1 Appraisal Committee's preliminary recommendations

1.1 Abatacept in combination with methotrexate is not recommended for the treatment of moderate to severe active rheumatoid arthritis in adults whose disease has responded inadequately to one or more conventional non-biological disease-modifying anti-rheumatic drugs (DMARDs) including methotrexate.

2 The technology

2.1 Abatacept (Orencia, Bristol-Myers Squibb) is a selective T-cell co-stimulation modulator that blocks a co-stimulatory signal required to activate T-cells. Abatacept has a marketing authorisation for use in combination with methotrexate for the treatment of moderate to severe active rheumatoid arthritis in adults whose disease has responded inadequately to previous therapy with one or more DMARDs including methotrexate or a tumour necrosis factor (TNF) inhibitor.

2.2 Common adverse effects of abatacept therapy include infections, including sepsis and pneumonia. Abatacept is contraindicated in people with severe, uncontrolled or opportunistic infections. Before initiating therapy, clinicians should evaluate people for both active and inactive (latent) tuberculosis infection. For full details of adverse effects and contraindications, see the summary of product characteristics.

2.3 Abatacept is administered as a 30-minute intravenous infusion. After an initial infusion (week 0), a person receives an infusion at week 2, week 4 and every 4 weeks thereafter. Abatacept is available in 250-mg vials at a cost of £242.17 per vial (excluding VAT; ‘British national formulary’ [BNF] edition 61). Fourteen infusions are required in the first year, and 13 infusions in subsequent years. The dose of abatacept depends on body weight: people weighing less than 60 kg, 60–100 kg and over 100 kg require 500 mg, 750 mg and 1000 mg respectively. The annual drug costs associated with abatacept vary according to body weight and the number of infusions required. For a person weighing between 60 and 100 kg, the annual drug cost is £10,171.14 in the first year and £9444.63 in subsequent years. Costs may vary in different settings because of negotiated procurement discounts.

3 The manufacturer's submission

The Appraisal Committee (appendix A) considered evidence submitted by the manufacturer of abatacept and a review of this submission by the Evidence Review Group (ERG; appendix B).

 Clinical effectiveness

3.1 The decision problem defined in the scope asked whether, compared with conventional non-biological DMARDs (from now on referred to as conventional DMARDs) or with biological DMARDs (as recommended in NICE technology appraisal guidance 130 and 186; TA130, TA186), abatacept plus methotrexate is clinically and cost effective in adults whose disease has not responded adequately to one or more conventional DMARDs, including methotrexate. The manufacturer addressed this decision problem, but in addition, focussed its submission specifically on a subpopulation of people for whom self-administration of subcutaneously-injected biological DMARDs was not appropriate. The manufacturer stated that these people would include people with problems handling the injection devices, with mental health problems, or with an aversion to, or phobia of, needles. The manufacturer estimated that approximately 10% of people eligible for a biological DMARD would not be able to self-inject subcutaneously. Of the biological DMARDs recommended by NICE as treatment options after an inadequate response to conventional DMARDs, only infliximab is administered intravenously. Therefore, the manufacturer focussed on two comparators for abatacept plus methotrexate: infliximab plus methotrexate, and conventional DMARDs, which have not yet been tried in the treatment pathway.

3.2 To establish the efficacy of abatacept plus methotrexate compared with infliximab plus methotrexate, the manufacturer performed a systematic review to identify randomised controlled trials (RCTs) comparing abatacept plus methotrexate with infliximab plus methotrexate, or placebo plus methotrexate. The manufacturer sought head-to-head trials to provide direct evidence, and placebo-controlled trials to provide indirect evidence. The manufacturer identified four RCTs; of these one compared 2 mg/kg or 10 mg/kg abatacept plus methotrexate with placebo plus methotrexate (Kremer phase 2b [n = 339]), two compared 10 mg/kg abatacept plus methotrexate with placebo plus methotrexate (AIM [n = 652] and IM101-119 [n = 50]), and one compared abatacept 10 mg/kg plus methotrexate with infliximab 3 mg/kg plus methotrexate or placebo plus methotrexate (ATTEST [n = 431]). Three trials lasted 1 year (the Kremer phase 2b, AIM and ATTEST trials). The AIM and ATTEST trials enrolled people who had active rheumatoid arthritis for at least 1 year, and the IM101-119 trial lasted 4 months and enrolled people who hadactive rheumatoid arthritis according to disease activity score 28 (see section 3.3). The Kremer phase 2b trial enrolled people who had active rheumatoid arthritis, but the duration of their disease was not stated. The average number of DMARDs previously received by the trial participants ranged from 1.2 to 1.8 (reported in AIM and ATTEST trials). The mean time since first diagnosis was approximately 2.3 years in the IM101-119 trial and between 7.3 and 9.7 years in the three other trials.

3.3 Several tools were used to assess the response to treatment in rheumatoid arthritis. The Stanford Health Assessment Questionnaire (HAQ) and the shorter HAQ-Disability Index (HAQ-DI) or Modified Health Assessment Questionnaire (M-HAQ) score the ability to perform daily activities; ranging from 0 (least disability) to 3 (most severe disability). The American College of Rheumatology (ACR) response criteria (ACR20, ACR50 and ACR70) require improvement in a percentage (20, 50 or 70% respectively) of tender joints, swollen joints, global assessments, pain, disability and circulating inflammatory markers. The disease activity score (DAS) is calculated using a formula that includes counts for tender and swollen joints (53 and 44 joints respectively), an evaluation of general health (on a scale of 0 to 100) and a measure of circulating inflammatory markers. DAS28 is similar to DAS, but assesses 28 joints only. A DAS28 score greater than 5.1 indicates high disease activity, between 3.2 and 5.1 moderate disease activity, less than 3.2 low disease activity, and a score of less than 2.6 indicates remission. An improvement in DAS28 score of 0.6 or less indicates a poor response, and an improvement greater than 1.2 indicates a good response.

3.4 Primary outcomes differed by study. The AIM and Kremer phase 2b trials measured ACR20 at 6 months. The AIM trial also measured change from baseline in radiographic progression of joint erosions and of physical function of at least 0.3 in HAQ-DI at 1 year as primary outcomes. DAS28 at 6 months was the primary endpoint in the ATTEST study, while change in wrist synovitis score at 4 months was the primary outcome measure in study IM101-119. Secondary outcomes in the trials included: physical function measured using either the HAQ-DI or M-HAQ (referred to from now on as HAQ), ACR response, health-related quality of life as measured by the short-form 36 (SF-36) questionnaire, global assessment scales and adverse events.

3.5 At 6 months, three of four trials reported statistically significant differences in mean improvements in HAQ score, ranging from −0.19 to −0.38 with abatacept plus methotrexate compared with placebo plus methotrexate. At 1 year, two trials reported statistically significant differences in mean improvements in HAQ score of −0.29 (p < 0.001) and −0.36 (p < 0.001) compared with placebo plus methotrexate.

3.6 At 6 months and 1 year, the three trials reported a higher likelihood of achieving a clinically meaningful improvement in physical function (defined as a reduction in HAQ score of 0.22 in the Kremer phase 2b study, and of 0.3 in the AIM and ATTEST trials) with abatacept plus methotrexate compared with placebo plus methotrexate. The relative risks for a clinically meaningful improvement in physical function at 6 months with abatacept plus methotrexate compared with placebo plus methotrexate in the individual trials were 1.34 (95% confidence interval [CI] 1.14 to 1.58), 1.73 (95% CI 1.29 to 2.33) and 1.50 (95% CI 1.16 to 1.94) and at 1 year were 1.61 (95% CI 1.35 to 1.94) and 1.79 (95% CI 1.27 to 2.52).

3.7 At 6 months and 1 year, abatacept plus methotrexate led to significantly higher probability of having low disease activity and of achieving remission than placebo plus methotrexate as measured by DAS28 relative to baseline. The ATTEST study, which was not adequately powered to detect differences between abatacept and infliximab, reported non-significant mean improvements in DAS28 score at 6 months (−0.28) and at 1 year (−0.62) with abatacept plus methotrexate compared with infliximab plus methotrexate. The ATTEST study also reported that abatacept plus methotrexate was associated with a significantly higher probability of having low disease activity at 6 months and 1 year, and a higher likelihood of remission at 6 months and 1 year than infliximab plus methotrexate.

3.8 The outcome of ACR responses at 6 months and/or 1 year were reported in all three 1-year trials as a primary or secondary outcome. At both 6 months and 1 year, abatacept 10 mg/kg and infliximab were associated with a significantly higher probability of achieving an ACR20, ACR50 or ACR70 response than placebo. There were no reported statistically significant differences between the groups randomised to abatacept or infliximab in ACR20, ACR50 or ACR70 at 6 months or 1 year.

3.9 The AIM, ATTEST and Kremer phase 2b trials measured health-related quality of life using the SF-36 questionnaire at baseline, 6 months and 1 year. Quality of life data were not collected in the IM101-119 trial. None of the trials collected health related quality of life data using the EuroQol 5-D (EQ-5D) questionnaire. In the Kremer phase 2b and AIM trials, abatacept plus methotrexate led to statistically significant improvements from baseline in the physical and mental components of the SF-36 at 6 months compared with placebo plus methotrexate. The ATTEST study reported significant improvements in the physical and mental components of the SF-36 in both the groups randomised to abatacept and to infliximab compared with placebo.

3.10 The manufacturer presented data on adverse events from the ATTEST, Kremer phase 2b, AIM and IM101-119 trials. In the three 1-year trials, abatacept was not associated with a significantly higher rate of serious adverse events compared with placebo at 6 months or 1 year. The ATTEST study reported that abatacept plus methotrexate compared with infliximab plus methotrexate was associated with lower rates of serious adverse events at 1 year (9.6% versus 18.2%), lower rates of discontinuing the drug because of adverse events (3.2% versus 7.3%), lower rates of discontinuing the drug because of serious adverse events (2.6% versus 3.6%), lower rates of serious infections (1.9% versus 8.5%) and acute infusion events (7.1% versus 24.8%). Longer term data incorporated into the safety analyses of abatacept indicated that the incidence of serious adverse events did not increase over time.

3.11 To compare the efficacy of abatacept plus methotrexate with placebo plus methotrexate, the manufacturer carried out a series of pairwise meta-analyses using data from the Kremer phase 2b, AIM and ATTEST trials. The manufacturer’s fixed effects meta-analyses reported a mean reduction (improvement) from baseline in HAQ score for abatacept plus methotrexate compared with placebo plus methotrexate at 24 or 26 weeks (−0.2524, 95% CI −0.3253 to −0.1794) and 52 weeks (−0.3105, 95% CI −0.3934 to −0.2275). A mean reduction (improvement) from baseline in DAS28 score was reported for abatacept plus methotrexate compared with placebo plus methotrexate at 24 or 28 weeks (−1.123, 95% CI −1.3275 to −0.9186).

3.12 The manufacturer’s systematic review also identified trials for a mixed treatment comparison intended to compare abatacept plus methotrexate with five biological DMARDs, all plus methotrexate (adalimumab, certolizumab pegol, etanercept, golimumab, infliximab), and with placebo plus methotrexate. The mixed treatment comparison included 11 trials, and focussed on the change in the HAQ score from baseline to 24 or 26 weeks. The absolute change from baseline for biological agents plus methotrexate was −0.46 for infliximab, −0.55 for etanercept, −0.57 for abatacept, −0.60 for adalimumab, −0.61 for golimumab and −0.65 for certolizumab pegol. The absolute change from baseline for placebo plus methotrexate was −0.27. The mixed treatment comparison showed that abatacept plus methotrexate displayed efficacy comparable with that of most other biological DMARDs. The mixed treatment comparison also reported the mean rate of serious adverse events for the interventions from the included trials. The average percentage of people discontinuing treatment because of serious adverse events ranged from 0.0% for palliative care to 12.8% for certolizumab pegol.

 Cost effectiveness

3.13 The manufacturer systematically reviewed economic evaluations of abatacept plus methotrexate for people with moderate to severe active rheumatoid arthritis whose disease had not responded adequately to methotrexate monotherapy, or who were intolerant to methotrexate monotherapy. Fourteen economic evaluations met the inclusion criteria for the systematic review; however, the manufacturer considered that none addressed the decision problem of this appraisal.

3.14 The manufacturer developed a new economic model that was based on cost−utility analyses run over a lifetime horizon and from the perspective of the healthcare provider. The model addressed the cost effectiveness of abatacept compared with three comparators: all other biological DMARDs, conventional DMARDs and infliximab plus methotrexate. Infliximab, like abatacept, is administered intravenously, and is the only intravenously administered biological DMARD recommended by NICE at this point in the clinical pathway for the treatment of rheumatoid arthritis. All people entering the model are assumed to either have severe active rheumatoid arthritis that has responded inadequately, or to have intolerance to methotrexate and also possibly to other conventional DMARDs, but not to a biological DMARD. 

3.15 The model begins with the person entering an initial treatment phase and receiving treatment with abatacept plus methotrexate, or a different biological DMARD plus methotrexate (that is, adalimumab, certolizumab pegol, etanercept, golimumab or infliximab), or another conventional DMARD. The model assumes that people receive this initial treatment for 6 months unless their rheumatoid arthritis did not respond (defined as a HAQ change from baseline of less than 0.3), they experienced a serious adverse event, or they died. The model assumes that people whose disease responded did not have an adverse event, or people who did not die remained on their allocated treatment and entered a long-term maintenance phase. As modelled, people leave this phase if the treatment becomes ineffective (if there is no longer a HAQ reduction of 0.3) or if they die. People who discontinue their allocated treatment either in the initial phase or the long-term phase, regardless of their initial treatment, enter the next phase of treatment with a sequence of conventional DMARDs (leflunomide, gold, azathioprine, ciclosporin, penicillamine), and then palliative care. The sequence continues until the person dies.

3.16 The simulated cohort of people in the model had a mean age of 51.5 years with a standard deviation of 12.90; 77.8% were women. The simulated cohort had a mean baseline HAQ of 1.71 with a standard deviation of 0.70.

3.17 The model assumed that a person receiving a particular drug could experience serious adverse events and that these occur only within the first 6 months of treatment (during the initial treatment phases). The manufacturer obtained the rates for serious adverse events for each treatment from its mixed treatment comparison. The model assumed that if a serious adverse event did occur, the person discontinued treatment and their HAQ score returned to the value at which they began treatment. The manufacturer did not model any costs or decreases in utility associated with experiencing an adverse event.

3.18 The manufacturer assumed that during the initial 6-month treatment phase, a person continued treatment if their HAQ score dropped (improved) by at least 0.3 from baseline. The manufacturer derived the figure of 0.3 from the AIM and ATTEST trials. The manufacturer estimated how much a person’s disease would improve with a given drug, using the HAQ changes from the mixed treatment comparison. The manufacturer assumed that all changes in HAQ would occur gradually and increase over 3 months, at which point the HAQ score would reflect all the possible improvement. If the person’s HAQ scores did not improve by 0.3, the person discontinued treatment and the manufacturer assumed that their HAQ score returned (rebounded) immediately to the value at which their treatment began. For people who continued into the maintenance phase, the manufacturer assumed that people on a given biological DMARD were no more or less likely to discontinue the drug. The manufacturer estimated the time to discontinuing a biological DMARD from a Weibull distribution, with a mean value of 8.82 years and median value of 4.21 years.

3.19 The economic model assumed that while people receive biological DMARDs, after initially improving over 3 months, their HAQ score remains constant. The model also assumes that while people receive conventional DMARDs, their HAQ score increases (worsens) by 0.045 each year, or while people receive palliative care, their HAQ score increases by 0.06 each year. Therefore, the HAQ scores of people who receive biological DMARDs would be the same at the end of treatment as at the start of treatment; while the HAQ scores of people who received conventional DMARDs and palliative care would be higher. The manufacturer assumed that the worse the HAQ score, the greater the risk of death, and that a person could die at any phase of the model. The manufacturer assumed that the magnitude of this increase (relative risk) was 1.33 (95% CI 1.10 to1.61) for each unit increase in HAQ score. The manufacturer derived this estimate from a North American paper published in 1994 by Wolfe et al.

3.20 The manufacturer obtained the costs in the model from UK sources and publications identified in the systematic review and used values from 2009. The manufacturer took the costs associated with biological DMARDs from the BNF 60. The manufacturer did not give separate costs to either conventional DMARDs or palliative care, but rather incorporated these into the costs associated with the disease (see below). The manufacturer defined the dosage of the biological DMARDs in accordance with each drug’s SPC. The manufacturer assumed that a person received abatacept infusions at the start of treatment, 15 days later, 29 days later, and thereafter every 4 weeks. The dose of abatacept depended on a person’s weight: people who weighed less than 60 kg received two vials (£484.34), people who weighed between 60 kg and 100 kg received three vials (£726.51) and people who weighed over 100 kg received four vials (£968.68). The manufacturer did not assume in the base-case scenario that people could share a vial of abatacept. The manufacturer assumed that the doses of either infliximab or etanercept could increase over time: 29% of people receiving infliximab had their dose increased to 5 mg/kg at 1 year and 1% of people receiving etanercept had their dose increased to 37.5 mg at 1 year. In contrast to this, the manufacturer assumed that people receiving abatacept would not require higher doses over time. The manufacturer incorporated costs associated with rheumatoid arthritis into the model by relating a cost to an interval in the HAQ score. For example, a HAQ score less than 0.6 was associated with a cost of £2733 and a HAQ score between 0.6 and less than 1.1 was associated with a cost of £3668. These costs included hospitalisation, surgical interventions, ambulatory and community care, monitoring, conventional DMARDs and palliative care.

3.21 The manufacturer assumed that utility values applied in the model were dependent on the HAQ score. The higher (worse) the HAQ score, the lower the utility. In the base case, the manufacturer used a quadratic (a non-linear) equation to map HAQ score to utility. The manufacturer stated that in a previous NICE technology appraisal (‘Adalimumab, etanercept, infliximab, rituximab and abatacept for the treatment of rheumatoid arthritis after the failure of a TNF inhibitor’ [NICE technology appraisal guidance 195]) the manufacturers and the assessment group mapped utility values from the HAQ score.

3.22 The manufacturer presented pairwise and fully incremental results. The manufacturer presented deterministic base-case results in two ways: firstly, comparing abatacept plus methotrexate with all treatments (including other biological DMARDs and conventional DMARDs) and secondly, by comparing abatacept plus methotrexate with infliximab and conventional DMARDs. When the manufacturer presented results for abatacept compared with all other treatments, abatacept was dominated by adalimumab and certolizumab pegol (that is, abatacept was less effective but cost more than adalimumab and certolizumab pegol). When abatacept, infliximab and conventional DMARDs were compared (assuming that some people could not receive subcutaneous interventions), then infliximab was extendedly dominated (that is, a combination of abatacept and conventional DMARDs would provide the same health gain as for infliximab, but at a reduced cost). There were 6.16 quality-adjusted life years (QALYs) gained with abatacept plus methotrexate compared with 4.88 QALYs gained with conventional DMARDs. Total costs were £114,548 with abatacept plus methotrexate and £76,276 with conventional DMARDs. The manufacturer estimated the incremental cost-effectiveness ratio (ICER) to be £29,916 per QALY gained for abatacept plus methotrexate compared with conventional DMARDs.

3.23 The manufacturer conducted a range of one-way sensitivity analyses to test whether the changes in values within the model materially changed the ICER. Reducing the time horizon to 5 years had the greatest effect on the ICER, which increased to £84,390 per QALY gained for abatacept plus methotrexate compared with conventional DMARDs. Changes that had a small effect on the ICER included using a linear rather than a non-linear function to map HAQ to utility, and changing the decrease in HAQ required to define whether a person had responded to a drug.

 Comments from the Evidence Review Group (ERG)

3.24 The ERG noted that the manufacturer’s evidence base for assessing clinical effectiveness may not be complete because the manufacturers Medline search strategy failed to identify at least one relevant publication, excluded relevant databases and was restricted to trials published in English.

3.25 The ERG stated that the manufacturer presented the results of the trials inconsistently and with omissions, and did not present all the relevant data available in the public domain. The ERG commented that the manufacturer did not always provide an explanation when it presented data in its submission that differed from published data.

3.26 The ERG noted that people in the included trials had not had rheumatoid arthritis for as many years, or had taken as many conventional DMARDs as people in UK clinical practice starting a biological DMARD. Therefore, although the evidence submitted largely reflected the decision problem defined in the scope, the ERG considered that the difference between the populations may translate to a smaller actual benefit from abatacept in UK clinical practice than was observed in the trial populations. This was because people with disease of longer duration or who have received a larger number of treatments may respond less well than people with disease of shorter duration or who have received fewer treatments.

3.27 The ERG considered that the included trials were of reasonable methodological quality. However, the ERG noted that the Kremer phase 2b may have been biased, because the discontinuation rates differed in people randomised to placebo and active treatment. The AIM trial may also have been biased because people who did not adhere to treatment were excluded. In addition, the ERG noted that no efficacy data were available relating to the extra-articular manifestations of disease, and data relating to some outcomes that are important to people with rheumatoid arthritis (pain, fatigue, sleep quality, and health-related quality of life) were poorly presented.

3.28 The ERG undertook its own meta-analyses to compare the efficacy of abatacept plus methotrexate with placebo plus methotrexate based on relative risk. The analyses confirmed the overall results from the trials identified by the manufacturer.

3.29 The ERG considered that the manufacturer’s economic model was relatively complex in its programming. The ERG noted a number of concerns including: the model structure; the population of the model; the internal validity of the model; and the probabilistic sensitivity analyses.

3.30 Regarding the model structure, the ERG considered that the design did not reflect current standard practice in the UK. For example, the model does not allow use of multiple biological DMARDs (that is a person receives only one biological DMARD and, if discontinued, conventional DMARDs are restarted), which is not in line with NICE guidance. The ERG stated that a comparison of abatacept compared with sequences of drugs including multiple biological DMARDs would not favour abatacept.

3.31 The ERG noted that the base case in the model included escalating the dose of infliximab and etanercept if required, but not of abatacept. The ERG noted that in the manufacturer’s sensitivity analyses, which assumed that clinicians do not increase the dose of infliximab, the ICER for abatacept increased to £37,025 per QALY gained compared with infliximab.

3.32 In the base-case analyses, the manufacturer assumed that people do not share vials and generally go to hospital to receive intravenous infusions. The ERG stated that it may be possible for larger hospital units to share vials. The ERG noted that in the manufacturer’s sensitivity analysis that incorporated sharing of infliximab vials, the ICERs for abatacept against infliximab increased substantially to £57,843 per QALY gained.

3.33 The HAQ score was used by the manufacturer to determine a number of factors throughout the model including utility, costs, whether a treatment continues, and the risk of dying. However, the ERG noted that DAS28, rather than the HAQ score, is routinely used in clinical practice and is likely to be more useful. The ERG noted that the model considered that a sufficient response for a person to continue treatment was a change in HAQ score of at least 0.3 from baseline. The ERG highlighted that although based on the endpoints of the key trials, an improvement of 0.3 in HAQ score may not reflect a clinically meaningful improvement. This means that the model does not account for the possibility that a person’s rheumatoid arthritis could worsen despite receiving biological DMARDs.

3.34 The ERG was concerned that the probabilistic sensitivity analysis did not include any variation around HAQ score. The ERG stated that this resulted in the confidence interval for the mean probabilistic ICERs being smaller than would be expected, and that the results conducted by the manufacturer were incorrect.

3.35 The ERG stated that a random effects meta-analysis of serious adverse events, rather than the fixed effects meta-analysis used by the manufacturer, would have been more appropriate, because it quantifies the degree of heterogeneity between trials.The ERG was aware that the manufacturer incorporated no costs or decrements in utility into the model when people experienced serious adverse events. The ERG considered that including these could reduce the ICERs for abatacept compared with infliximab, because people on abatacept experience fewer serious adverse events than people on infliximab. In addition, the ERG was unclear what impact the omission of non-serious adverse events from the model might have on the ICERs. The manufacturer’s model did not include any decrease in utility that a person receiving an intravenous infusion might experience; the ERG considered that because people receive abatacept more often than infliximab, then this could potentially increase the ICERs for abatacept relative to infliximab.

3.36 The ERG considered that in using the HAQ score, the manufacturer appropriately approximated the costs associated with rheumatoid arthritis. However, the ERG was concerned that the actual costs chosen by the manufacturer may have included costs associated with decreased productivity from a person’s decreased ability to work, and as such would fall outside the NICE reference case, which includes only costs to the NHS and carers. The ERG considered that more appropriate costs would be £1120 per HAQ unit, which was used in TA195.

3.37 The ERG undertook exploratory univariate sensitivity analyses to investigate the impact of some of the key assumptions on the cost-effectiveness estimates of abatacept. The exploratory analyses focussed on comparing abatacept with infliximab and conventional DMARDs in people for whom subcutaneous injections may be inappropriate. The exploratory sensitivity analysis showed that the ICER for abatacept plus methotrexate varied from £28,464 to £32,077 per QALY gained compared with conventional DMARDs, and £29,322 to £53,534 per QALY gained compared with infliximab plus methotrexate. The key parameter affecting the ICER for abatacept plus methotrexate compared with conventional DMARDs was reducing the time horizon to 5 years. The key parameters affecting the ICER for abatacept plus methotrexate when compared with infliximab were whether vial sharing or dose escalation were assumed.

3.38 The ERG undertook multivariate scenario analyses as follows:

• Objective analysis: The ERG corrected arithmetic errors in the manufacturer’s submission; applied costs of £1120 for each unit of HAQ that excluded losses in productivity associated with rheumatoid arthritis (instead of different costs for every half interval of HAQ change as was assumed by the manufacturer); sampled the time to discontinuing infliximab and abatacept independently (instead of dependently); set the standard deviation of response to treatment to 0.3; set the rates of serious adverse events as equal for abatacept and infliximab; and excluded the possibility of escalating the dose of infliximab.
• Favourable analysis: The ERG used its objective analysis but also took the rate of serious adverse events from the manufacturer’s submission; and set the HAQ increase defining a response at 0.5 rather than 0.3.
• Optimistic analysis: The ERG used its favourable analysis but also assumed that clinicians escalate the dosages of infliximab, but not abatacept.
• Pessimistic analysis: The ERG used its objective analysis but did not assume that vial sharing for infliximab occurred, and used a linear approach to mapping HAQ to utility as used by Bansback et al (2005).
• Hybrid analysis: The ERG weighted the results of its optimistic and pessimistic scenarios in the ratio of 37:63, with vial sharing in 63% of cases (taken from NICE technology appraisal guidance 195).

3.39 The ERG’s analyses showed that the ICER for abatacept plus methotrexate compared with conventional DMARDs varied from £29,552 to £36,045 per QALY gained, and compared with infliximab varied from £29,661 to £63,208 per QALY gained for the ‘favourable’ and ‘pessimistic’ scenarios respectively. The ERG noted that the ‘objective’ analysis showed that the ICER for abatacept plus methotrexate compared with conventional DMARDs was £32,255 per QALY gained, and compared with infliximab was £39,748 per QALY gained.

3.40 Full details of all the evidence are in the manufacturer’s submission and the ERG report, which are available from www.nice.org.uk/guidance/TAXXX

4  Consideration of the evidence

4.1 The Appraisal Committee reviewed the data available on the clinical and cost effectiveness of abatacept plus methotrexate, having considered evidence on the nature of rheumatoid arthritis and the value placed on the benefits of abatacept plus methotrexate by people with the condition, those who represent them, and clinical specialists. It also took into account the effective use of NHS resources.

4.2 The Committee considered the current clinical practice for the treatment of rheumatoid arthritis in the UK. The Committee heard from patient experts that rheumatoid arthritis can have a severe impact on quality of life, and that pain and depression are common among people with the disease. The Committee heard that rheumatoid arthritis often affects people’s ability to work, and places a considerable burden on the carers of people with the disease. Clinical specialists and patient experts emphasised the importance of having a choice of treatment for people whose disease has not responded adequately to initial treatment with conventional DMARDs. The clinical specialists expressed that the choice of a biological agent with a mechanism other than inhibiting TNF was especially important for people who cannot be treated with a TNF inhibitor. The Committee heard from clinical specialists that it is difficult to predict whether a person’s disease will respond to a given class of drugs, or to a drug within a given class.

4.3 The Committee heard that the management of rheumatoid arthritis has been changing in line with NICE guidance, and that clinicians start treatment with conventional DMARDs or TNF inhibitors sooner after a person’s diagnosis of rheumatoid arthritis than in the past. Therefore, the characteristics of the people being treated with biological DMARDs in the UK have changed over time. The Committee heard from the clinical specialists that starting treatment with a biological DMARD sooner may increase a person’s potential to benefit from treatment because he or she will have accumulated less irreversible joint damage. The Committee noted that the ERG’s view that the population in the trials had disease of shorter duration than the average UK patient was inconsistent with the view of the clinical specialists that the population in the trials had disease of longer duration than UK patients. The Committee concluded that the duration and severity of rheumatoid arthritis could limit the generalisability of estimates of effectiveness from the included trials to the UK population.

4.4 The Committee discussed the most appropriate method of assessing response to treatment with biological DMARDs. The Committee was made aware that DAS is more often used clinically to assess response to treatment, and HAQ more often used in the research setting. The Committee was aware that the DAS28 measure incorporated assessment of tender and swollen joints as well as biochemical measures of disease activity. The Committee heard from the clinical specialists that HAQ score is affected by both reversible and irreversible components of the disease process, and that longstanding disease lessens the potential for the HAQ score to improve because of irreversible damage. The Committee also heard from clinical specialists that they do not routinely use the HAQ score in clinical practice. The Committee heard from the manufacturer that it used HAQ for consistency because previous submissions for other NICE technology appraisals related to rheumatoid arthritis also used HAQ. The Committee considered that consistency had merits, but making a decision based on clinically meaningful outcomes was more important. The Committee expressed a preference for DAS28 as an outcome measure in economic models of rheumatoid arthritis, noting also that clinicians decide to stop or change treatment based on DAS. In addition, the Committee was concerned about the methodological quality and presentation of the manufacturer’s economic evaluation.

4.5 The Committee discussed the clinical effectiveness of abatacept with regard to the appropriate comparators and subpopulations. The Committee was aware that the scope specified that in a population who had taken at least one conventional DMARD, abatacept should be compared with other biological DMARDs or conventional DMARDs. The Committee was also aware the manufacturer specifically focussed on an analysis of abatacept compared with infliximab in a subpopulation of people who may not be able to use subcutaneous therapy. Lastly, the Committee was aware of a potential additional decision problem expressed by the clinical specialists, which compares abatacept with conventional DMARDs, but only in the subpopulation of people for whom clinicians consider TNF inhibitor treatment inappropriate because of a contraindication. The Committee noted that the mixed treatment comparison showed that abatacept plus methotrexate had similar efficacy to most of the other biological DMARDs. The Committee also noted that the results of the mixed treatment comparison showed slightly better results for abatacept than those observed in the trials included in the meta-analysis (see sections 3.11 and 3.12). In addition, the Committee noted that the manufacturer omitted key trials from the network and included trials that included people with different baseline characteristics. Therefore the Committee decided to view the results of the mixed treatment comparison with caution.

4.6 The Committee considered the cost effectiveness of abatacept plus methotrexate compared with other biological DMARDs. The Committee was aware that NICE recommends adalimumab, etanercept, infliximab (TA130) or certolizumab pegol (TA186) as treatment options at the same point in the clinical pathway at which abatacept is being considered in this appraisal. The Committee heard from the manufacturer that abatacept plus methotrexate was not a cost-effective use of NHS resources when compared with subcutaneously delivered biological DMARDs. The Committee noted that in all individual comparisons, abatacept plus methotrexate was dominated by subcutaneous therapies and agreed with the manufacturer that, compared with subcutaneously delivered biological DMARDs, abatacept plus methotrexate would not provide a cost-effective use of NHS resources.

4.7 The Committee then considered the decision problem added by the manufacturer, that is, the clinical and cost effectiveness of abatacept compared with infliximab in people for whom subcutaneous biological DMARDs were unsuitable. The Committee considered the data on clinical effectiveness presented by the manufacturer and noted that abatacept plus methotrexate and infliximab plus methotrexate showed statistically significant differences in HAQ score and DAS28 score at 6 months and 1 year compared with placebo plus methotrexate. The Committee noted that there was no significant difference between infliximab plus methotrexate compared with abatacept plus methotrexate, but also noted that although the ATTEST study included separate arms for abatacept, infliximab and placebo, this study was not powered to detect statistically significant differences between abatacept and infliximab.

4.8 The Committee discussed whether people for whom subcutaneous therapies are unsuitable as defined by the manufacturer represented a clearly defined and identifiable population relevant for clinical practice in the NHS. The Committee heard from the clinical specialists that they were disappointed that the manufacturer focused on this question, which they felt did not reflect clinical practice. The Committee heard from the clinical specialists that in current practice, route of administration rarely determines which drug to prescribe. The Committee heard that the devices used to self-administer subcutaneous injections had improved considerably and few people experience problems handling the injection devices. The Committee heard that subcutaneous interventions could be administered at home by a nurse or a family member, subject to local decision making, or in hospitals (as with intravenous infusions), where clinicians could monitor people more closely if required. However, the Committee heard from patient experts that people do care whether therapies are injected intravenously or subcutaneously. The Committee was aware that the manufacturer proposed that the population for whom subcutaneous therapy was not appropriate would include people with needle phobia. However, the Committee concluded that people with needle phobia would have the same problem with intravenous therapy. The Committee agreed that there was no clinically plausible reason related to route of administration that supports limiting the decision problem to this population. Therefore the Committee concluded that the question of the cost effectiveness of abatacept plus methotrexate compared with infliximab plus methotrexate for the population of people for whom self-administration of subcutaneously injected biological agents is inappropriate, is not relevant for the NHS.

4.9 The Committee then considered the population of people for whom clinicians felt that treatment with a TNF inhibitor was contraindicated, as proposed by the clinical specialists. The Committee heard from the clinical specialists that some people have one or more contraindications to treatment with a TNF inhibitor including heart failure, sepsis, or malignancy. The Committee was aware that abatacept has a different mechanism of action to TNF inhibitors, because it affects the co-stimulation of T-cells. The Committee identified that the appropriate comparator for this decision problem would be conventional DMARDs, because existing NICE guidance does not recommend any biological DMARDs with a different mechanism of action to TNF inhibition at the stage of rheumatoid arthritis for which abatacept is considered in this appraisal. The Committee was aware that in practice these people may receive rituximab and that the number of people for whom TNF inhibitors are contraindicated is likely to be small. However, the Committee acknowledged that in strict accordance with the marketing authorisation and the NICE recommendations for rituximab, people must have disease that has shown an inadequate response, or be intolerant to, TNF inhibitors to receive rituximab (that is, people must have received treatment with a TNF inhibitor before they can be considered for treatment with rituximab). 

4.10 The Committee considered the clinical effectiveness of abatacept plus methotrexate compared with placebo plus methotrexate (that is, conventional DMARDs alone). The Committee was aware that it had not been presented with any clinical evidence regarding the use of abatacept in people for whom a TNF inhibitor was contraindicated. However, taking into account its earlier conclusions to view with caution the results of the mixed treatment comparison, the Committee noted the results that suggested that abatacept plus methotrexate statistically significantly improved ACR, DAS and HAQ scores compared with placebo plus methotrexate. The Committee also noted that placebo plus methotrexate was also associated with potentially clinically meaningful improvements. However, on balance the Committee agreed that abatacept plus methotrexate is clinically effective compared with placebo plus methotrexate.

4.11 The Committee then considered the cost effectiveness of abatacept plus methotrexate for people for whom TNF inhibitors are contraindicated. The Committee noted that no ICERs for this defined population had been presented, but considered that the pairwise ICERs of abatacept plus methotrexate compared with conventional DMARDs could be used as a proxy to estimate cost-effectiveness in this group. The Committee noted that the manufacturer’s estimates contained arithmetical errors and that the ERG exploratory analyses corrected for these errors; the base-case ICER for this group was therefore £29,700 per QALY gained.

4.12 The Committee considered the use in the model of effectiveness estimates derived from the HAQ score from the mixed treatment comparison. The Committee noted that constructing a mixed treatment comparison and modelling data from the HAQ score instead of DAS28 score was likely to have disproportionately favoured abatacept compared with conventional DMARDs because the results obtained for abatacept plus methotrexate were better than for conventional DMARDs plus methotrexate when using the HAQ score compared with the DAS28 score.

4.13 The Committee noted that the economic model had not included health-related quality of life measured using a generic preference-based measure, but had instead mapped a disease-specific measure (HAQ) to a generic measure (EQ-5D). The Committee noted that the manufacturer had chosen to do this because mapping HAQ to utilities had been used in previous NICE technology appraisals of treatments for rheumatoid arthritis in the absence of directly elicited EQ-5D data. The Committee noted that the manufacturer’s mapping of HAQ scores to EQ-5D utility values resulted in the possibility of clinical scenarios where having rheumatoid arthritis would be worse than being dead. The Committee heard from the patient experts that it was possible that some people with rheumatoid arthritis may experience such severe disease. The Committee noted that estimates using a non-linear approach to mapping were more favourable to abatacept, and was aware of the manufacturer’s sensitivity analysis that showed that using a linear utility mapping increased the ICER for abatacept plus methotrexate compared with conventional DMARDs plus methotrexate from £29,700 per QALY gained in the base case to £32,100 per QALY gained.

4.14 The Committee noted that the economic model assumed a 33% increase in mortality rate for each unit increase in HAQ score. The Committee noted that the manufacturer took these estimates from a study published over 15 years ago. The Committee heard from the clinical specialists that the treatment of the disease has improved since then, in particular the ability to reduce inflammation. The Committee concluded that the value assumed by the manufacturer may overestimate mortality associated with the HAQ score and that if a lower rate of mortality were assumed, then this could increase the ICER of abatacept compared with conventional DMARDs.

4.15 The Committee discussed the manufacturer’s assumptions of how rheumatoid arthritis progresses (represented by HAQ score) on and off treatments. The model assumed that while on a biological DMARD, after an initial improvement, the disease does not get better or worse. The Committee heard from the clinical specialists that this assumption was clinically plausible. However, the Committee heard from the clinical specialists that some people’s disease continues to improve, and heard from the patient experts that some people’s disease worsens despite treatment. The model also assumed that if a person stopped taking a biological DMARD (either because the disease did not respond to the biological DMARD or because the person experienced a serious adverse event), then the HAQ score worsened (rebounded) to the pre-treatment level, after which the disease progressed at a rate equal to that of a person on a sequence of conventional DMARDs followed by palliative care. The Committee heard that the clinical specialists believed that the rebound assumption relating to biological DMARDs is likely to underestimate the underlying disease progression. The Committee concluded that the assumptions incorporated into the model by the manufacturer disproportionately favoured abatacept when compared with conventional DMARDs.

4.16 The Committee noted that the manufacturer excluded costs or disutilities associated with adverse events from the model. The Committee considered that if adverse events occur at different rates between interventions, then these should have been included in the model. The Committee was aware that trial data showed that the overall rates of adverse events were similar for abatacept plus methotrexate and placebo plus methotrexate, but that no long-term data were available. The Committee considered that adverse events would be expected with abatacept, as with any biological DMARD, and that these could occur more frequently than with placebo plus methotrexate. The Committee therefore agreed that were adverse events included, the ICER would be likely to increase for abatacept plus methotrexate compared with placebo plus methotrexate.

4.17 The Committee considered the costs included in the economic model. The Committee heard the manufacturer acknowledge that it had used costs that included loss of productivity, and that this was outside the reference case defined by NICE. The Committee agreed that the costs proposed by the ERG were more appropriate. The Committee noted that including these costs increased the ERG’s corrected base-case ICER from £29,700 to £29,900 per QALY gained. The Committee was also aware that costs of escalating the dose of abatacept were not included in the model. The Committee agreed that there was no evidence currently to suggest that people had a decreased response to abatacept over time; however, it considered that it was biologically plausible that this may occur in the future and in the long term, given the experience from other biological DMARDs. The Committee concluded that if people required increasing doses of abatacept over time, then this would increase the ICERs for abatacept plus methotrexate compared with conventional DMARDs.

4.18 The Committee discussed the time horizon of the model presented in analyses of cost effectiveness by the manufacturer, and that reducing the time horizon from lifetime to 5 years increased the ICER from £29,900 to £84,400 per QALY gained. Bearing in mind this sensitivity of the ICER to the time horizon, the Committee concluded that this indicated that the ICER was also sensitive to changes in mortality and HAQ change over time, because these were the key factors that would impact on an ICER over a longer term time horizon.

4.19 In summary, the Committee believed the manufacturer’s model did not use the most plausible assumptions and inputs for the following parameters:

• omitting trials from the mixed treatment comparison (see section 4.5)
• modelling data from the HAQ score instead of DAS28 score (see section 4.12)
• the approach to mapping HAQ score to utilities (see section 4.13)
• the increase in mortality rate for each unit increase in HAQ score (see section 4.14)
• the underlying disease progression rebound in HAQ following discontinuation of treatment (see section 4.15)
• the exclusion of costs or disutilities associated with adverse events from the model (see section 4.16)
• the inclusion of productivity costs (see section 4.17).

The Committee was aware that for each of these parameters, the manufacturer’s assumptions favoured abatacept. It noted that whenever the assumptions that the Committee considered more realistic (either in the manufacturer’s sensitivity analyses or in the ERG’s analyses) could be quantified, the ICERs for abatacept plus methotrexate always increased above the base case of £29,700 per QALY, and when the alternative assumptions could not be quantified, the ICERs would be likely to increase. Therefore, the Committee concluded that a model that combined plausible assumptions would produce ICERs that exceeded the range it could consider to represent an effective use of NHS resources. The Committee therefore concluded that abatacept plus methotrexate was not recommended for the treatment of moderate to severe rheumatoid arthritis in adults whose disease has responded inadequately to one or more conventional non-biological DMARDs including methotrexate.

4.20 The Committee considered whether NICE’s duties under the equalities legislation required it to alter or to add to its recommendations in any way. The Committee noted the manufacturer’s exploration of potential equality issues in its submission and that the manufacturer indicated that patients who require or reasonably request intravenous infusion may do so because of their age, a disability or their race and that denying them intravenous treatment would be unfair. The Committee had earlier concluded that devices used to self-administer subcutaneous injections had improved considerably, that few people experience problems handling the injection devices and that subcutaneous interventions could be administered at home by a nurse or a family member (see section 4.8). Therefore the Committee agreed that the manufacturer’s definition of this group was not relevant for clinical practice in the NHS and that this did not present an equality issue. The Committee therefore concluded that its preliminary recommendation would not have a particular impact on any of the groups whose interests are protected by the equalities legislation and that there was no need to alter or add to its preliminary recommendations.

Summary of Appraisal Committee’s key conclusions

TAXXX	Appraisal title: Abatacept for the treatment of rheumatoid arthritis after the failure of conventional disease-modifying anti-rheumatic drugs		Section
Key conclusion
Abatacept in combination with methotrexate is not recommended for the treatment of moderate to severe active rheumatoid arthritis in adults whose disease has responded inadequately to one or more conventional non-biological disease-modifying anti-rheumatic drugs (DMARDs) including methotrexate.			1.1
Current practice
Clinical need of patients, including the availability of alternative treatments		Clinical specialists and patient experts emphasised the importance of having a choice of treatment for people whose disease has not responded adequately to initial treatment with conventional DMARDs. The clinical specialists expressed that the choice of a biological agent with a mechanism other than inhibiting TNF was especially important for people who cannot be treated with a TNF inhibitor. The Committee heard from clinical specialists that it is difficult to identify people whose disease might respond better to a given class of drugs, or to a drug within a given class.	4.2 4.2
The technology
Proposed benefits of the technology How innovative is the technology in its potential to make a significant and substantial impact on health-related benefits?		The clinical specialists expressed that the choice of a biological agent with a mechanism other than inhibiting TNF was especially important for people who cannot be treated with a TNF inhibitor. No specific claim of innovation was made in this appraisal.	4.2
What is the position of the treatment in the pathway of care for the condition?		The Committee was aware that NICE recommends adalimumab, etanercept, infliximab (TA130) or certolizumab pegol (TA186) as treatment options at the same point in the clinical pathway at which abatacept is being considered in this appraisal. The Committee concluded that the question of the cost effectiveness of abatacept plus methotrexate compared with infliximab plus methotrexate for the population of people for whom self-administration of subcutaneously injected biological agents is inappropriate, is not relevant for the NHS. The Committee identified that the appropriate comparator for people for whom clinicians felt that treatment with a TNF inhibitor was contraindicated would be conventional DMARDs, because existing NICE guidance does not recommend any biological DMARDs with a different mechanism of action to TNF inhibition at the stage of rheumatoid arthritis for which abatacept is considered in this appraisal.	4.6 4.8 4.9
Adverse effects		Common adverse effects of abatacept therapy include infections, including sepsis and pneumonia. Abatacept is contraindicated in people with severe, uncontrolled or opportunistic infections. Before initiating therapy, clinicians should evaluate people for both active and inactive (latent) tuberculosis infection.	2.2
Evidence for clinical effectiveness
Availability, nature and quality of evidence		The clinical evidence presented in the manufacturer’s submission was derived mainly from four RCTs that compared treatment with abatacept plus methotrexate with placebo plus methotrexate or infliximab plus methotrexate. A mixed treatment comparison was also conducted to compare abatacept plus methotrexate with five biological DMARDs, all plus methotrexate (adalimumab, certolizumab pegol, etanercept, golimumab, infliximab), and with placebo plus methotrexate. The Committee noted that the mixed treatment comparison showed that abatacept plus methotrexate had similar efficacy to most of the other biological DMARDs. The Committee also noted that the results of the mixed treatment comparison showed better results for abatacept than those observed in the trials included in the meta-analysis. In addition, the Committee noted that the manufacturer omitted key trials from the network and included trials that included people with different baseline characteristics. Therefore the Committee decided to view the results of the mixed treatment comparison with caution.	3.2 3.12 4.5
Relevance to general clinical practice in the NHS		The Committee heard that the management of rheumatoid arthritis has been changing in line with NICE guidance, and that clinicians start treatment with conventional DMARDs or TNF inhibitors sooner after a person’s diagnosis of rheumatoid arthritis than in the past. The Committee concluded that the duration and severity of rheumatoid arthritis could limit the generalisability of estimates of effectiveness from the included trials to the UK population.	4.3
Uncertainties generated by the evidence		The Committee was made aware that DAS is more often used clinically to assess response to treatment, and HAQ more often used in the research setting. The Committee expressed a preference for DAS28 as an outcome measure in economic models of rheumatoid arthritis, noting also that clinicians decide to stop or change treatment based on DAS.	4.4
Are there any clinically relevant subgroups for which there is evidence of differential effectiveness?		Not applicable	-
Estimate of the size of the clinical effectiveness including strength of supporting evidence		The Committee agreed that abatacept plus methotrexate is clinically effective compared with placebo plus methotrexate The Committee noted that the mixed treatment comparison showed that abatacept plus methotrexate had similar efficacy to most of the other biological DMARDs. However, the Committee noted that the manufacturer omitted key trials from the network and included trials that included people with different baseline characteristics. Therefore the Committee decided to view the results of the mixed treatment comparison with caution.	4.10 4.5
Evidence for cost effectiveness
Availability and nature of evidence		The Committee was concerned about the methodological quality and presentation of the manufacturer’s economic evaluation.	4.4
Uncertainties around and plausibility of assumptions and inputs in the economic model		The Committee noted that the manufacturer’s mapping of HAQ scores to EQ-5D utility values resulted in the possibility of clinical scenarios where having rheumatoid arthritis would be worse than being dead. The Committee noted that the economic model assumed a 33% increase in mortality rate for each unit increase in HAQ score. The Committee concluded that the value assumed by the manufacturer overestimates mortality associated with HAQ score. The Committee discussed the manufacturer’s assumptions of how rheumatoid arthritis progresses (represented by HAQ score) on and off different treatments. The Committee concluded that the assumptions incorporated into the model by the manufacturer disproportionately favoured abatacept when compared with conventional DMARDs. The Committee agreed that were adverse events included, the ICER would be likely to increase for abatacept plus methotrexate compared with placebo plus methotrexate. The Committee concluded that if people required increasing doses of abatacept over time, then this would increase the ICERs for abatacept plus methotrexate compared with conventional DMARDs. The Committee, bearing in mind the sensitivity of the ICER to the time horizon, concluded that this demonstrated that the ICER was also sensitive to changes in mortality and HAQ change over time, because these were the key factors that would impact on an ICER over a longer term time horizon.	4.13 4.14 4.15 4.16 4.17 4.18
Incorporation of health-related quality-of-life benefits and utility values Have any potential significant and substantial health-related benefits been identified that were not included in the economic model, and how have they been considered?		The Committee noted that the economic model had not included health-related quality of life measured using a generic preference-based measure, but had instead mapped a disease-specific measure (HAQ) to a generic measure (EQ-5D). The Committee noted that the manufacturer had chosen to do this because mapping HAQ to utilities had been used in previous NICE technology appraisals of treatments for rheumatoid arthritis in the absence of directly elicited EQ-5D data. The Committee noted that the manufacturer’s mapping of HAQ scores to EQ-5D utility values resulted in the possibility of clinical scenarios where having rheumatoid arthritis would be worse than being dead. The Committee heard from the patient experts that it was possible some people with rheumatoid arthritis may experience such severe disease. The Committee noted that estimates using a non-linear approach to mapping were more favourable to abatacept, and was aware of the manufacturer’s sensitivity analysis that showed that using a linear utility mapping increased the ICER for abatacept plus methotrexate from £29,700 (base case) to £32,100 per QALY gained compared with conventional DMARDs plus methotrexate. No additional health related benefits were identified that were not in the model.	4.13
Are there specific groups of people for whom the technology is particularly cost effective?		Not applicable	-
What are the key drivers of cost effectiveness?		The ICER was most sensitive to assumptions relating to the time horizon of the model. Reducing the time horizon from lifetime to 5 years increased the ICER from £29,900 to £84,400 per QALY gained, indicating that the ICER was also sensitive to changes in mortality and HAQ change over time, because these were the key factors that would impact on an ICER over a longer term time horizon.	4.18
Most likely cost-effectiveness estimate (given as an ICER)		The Committee was aware that for many model inputs the manufacturer’s assumptions favoured abatacept. It noted that whenever the assumptions that the Committee considered more realistic (either in the manufacturer’s sensitivity analyses or in the ERG’s analyses) could be quantified, the ICERs for abatacept plus methotrexate always increased above £29,700 per QALY, and when the alternative assumptions could not be quantified, the ICERs would be likely to increase. Therefore, the Committee concluded that a model that combined plausible assumptions would produce ICERs that exceeded the range it could consider to represent an effective use of NHS resources.	4.19
Additional factors taken into account
Patient access schemes (PPRS)		Not applicable to this appraisal.	-
End-of-life considerations		Not applicable to this appraisal.	-
Equalities considerations and social value judgements		The Committee noted that the manufacturer indicated that patients who require or reasonably request intravenous infusion may do so because of their age, a disability or their race and that denying them intravenous treatment would be unfair. The Committee agreed that the manufacturer’s definition of this group was not relevant for clinical practice in the NHS and that this did not present an equality issue. The Committee concluded that its preliminary recommendation would not have a particular impact on any of the groups whose interests are protected by the equalities legislation and that there was no need to alter or add to its preliminary recommendations.	4.20

5 Implementation

5.1 The Secretary of State and the Welsh Assembly Minister for Health and Social Services have issued directions to the NHS in England and Wales on implementing NICE technology appraisal guidance. When a NICE technology appraisal recommends use of a drug or treatment, or other technology, the NHS must usually provide funding and resources for it within 3 months of the guidance being published. If the Department of Health issues a variation to the 3-month funding direction, details will be available on the NICE website. When there is no NICE technology appraisal guidance on a drug, treatment or other technology, decisions on funding should be made locally.

5.2 NICE has developed tools to help organisations put this guidance into practice (listed below). These are available on our website (www.nice.org.uk/guidance/TAXXX). [NICE to amend list as needed at time of publication]

• Slides highlighting key messages for local discussion.
• Costing template and report to estimate the national and local savings and costs associated with implementation.
• Implementation advice on how to put the guidance into practice and national initiatives that support this locally.
• A costing statement explaining the resource impact of this guidance.
• Audit support for monitoring local practice.

6 Related NICE guidance

Published

• Tocilizumab for the treatment of rheumatoid arthritis. NICE technology appraisal guidance 198 (2010). Available from www.nice.org.uk/guidance/TA198
• Adalimumab, etanercept, infliximab, rituximab and abatacept for the treatment of rheumatoid arthritis after the failure of a TNF inhibitor. NICE technology appraisal guidance 195 (2010). Available from www.nice.org.uk/guidance/TA195
• Certolizumab pegol for the treatment of rheumatoid arthritis in adults. NICE technology appraisal guidance 186 (2010). Available from www.nice.org.uk/guidance/TA186
• Rheumatoid arthritis: the management of rheumatoid arthritis in adults. NICE clinical guideline 79 (2009). Available from www.nice.org.uk/guidance/CG79
• Adalimumab, etanercept and infliximab for the treatment of rheumatoid arthritis. NICE technology appraisal guidance 130 (2007). Available from www.nice.org.uk/guidance/TA130

Under development

NICE is developing the following guidance (details available from www.nice.org.uk):

• Golimumab for the treatment of rheumatoid arthritis after the failure of previous disease-modifying anti-rheumatic drugs. NICE technology appraisal guidance (publication expected June 2011).

7 Proposed date for review of guidance

7.1 NICE proposes that the guidance on this technology is considered for review by the Guidance Executive in July 2014. NICE welcomes comment on this proposed date. The Guidance Executive will decide whether the technology should be reviewed based on information gathered by NICE, and in consultation with consultees and commentators.

Amanda Adler
Chair, Appraisal Committee
March 2011

Appendix A: Appraisal Committee members and NICE project team

A Appraisal Committee members

The Appraisal Committee is one of NICE’s standing advisory committees. Its members are appointed for a 3-year term. A list of the Committee members who took part in the discussions for this appraisal appears below. The Appraisal Committee meets three times a month except in December, when there are no meetings. The Committee membership is split into three branches, each with a chair and vice chair. Each branch considers its own list of technologies, and ongoing topics are not moved between the branches.

Committee members are asked to declare any interests in the technology to be appraised. If it is considered there is a conflict of interest, the member is excluded from participating further in that appraisal.

The minutes of each Appraisal Committee meeting, which include the names of the members who attended and their declarations of interests, are posted on the NICE website.

Dr Amanda Adler (Chair)
Consultant Physician, Addenbrooke's Hospital

Dr Ray Armstrong
Consultant Rheumatologist, Southampton General Hospital

Dr Jeff Aronson
Reader in Clinical Pharmacology, University Department of Primary Health Care, University of Oxford

Dr Peter Barry
Consultant in Paediatric Intensive Care, Leicester Royal Infirmary

Professor John Cairns
Professor of Health Economics Public Health and Policy, London School of Hygiene and Tropical Medicine

Dr Mark Chakravarty
External Relations Director - Pharmaceuticals & Personal Health, Oral Care Europe

Mrs Eleanor Grey
Lay member

Mr Sanjay Gupta
YPD Service Case Manager, Southwark Health and Social Care, Southwark PCT

Dr Neil Iosson
General Practitioner

Mr Terence Lewis
Lay Member

Dr Rubin Minhas
General Practitioner and Clinical Director, BMJ Evidence Centre

Dr Peter Norrie
Principal Lecturer in Nursing, DeMontfort University

Professor Stephen Palmer
Professor of Health Economics, Centre for Health Economics, University of York

Dr Sanjeev Patel
Consultant Physician & Senior Lecturer in Rheumatology, St Helier University Hospital

Dr John Pounsford
Consultant Physician, Frenchay Hospital, Bristol

Dr Casey Quinn
Lecturer in Health Economics, Division of Primary Care, University of Nottingham

Mr Alun Roebuck
Consultant Nurse in Critical and Acute Care, United Lincolnshire NHS Trust

Dr Florian Alexander Ruths
Consultant Psychiatrist & Cognitive Therapist at the Maudsley Hospital, London

Mr Navin Sewak
Primary Care Pharmacist, NHS Hammersmith and Fulham

Mr Roderick Smith
Finance Director, West Kent Primary Care Trust

Mr Cliff Snelling
Lay Member

Professor Ken Stein (Vice Chair)
Professor of Public Health, Peninsula Technology Assessment Group (PenTAG), University of Exeter

Professor Andrew Stevens
Professor of Public Health, Department of Public Health and Epidemiology, University of Birmingham

B NICE project team

Each technology appraisal is assigned to a team consisting of one or more health technology analysts (who act as technical leads for the appraisal), a technical adviser and a project manager.

Scott Goulden
Technical Lead

Rebecca Trowman
Technical Adviser

Jeremy Powell
Project Manager

 Appendix B: Sources of evidence considered by the Committee

A The Evidence Review Group (ERG) report for this appraisal was prepared by School of Health & Related Research, Sheffield:

• Lloyd Jones M, Stevenson M, Stevens J et al. Abatacept for the treatment of rheumatoid arthritis after the failure of conventional disease-modifying anti-rheumatic drugs, February 2011

B The following organisations accepted the invitation to participate in this appraisal as consultees and commentators. They were invited to comment on the draft scope, the ERG report and the appraisal consultation document (ACD). Organisations listed in I were also invited to make written submissions. Organisations listed in II and III had the opportunity to give their expert views. Organisations listed in I and II also have the opportunity to appeal against the final appraisal determination.

I Manufacturer/sponsor:

• Bristol-Myers Squibb

II Professional/specialist and patient/carer groups:

• British Health Professionals in Rheumatology
• British Society for Rheumatology
• National Rheumatoid Arthritis Society
• Primary Care Rheumatology Society
• Royal College of Nursing
• Royal College of Pathologists
• Royal College of Physicians

III Other consultees:

• Department of Health
• Heart of Birmingham Teaching Primary Care Trust
• Telford and Wrekin Primary Care Trust
• Welsh Assembly Government

IV Commentator organisations (did not provide written evidence and without the right of appeal):

• Abbott Laboratories
• Astrazeneca UK
• British National Formulary
• Commissioning Support Appraisals Service
• Department of Health, Social Services and Public Safety - Northern Ireland
• GlaxoSmithKline
• Medicines and Healthcare products Regulatory Agency
• NHS Quality Improvement Scotland
• Pfizer
• sanofi-aventis
• Schering-Plough

C The following individuals were selected from clinical specialist and patient expert nominations from the non-manufacturer/sponsor consultees and commentators. They gave their expert personal view on abatacept by attending the initial Committee discussion and providing written evidence to the Committee. They are invited to comment on the ACD.

• Professor Peter C. Taylor, Professor of Experimental Rheumatology and Head of Clinical Trials, nominated by Bristol Myers Squibb Pharmaceuticals – clinical specialist
• Dr Patrick Kiely, Consultant Physician and Rheumatologist, nominated by the British Society for Rheumatology – clinical specialist
• Mrs Ailsa Bosworth, Chief Executive, National Rheumatoid Arthritis Society, nominated by the National Rheumatoid Arthritis Society - patient expert
• Ms Jean Burke, nominated by the National Rheumatoid Arthritis Society - patient expert

D Representatives from the following manufacturer/sponsor attended Committee Meetings. They contributed only when asked by the Committee chair to clarify specific issues and comment on factual accuracy.

• Bristol-Myers Squibb