The cost effectiveness and budget impact of natalizumab for formulary inclusion

Abstract

Background: Crohn's disease (CD) and multiple sclerosis (MS) are debilitating autoimmune diseases, which represent a substantial cost burden in the context of managed care. As a corollary, there is an unmet pharmacotherapeutic need in patient populations with relapsing forms of MS, in addition to populations with moderately to severely active CD with evidence of inflammation who have experienced an inadequate response to other mainstream therapies. The purpose of this study was to analyze the clinical and economic data associated with natalizumab (Tysabri) and to determine the potential impact of its formulary inclusion in a hypothetical health plan.

Findings: Regarding MS, the implemented cost-effectiveness and budget-impact models demonstrated an anticipated reduction in relapse rate of 67% over 2 years, and a total therapy cost of $72,120 over 2 years, equating to a cost per relapse avoided of $56,594. With respect to the model assumptions, the market share of natalizumab would experience an increase to 8.5%, resulting in a total per-member, per-month healthcare cost increase of $0.003 ($0.002 for pharmacy costs and $0.001 for medical costs).

Regarding CD, over a 2-year period outlined by the model, natalizumab produced the highest average time in remission, steroid-free remission, and remission or response in comparison to the other agents. The mean total costs associated with the initiation of natalizumab, infliximab, and adalimumab were $68,372, $62,090, and $61,796, respectively. Although natalizumab's costs were higher, the mean time spent in remission while on this medication was 4.5 months, as opposed to 2.4 months for infliximab and 2.9 months with adalimumab. This shift in market share was used to estimate the change in total costs (medical + pharmacy), and the per-member per-month change for the model's base case was calculated to be $0.035.

Limitations: The aforementioned cost-effectiveness results for natalizumab in the treatment for CD and MS were limited by the model's predetermined assumptions. These assumptions include anticipated reduction in relapse rate after 2 years of therapy and acquisition costs in the MS model, as well as assuming a certain percentage of patients were primary and secondary failures of TNFα inhibitor therapy in the CD model.

Conclusion: The evidence presented here demonstrates that natalizumab provides clinical practitioners with another tool in their fight against both MS and CD, albeit by way of a different mechanism of action. After a thorough review of the evidence, the authors find that natalizumab has been shown to be relatively cost effective in the treatment of both conditions from a payer perspective; the therapy adds a new option for those patients for whom conventional treatment was unsuccessful.

Key words::

• cost effectiveness
• Crohn's disease
• multiple sclerosis
• natalizumab

Introduction

Multiple sclerosis (MS) is an autoimmune disease of the central nervous system that is characterized by inflammation, demyelination, and axonal degeneration. Even in its early stages, MS involves neurologic damage at the time of relapse, some of which may be irreversible¹. In 1996, the National Multiple Sclerosis Society established the following standardized subtypes: primary progressive, secondary progressive, progressive-relapsing, and relapsing-remitting, the latter comprising 85–90% of individuals with the disease². Etiologically speaking, various theoretical propositions have been made which include, but are not limited to, genetic predisposition, infection, and environmental characteristics (i.e. geographical segmentation, climate, sunlight); while the exact cause remains to be precisely identified, MS etiology is most likely a combination of these proposed mechanisms³. Disease-modifying therapies (DMTs) such as natalizumab (Tysabri) have demonstrated efficacy in treating MS by reducing clinical relapses, slowing disease disability and progression, minimizing MRI lesion activity, and improving quality of life measures¹.

Crohn's disease (CD) is an autoimmune disease which produces inflammation of the gastrointestinal tract. In North America, CD affects between 400,000 and 600,000 individuals, with an incidence rate of 26–198.5 cases per 100,000 persons⁴. Moreover, CD exhibits a bimodal distribution in incidence as a function of age, afflicting individuals in their teens and 20s, and those in their 50s through 70s⁵. Like MS, CD is classified into standardized subtypes, namely ileal, ileocolic, and colonic, which ostensibly refer to the respective regions of the GI tract which may be affected by the disease. Clinically, 50% of the cases observed are characterized as ileocolic, whereby skip lesions and fistulae are observed throughout the entire lower GI tract⁶.

In the US alone, the cost burden of CD proves to be substantial – in 2006, estimated direct medical costs were approximately $18,500 per patient on an annualized basis. Hospitalizations accounted for 53–66% of direct medical costs, with an average cost-per-hospitalization of $37,459. Moreover, indirect costs represented 28% of the total costs associated with Crohn's disease⁷.

In a recent study performed by Taylor et al, an analysis demonstrated that in 2007, the average annual direct and indirect costs per patient were $ 17,133 and $12,671, respectively⁸. In another study published in the Journal of Neurology, Neurosurgery, and Psychiatry, the total mean annual costs per MS patient in Europe varied between $13,500 and $46,600 in 2006⁹.

Natalizumab, a recombinant immunoglobulin-4 (IgG4) monoclonal antibody that antagonizes α4 integrin, has recently been approved for the treatment of MS, in addition to its longstanding approval for the treatment of refractory CD. In the context of MS, natalizumab inhibits the interaction of α4-β1 integrins with vascular cell adhesion molecule-1 (VCAM-1), thus preventing inflammatory mediators from crossing the blood–brain barrier¹⁰. In the context of CD, natalizumab inhibits the interaction of α4-β7 integrins with mucosal addressin cell adhesion molecule (MadCAM-1), ultimately resulting in a reduction in leukocyte migration into the parenchyma and leukocyte activation within inflammatory sites¹⁰. Natalizumab's novel mechanism of action suggests its potential in the treatment of these debilitating diseases; it also underscores the need for considerable deliberation in the context of health-plan formulary consideration, given the direct and indirect costs associated with this medication. Natalizumab is indicated as monotherapy for the treatment of patients with relapsing forms of multiple sclerosis to delay the accumulation of physical disability and reduce the frequency of clinical exacerbations¹⁶. The safety and efficacy of natalizumab use beyond 2 years is unknown. It is indicated for inducing and maintaining clinical response and remission in adult patients with moderate to severe Crohn's disease with evidence of active inflammation who have had an inadequate response to, or are unable to tolerate, conventional CD therapies and inhibitors of TNFα¹⁰.

The purpose of this study was to analyze the clinical and economic data associated with natalizumab and to determine the potential impact of its formulary inclusion in a hypothetical health plan. An assessment of the cost effectiveness of natalizumab is provided through the independent analysis of the manufacturer's value argument in the context of both MS and CD.

Methods

Combined cost-effectiveness and budget-impact analysis of natalizumab for MS

To assess the cost effectiveness of natalizumab for Crohn's disease in comparison with comparator MS therapies, the authors applied a retrospective combined cost-effectiveness and budget-impact analysis model originally designed by Elan Pharmaceuticals³⁰. This model assumed the perspective of a third-party US managed-care payer and was applied to pooled data from five randomized, controlled, double-blind studies, in addition to claims data from a proposed hypothetical health plan of 1 million members and a total of 592 patients with relapsing forms of MS receiving immunomodulatory therapy. The time horizon or enrollment period was set to 2 years^10–14.

Due to a lack of head-to-head comparisons with natalizumab, conclusions were essentially limited to indirect comparisons. This study estimated the incremental cost effectiveness of natalizumab, given as an intravenous (IV) dose of 300 mg once every 4 weeks, in the treatment of patients with relapsing forms of MS relative to the following drugs: interferon β-1a (Avonex, Biogen Idec, Inc, Research Triangle Park, NC, USA), given as an intramuscular (IM) dose of 30 μg once weekly; interferon β-1b (Betaseron, Bayer Healthcare Pharmaceuticals, Inc., Montville, NJ, USA), given as a subcutaneous (SC) dose of 0.25 mg every other day; glatiramer acetate (Copaxone, Teva Pharmaceuticals, Ltd., North Wales, PA, USA), given as a SC dose of 20 mg once daily; and interferon β-1a (Rebif, EMD Serono, Inc., Darmstadt, Germany), given as a SC dose of 44 μg three times a week, to determine budgetary impact of increased utilization of natalizumab from a payer perspective. In addition, this study further evaluated the impact of incremental pharmacy costs.

In order to generate 2-year costs per patient and to effectively compare the relapse rate reduction between natalizumab and other comparator treatments, the model utilized the 2-year relapse rate occurring in the placebo groups of the immunomodulatory phase III clinical trials in addition to the number of patients in each placebo group to calculate a weighted average number of relapses per patient prior to treatment with each of the therapies. The relapse rate after immunodulator therapy was 67% over 2 years for natalizumab (p < 0.001); 32% for IM interferon β-1a (p = 0.04); 34% for interferon β-1b (p = 0.0101); 29% for glatiramer acetate (p = 0.007); and 32% for SC interferon β-1a (p < 0.005)^10–14. Data acquisition costs were then incorporated, and the difference in frequency of administration, information regarding the number of doses per package, and the number of packages per month was incorporated into the model by calculating the net monthly drug costs and net 2-year drug costs.

The model assumed that only one complete blood count (CBC) at $42 per test, and one liver function test (LFT) at $51 per test, were conducted at the initiation of therapy, and incorporated laboratory tests and imaging studies required or explicitly recommended by immunomodulator prescribing information during treatment^15–20. The model assumed one MRI ($1,966 per MRI) for natalizumab patients over the 2-year period and no MRIs for any of the interferon therapies or glatiramer acetate^15,16. Although no guidance on the frequency of regular follow-up visits was suggested by the prescribing information, the model assumed a conservative approach and incorporated one neurologist office visit ($159 per visit) every 6 months during therapy, for a total of four visits over the 2-year period^15,16. However, the model assumed a total of five visits over the 2-year period for patients receiving natalizumab. Prescribing information for natalizumab required prescribers to evaluate a patient receiving the drug at 3 and 6 months after the first infusion and every 6 months thereafter^16,21.

To account for the varying costs of managing MS relapse based on the intensity of relapse, the model accounted for distribution of patients requiring low (40% of patients, $310 per relapse), medium (40% of patients, $2,357), and high-intensity (20% of patients, $16,422 per relapse) medical management^22,23.

Cost effectiveness analysis and budget impact analysis of natalizumab therapy for CD

To assess the cost effectiveness of natalizumab for Crohn's disease with comparator TNFα inhibitor therapies, we applied a retrospective budget impact analysis model designed by Elan Pharmaceuticals²⁸. This model assumed the perspective of a third-party US managed-care payer and was applied to pooled data from three randomized, controlled, double-blind studies, and claims data from a proposed hypothetical health plan of 1 million members and a total of 1,014 patients with moderate-to-severe CD with active inflammation who were eligible for second-line biologic therapy. The time horizon or enrollment period was set to 2 years^24–27. Due to a lack of head-to-head comparisons with natalizumab, conclusions from indirect comparisons were limited. This study estimated the incremental cost effectiveness of natalizumab (300 mg IV once every 4 weeks) in the treatment of patients with moderately to severely active CD relative to the TNFα inhibitors, infliximab (Remicade, Centocor Ortho Biotech Inc., Horsham, PA, USA), given at 5 mg/kg IV on 0, 2, and 6 weeks, followed by 5–10 mg/kg thereafter; and adalimumab (Humira, Abbott Labs, Abbott Park, IL, USA), given at 160 mg SC on week 0, then 80 mg on week 2, and 40 mg every other week thereafter, to determine the budgetary impact of increased utilization of natalizumab from a payer perspective. In addition, this study further evaluated the impact of incremental pharmacy costs^16,26,28.

In order to generate 2-year costs per patient and compare the response rate between natalizumab and competitor treatments, the model utilized placebo-adjusted steroid-free remission, response and non-response rates for primary and secondary failures during maintenance treatments, which were reported from previous randomized controlled clinical trials (CD1 and CD3 for natalizumab, ACCENT-1 for infliximab, and CHARM for adalimumab)^24,25,27. Rates were adjusted to account for variable placebo responses, and were calculated as proportions of patients entering the model at induction. Model inputs for steroid use in the natalizumab arm at months 6 and 12 were estimated as 26.9% and 26.0%, respectively. Model inputs for steroid use in the infliximab arm at months 6 and 12 were estimated as 29.0% and 29.7%, respectively. Model inputs for steroid use in the adalimumab arm at months 6 and 12 were estimated as 25.0% and 29.7%, respectively. Since no data were available for steroid use rate beyond 12 months, the model assumed that rates at 18 and 24 months for each product were identical to the steroid use rates at month 12^24,25,27.

The model assumed that the cost of hospitalization in each of the 6 months over 2 years corresponded with costs incurred with hospitalization by response rate and applied data from CD Database Study by Elan Pharmaceuticals to determine hospitalization costs. The model assumed hospitalization costs during remission/response at $11,987, $3,044 for medical costs associated with steroid use, and $3,487 for non-hospitalization costs. The model also assumed hospitalization costs incurred during non-response as $11,987, $7,421 for medical costs associated with steroid use, and $3,487 for non-hospitalization medical costs³¹.

Results

Multiple sclerosis

Based on the model assumptions, the results of the cost effectiveness analysis demonstrated that, in comparison with natalizumab, the comparator treatments (IM interferon β-1a, interferon β-1b, glatiramer acetate and SC interferon β-1a) were less costly over 2 years of therapy. However, these agents were also less effective when compared to natalizumab as assessed by the number of relapses avoided over 2 years. The incremental cost-effectiveness ratios (ICERs) of natalizumab relative to the comparator treatments were: natalizumab versus IM interferon β-1a: $23,029 incremental cost per additional relapse avoided; natalizumab versus interferon β-1b: $24,452 incremental cost per additional relapse avoided; natalizumab versus glatiramer acetate: $20,671 incremental cost per additional relapse avoided; natalizumab versus SC interferon β-1a: $20,403 incremental cost per additional relapse avoided³⁰.

The result of the threshold analysis determined the increase in cost of natalizumab at which point the cost-effectiveness ratios of natalizumab and a comparative agent would be equal. The results were: natalizumab versus IM interferon β-1a: increase natalizumab cost from $36,060 to $59,452 per year. Natalizumab versus interferon β-1b: increase natalizumab cost from $36,060 to $55,938 per year. Natalizumab versus glatiramer acetate: increase natalizumab cost from $36,060 to $66,052 per year. Natalizumab versus SC interferon β-1a: increase natalizumab cost from $36,060 to $61,282³⁰.

One-way sensitivity analyses were carried out to evaluate the impact of varying model parameters on the cost per relapse avoided for natalizumab. The two parameters with the most influence on the cost per relapse avoided for natalizumab were the anticipated relative reduction in relapse rate and the average number of relapses a patient suffers from prior to treatment. The analysis indicated that, even if the anticipated reduction in relapse rate decreased from 67% to 50% relative to placebo, the cost per relapse avoided for natalizumab would increase to $77,316, making natalizumab still one of the most cost-effective agents in reducing the rate of relapses³⁰.

Crohn's disease

Based on the assumptions and inputs used in the model, the projected average time a patient would spend in remission over the 2-year modeled period was 0.38 years (4.56 months) with natalizumab, 0.20 years (2.4 months) with infliximab, and 0.24 years (2.88 months) with adalimumab³¹.

The budget impact model estimated the potential budget impact of post-natalizumab launch in which natalizumab was assumed to have a 13% share of the biologics market of eligible patients, with a shift in market share away from infliximab and adalimumab occurring in equal proportions. The introduction of natalizumab produced a market acquisition in the 2-year treatment period of roughly 7.5% from those other agents which results in a 13% net garner of biologics market share among the subset of patients aligned with the labeled indication of natalizumab for patients who are failures of both standard therapy and prior TNF inhibitor therapy³¹. The budget impact was estimated as the change in total costs (medical costs + drug costs) resulting from shifts in market share between the pre- and post-natalizumab launch scenarios. Table 1 summarizes the estimated change in costs in pre- and post- natalizumab launch scenarios. The results demonstrated an increase in total drug costs of $924,643, which was offset by a decrease in medical costs of $77,131³¹.

Table 1. Budget impact model results³¹.

	Pre-TYSABRI launch	Post-TYSABRI launch	Change
n	1,014	1,014	—
Distribution by treatment
TYSABRI	—	13.0%	13.0%
Infliximab	50%	43.5%	(6.5%)
Adalimumab	50%	43.5%	(6.5%)
Population costs over induction and maintenance
Drug costs	$26,731,266	$27,655,910	$924,643
Medical costs	$36,079,095	$36,001,965	–$77,131
Total costs	$62,810,362	$63,657,874	$847,513
PMPM costs
Drug costs	$1.11	$1.15	$0.039
Medical costs	$1.50	$150	$(–0.003)
Total costs	$2.62	$2.65	$0.035

Discussion

Multiple sclerosis and Crohn's disease leave many patients suffering from unrestrained physical or cognitive impairment and uncontrolled gastrointestinal inflammation, respectively³⁰. Many approaches to therapy result in failure or unbearable adverse effects leading to noncompliance. There are several contributing factors that lead to patients discontinuing their treatment³⁰. It has been observed that, with currently available first-line therapies, 4% of patients discontinue their therapy and switch to another, while the remaining 96% discontinue altogether³⁰. The reasons discovered for discontinuing therapy include intolerance and adverse events, lack of efficacy perceived by the patient, self-administration difficulty, or perception that one's disease status is not severe³⁰. Clearly, new medications such as natalizumab potentially have positive effects in bridging this gap in needed therapy through reduced frequency of administration, and more clinically significant and notably better patient-perceived outcomes.

A cost-effectiveness analysis was performed to evaluate natalizumab's possible use compared to other therapies. In Biogen's economic model, four primary inputs were employed to drive the model's estimates: (1) drug acquisition costs, (2) drug administration costs, (3) anticipated reduction in relapse rate after 2 years of therapy, and (4) estimated market utilization of natalizumab³⁰. Each primary input consisted of different assumptions as determined by Biogen Idec and Elan. Additional key assumptions included the following: injectable administration, constant sample population, and 100% compliance. The cost-effectiveness portion of the model further assumed that one patient began therapy with each immunomodulator agent in order to assess the per-patient costs of 2 years of therapy.

In calculating natalizumab's comparative effectiveness, ICERs illustrated that the other currently available drugs on the market were not as expensive as natalizumab, but these drugs were also not as effective. When comparing natalizumab versus IM interferon β-1a, the incremental cost per additional relapse avoided was $23,029; natalizumab versus interferon β-1b was $24,452, natalizumab versus glatiramer acetate was $20,671, and natalizumab versus SC interferon β-1a was $20,403³⁰.

Included in the economic model was an analysis comparing the 1-year total costs of therapy for each of the aforementioned immunomodulator agents to natalizumab, and examining at what point the cost effectiveness of both agents were equal (i.e. threshold analysis). Compared to IM interferon β-1a, natalizumab's 1-year therapeutic costs would have to increase from $36,060 to $59,452 in order to be equally cost effective². Furthermore, natalizumab's 1-year therapeutic costs would have to increase to $55,938 to be equal to interferon β-1b, $66,052 to be equal to glatiramer acetate, and $61,282 to be equal to SC interferon β-1a³⁰. The effectiveness of natalizumab was so superior that the cost-effectiveness ratio outperformed the other agents.

A sensitivity analysis illustrated that the anticipated relative reduction in relapse rate and the average number of relapses a patient suffered prior to treatment were the most prominent factors affecting the cost per relapse avoided²⁷. It further tested the result of decreasing the reduction of relapse rate achieved by natalizumab to 17%. The resultant cost per relapse avoided would increase from $56,594 to $71,316; however, this estimate still proved natalizumab to be more cost effective²⁷. With respect to the model assumptions, the market share of natalizumab would experience an increase to 8.5%, leading to a total per-member per-month healthcare cost increase of $0.003 ($0.002 for pharmacy costs and $0.001 for medical costs)²⁷.

The same pharmacoeconomic analysis was performed for natalizumab in the treatment of Crohn's disease. The analysis compared the cost effectiveness of natalizumab, infliximab, and adalimumab in patients who were not responsive to TNFα inhibitor treatment. Specific model assumptions included the following: all patients matched the indication use for natalizumab, 2-year maintenance period, three definitive states of health (remission, response, and non-response as characterized by CDAI), non-responders switched to another biologic agent, patients were concomitantly taking other immunosuppressive medications, and patients remained non-responders once enrolled²⁸. A decision tree was constructed to demonstrate the model's separation of two distinct phases: induction and maintenance. Additional key assumptions based upon population data were that 51.3% of model patients were primary failures of TNFα inhibitor therapy, and 48.7% were secondary failures²⁸.

In the cost-effectiveness analysis, the model evaluated three parameters for all three of the aforementioned therapeutic agents: time spent in remission, time spent in steroid-free remission and time spent in either remission or response. Over a 2-year period outlined by the model, natalizumab produced the highest average time in each health state in comparison to the other agents. The mean total costs associated with the initiation of natalizumab, infliximab, and adalimumab were $68,372, $62,090, and $61,796, respectively³¹. Although natalizumab's costs were higher, the mean time spent in remission while on this medication was 4.5 months, as opposed to 2.4 months for infliximab and 2.9 months with adalimumab³¹. In addition, natalizumab demonstrated the lowest cost-effectiveness ratio – the average cost for a patient-year in remission was 70% and 45% higher with infliximab and adalimumab, respectively; 61% and 29% higher for a patient-year in steroid-free remission, respectively; and 44% and 29% higher for a patient-year in remission or response, respectively³¹. In the sensitivity analysis, all average cost-effectiveness ratios were favorable in spite of significant alterations in model inputs such as a theoretical 100% increase in the cost of natalizumab and reductions of remission free patients by 50%³¹.

The budget impact analysis assumed an equal market distribution among the other TNFα agents. The introduction of natalizumab produced a market acquisition in the 2-year treatment period of roughly 7.5% from those other agents. This resulted in a 13% net garner of biologics market share among the subset of patients aligned with the labeled indication of natalizumab for patients who are failures of both standard therapy and prior TNF inhibitor therapy³¹. This shift in market share was used to estimate the change in total costs (medical + pharmacy),and the per-member per-month change for the models base case was calculated to be $0.035³¹.

Natalizumab's route of administration and required RiskMAP undoubtedly designates it as a specialty pharmaceutical product. The difficulty with specialty pharmaceuticals is that insurers do not know whether to cover them under the pharmacy or medical benefit structure. The TOUCH Program – a RiskMAP – necessitates specialized healthcare professionals to administer natalizumab, provide adequate monitoring of adverse events, quality of life assessments, and patient education³⁰. Traditionally, the use of such services would cover such a drug under a patient's medical benefit. Pharmacy benefits generally follow more stringent coding, billing logistics, contractual definitions, and tracking³². Obviously, a rearrangement of specialty pharmaceuticals under the pharmacy benefit may allow drugs like natalizumab to be eligible for manufacture discounts or rebates, facilitate processing, and provide detailed information for future analysis³².

Conclusion

Due to the complexity of both diseases and their respective therapies, there continues to be an unmet need for the treatment of MS and CD. The reintroduction of natalizumab, however, provides clinical practitioners with another tool in their fight against both conditions by way of a different mechanism of action. After a thorough review of the evidence, the authors find that natalizumab has been shown to be relatively cost effective in the treatment of both MS and CD from a payer perspective; the therapy adds a new option for those patients for whom conventional treatment was unsuccessful.

The evidence presented here shows that natalizumab provides clinical practitioners with another tool in their fight against both MS and CD, albeit by way of a different mechanism of action. After a thorough review of the evidence, the authors find that natalizumab has been shown to be relatively cost effective in the treatment of both conditions from a payer perspective; the therapy adds a new option for those patients for whom conventional treatment was unsuccessful. Payers should consider these conclusions along with the fact that the long-term effects of natalizumab on morbidity and mortality have yet to be established; thus, additional research should be performed before fully assessing the benefits of natalizumab use in both MS and CD. Further cost-effectiveness research must also be performed in regard to the financial impact of severe adverse events on natalizumab therapy, such as the cost of monitoring and possible treatment of progressive multifocal leukoencephalopathy. Additionally, thorough analyses of payers' respective populations must be implemented to determine whether including natalizumab as a treatment option for their MS and CD patients is the best strategy to maximize their resources.

Transparency

Declaration of funding:

This study was not funded.

Declaration of financial/other relationships:

J.B., R.B-L., M.J., J.L., C.R., L.S., C.R., and F.T.S. have disclosed that they have no relevant financial relationships.

The JME peer reviewers 1 and 2 have not received an honorarium for their review work on this manuscript. Both have disclosed that they have no relevant financial relationships.

Acknowledgements

The authors would like to acknowledge Richard N. Fry, RPh, Director of Programs, Foundation for Managed Care Pharmacy for his contributions toward this work and his giving the opportunity to the students to pursue this project. The authors are also grateful to Biogen Idec and Elan Pharmaceuticals Inc. for the use of their dossier for natalizumab for this competition.

Notes

* Elan Pharmaceuticals, Inc. Dublin, Ireland; Biogen Idec, Inc. Wellesley, MA

* Elan Pharmaceuticals, Inc. Dublin, Ireland; Biogen Idec, Inc. Wellesley, MA