Multiple sclerosis (MS) is a chronic and debilitating inflammatory autoimmune disorder of the central nervous system that afflicts approximately 400,000 people in the United States and 2.1 million people worldwide [
Treatment of MS has advanced significantly over the past several decades. Historically, MS was treated in a solely supportive manner through symptomatic pharmacological treatment in the event of disease exacerbations (MS attacks), generally via powerful doses of short-course steroids [
Characteristics of disease-modifying therapies for multiple sclerosis.
DMT brand name (Generic name) | Manufacturer | FDA approval year | Dose frequency administration [ |
2010 Annual Tx cost [ |
Significant risks listed in package insert |
---|---|---|---|---|---|
Betaseron [ |
Bayer Health Care Pharmaceuticals, Inc | 1993 | 250 ug 2 days SC | $38,369 | |
Avonex [ |
Biogen Idec, Inc | 1996 | 30 ug weekly IM | $38,532 | |
Copaxone [ |
Teva Neuroscience, Inc | 1996 | 20 mg daily SC | $42,940 | |
Novantrone [ |
Novartis Pharmaceuticals Corporation | 2000 | 12 mg/m2 |
$6,344 | Cardiotoxicity |
Rebif [ |
EMD Serono, Inc | 2000 | 44 ug 3x weekly SC | $38,646 | |
Tysabri [ |
Elan Pharmaceuticals, Inc and Biogen Idec, Inc | 2004 | 300 mg 4 weeks IV | $40,426 | Increased risk of PML, |
Extavia [ |
Novartis Pharmaceuticals Corporation | 2009 | 250 ug 2 days SC | $38,368 | |
Gilenya [ |
Novartis Pharmaceutical Corporation | 2010 | 0.5 mg daily oral | $47,944 | Cardiotoxicity, Elevated risk for infections |
IFN: Interferon; IM: intramuscular; IV: intravenous; PML: progressive multifocal leukoencephalopathy; SC: subcutaneous; Tx: treatment.
Three review studies from 2003-4 focused on cost-effectiveness of DMTs in the treatment of MS and included multiple current therapies [
Over the past ten years, there has been a lack of comprehensive and systematic reviews of multiple sclerosis cost-effectiveness studies. Previous reviews did not address DMTs approved in recent years (natalizumab or fingolimod) and none have been recent and comprehensive enough to detail the current state of the cost-effectiveness of DMTs in treating multiple sclerosis. The aim of this systematic review was to provide a current and comprehensive understanding of the cost-effectiveness of DMTs for the treatment of multiple sclerosis by quantitatively evaluating the quality of cost-effectiveness studies and exploring how the multiple sclerosis cost-effectiveness field has progressed from past recommendations.
Cost-effectiveness analysis (CEA) is a full form of health economic analysis, where both costs and consequences (outcomes) of health programs or treatments are examined. CEA is used in situations where decision makers are considering a limited range of options within a given field, and within a given operating budget [
CEA studies follow a decision-analytic modeling (DAM) approach where currently available evidence concerning the effectiveness and costs of alternative healthcare interventions is synthesized in order to inform decision makers about the relative value of competing alternatives [
We performed a systematic search in September of 2012 using MEDLINE (PubMed) querying for the MeSH term, “cost-benefit analysis,” and the general search term, “multiple sclerosis.” The MeSH term “cost-benefit analysis” includes the following nested entry terms: cost-benefit, cost-effectiveness, and benefits and costs. There is no other MeSH term comprehensive enough to require elements of both costs and health outcomes. To control for the impact of including DMT generic and brand names in the query, we included (mitoxantrone OR interferon OR glatiramer acetate OR natalizumab OR fingolimod OR avonex OR betaseron OR extavia OR rebif OR novantrone OR copaxone OR gilenya OR tysabri) in addition to the initial search terms as part of an additional search for comparative purposes. We utilized PubMed search filters to limit the initial search to articles published in the English language from January 1, 2004 to August 31, 2012. In addition to the PubMed filters applied, we imposed a three-point exclusion criterion to the search results. The exclusion criteria were mutually exclusive, exhaustive, and hierarchical in the following manner (see Figure not original research; not comparative; does not include both costs and outcomes.
Flowchart of systematic search methodology and yield.
If an article was excluded under the first criterion, it would not be further scrutinized for adherence to the second or third criteria, and if it adhered to the first but not to the second, it would not be further scrutinized for adherence to the third criterion. Those that reached this final exclusion criterion and adhered to all three criteria were included in the final systematic review database. To ensure the comprehensiveness of our database, we performed two additional searches using the same queries and criteria via the Tufts Medical Center Cost-Effectiveness Analysis Registry (TUFTS) and the National Health Service Economic Evaluation Database (NHSEED). Additional articles that were not found in the initial search and adhered to our three-point criterion were added to the final database.
We quantitatively evaluated the quality of the MS DMT cost-effectiveness studies in a systematic and transparent manner through the use of the 16-item Quality of Health Economic Studies (QHES) validated instrument [
We identified four main topics to focus our discussion of the composition and quality of the included cost-effectiveness studies: (i) model structure (QHES items 1, 2, 8, 12, and 13), (ii) model inputs (QHES items 3, 7, 9, 10, and 11), (iii) results/conclusions (QHES items 6, 11, and 15), and (iv) study integrity (QHES items 5, 13, 14, and 16) [
Of the 82 total studies that met our initial MEDLINE search criteria [
Table
Cost-effectiveness analyses of multiple sclerosis disease modifying therapies (2004–2012).
First author year | Model type |
Study pop |
Comparators | Primary effectiveness outcome | Results | Stated conclusion | QHES |
Sponsor |
---|---|---|---|---|---|---|---|---|
Prosser |
CSM1 |
Non-primary progressive MS |
(1) IFNB-1a (IM) |
QALYs | 1 versus 4: Not CE |
1–3 less CE than 4 under a wide range of assumptions. | 87 | National MS Society |
| ||||||||
Iskedjian |
CSM1 |
SDE at risk for CDMS |
(1) IFNB-1a (IM) + MPS (iv) |
QAMLY | $227,586/QAMLY (MoH) |
1 is a reasonably CE approach to Tx of patients with a SDE. ICER improves if Tx is initiated in pre-CDMS. | 69 | Biogen Idec, Inc. |
| ||||||||
Perini |
OLH2HCS |
SPMS |
(1) Mitoxantrone |
Relapse rate decrease/EDSS decrease | 1: 88%/0.9 (€8171) |
2 should be considered as a first-line rescue therapy for MS | 34 | Not stated |
| ||||||||
Bell |
CSM1 |
RRMS |
(1) GA |
QALYs | 1 versus 5: $258,465/QALY |
1 is the best DMT and resulted in better outcomes than 5 alone. | 90 | Teva Neuroscience, Inc. |
| ||||||||
Gani |
CSM1 |
HARRMS |
(1) Natalizumab |
QALYs | 1 versus 2: |
1 is more CE than 2, 3 and 4 for HARRMS. | 100 | Biogen Idec, Inc. |
| ||||||||
Kobelt |
CSM1 |
RRMS |
(1) Natalizumab |
QALYs | 1 dominates 2 | For this population, 1 provides an additional health benefit at a similar cost to 2. | 90 | Biogen Idec Inc./Elan Pharma. |
| ||||||||
Castelli-Haley |
RMCA |
ITT: CDMS, GA or IFNB-1b Rx, Ins. cov. |
(1) GA |
Risk of Relapse |
ITT: |
RRMS patients treated with 1 have significantly lower relapse rate. Costs are lower for 1 in CU cohort. | 43 | Teva Neuroscience, Inc. |
| ||||||||
Chiao |
CSM2 |
DMT candidates with relapsing MS |
(1) Natalizumab |
Relapses avoided | 1 versus 2: $23,029/RA |
1 was the most CE. | 78 | Biogen Idec, Inc./Elan Pharma. |
| ||||||||
Earnshaw |
CSM1 |
RRMS |
(1) Natalizumab |
QALYs | Healthcare: |
1 and 2 are associated with increased benefits compared with 3 at higher costs. | 88 | Teva Neuroscience, Inc. |
| ||||||||
Goldberg |
CSM3 |
RRMS |
(1) GA |
Relapses avoided | 1 versus 5: $88,310/RA |
1, 3 and 4 represent the most CE DMDs for Tx of RRMS. | 86 | EMD Serono, Inc |
| ||||||||
Guo |
PLS4 |
Relapsing MS |
(1) IFNB-1a (SC) |
Relapses avoided/ |
1 versus 2: |
1 versus 2 yields greater health benefits over 4 years at a reasonable cost. | 78 | EMD Serono, Inc |
| ||||||||
Jankovic |
CSM1 |
RRMS |
(1) GA (SC) |
QALYs | >$20,000,000/QALY |
IMT of RRMS in a Balkan country is not CE. | 46 | Serbian Ministry of Science and Ecology |
| ||||||||
Kobelt |
CSM1 |
Relapsing forms |
(1) DMTs |
QALYs | 1 versus 2: €15,385/QALY |
Cost increase with DMTs is moderate for health gained. | 71 | Authors declare none |
| ||||||||
Lazzaro |
CSM5 |
CIS patients |
IFNB-1b: |
QALYs | INHS: €2,575 1 versus 2 |
1 significantly delays conversion to CDMS and is economically advantageous. | 75 | Bayer Schering Pharma, Italy |
| ||||||||
Tappenden |
CSM1 |
Medicare |
(1) IFNB-1a (PA) |
QALYs | 1 versus 7: $66–234k/QALY |
Suggests prudent use of a discontinuation rule may improve CE. | 100 | USDHHS AHRQ |
| ||||||||
Bakshai |
CSM2 |
Relapsing forms receiving IMT |
(1) Natalizumab |
Relapses avoided | 1 versus 2: $23,029/RA |
1 is relatively CE and adds a new option for those patients for whom conventional Tx was unsuccessful. | 50 | Not funded |
| ||||||||
Nuijten |
CSM1 |
RRMS |
(1) IFNB-1a (SC) |
Relapses avoided | 1 versus 5: €51,250/RA |
1 versus 5 had favorable overall CE compared with all other available DMDs for the Tx of RRMS. | 79 | Merck Pharma |
| ||||||||
Tappenden |
CSM1 |
SPMS |
(1) AHSCT |
QALYs | 1 versus 2: |
1 could potentially achieve an acceptable CE, however RCTs are needed to confirm this. | 100 | No commercial or research funding |
| ||||||||
Noyes |
CSM1 |
RRMS and SPMS |
(1) IFNB-1a (IM) |
QALYS | 1–4 versus 5: |
DMTs resulted in small health gains for the cost. Starting DMTs earlier resulted in more favorable CE results. | 83 | Biogen, Boston Scientific, NIH, NMSS, and USDOD |
| ||||||||
O'Day |
CSM3 |
Relapsing MS |
(1) Natalizumab |
Relapses avoided | 1 dominates 2 | 1 dominates 2 in terms of incremental cost per relapse avoided. | 86 | Biogen Idec, Inc. |
| ||||||||
Caloyeras |
CSM1 |
First initial event suggestive of MS and CDMS |
(1) IFNB-1b early treatment |
QALYS | 1 dominates 2 | Early treatment improved patient outcomes while controlling costs. 1 dominates 2 | 99 | Abt Bio Pharma Solutions, Inc. |
| ||||||||
Pan |
CSM1 |
RRMS |
(1) IFNB-1b early treatment |
Life-Years Gained; QALYS | 1 versus 2: $30,967/LY; |
Treatment during early disease phase increased patient life-years and QALYs. Early treatment with IFNB-1b likely cost-effective | 90 | Bayer Health-Care Pharmaceuticals |
Models: 1Markov, 2Cost-effectiveness/budget impact, 3Average patient simulation, 4Discrete event simulation, 5Open cohorts epidemiology model.
Comparators: No Tx: No physician care, Supportive: Symptom management alone.
Acronyms (alphabetical): AHSCT: Autologous haematopoietic stem cell transplantation; CE: Cost effective; CDMS: Clinically diagnosed multiple sclerosis; CIS: Clinically isolated syndrome; CP: Cyclophosphamide; CSM: Cohort simulation model; CU: Continuous use; DMD: Disease-modifying drug; DMT: Disease-modifying therapy; HARRMS: Highly active relapsing-remitting multiple sclerosis; IM: Intramuscular; IMT: Immunomodulatory treatment; INHS: Italian National Health Service; ITT: Intent to treat; MLY: Monosymptomatic life years gained; MoH: Ministry of Health; MPS: Methylprednisolone; NS: Not stated; OLH2HCS: Open-labeled head-to-head clinical study; PA: Physician-administered subcutaneous; PLS: Patient-level simulation; PMPM: Per member per month; PSS: Personal Social Services; QAMLY: Quality-adjusted monosymptomatic life years gained; RA: relapse avoided; RMCA: Retrospective multivariate cohort analysis; RRMS: Relapsing-remitting multiple sclerosis; SA: Self-administered subcutaneous; SC: Subcutaneous; SDE: Single demyelinating event; SoC: Societal; SPMS: Secondary progressive multiple sclerosis; UKNHS: United Kingdom National Health Service.
We identified two patient-level trial-based studies: one based on an open-labeled head-to-head clinical trial [
12 of the 22 studies were based upon data from the USA [
Of the included studies, 16 included an interferon product [
The most prominent primary effectiveness outcome chosen was QALYs, which was represented in 13 of the 22 studies [
Table
In 21 of the 22 studies, the objective was “presented in a clear, specific, and measurable manner” (QHES item 1) [
Inputs were drawn from the best available source in 20 of the 22 studies (QHES item 3) [
20 of the 22 studies performed an incremental analysis for costs and outcomes between alternatives (QHES item 6) [
20 of the 22 studies addressed uncertainty (QHES item 5) [
Clinical guidelines for the treatment of MS with DMTs remain underdeveloped and lacking in comprehensive understanding and consensus regarding what DMT should be used for what type of MS patient. The American Academy of Neurology (AAN) and the MS Council for Clinical Practice Guidelines have not published comprehensive guidelines including all current DMTs for the US since 2002 (guidelines were reaffirmed in 2008) [
There was a preponderance of simulation models within our included studies. Simulation modeling allows for the projection of short-term data to reflect the chronic nature of MS. The use of simulation models in lieu of long-term DMT studies indicates progress in the field and is consistent with previous recommendations regarding the use of simulation modeling when long-term cost and outcome studies are lacking. The large majority of these simulation models employed a Markov structure, which allows for long-term analysis, up to lifetime in scope. Markov modeling also allows for transition between disease states for cohorts of patients, which reflects the natural disease progression within MS. This further indicates a positive trajectory in the cost-effectiveness evaluation of DMTs for MS.
The past recommendations supporting use of the long-term time horizons and societal perspectives harmonize well with the use of QALYs as the primary outcome measure, as was suggested by Kobelt [
The assumptions employed in the simulation models were diverse and inconsistent between studies indicating much room for improvement in conforming to past recommendations with respect to consistency in modeling. This diversity is a factor in the wide range of cost-effectiveness estimates for the same DMTs across studies. The generally poor performance of studies in explicitly stating and properly justifying model assumptions, discussing the magnitude and direction of potential biases, employing transparent and repeatable models/methods, and justifying their choice of model (QHES items 12, 13, and 14) further indicates a lack of adherence to proper guidelines within the field for consistent and appropriate modeling methodology. We call for more efforts in the standardization of cost-effectiveness studies within MS and for studies to provide rationale for why the design and assumptions may differ from previous cost-effectiveness studies.
The lack of head-to-head randomized controlled trials (RCTs) between DMTs and absence of long-term observational data serve to drive the need for cost-effectiveness studies to employ multiple assumptions since the appropriate long-term comparative data do not exist. The currently available RCT data are generally specific to individual DMTs and are based upon different patient populations and different study characteristics such as treatment adherence, dropout rates, and adverse outcomes. Therefore, the results of individual studies are contingent upon which RCT data are employed, along with the assumptions included in the model. This heterogeneity makes comparing the results across studies difficult. More comparative head-to-head RCTs among DMTs and prospective observational studies are needed and will generate less heterogeneity in cost-effectiveness model structures and inputs yielding less uncertainty in cost-effectiveness results. It is important that proper guidelines for cost-effectiveness modeling [
Finally, there was a lack of cost-effectiveness studies on fingolimod. There is a need for studies including fingolimod as a comparator, specifically coupled with traditional DMTs (interferons and glatiramer acetate) to determine this new treatment’s value as a therapy option in MS. Given the high annual treatment cost of fingolimod, the cost-effectiveness evidence as compared to other DMTs becomes even more important. There were no studies that included all approved DMTs as comparators indicating a need for a comprehensive study including all appropriate DMTs for particular MS patient populations. A comprehensive cost-effectiveness study of this nature would help to alleviate the problems of comparing cost-effectiveness across different studies that employ different methodologies and study assumptions.
The cost-effectiveness body of evidence of DMTs for the treatment of MS has shown progress in responding to the recommendation of past reviews, while there remains room for improvement in many areas. The area of greatest advancement is in the use of simulation models that represent the chronic nature of the disease. This appears to be the dominant trend in current studies. To a lesser degree, the field has shown progress in adhering to the recommendation that long-term time horizons, societal perspectives, and QALYs are utilized, albeit with room for further improvement. We recommend that studies utilize lifetime horizons, societal perspectives, and QALYs as the primary standards in CEA studies of DMTs for MS, with the option to supplement the base-case analysis by including short-term horizons, payer perspectives, and a cost-consequences approach. This recommendation conforms to the suggestions of past reviewers, while adding the ability for individual studies to inform multiple decision makers. There is a great need for improvement in the standardization of modeling procedures and data inputs. We recommend that modeling and input assumption standards are developed within MS cost-effectiveness studies to aid comparability across future studies. The recommendation to perform head-to-head RCTs between DMTs and collect long-term prospective observational data would improve study consistency in the future. Finally, we recommend that comprehensive studies comparing all approved DMTs in concert are performed to help control for the inconsistencies between studies and provide meaningful results for decision makers.
This research was funded in part by the Agency for Healthcare Research and Quality as part of the Colorado Comparative Effectiveness Research K12 Training Program: K12 HS019464. No other funding source was directly tied to this study.