the efficacy and safety of fingolimod compared with placebo over 24 months (FREEDOMS trial), and with beta interferon-1a (Avonex) over 12 months (TRANSFORMS trial) in adults with relapsing–remitting multiple sclerosis. The Committee noted that the populations in the FREEDOMS and TRANSFORMS trials were broader than those in the marketing authorisation for fingolimod and therefore after the trials finished the manufacturer had to identify subgroups that approximated the populations in the marketing authorisation. The Committee heard from the ERG that the subgroups identified by the manufacturer were reasonable approximations to the populations in the marketing authorisation but that the subgroups were not mutually exclusive. The Committee noted that in response to clarification the manufacturer provided some revised analyses for people in population 1b who did not have rapidly evolving severe relapsing–remitting multiple sclerosis (that is population 1b, with people who also met the criteria for population 2 excluded). The Committee noted that these analyses generated lower ICERs than those for the whole of population 1b, but it was aware of reservations expressed by the manufacturer and the ERG about the small samples on which the subgroup analysis was based. Separate analyses were not provided for populations 1a or 2. The Committee also considered the manufacturer's mixed treatment comparison, which assessed treatment effects between comparators in the absence of direct evidence. It heard from the manufacturer and the ERG that there was considerable heterogeneity between the studies and that none of the studies in the analysis included populations that closely and consistently match those described in the marketing authorisation for fingolimod. It also heard that the TRANSFORMS and FREEDOMS trials were not powered to assess the efficacy of fingolimod in the subgroups defined by the marketing authorisation. The Committee concluded that the available evidence shows that people with relapsing–remitting multiple sclerosis who are treated with fingolimod have lower relapse rates than people treated with Avonex or placebo. The Committee also agreed that fingolimod was shown to reduce disability progression in people with relapsing–remitting multiple sclerosis compared with placebo in the whole population of the FREEDOMS trial; however, there was no significant impact on disability progression compared with Avonex in the TRANSFORMS trial. The Committee heard from the manufacturer that the European Medicines Agency had judged that the clinical trial data demonstrated consistent treatment effects with fingolimod in all of the subgroups with highly active disease covered by the marketing authorisation. The Committee concluded that the available evidence shows that fingolimod improves outcomes for the whole population in the clinical trials, and in those with highly active disease defined by the marketing authorisation.

4.8 The Committee discussed the health-related quality-of-life data from the FREEDOMS and TRANSFORMS trials. It noted that there were no statistically significant changes from baseline for EQ-5D measures observed for people with relapsing–remitting multiple sclerosis treated with fingolimod or placebo in the FREEDOMS trial, and that only a slight non-significant improvement in health-related quality of life according to the PRIMUS-QoL scale was observed for people treated with fingolimod or Avonex in the TRANSFORMS trial. The Committee heard from the clinical specialists that a patient's quality of life may not be affected by treatment because multiple sclerosis is a chronic disease with fluctuating symptoms, particularly in the relatively early stages of the condition. The Committee noted that the populations in the FREEDOMS and TRANSFORMS trials did not experience significant progression of disease and this may also have reduced the apparent impact of treatment on quality of life measures. The Committee concluded that it was therefore clinically plausible to see non-statistically significant changes in health-related quality-of-life measures in the FREEDOMS and TRANSFORMS trials, and that the impact of fingolimod on health-related quality of life remained uncertain.

Cost effectiveness

4.9 The Committee discussed the cost-effectiveness estimates from the manufacturer's original economic model, the assumptions on which these were based, the revised analyses from the manufacturer in response to the first and second appraisal consultation documents and the ERG's critiques and exploratory analyses. The Committee noted that the manufacturer provided a probabilistic Markov model that was structurally similar to other models used in previous NICE technology appraisal guidance on treatments for multiple sclerosis. The Committee noted that the manufacturer's original base-case analysis presented deterministic results for fingolimod compared with Avonex, which were substantially lower than the probabilistic results estimated by the ERG. The Committee acknowledged the concerns of the ERG that deterministic results should not be presented from a probabilistic model, and concluded that it would base its decision on the probabilistic results.

4.10 The Committee noted the concerns of the clinical specialists that the manufacturer's model may not reflect the natural history of multiple sclerosis because it does not allow for improvement in EDSS scores. The Committee heard from the manufacturer that the ability to include improvements in EDSS scores had been intentionally removed from the model to produce a conservative estimate of the cost effectiveness of fingolimod. The Committee heard from the clinical specialists that few people experience an improvement in EDSS score and therefore it concluded that the manufacturer's approach was reasonable.

4.11 The Committee heard from the manufacturer that disability progression rates in the model were derived from a longitudinal data set from a population with multiple sclerosis in Ontario, Canada. This data source was chosen because it has been considered previously in other NICE technology appraisal guidance on treatments for multiple sclerosis. The Committee heard from the clinical specialists and the ERG that this may have given more rapid disability progression rates than those seen in the clinical trials and in the current UK patient population. However, the Committee noted that no alternative data sources had been made available. The Committee was concerned that the manufacturer did not explain the divergence between model predictions and the trial observations for disease progression in its original submission, because the ERG's exploratory analyses showed that the model predictions were highly sensitive to the natural history progression data used. The Committee noted from subsequent sensitivity analyses carried out by the manufacturer that the ICERs increased only slightly with changes in the assumptions on natural history of disease progression (section 3.20). The Committee was persuaded that disease progression in people initially treated with disease-modifying treatments may be less rapid in current clinical practice than in the Ontario data set and concluded that data on the natural history of disability progression were a source of uncertainty in the model.

4.12 The Committee heard from the manufacturer that the utility data from the FREEDOMS or TRANSFORMS trials were not used in the original model because the study populations included only people with an EDSS score up to 6, and therefore utility data for higher EDSS scores were not available. The Committee heard from the ERG that the manufacturer used EDSS-based EQ-5D scores from a published source instead, despite this study being criticised in technology appraisal guidance 127 for having low response rates, not representing the appropriate population and being prone to selection bias. The Committee agreed that because the manufacturer's base case targeted a very specific group (population 1b), it would have been more appropriate to use utility data for that group from the trials where possible, and to use data from other sources only for EDSS scores above 6. The Committee considered additional analyses from the manufacturer in response to the appraisal consultation documents, which indicated that the base-case ICER for population 1b slightly decreased when utility data from the FREEDOMS and TRANSFORMS trials were used for EDSS states up to 6 and published data from Orme et al. (2007) were only used for EDSS scores above 6. The Committee concluded that the manufacturer's revised approach to incorporating utility estimates in the model was reasonable.

4.13 The Committee was concerned about the assumption in the manufacturer's original model that the treatment effect observed for the duration of the trials (1 or 2 years) was maintained at the same level during the on-treatment periods. It noted that sensitivity analyses carried out by the manufacturer and the ERG showed that a reduction in the assumed duration of treatment effect increased the ICERs substantively (see sections 3.15 and 3.32). The Committee heard from the manufacturer that there is currently no evidence to support the hypothesis that the efficacy of fingolimod will reduce over time and preliminary results from the FREEDOMS extension study show that there is no loss of efficacy over 4 years. In the absence of data beyond 4 years, the Committee decided to be cautious and considered it appropriate to include a 50% waning of treatment effect after 5 years in the base-case analysis. However, it acknowledged that if the treatment effect did not wane over time then this would overestimate the base-case ICER.

4.14 The Committee noted potential inaccuracies in some of the administration costs included in the manufacturer's original model. In particular, it heard from the manufacturer that it was assumed that people treated with Avonex had more visits to a neurologist than people treated with fingolimod. The Committee heard from the clinical specialists that this assumption was probably not correct and that it is more plausible that people receiving fingolimod would have three visits during the first year of treatment, compared with two visits for people receiving Avonex. The Committee noted that the manufacturer had corrected this assumption in the revised analyses submitted in response to the appraisal consultation documents and was persuaded that revising the costs in the model had a minimal impact on the ICER.

4.15 The Committee acknowledged that based on current practice in the NHS it would be inappropriate to use Avonex alone as a comparator for fingolimod. The Committee considered exploratory analyses conducted by the ERG on the manufacturer's original model, which incrementally compared fingolimod with best supportive care and Rebif-44. The Committee noted that these analyses were based on indirect comparisons of limited data and that in population 1b Avonex was dominated by Rebif-44. The Committee noted comments from the manufacturer in response to the appraisal consultation documents, which suggested that the Rebif-44 data used in the comparison with fingolimod were from patients with relapsing–remitting multiple sclerosis regardless of previous treatment, rather than from those whose disease had a suboptimal response to disease-modifying therapy (that is, population 1b). The Committee was persuaded that this may have resulted in an overestimation of the ICER for fingolimod compared with Rebif-44.

4.16 The Committee acknowledged that although the manufacturer had tried to address some of the concerns raised during the first and second Committee meetings, the manufacturer and the ERG still had divided opinions on the most appropriate methodological approaches to evaluate the cost effectiveness of fingolimod in population 1b. The Committee considered the manufacturer's revised probabilistic base-case ICER of £17,300 per QALY gained for fingolimod compared with Avonex. The Committee noted that in the ERG's incremental analysis, Avonex was extendedly dominated and the probabilistic ICER for fingolimod compared with best supportive care was £58,000 per QALY gained. The Committee acknowledged that the ERG's analyses demonstrated that beta-interferon treatment may not be cost effective compared with what is considered an appropriate use of NHS resources as defined in the NICE Guide to the methods of technology appraisal (2008). However it was mindful of the need to take account of current NHS practice, including the risk-sharing scheme (which currently funds beta-interferon treatments for people with multiple sclerosis), when defining the appropriate comparator(s) for assessment.

4.17 The Committee acknowledged that the choice of comparator in the manufacturer's model was a key driver of cost effectiveness. It also acknowledged that there was variation in current practice and therefore concluded that fingolimod should be compared with a weighted average of the comparators currently used in UK clinical practice to manage relapsing–remitting multiple sclerosis. This includes best supportive care together with a mix of beta interferons (with the proportions for the beta interferons based on market share data from the Prescription Pricing Authority). The Committee noted that the manufacturer's probabilistic ICER for fingolimod compared with the weighted average of the comparators was £27,800 per QALY gained. The Committee acknowledged that the manufacturer had assumed that best supportive care contributes only 5% to the weighted average in the base case, and that sensitivity analyses showed that if a higher proportion was assumed, such as 10%, the ICER would increase to approximately £30,000 per QALY gained (see section 3.19). The Committee noted from the manufacturer's and the ERG's sensitivity analyses that the ICER for fingolimod compared with the weighted average of the comparators depends on the proportions assumed for the comparator treatments, and the assumptions about the natural history of disability progression and the waning of treatment effect after 5 years. The Committee concluded that the most plausible ICER for fingolimod compared with the weighted average of the comparators was likely to be in the range of £25,000 to £35,000 per QALY gained.

4.18 The Committee noted that the most plausible ICER for fingolimod could be higher than what is normally considered an effective use of NHS resources. It was mindful that the NICE Guide to the methods of technology appraisal (2008) states that a strong case should be identified for an ICER that is higher than £30,000 per QALY gained. The Committee noted that in these circumstances the NICE Guide to methods of technology appraisal states that judgements about the acceptability of the technology as an effective use of NHS resources will specifically take account of:

4.19 The Committee discussed whether the assessment of the change in health-related quality of life had been inadequately captured in the economic analysis. It heard from the patient experts that people who receive fingolimod have fewer adverse reactions than those who receive beta-interferon therapy. In addition, treatment with fingolimod significantly reduces relapses and could allow people to lead an active and fulfilling life and contribute more fully to society. The Committee also heard from the manufacturer that any impact of treatment with fingolimod on the severity of relapses had not been captured in the model. In addition, the benefits from a decreased need for informal care provided by family and friends of people with multiple sclerosis had not been considered (because it is not in line with NICE's reference case). In the manufacturer's view, inclusion of these factors would decrease the ICER. The Committee accepted that fingolimod is a valuable new therapy and that its oral formulation represents innovation in the treatment of multiple sclerosis. The Committee recognised that including all of the benefits suggested by the manufacturer and patient experts in the manufacturer's model could decrease the ICER to a level that would be considered a cost-effective use of NHS resources.

4.20 The Committee noted that the current risk-sharing scheme allows beta interferons to be purchased at a price which the Department of Health considers to be a cost-effective use of NHS resources. However, it regretted that published outcome data from the scheme to justify the negotiated procurement price for these treatments are lacking. Taking these difficulties into consideration, the Committee made an exceptional case and recommended fingolimod as an option for the treatment of highly active relapsing–remitting multiple sclerosis in adults who have an unchanged or increased relapse rate, or ongoing severe relapses, compared with the previous year despite previous treatment with beta interferon (population 1b). The Committee also emphasised that it is important that a new model for multiple sclerosis is developed for any future appraisals of treatments for multiple sclerosis. The new model should ideally be based on UK patient cohorts, should use the best available evidence (including experience to date from the risk-sharing scheme) and should include all currently available treatments, so that future appraisals of treatments for multiple sclerosis are directly relevant to UK clinical practice.

Summary of Appraisal Committee's key conclusions