Cost-Utility of First-Line Disease-Modifying Treatments for Relapsing-Remitting Multiple Sclerosis

Clinical Therapeutics/Volume 39, Number 3, 2017 Cost-utility of First-line Disease-modifying Treatments for Relapsing–Remitting Multiple Sclerosis Erkki Soini, MSc 1 ; Jaana Joutseno, MSc 2 ; and Marja-Liisa Sumelahti, MD 3 1 ESiOR Oy, Kuopio, Finland; 2 Genzyme (a Sanofi Company), Helsinki, Finland; and 3 School of Medicine, University of Tampere, Tampere, Finland ABSTRACT Purpose: This study evaluated the cost-effectiveness of first-line treatments of relapsing–remitting multiple sclerosis (RRMS) (dimethyl fumarate [DMF] 240 mg PO BID, teriflunomide 14 mg once daily, glatiramer acetate 20 mg SC once daily, interferon [IFN]-β1a 44 mg TIW, IFN-β1b 250 mg EOD, and IFN-β1a 30 mg IM QW) and best supportive care (BSC) in the health care payer setting in Finland. Methods: The primary outcome was the modeled incremental cost-effectiveness ratio (ICER; €/qualityadjusted life-year [QALY] gained, 3%/y discounting). Markov cohort modeling with a 15-year time horizon was employed. During each 1-year modeling cycle, patients either maintained the Expanded Disability Status Scale (EDSS) score or experienced progression, developed secondary progressive MS (SPMS) or showed EDSS progression in SPMS, experienced relapse with/without hospitalization, experienced an adverse event (AE), or died. Patients' characteristics, RRMS progression probabilities, and standardized mortality ratios were derived from a registry of patients with MS in Finland. A mixedtreatment comparison (MTC) informed the treatment effects. Finnish EuroQol Five-Dimensional Questionnaire, Three-Level Version quality-of-life and direct-cost estimates associated with EDSS scores, relapses, and AEs were applied. Four approaches were used to assess the outcomes: cost-effectiveness plane and efficiency frontiers (relative value of efficient treatments); costeffectiveness acceptability frontier, which demonstrated optimal treatment to maximize net benefit; Bayesian treatment ranking (BTR); and an impact investment assessment (IIA; a cost-benefit assessment), which increased the clinical interpretation and appeal of modeled outcomes in terms of absolute benefitgained with fixed drug-related budget. Robustness of results was tested extensively with sensitivity analyses. Findings: Based on the modeled results, teriflunomide was less costly, with greater QALYs, versus glatiramer acetate and the IFNs. Teriflunomide had the lowest ICER (24,081) versus BSC. DMF brought marginally more QALYs (0.089) than did teriflunomide, with greater costs over the 15 years. The ICER for DMF versus teriflunomide was 75,431. Teriflunomide had 450% cost-effectiveness probabilities with a willingness-to-pay threshold of o€77,416/QALY gained. According to BTR, teriflunomide was first-best among the disease-modifying therapies, with potential willingness-to-pay thresholds of up to €68,000/QALY gained. In the IIA, teriflunomide was associated with the longest incremental quality-adjusted survival and time without cane use. Generally, primary outcomes results were robust, based on the sensitivity analyses. The results were sensitive only to large changes in analysis perspective or mixed-treatment comparison. Implications: The results were sensitive only to large changes in analysis perspective or MTC. Based on the analyses, teriflunomide was cost-effective versus BSC or DMF with the common threshold values, was dominant versus other first-line RRMS treatments, and provided the greatest impact on investment. Teriflunomide is potentially the most cost-effective option among first-line treatments of * Selected data from this article were presented in poster format at the 31st Congress of the European Committee for Treatment and Research in Multiple Sclerosis, Barcelona, Spain, October 7–10, 2015; and in poster format at the 18th Annual European Congress of the International Society for Pharmacoeconomics and Outcomes Research, Milan, Italy, November 7–11, 2015 (Value Health 2015;18:A756). Accepted for publication January 18, 2017. http://dx.doi.org/10.1016/j.clinthera.2017.01.028 0149-2918/$ - see front matter &2017 The Authors. Published by Elsevier HS Journals, Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). March 2017 537 RRMS in Finland. (Clin Ther. 2017;39:537–557) &2017 The Authors. Published by Elsevier HS Journals, Inc. Key words: cost-effectiveness, dimethyl fumarate, economic evaluation, glatiramer acetate, interferon-β, teriflunomide. INTRODUCTION Multiple sclerosis (MS)—a chronic progressive, autoimmune, inflammatory disease—affects 42 million people worldwide. Approximately 89% of cases are classified as relapsing–remitting MS (RRMS) at the time of diagnosis. 1 MS prevalence is particularly high in the United Kingdom, the United States, Canada, Germany, and Scandinavia. 2,3 In Finland, MS prevalence varies regionally, from 100 to 200 per 100,000 inhabitants. 4–7 In young adults with MS, prognosis is based on an individual’s factors. 1 The progression and accumulating disability cause a significant human and economic burden 8–15 and the need for support. 16 The risk for death among Finnish patients with MS is 2.8-fold compared with that in the general population, being 3.4-fold in women and 2.2-fold in men as early as 2 to 10 years after diagnosis. 17 Relapse, MS progression, and disability level (eg, higher Expanded Disability Status Scale [EDSS] score 18 ) are associated with a higher risk for mortality, 17,19,20 additional costs, 9–14 and quality of life (QoL) losses. 9,10,12,14,21–24 MS treatment with disease-modifying therapies (DMTs) is aimed at decreasing the inflammatory activity leading to relapses, stopping or slowing progression of residual disability, and, eventually, delaying the progression to the secondary progressive phase. However, long-term prognosis among treated patients is largely unknown. Based on Finnish drug reimbursement and sales data, 25 commonly used firstline DMTs include injectable DMTs, namely glatiramer acetate (GA), interferon (IFN)-β1a IM, IFN-β1a SC, and IFN-β1b SC. Dimethyl fumarate (DMF) and teriflunomide are new oral DMTs reimbursed as the first-line treatment of RRMS in Finland. The efficacy and safety of DMF 240 mg BID for established MS have been studied in the Phase III CONFIRM (Comparator and an Oral Fumarate in Relapsing-Remitting Multiple Sclerosis) 26,27 and DEFINE (Determination of the Efficacy and Safety of Oral Fumarate in Relapsing– Remitting MS) 28,29 trials (ClinicalTrials.gov identifiers: NCT00451451 and NCT00420212, respectively). The efficacy and safety of teriflunomide 14 mg once daily for established MS have been demonstrated in the Phase III TEMSO (Teriflunomide Multiple Sclerosis Oral Teriflunomide for Relapsing Multiple Sclerosis) 30–33 and TOWER (Teriflunomide Oral in People With Relapsing Multiple Sclerosis) 34,35 trials (NCT00134563 and NCT00751881, respectively), and in patients with a first clinical episode suggestive of MS in the TOPIC (Oral Teriflunomide for Patients with a First Clinical Episode Suggestive of Multiple Sclerosis) 36 trial (NCT00622700). Effectiveness of teriflunomide compared with IFN-β1b SC has been demonstrated in the Phase III TENERE (Teriflunomide and Rebif® in Patients with Relapsing Multiple Sclerosis) 37 trial (NCT00883337). We evaluated the cost-utility of injectable and oral first-line DMTs in the Finnish population of patients with RRMS, based on a decision-analytical model. To our knowledge, there are no previously published journal articles on the cost-utility of first-line oral DMTs in a European setting or on oral and injectable DMTs for first-line treatment of RRMS. In addition, progression of RRMS in Finnish patients has not been assessed before, and the 4 different approaches elaborating the key results from MS cost-utility analysis have not been previously reported. MATERIALS AND METHODS The cost-utility of the first-line DMTs in the Finnish RRMS population was assessed in a decisionanalytical modeling framework 38 by implementing a Markov cohort model with mutually exclusive health states in Excel 2007, including Visual Basic for Applications (Microsoft Corporation, Redmond, Washington). The modeling approach followed the Finnish guidance for health economic analyses. 39 The primary outcome of analysis was the modeled incremental cost-effectiveness ratio (ICER), reported as Euros per quality-adjusted life-year (€/QALY) gained. The interpretation of ICER is challenging in Finland because the decision maker’s willingness-topay (WTP) threshold per QALY gained has not been publicly declared, 40 and significant variation in Clinical Therapeutics 538 Volume 39 Number 3 decision maker WTP between diseases may exist. 41 Based on our experience, the UK thresholds 42,43 could be applicable in Finland, so that values of o€25,000 or €25,000–37,000/QALY gained would indicate most plausible or plausible cost-effectiveness, respectively; and, on average, €55,000/ QALY gained could be acceptable for end-of-life treatment based on the UK population-weighted decisions. This applicability of UK thresholds is based on the observation that many articles from Finland 41,44–55 have referred to a WTP threshold of €50,000/QALY gained, which is probably based on the so-called "dialysis argument." 41 The Finnish Medicines Agency has considered that €68,000/QALY gained approaches the maximum cost-effectiveness threshold for a life-threatening cancer 56 —a result well in line with earlier Finnish average findings. 41 The health care payer setting, which is recommended in the Finnish guidance for health economic analyses, 39 was used in the modeling. This model includes direct health and social care costs, and excludes income transfers (taxes) and indirect costs (eg, time costs, disability payments, presenteeism, absenteeism, and informal care). A scenario analysis, including productivity losses, 14 was performed to assess the robustness of this direct-costing perspective. A summary of the modeled key research questions is given in Table I as an extended PICO framework, which is used to capture and clarify the essential parts of complicated cost-effectiveness assessment in a sensible order (namely, PICOSTEPS: P, patients; I, interventions; C, comparator; O, outcomes; S, setting; T, time horizon; E, effects; P, perspective; and S, sensitivity analyses). A relatively straightforward, limited cost–benefit analysis (clinical value analysis) approach was recently developed. 46 As a secondary complementary analysis, an impact investment assessment (IIA) was carried out to increase the clinical appeal and interpretation of the primary outcome results. 46 The IIA here covered a fixed drug-related budget based on the most affordable DMT and incremental quality-adjusted survival or time to cane use (EDSS score, 6) versus best supportive care (BSC; trial comparator). The outcome (impact on investment [II]) of the IIA was the duration of benefit obtained in comparison with BSC with the fixed budget. This IIA incorporated an explicit minimal willingness-to-invest (WTI) value for DMT based on the most affordable DMT and, thus, demonstrated the mean absolute cost–benefit in terms of a single unit: II¼Drug health benefitivs BSCðÞ Assumed drugrelated minimal WTIðÞ Drugrelated costi ðÞ (Equation 1) where iindicates a particular drug treatment. Consequently, the result of the IIA is a standardized benefit (II) obtained with the given WTI (in fact, the WTI can be greater than the minimum assumed here, and the benefit increases accordingly). Patients Finland’s MS research registry data were used to define the cohort characteristics in the model. Based on the MS research registry data (713 ambulatory patients from Finland, with MS diagnosed in 1991– 2010 and an EDSS score of 0–6.5 observed at baseline; see Supplemental Material A in the online version at http://dx.doi.org/10.1016/j.clinthera.2017.01.028), the mean age of modeled patients was 35.64 years, and the female/male ratio was 2.57. The distribution of EDSS scores at baseline is shown in Figure 1. Model The clinical course of MS was modeled (Figure 2) 73,74 to capture all relevant evidence, 38,39,43 as no direct comparison is currently available. Models are always hypothetical and contain an element of uncertainty, but when relying on conservative and fair structure and estimates—and keeping the modeling assumptions in mind—they can produce useful information for decision making. In the model shown in Figure 2,patients with RRMS either maintained the same EDSS or transited to another EDSS health state as the disease progressed, developed secondary progressive MS (SPMS), transited to another EDSS state in SPMS, or died (EDSS score, 10; absorbing state) within the 1-year model cycles. Within each cycle, patients experienced a relapse (with/without hospitalization) and/or an adverse event (AE). The relative effects of DMTs were implemented as modifiers of the modeled clinical course of MS. Midcycle estimates (life-table method of half-cycle correction 75–77 ) were used to avoid overor underestimation of modeled outcomes. E. Soini et al. March 2017 539 Disease Progression Disease progression and relapses were modeled independently. Disease progression in terms of the EDSS score development during RRMS was estimated from Finland’s MS research registry data, consisting of 2299 EDSS measurements. The probability of transiting from RRMS to SPMS was estimated, and EDSS development during SPMS was based on results Table I. PICOSTEPS: Summary of the research questions. PICOSTEPS Description P: Patients Finnish adults with incident RRMS and EDSS scores 0.0–6.5 at baseline based on data from a Finnish MS registry I: Interventions DMTs: DMF 240 mg PO BID, teriflunomide 14 mg once daily, GA 20 mg SC once daily, IFNβ1a 44 mg SC TIW, IFN-β1b 250 mg SC EOD, IFN-β1a 30 mgIMQW C: Comparator Common comparator: BSC (trial placebo) O: Outcomes Primary: ICER given as the cost/QALY gained based on the direct cost Secondary: disaggregated and total QALYs (based on EQ-5D-3L) and costs, life-years, years without impaired mobility (EDSS o6; ie, years without cane use), cost-effectiveness plane and efficiency frontiers, cost-effectiveness acceptability frontiers, Bayesian treatment ranking, and cost–benefit assessment. Discounting: 3%/y S: Setting Probabilistic decision analytical modeling (Markov cohort model), including 21 health states reflecting the disease progression (modified by treatment efficacy); and events reflecting relapses, AEs, and withdrawals T: Time horizon 15 years, based on the follow-up data from the Finnish registry, time since diagnosis in a Finnish cost and EQ-5D-3L MS study, 14 years covered by the British Columbia, Canada, registry, 57,58 and approximate time from RRMS to SPMS in the London Ontario MS registry database. For the London Ontario MS registry origins, see Weinshenker et al. 59 E: Effects RRMS progression: Finnish MS registry data (see Supplemental Material A in the online version at http://dx.doi.org/10.1016/j.clinthera.2017.01.028). SPMS progression: London Ontario MS registry (see Supplemental Material A in the online version at http://dx.doi. org/10.1016/j.clinthera.2017.01.028). Relapse rates: published elsewhere. 21,60 Relapseassociated hospitalizations: published elsewhere. 30,32,33 Mortality: Finnish MS registry data and statistics 61 with EDSS-related 17 adjustment. EDSS-associated costs and quality of life: estimated from a Finnish study. 14 Relapse costs: Finnish MS registry data. Relapse disutility: Finnish study 14 accounting for hospitalization status and duration. 23,24 12-wk responses with DMT, annual relapse rates, and withdrawals: mixed-treatment comparison. 62,63 DMT effects on relapses resulting in hospitalizations: published elsewhere. 32,64,65 DMT costs: drugs, 66 monitoring. 67–71 AEs: disutility, 72 duration, costs, and occurrence (see Supplemental Material B in the online version at http://dx.doi.org/ 10.1016/j.clinthera.2017.01.028). P: Perspective Finnish payer perspective. A scenario analysis with a societal perspective. S: Sensitivity analyses 25 deterministic scenarios: impact of modeling assumptions, result robustness, and generalizability Probabilistic sensitivity analysis: joint uncertainty of the input estimates AE ¼adverse event; BSC ¼best supportive care; DMF ¼dimethyl fumarate; DMT ¼disease-modifying therapy; EDSS ¼ Expanded Disability Status Scale; EQ-5D-3L ¼EuroQol Five-Dimensional Questionnaire, Three-Level Version; GA ¼ glatiramer acetate; ICER ¼incremental cost-effective ratio; IFN ¼interferon; MS ¼multiple sclerosis; QALY ¼qualityadjusted life-year; RRMS ¼relapsing–remitting multiple sclerosis; SPMS ¼secondary progressive multiple sclerosis. Clinical Therapeutics 540 Volume 39 Number 3 from the London Ontario registry of MS (see Supplemental Material A in the online version at http:// dx.doi.org/10.1016/j.clinthera.2017.01.028). For the origins of registry, see Weinshenker et al. 59 The relapse rates in patients not receiving DMTs were taken from published references. 21,60 The percentage of relapses leading to hospitalization (30.7%) was estimated from the TEMSO trial. 30,32,33 The annual probability of death was modeled based on Finland’s general population mortality rates by applying the observed MS female/male ratio of 2.57 from Finland’s MS research registry data to Finland’s all-cause ageand sex-specific mortality rates from the year 2014, 61 multiplying the sexweighted general population mortality rate by the EDSS-specific standardized mortality ratio, and converting the result to give the probability. 78 The EDSSspecific standardized mortality ratio was estimated from Finland’s MS research registry results 17 by using linear interpolation: Standardized mortality ratio ¼0:515 EDSSþ1:000 (Equation 2) Treatment Efficacy and Tolerability Treatment efficacy was assessed by common MS study outcomes: sustaining the same disability status for 12 weeks, annualized relapse rate (ARR), and relapses. Persistence was assessed by withdrawal rates, and tolerability, by AEs. Relative rates of hospitalization in the model were derived from the following clinical trials: IFN-β1a SC, CARE MS I (Comparison of Alemtuzumab and Rebif Efficacy in Multiple Sclerosis) 64 (assumed to apply to GA and IFN-β1b SC); IFN-β1a IM, TRANSFORMS (Trial Assessing Injectable Interferon versus FTY720 Oral in Relapsing–Remitting Multiple Sclerosis) 65 ; and teriflunomide, TEMSO 32 (assumed to apply to DMF). Withdrawals were assumed to happen at the initiation of a new model cycle (but not at the start of the first cycle), and patients were assumed to discontinue their current treatment when they progressed from RRMS to SPMS. Disability progression, ARR, and withdrawal rates were modeled based on a mixed-treatment comparison assessed by the National Institute for Health and Care Excellence. 62,63 To account for new MS diagnostics, earlier treatment, and evidence of decreased ARR over time, the base case analysis included trials that enrolled Z80% of patients who had RRMS and had been recruiting patients since 2000. In addition, multiway sensitivity analyses (disability progression, ARR, withdrawal rates) of mixed-treatment comparison without year limit and with or without adjustment for placebo relapses were performed. Treatment safety was modeled using reported AEs from clinical trials or earlier health technology assessments, their costs, and QoL effects (see Supplemental Material B in the online version at http://dx.doi.org/ 10.1016/j.clinthera.2017.01.028). AEs reported with 35 30 25 20 15 10 5 0EDSS 0 EDSS 1 EDSS 2 EDSS 3 EDSS 4 EDSS 5 EDSS 6 Proportion of Patients, % 26.79 33.10 12.06 5.47 2.95 3.51 16.13 Figure 1. Expanded Disability Status Scale (EDSS) score distribution at the initiation of modeling. transitions may happen between EDSS 0-9 and to death SPMS Death 0123456789 0123456789 RRMS transitions may happen between EDSS 0-9, to SPMS and to death Figure 2. Simplified presentation of the Markov model and its key health states. Relapses and adverse events are not depicted. EDSS ¼Expanded Disability Status Scale; RRMS ¼relapsing–remitting multiple sclerosis; SPMS ¼secondary-progressive multiple sclerosis. E. Soini et al. March 2017 541 similar terms were assumed to be treated similarly and to result in similar QoL loss. Quality-adjusted Survival The EuroQol Five-Dimensional Questionnaire, Three-Level Version (EQ-5D-3L) QoL for EDSS scores was modeled on the basis of data from DEFENSE (Burden of Illness in Multiple Sclerosis), 14 a recent cross-sectional survey from Finland. The occurrence and impact 72 of AEs (see Supplemental Material B in the online version at http://dx.doi.org/10.1016/j.clinthera.2017.01.028) and relapses 14,24 were accounted for. Finland’s EDSS-related QoL values 14 were deemed acceptable because the mean EQ-5D-3L score in EDSS 0-1 was in line with values from the general population of Finland. 79 However, the study from Finland 14 did not specify QoL related to relapse with and without hospitalizations. Findings from studies suggest greater disutility for relapse with hospitalization compared with relapses without hospitalization. 23,24 In a US study, the QoL losses in relapsed patients with and without hospitalization were reported as –0.302 and –0.091, respectively. 24 The latter estimate is similar to the Finnish relapse loss, that is, –0.064, 14 which used an extensive 1-year recall period and did not make a distinction between hospitalized and nonhospitalized patients or number of relapses. To approximate the QoL loss associated with hospitalizations, the Finnish QoL loss was weighted with the observed ratio between the QoL losses for hospitalized and nonhospitalized relapses in the US study 24 (ratio –0.302/–0.091 ¼3.3187) to obtain disutility for hospitalized patients in Finland. The applied QoL losses in relapsed patients with and without hospitalization in the model were –0.212 and –0.064, respectively. The QoL effect of relapse was assumed to last for 3 months. 23 Costs Annual DMT cost was calculated using the indicated mean dose of each drug and number of doses per year (365.25 d/y), determined for each treatment regimen based on the product labeling. For drugs with multiple package sizes, the drug costs were estimated by weighting of the package costs by their estimated market share (Table II). A 100% dose intensity and adherence were assumed. Administration, monitoring (Table III), and AE costs (see Supplemental Material B in the online version at http://dx.doi.org/10.1016/j.clinthera.2017. 01.028) were calculated on the basis of resource consumption multiplied by the associated unit costs. DMT-associated resources were based on the product labeling, recommendations in Finland, 1,80,81 publications or earlier assessments (see Supplemental Material B in the online version at http://dx.doi.org/10.1016/ j.clinthera.2017.01.028), and clinical practice. In addition to the EQ-5D-3L QoL scores, which are hard to predict with common regressions, 82,83 the DEFENSE survey 14 assessed the costs of patients with MS in Finland. The EDSS-related direct costs were estimated based on data from the DEFENSE survey 14 and are reported in Table III. Because of limitations in the assessment of DEFENSE-derived relapse costs, the costs of relapses were estimated from other patients with RRMS in Finland (Tampere; N ¼581; data included procedures, hospital visits, hospital stays, and unit cost 70 ) using semilog multivariate methodology explained elsewhere. 47,84 Based on this analysis, the additional costs per relapse with and without hospitalization were €5537.57 and €1297.41, respectively. In a scenario analysis, the relationship between EDSS and annual direct care costs (excluding DMT costs) was estimated based on a nonlinear interpolation of findings reported in a study from Finland, 13 as follows: Annual direct ðDMTs excl:Þcosts ¼€ð128:44 EDSS2þ4266:60 EDSS–2480:10Þ; (Equation3) converted to 2014 real value 71 and with EDSS 0 set to €0. The costs applied in this sensitivity analysis were well in line with those from other MS cost studies from Finland 15 and elsewhere. 9–11 Apart from the drugs, which were valued at January 2016 prices, 66 health care costs were valued at 2013–2014 real prices. The required inflation adjustments were performed using Finland’sofficial price index for communal health care expenditures or income index. 71,85 The modeled costs and health outcomes were discounted at 3%/y. Clinical Therapeutics 542 Volume 39 Number 3 Sensitivity and Generalizability of Results The robustness and generalizability of the base case results were assessed using various deterministic and probabilistic sensitivity analyses (DSA and PSA, respectively). The base case was based on most credible inputs. DSAs were based on 25 different scenarios, including major or noncredible changes in methods, health risks, treatment, costs, QoL, population, and settings. Means based on all 25 DSA scenarios were also calculated. The details of the DSAs are shown in Table IV. Probabilistic Sensitivity Analysis For PSA, a second-order Monte Carlo simulation was used to take into account the joint variation in the economic and clinical outcomes due to sampling uncertainty related to model parameters. The following distributions were used: βfor ARR and withdrawal rates, γfor EDSS-related and treatment costs, log-normal for EDSS transitions, disease progression hazard rates, treatment effect on ARR, treatment effect on hospitalization relapse percentage and QoL, and Dirichlet distribution for the percentage of relapses involving hospitalization (see Supplemental Material C in the online version at http://dx.doi.org/ 10.1016/j.clinthera.2017.01.028). Based on the PSA, cost-effectiveness acceptability frontiers demonstrated optimal treatment to maximize net benefit with different WTP thresholds, and Bayesian treatment ranking ranked the best treatments. RESULTS The average modeled base case results are reported in Table V. The mean projected 15-year total payer’s direct costs differed considerably (by 17.2%) between the most affordable (teriflunomide) and the most costly (IFN-β1b SC) DMT. The respective relative QALY gain difference was 9.3%. The maximum relative QALY difference was 10.6% between the 2 DMTs (DMF and IFN-β1b SC). The modeled key outcome (ICERs €/QALY gained in comparison with BSC alone) ranged considerably, from Table II. Drug-related use and costs. DMT Dose/Amount per Package Cost per Package,€ * Dosage (SPCs) Use, % Cost, € DMF 120 mg † 120 mg, 14 tablets 188.37 120 mg PO BID 1.92 14,435/1st y DMF 240 mg † 240 mg, 56 tablets 1151.56 240 mg PO BID 15.33 240 mg, 168 tablets 3319.33 82.75 DMF 240 mg † 240 mg, 56 tablets 1151.56 240 mg PO BID 15.33 14,523/2nd y 240 mg, 168 tablets 3319.33 84.67 GA 20 mg ‡ 20 mg/mL, 28 1 mL 836.11 20 mg SC once daily 100.00 10,907 IFN-β1a 30 mgIM § 30 mg/0.5 mL, 4 0.5 mL 814.90 30 mg SC QW 100.00 10,630 IFN-β1a 44 mgSC ‖ 44 mg/0.5 mL, 12 0.5 mL 897.83 44 mg SC TIW 100.00 11,712 IFN-β1b 250 mgSC ¶ 250 mg/mL, 15 1 mL 793.08 250 mg SC EOD 100.00 9656 Teriflunomide 14 mg # 14 mg, 28 tablets 1017.89 14 mg PO once daily 15.33 12,023 14 mg, 84 tablets 2712.79 84.67 DMT ¼disease-modifying therapy; DMF ¼dimethyl fumarate; GA ¼glatiramer acetate; IFN ¼interferon; SPC ¼summary of product characteristics. * Drug costs are at January 2016 values. † Trademark: Tecfidera s (Biogen, Weston, Massachusetts). ‡ Trademark: Copaxone s (Teva, Ulm, Germany). § Trademark: Avonex s (Biogen). ‖ Trademark: Rebif s (EMD Serono, Rockland, Massachusetts). ¶ Trademark: Betaferon s (Bayer Pharmaceuticals, West Haven, Connecticut). # Trademark: Aubagio s (Genzyme [a SanofiCompany], Cambridge, Massachusetts). E. Soini et al. March 2017 543 24,081 (teriflunomide) to 248,652 (GA) per QALY gained, and BSC dominated IFN-β1b SC in the base case. Teriflunomide was estimated to be less costly and more effective (dominant) than injectable first-line DMTs, and DMF had a high ICER of 75,431 versus teriflunomide, resulting from the marginally more QALYs (0.089) with DMF and higher costs versus teriflunomide over 15 years. (Table V and Figure 3). If the WTP threshold for additional QALY gained is set to the most plausible level (€25,000), only teriflunomide represents a cost-effective alternative to BSC alone, based on the modeling. If the WTP is between €37,000 (plausible) and €55,000 (end of life) per QALY gained, only teriflunomide and DMF represent cost-effective alternatives to BSC alone. However, with a modeled ICER of 75,414 for DMF versus teriflunomide, DMF is unlikely to be considered cost-effective in the Finnish setting given the unofficial assumed WTP thresholds detailed in Materials and Methods. The cost–benefit analysis type IIA utilized the minimal mean expected DMT-related discounted Table III. Monitoring and disability (EDSS)-related resource use and costs. Monitoring Unit Cost, € * Resources, First Year/Later Year † DMF GA IFNs Teriflunomide BSC Specialist visit 340.76, Including 5% copayment 69 2/1 2/1 2/1 2/1 0/0 SC training 50.97 Nurse visit 69 0/0 1/0 1/0 0/0 0/0 Laboratory fee ‡ 5.47 68 4/4 0/0 4/1 17/6 0/0 ALT 1.00 67 4/1 0/0 4/1 17/6 0/0 GGT, creatinine 2.00 67 4/1 0/0 0/0 0/0 0/0 BC 1.55 67 0/0 0/0 4/1 0/0 0/0 FBC 6.60 67 4/4 0/0 0/0 4/1 0/0 MxA 92.50 70 0/0 0/0 1/1 0/0 0/0 TSH 2.50 67 0/0 0/0 1/0 0/0 0/0 UT 5.84 68 4/1 0/0 0/0 0/0 0/0 MRI, head 335.58 69 1/0.5 1/0.5 1/0.5 1/0.5 0/0 Phone call § 9.56 After tests 69 2/3 0/0 2/0 15/5 0/0 Disability related EDSS score 14 –0/1 2/3 4/5 6/7 8–9 Direct health care costs, € ‖ –1108/1446 2890/3470 3909/5656 7919/12,185 15,718 Direct non–health care costs, € –49/834 1693/4526 5767/15,289 18,749/32,364 68,852 ALT ¼alanine aminotransferase; BC ¼blood count; BSC ¼best supportive care; DMF ¼dimethyl fumarate; EDSS ¼ Expanded Disability Status Scale; FBC ¼full blood count; GA ¼glatiramer acetate; GGT ¼gamma-glutamyl transferase; MRI ¼magnetic resonance imaging; MxA ¼protein induced by interferon-alfa/β;TSH¼thyroid-stimulating hormone; UT ¼urine test. * Pre-2013 nontariff costs 14,68,69 were indexed to the 2014 price level using official communal price index for health care services. 71 † Unless otherwise noted. ‡ Fixed laboratory fee for each test taking time. § Phone call after laboratory tests if specialist visit not arranged. ‖ Estimated costs of disease-modifying therapies (DMTs) were excluded based on the digitalization and estimation of DMT costs in Figure 4 in Ruutiainen et al. 14 Clinical Therapeutics 544 Volume 39 Number 3 budget per patient (minimum WTI) of €42,077 based on the drug-related costs of IFN-β1a SC. The consequent discounted IIs in terms of incremental quality-adjusted survivals versus BSC were: teriflunomide, 0.337 QALYs gained; DMF, 0.314; IFN-β1a SC, 0.264; GA, 0.120; IFN-β1a IM, 0.119; and IFN-β1b SC, –0.239, all with the assumed WTI. The respective incremental time to cane uses were Table IV. Details of deterministic sensitivity analyses. Category Scenario Discounting No discounting Discounting with 5%/y Health risks British Columbia, Canada, RRMS EDSS development, based on patients more than 28 years old 57 Alternative natural relapse source 86 Rate for relapses leading to hospitalization based on the 1:2.75 ratio from Tampere data (26.7% of annual relapses result in hospitalization when adjusting for covariates including also EDSS score; N ¼581; mean age at relapse, 40 y) Relapse time, 2 mo Relapse time, 4 mo Treatment DMT discontinuation when EDSS 7 and over was reached, based on reimbursement criteria Disability progression and ARR set to the lower 95% credibility interval threshold of MTC results Disability progression and ARR set to the higher 95% credibility interval threshold of MTC results Alternative source disability progression, ARR, and withdrawal rates from the MTC: no year limit and adjustment for placebo relapses Alternative source disability progression, ARR, and withdrawal rates from the MTC: no year limit Time with AEs doubled (same as doubling AE disutility for those AEs that last a shorter time than the model cycle) Time with AEs halved (same as halving AE disutility) Costs EDSS costs based on the other Finnish source 13 at 2014 values 61 Monitoring costs doubled Monitoring costs halved Relapse cost doubled Relapse cost halved AE costs doubled AE costs halved Societal approach (productivity loss included) 14,85 QoL Alternative EDSS QoL source 10 Similar QoL loss assumed for all relapses 14 Result generalizability TEMSO 30,32,33 patient characteristics and placebo transition probabilities for RRMS EDSS AE =adverse event; ARR =annualized relapse rate; EDSS =Expanded Disability Status Scale; MTC =mixed-treatment comparison; QoL =quality of life; RRMS =relapsing–remitting multiple sclerosis; TEMSO =Randomized Trial of Oral Teriflunomide for Relapsing Multiple Sclerosis) Oral Teriflunomide for Patients with Relapsing Multiple. E. Soini et al. March 2017 545 SUPPLEMENTAL MATERIAL Supplemental material accompanying this article can be found in the online version at http://dx.doi.org/ 10.1016/j.clinthera.2017.01.028. REFERENCES 1. Finnish Medical Society Duodecim, Finnish Neurological Society Working Group. MS Disease. Current Care Criteria. Helsinki, Finland: Current Care; 2015. 2. Pugliatti M, Rosati G, Carton H, et al. The epidemiology of multiple sclerosis in Europe. European J Neurology.2006; 13:700–722. 3. World Health Organization. Atlas multiple sclerosis resources in the world 2008. Geneva, Switzerland: World Health Organization; 2008:1–56. 4. Holmberg M, Murtonen A, Elovaara I, Sumelahti ML. Increased female MS incidence and differences in gender-specific risk in mediumand high-risk regions in Finland from 1981-2010. Multiple Sclerosis International. 2013;2013:182516. 5. Sumelahti ML, Tienari PJ, Hakama M, Wikstrom J. Multiple sclerosis in Finland: incidence trends and differences in relapsing remitting and primary progressive disease courses. J Neurol Neurosurg Psychiatry. 2003; 74:25–28. 6. Sumelahti ML, Tienari PJ, Wikstrom J, et al. Regional and temporal variation in the incidence of multiple sclerosis in Finland 1979-1993. Neuroepidemiology. 2000; 19:67–75. 7. Sumelahti ML, Tienari PJ, Wikstrom J, et al. Increasing prevalence of multiple sclerosis in Finland. Acta Neurol Scand. 2001;103:153–158. 8. Brunet DG, Hopman WM, Singer MA, et al. Measurement of health-related quality of life in multiple sclerosis patients. The Canadian J Neurological Sciences. 1996;23:99– 103. 9. Karampampa K, Gustavsson A, van Munster ET, et al. Treatment experience, burden, and unmet needs (TRIBUNE) in Multiple Sclerosis study: the costs and utilities of MS patients in The Netherlands. J Medical Economics. 2013;16:939–950. 10. Kobelt G, Berg J, Atherly D, Hadjimichael O. Costs and quality of life in multiple sclerosis: a cross-sectional study in the United States. Neurology. 2006;66: 1696–1702. 11. Kobelt G, Berg J, Lindgren P, et al. Modeling the costeffectiveness of a new treatment for MS (natalizumab) compared with current standard practice in Sweden. Mult Scler. 2008;14:679–690. 12. Kobelt G, Kasteng F. Access to Innovative Treatments in Multiple Sclerosis in Europe. Brussels, Belgium: European Federation of Pharmaceutical Industries and Associations; 2009:1–89. 13. Martikainen J, Sintonen H. Treatment costs and quality of life losses. In: Elovaara I, Pirttilä T, Färkkilä M, Hietaharju A, editors. Clinical Neuroimmunology. Helsinki, Finland: Helsinki University Press Publishing; 2006:184–187. 14.Ruutiainen J, Viita AM, Hahl J, et al. Burden of illness in multiple sclerosis (DEFENSE) study: the costs and quality-of-life of Finnish patients with multiple sclerosis. J Medical Economics. 2016;19:21–33. 15. Sillanpaa M, Andlin-Sobocki P, Lonnqvist J. Costs of brain disorders in Finland. Acta Neurol Scand. 2008;117: 167–172. 16. Valjakka S, Nurmi-Koikkalainen P, Anttila H, Konttinen J. Living of individuals with neurological long-term illness and impairment: Helsinki: Aspa Foundation sr;2013. 17. Sumelahti ML, Hakama M, Elovaara I, Pukkala E. Causes of death among patients with multiple sclerosis. Mult Scler. 2010;16:1437–1442. 18. Kurtzke JF. Rating neurologic impairment in multiple sclerosis: an expanded disability status scale (EDSS). Neurology. 1983;33:1444–1452. 19. Kingwell E, van der Kop M, Zhao Y, et al. Relative mortality and survival in multiple sclerosis: findings from British Columbia, Canada. J Neurol Neurosurg Psychiatry. 2012;83:61–66. 20. Pokorski RJ. Long-term survival experience of patients with multiple sclerosis. J Insur Med. 1997;29:101–106. 21. Biogen Idec. Natalizumab (Tysabri) for the treatment of adults with highly active relapsing remitting multiple sclerosis: Single technology appraisal (STA) submission to the National Institute for Health and Clinical Excellence. 2006:1–269. 22. Naci H, Fleurence R, Birt J, Duhig A. Economic burden of multiple sclerosis: a systematic review of the literature. PharmacoEconomics. 2010;28:363–379. 23. Orme M, Kerrigan J, Tyas D, et al. The effect of disease, functional status, and relapses on the utility of people with multiple sclerosis in the UK. Value Health. 2007;10:54–60. 24. Prosser LA, Kuntz KM, Bar-Or A, Weinstein MC. Costeffectiveness of interferon beta-1a, interferon beta-1b, and glatiramer acetate in newly diagnosed non-primary progressive multiple sclerosis. Value Health. 2004;7:554– 568. 25. Kelasto. Number of individuals with reimbursements and prescriptions. Helsinki: Kela; 2015. 26. Fox RJ, Miller DH, Phillips JT, et al. Placebo-controlled phase 3 study of oral BG-12 or glatiramer in multiple sclerosis. N Engl J Med. 2012;367:1087–1097. 27. Hutchinson M, Fox RJ, Miller DH, et al. Clinical efficacy of BG-12 (dimethyl fumarate) in patients with relapsingremitting multiple sclerosis: subgroup analyses of the CONFIRM study. J Neurol. 2013;260:2286–2296. Clinical Therapeutics 552 Volume 39 Number 3 28. Bar-Or A, Gold R, Kappos L, et al. Clinical efficacy of BG-12 (dimethyl fumarate) in patients with relapsing-remitting multiple sclerosis: subgroup analyses of the DEFINE study. J Neurol.2013; 260:2297–2305. 29. Gold R, Kappos L, Arnold DL, et al. Placebo-controlled phase 3 study of oral BG-12 for relapsing multiple sclerosis. N Engl J Med. 2012;367:1098–1107. 30. Miller AE, O’Connor P, Wolinsky JS, et al. Pre-specified subgroup analyses of a placebo-controlled phase III trial (TEMSO) of oral teriflunomide in relapsing multiple sclerosis. Mult Scler. 2012;18: 1625–1632. 31. O’Connor P, Comi G, Freedman MS, et al. Long-term safety and efficacy of teriflunomide: Nineyear follow-up of the randomized TEMSO study. Neurology.2016; 86:920–930. 32. O’Connor P, Wolinsky JS, Confavreux C, et al. Randomized trial of oral teriflunomide for relapsing multiple sclerosis. N Engl J Med. 2011;365:1293–1303. 33. O’Connor PW, Lublin FD, Wolinsky JS, et al. Teriflunomide reduces relapse-related neurological sequelae, hospitalizations and steroid use. JNeurol. 2013;260: 2472–2480. 34. Confavreux C, O’Connor P, Comi G, et al. Oral teriflunomide for patients with relapsing multiple sclerosis (TOWER): a randomised, double-blind, placebocontrolled, phase 3 trial. The Lancet. Neurology. 2014;13:247– 256. 35. Miller AE, Macdonell R, Comi G, et al. Teriflunomide reduces relapses with sequelae and relapses leading to hospitalizations: results from the TOWER study. J Neurol. 2014;261:1781–1788. 36. Miller AE, Wolinsky JS, Kappos L, et al. Oral teriflunomide for patients with a first clinical episode suggestive of multiple sclerosis (TOPIC): a randomised, doubleblind, placebo-controlled, phase 3trial.The Lancet. Neurology.2014; 13:977–986. 37. Vermersch P, Czlonkowska A, Grimaldi LM, et al. Teriflunomide versus subcutaneous interferon beta-1a in patients with relapsing multiple sclerosis: a randomised, controlled phase 3 trial. Mult Scler. 2014;20:705–716. 38. Briggs A, Claxton C, Sculpher M. Decision modelling for health economic evaluation. Oxford, UK: Oxford University Press; 2008. 39. Pharmaceuticals Pricing Board. Composing health economic evaluation for the pricing and reimbursement application. Helsinki: Pharmaceuticals Pricing Board.2015. 40. Soini E, Hallinen T, Sokka AL, Saarinen K. Cost-utility of firstline actinic keratosis treatments in Finland. Adv Ther. 2015;32: 455–476. 41. Soini E, Kukkonen J, Myllykangas M, Ryynänen OP. Contingent valuation of eight new treatments: what is the clinician’s and politician’s willingness to pay? The Open Complementary Medicine Journal. 2012;4:1–11. 42. Collins M, Latimer N. NICE’s end of life decision making scheme: impact on population health. BMJ (Clinical research ed.).2013: f1363. 43. National Institute for Health and Care Excellence (NICE). Guide to the Methods of Technology Appraisal 2013. London, UK: NICE; 2013. 44. Hallinen T, Soini E. The impact of the pharmaceutical pricing system on cost-effectiveness results: Finnish analysis. The Open Pharmacoeconomics & Health Economics Journal. 2011;3:6–10. 45. Hallinen TA, Soini EJ, Eklund K, Puolakka K. Cost-utility of different treatment strategies after the failure of tumour necrosis factor inhibitor in rheumatoid arthritis in the Finnish setting. Rheumatology (Oxford). 2010;49:767–777. 46. Soini E, Hautala A, Poikonen E, et al. Cost-effectiveness of firstline chronic lymphocytic leukemia treatments when full-dose fludarabine is unsuitable. Clin Ther. 2016;38:889–904. 47. Soini E, Leussu M, Hallinen T. Administration costs of intravenous biologic drugs for rheumatoid arthritis. SpringerPlus.2013; 2:531. 48. Soini E, Martikainen JA, Nousiainen T. Treatment of follicular non-Hodgkin’s lymphoma with or without rituximab: costeffectiveness and value of information based on a 5-year followup. Annals of oncology: official journal of the European Society for Medical Oncology. 2011;22:1189–1197. 49. Martikainen J, Hallinen T, Soini E. Economic evaluation of medicines. Pharmaceutical economics and management: theory, research and practice [in Finnish]. Dosis farmaseuttinen aikakauskirja. 2006;22:289–300. 50. Soini E, Hallinen T, Brignone M, et al. Cost-utility analysis of vortioxetine versus agomelatine, bupropion SR, sertraline and venlafaxine XR after treatment switch in major depressive disorder in Finland. Expert Review PharMacoeconomics Outcomes Research. 2016:1–10. 51. Peura P, Martikainen J, Soini E, et al. Cost-effectiveness of statins in the prevention of coronary heart disease events in middleaged Finnish men. Curr Med Res Opin. 2008;24:1823–1832. 52. Hallinen T, Soini E, Asseburg C, et al. Cost-effectiveness of insulin glargine compared to other long-acting basal insulins in the treatment of Finnish type 1 and type 2 diabetes patients based on individual studies. Dosis E. Soini et al. March 2017 553 farmaseuttinen aikakauskirja.2012; 28:145–164. 53. Soini E. Economic evaluation of rare disorders: what support for decisions should be produced [in Finnish]? Harvinaiset Sairaudet. 2011;1:26–31. 54. Soini E. Cost-effectiveness of arthritis care in Finland [in Finnish]. BestPractice Reumasairaudet —La ¨a ¨ketieteen Asiantuntijoiden Ammattilehti. 2013;1:7–9. 55. Soini EJ, Davies G, Martikainen JA, et al. Population-based health-economic evaluation of the secondary prevention of coronary heart disease in Finland. Curr Med Res Opin. 2010;26:25– 36. 56. Fimea. Medicines HTA advisory board meeting memo. Helsinki: Fimea. 2014:1–3. 57. Palace J, Bregenzer T, Tremlett H, et al. UK multiple sclerosis risksharing scheme: a new natural history dataset and an improved Markov model. BMJ open.2014; 4:e004073. 58. Palace J, Duddy M, Bregenzer T, et al. Effectiveness and costeffectiveness of interferon beta and glatiramer acetate in the UK Multiple Sclerosis Risk Sharing Scheme at 6 years: a clinical cohort study with natural history comparator. The Lancet. Neurology. 2015;14:497–505. 59. Weinshenker BG, Bass B, Rice GP, et al. The natural history of multiple sclerosis: a geographically based study. I. Clinical course and disability. Brain: a Journal of Neurology. 1989;112: 133–146. 60. Patzold U, Pocklington PR. Course of multiple sclerosis: first results of a prospective study carried out of 102 MS patients from 1976-1980. Acta Neurol Scand. 1982;65:248–266. 61. Statistics Finland. Deaths: Statistics Finland [in Finnish]. Helsinki: Statistics Finland; 2016. 62. Cutter G, Fahrbach K, Huelin R, et al. Relative efficacy of teriflunomide 14 mg in relapsing MS: a mixed-treatment comparison. Orlando: Consortium of Multiple Sclerosis Centers - Americas Committee for Treatment and Research in Multiple Sclerosis; 2013. 63. National Institute for Health and Care Excellence (NICE). Final Appraisal Determination—Teriflunomide for Treating RelapsingRemitting Multiple Sclerosis. London, UK: NICE; 2013:1–62. 64. Cohen JA, Coles AJ, Arnold DL, et al. Alemtuzumab versus interferon beta 1a as first-line treatment for patients with relapsingremitting multiple sclerosis: a randomised controlled phase 3 trial. Lancet (London, England). 2012;380:1819–1828. 65. Haas J, Hartung HP, von Rosenstiel P, et al. Fingolimod reduces the number of severe relapses in patients with relapsing multiple sclerosis: results from phase III TRANSFORMS and FREEDOMS studies. Lisbon, Portugal: Presented at: European Neurological Society; 28-31 May 2011:28–31. Poster 902. 66. Pharmaceuticals Pricing Board. Reimbursable of Authorized Medicinal Products [in Finnish]. Helsinki: Pharmaceuticals Pricing Board; 2016. 67. FimLab. Pricelist for Laboratory Tests. Tampere: FimLab; 2013. 68. Hujanen T, Kapiainen S, Tuominen U, Pekurinen M. Health care unit costs in year 2006 in Finland. Helsinki: National Institute for Health and Welfare: STAKES Working Papers; 2008. 69. Kapiainen S, Väisänen A, Haula T. Health and Social Care Costs in Year 2011 in Finland. Helsinki, Finland: National Institute for Health and Welfare; 2014. 70. Hospital Tampere University. Services and Prices, 2013. Tampere, Finland: Tampere University Hospital; 2013. 71. Statistics Finland. Price Index for Public Expenditures. Helsinki: Statistics Finland; 2016. 72. Soini E, Hallinen T. Cost-utility of agomelatine, venlafaxine and placebo in the treatment of major depressive disorder (MDD) in Finland—economic modelling study using representative population data. Value Health. 2009; 12:A359. 73. Chilcott J, McCabe C, Tappenden P, et al. Modelling the cost effectiveness of interferon beta and glatiramer acetate in the management of multiple sclerosis. Commentary: evaluating disease modifying treatments in multiple sclerosis. BMJ (Clinical research ed.). 2003:522. discussion 522. 74. Gani R, Giovannoni G, Bates D, et al. Cost-effectiveness analyses of natalizumab (Tysabri) compared with other diseasemodifying therapies for people with highly active relapsingremitting multiple sclerosis in the UK. PharmacoEconomics. 2008; 26:617–627. 75.National Institute for Health and Care Excellence (NICE). Dimethyl fumarate for treating relapsingremitting multiple sclerosis. NICE technology appraisal guidance 320. London, UK: NICE; 2014. 76. Barendregt JJ. The life table method of half cycle correction: getting it right. Medical Decision Making: an International Journal of the Society for Medical Decision Making. 2014;34:283–285. 77. Naimark DM, Kabboul NN, Krahn MD. The half-cycle correction revisited: redemption of a kludge. Medical decision making: an international journal of the Society for Medical Decision Making.2013; 33:961–970. 78. Fleurence RL, Hollenbeak CS. Rates and probabilities in Clinical Therapeutics 554 Volume 39 Number 3 economic modelling: transformation, translation and appropriate application. PharmacoEconomics. 2007;25:3–6. 79. Saarni SI, Harkanen T, Sintonen H, et al. The impact of 29 chronic conditions on healthrelated quality of life: a general population survey in Finland using 15D and EQ-5D. Quality of Life Research: An International Journal of Quality of Life Aspects of Treatment, Care and Rehabilitation. 2006;15:1403–1414. 80. Elovaara I, Soilu-Hanninen M, Kuusisto H, et al. Magnetic resonance imaging of the brain in the monitoring of immune therapy of multiple sclerosis. Duodecim. 2015;131:1571–1580. 81. Remes A, Airas L, Atula S, et al. Update on current care guideline: multiple sclerosis. Duodecim. 2015;131:500–501. 82. Soini E, Koskela T, Ryynänen OP. International normalized ratio (INR) monitoring and percent time in therapeutic INR range (TTR) have impact on patient’s quality of life? Application of beta regressions in a prospective 3 months setting. Value Health. 2013;16:A535. 83. Soini E, Mäkirinne-Kallio N, Martikainen J, et al. Estimation of quality-adjusted survival with mild obstructive sleep apnoea and six quality of life (QoL) assessments: comparison between trapezoid rule, cross-sectional and mixed model estimates. Oslo, Norway: Presented at: 14th Biennial Society for Medical Decision Making (SMDM) European Meeting. June 10-12, 2012. Abstract 158. 84. Hallinen T, Martikainen JA, Soini EJ, et al. Direct costs of warfarin treatment among patients with atrial fibrillation in a Finnish health care setting. Curr Med Res Opin. 2006;22:683– 692. 85. Statistics Finland. Income index. Helsinki: Statistics Finland; 2016. 86. Held U, Heigenhauser L, Shang C, et al. Predictors of relapse rate in MS clinical trials. Neurology. 2005;65:1769–1773. 87. Bell C, Graham J, Earnshaw S, et al. Cost-effectiveness of four immunomodulatory therapies for relapsing-remitting multiple sclerosis: a Markov model based on long-term clinical data. Journal of Managed Care Pharmacy. 2007;13:245–261. 88. Dembek C, White LA, Quach J, et al. Cost-effectiveness of injectable disease-modifying therapies for the treatment of relapsing forms of multiple sclerosis in Spain. The European Journal of Health Economics: Health Economics in Prevention and Care. 2014;15: 353–362. 89. Imani A, Golestani M. Cost-utility analysis of disease-modifying drugs in relapsing-remitting multiple sclerosis in Iran. Iranian J Neurology. 2012;11:87–90. 90. Jankovic SM, Kostic M, Radosavljevic M, et al. Costeffectiveness of four immunomodulatory therapies for relapsingremitting multiple sclerosis: a Markov model based on data a Balkan country in socioeconomic transition. Vojnosanitetski Pregled. 2009;66:556–562. 91. Parkin D, Jacoby A, McNamee P, et al. Treatment of multiple sclerosis with interferon beta: an appraisal of cost-effectiveness and quality of life. J Neurol Neurosurg Psychiatry. 2000;68: 144–149. 92. Comi G, Freedman MS, Kappos L, et al. Pooled safety and tolerability data from four placebo-controlled teriflunomide studies and extensions. Multiple sclerosis and related disorders. 2016;5:97–104. 93. CoyleP,KhatriB,EdwardsK, et al. Teriflunomide real-world safety profile: results of the phase 4 Teri-PRO study. Presented at: 32nd Congress of the European Committee for Treatment and Research in Multiple Sclerosis. London, September 14-17, 2016. Poster P648. 94. Gold R, Khatri B, Edwards K, et al. Impact of teriflunomide treatment on real-world quality of life in the phase 4 Teri-PRO study. Presented at: 32nd Congress of the European Committee for Treatment and Research in Multiple Sclerosis. London, September 14-17, 2016. Poster P647. 95. Gold R, Khatri B, Edwards K, et al. Stable Disability and patient-reported performance outcomes over 48 weeks of teriflunomide treatment: results from the phase 4 Teri-PRO study. Presented at: 32nd Congress of the European Committee for Treatment and Research in Multiple Sclerosis. London, September 14-17, 2016. Poster P646. 96. Moisset X, Zuel M, Conde S, et al. Teriflunomide and dimethylfumarate: prescription and patients’satisfaction in a real life setting. Presented at: 32nd Congress of the European Committee for Treatment and Research in Multiple Sclerosis. London, September 14-17, 2016. Poster P650. 97. Guger M, Enzinger C, Leutmezer F, et al. Real Life Use of natalizumab, fingolimod, dimethylfumarate, teriflunomide and alemtuzumab in Austria: benefit-risk data from the Austrian Multiple Sclerosis Treatment Registry. Presented at: 32nd Congress of the European Committee for Treatment and Research in Multiple Sclerosis. London, September 14-17, 2016. Poster EP1510. 98. Annovazzi P, Mallucci G, Lo Re M, et al. Real life efficacy and tolerability of teriflunomide: a multicentre study. Presented at: 32nd Congress of the European Committee for Treatment and Research in Multiple Sclerosis. E. Soini et al. March 2017 555 London, September 14-17, 2016. Poster P616. 99. Di Tommaso V, Mancinelli L, Di Ioia M, et al. Efficacy and safety of first line oral therapies in relapsing-remitting multiple sclerosis: dimethylfumarate vs teriflunomide in the Chieti experience. Presented at: 32nd Congress of the European Committee for Treatment and Research in Multiple Sclerosis. London, September 14-17, 2016. Poster P1648. 100. Alnajashi H, Alshamrani F, Bakdache F, Freedman M. Differences in tolerability and discontinuation rates in teriflunomide-treated patients: a real world clinic experience. Presented at: 32nd Congress of the European Committee for Treatment and Research in Multiple Sclerosis. London, September 14-17, 2016. Poster P658. 101. Cohan SL, Edwards K, Chen C, et al. Rebound disease activity reduction in relapsing multiple sclerosis patients transitioned from natalizumab to teriflunomide. Presented at: 32nd Congress of the European Committee for Treatment and Research in Multiple Sclerosis. London, September 14-17, 2016. Poster P655. 102. Edwards K, O’Connor J, Siuta J. Evaluation of switching to teriflunomide in high risk natalizumab patients. Presented at: 32nd Congress of the European Committee for Treatment and Research in Multiple Sclerosis. London, September 14-17, 2016. Poster P706. 103. European Medicines Agency. Dimethyl fumarate [summary of product characteristics]. 2014. 104. US Food and Drug Administration. Dimethyl fumarate [prescribing information]. 2016. 105. Soini E, Holmberg M, Asseberg C, Sumelahti ML. Modelling the persistence of disease-modifying drug treatment (DMT) and its independent drivers in Finnish multiple sclerosis (MS) patients: parametric survival modelling. Value Health. 2014;17:A400. 106. Walther EU, Hohlfeld R. Multiple sclerosis: side effects of interferon beta therapy and their management. Neurology. 1999;53: 1622–1627. 107. Nikfar S, Kebriaeezadeh A, Dinarvand R, et al. Cost-effectiveness of different interferon beta products for relapsing-remitting and secondary progressive multiple sclerosis: decision analysis based on long-term clinical data and switchable treatments. Daru: journal of Faculty of Pharmacy, Tehran University of Medical Sciences.2013; 21:50. 108. Rubio-Terrés C, Arístegui Ruiz I, Medina Redondo F, Izquierdo Ayuso G. Cost-utility analysis of multiple sclerosis treatment with glatiramer acetate or interferon beta in Spain. Farm Hosp. 2003; 27:159–165. 109. Rubio-Terres C, Dominguez-Gil Hurle A. Cost-utility analysis of relapsing-remitting multiple sclerosis treatment with azathioprine or interferon beta in Spain. Rev Neurol. 2005;40:705–710. 110. Sanchez-de la Rosa R, Sabater E, Casado MA, Arroyo R. Costeffectiveness analysis of disease modifying drugs (interferons and glatiramer acetate) as first line treatments in remitting-relapsing multiple sclerosis patients. Journal of Medical Economics. 2012;15:424– 433. 111. Zhang X, Hay JW, Niu X. Cost effectiveness of fingolimod, teriflunomide, dimethyl fumarate and intramuscular interferon-beta1a in relapsing-remitting multiple sclerosis. CNS Drugs.2015;29:71–81. 112. Häkkinen U, Iversen T, Peltola M, et al. Health care performance comparison using a disease-based approach: the EuroHOPE project. Health Policy.2013;112:100–109. 113. Linna M, Hakkinen U, Peltola M, et al. Measuring cost efficiency in the Nordic hospitals—a crosssectional comparison of public hospitals in 2002. Health Care Manag Sci. 2010;13:346–357. 114. Medin E, Hakkinen U, Linna M, et al. International hospital productivity comparison: experiences from the Nordic countries. Health Policy. 2013;112:80–87. 115. Nyblin K. In which direction is the reimbursement system of pharmaceuticals going to evolve? Defensor legis. 2009;6:915. 116. Ministry of Social Affairs and Health, Ministry of Finance. Healthcare, social welfare and regional government reform package. Helsinki, Finland: Ministry of Social Affairs and Health, Ministry of Finance; 2016. 117. Soini E. Expert solutions in outcomes research. Legal Working Group Workshop, Secondary Use of Data: Information Management, Control and Monitoring [in Finnish]. Helsinki, Finland: Ministry of Social Affairs and Health; 2016. 118. Pelkonen L. Conditional Reimbursement [in Finnish]. Helsinki: Finnish Pharmaceuticals Pricing Board. 2016. 119. Riksdag Parliament. Government proposal 184/2016 vp [in Swedish]. 2016. 120. Belgian Health Care Knowledge Centre. Belgian Guidelines for Economic Evaluations and Budget Impact Analyses. Brussels: Belgian Health Care Knowledge Centre; 2012. 121. College of Health Economists. French guidelines for the economic evaluation of health care technologies [in French]. Paris: College of Health Economists; 2004:1–90. 122. Health Care Insurance Board. Guidelines for pharmacoeconomic research, updated version [in Dutch]. Diemen, The Netherlands: Health Care Insurance Board; 2006:1–14. Clinical Therapeutics 556 Volume 39 Number 3 123. Health Information and Quality Authority. Guidelines for the economic evaluation of health technologies in Ireland. Cork, Ireland: Health Information and Quality Authority; 2010:1–75. 124. Institute for Pharmaceutical Research. Guidelines on Health Economic Evaluation [in German]. Vienna, Austria: Institute for Pharmaceutical Research; 2006:1–11. 125. Baltic guideline for economic evaluation of pharmaceuticals (pharmacoeconomic analysis). Latvia, Estonia and Lithuania: 8 August 2002. International Society for Pharmacoeconomics and Outcomes Research; 2002:1–6. 126. López-Bastida J, Oliva J, Antonanzas F, et al. Spanish recommendations on economic evaluation of health technologies. The European Journal of Health Economics: Health Economics in Prevention and Care. 2010;11:513– 520. 127. Norwegian Medicines Agency. Guidelines on How To Conduct Pharmacoeconomic Analyses [in Norwegian]. Oslo: Norwegian Medicines Agency; 2012:1–27. 128. Scottish Medicines Consortium. Guidance to Manufacturers for Completion of New Product Assessment Form. Glasgow: Scottish Medicines Consortium; 2016. 129. Soini E, Hallinen T, Kauppi M, et al. Comprehensive health economic assessment of sequenced treatment with biologics in moderate-to-severe rheumatoid arthritis: analysis based on ACR50 and ACR70 responses. Arthritis Rheum. 2010;62:759. 130. Soini E, Martikainen JA, Vihervaara V, et al. Economic evaluation of sequential treatments for follicular non-Hodgkin lymphoma. Clin Ther. 2012;34:e912. 131. Shirani A, Zhao Y, Kingwell E, et al. Temporal trends of disability progression in multiple sclerosis: findings from British Columbia, Canada (1975-2009). Mult Scler. 2012;18:442–450. 132. Tremlett H, Zhao Y, Devonshire V. Natural history comparisons of primary and secondary progressive multiple sclerosis reveals differences and similarities. J Neurol. 2009;256:374–381. 133. Confavreux C, Vukusic S, Adeleine P. Early clinical predictors and progression of irreversible disability in multiple sclerosis: an amnesic process. Brain: A Journal of Neurology. 2003;126:770–782. 134. Leray E, Yaouanq J, Le Page E, et al. Evidence for a two-stage disability progression in multiple sclerosis. Brain: a Journal of Neurology. 2010;133:1900–1913. 135. Minden SL, Frankel D, Hadden L, et al. The Sonya Slifka longitudinal multiple sclerosis study: methods and sample characteristics. Mult Scler. 2006;12:24–38. Address correspondence to: Mr. Erkki Soini, ESiOR Ltd, Tulliportinkatu 2 LT4, Kuopio 70100, Finland. E-mail: erkki.soini@esior.fi E. Soini et al. March 2017 557 SUPPLEMENTARY MATERIAL Supplement A. EDSS-based RRMS and SPMS transition matrices EDSS 1 is the key outcome in the assessment of MS disability progression. In the Finnish PirkanmaaSeinäjoki-Vaasa MS registry, there were 1359 patients with MS with EDSS assessment data available, with altogether 2458 measurements. These patients were identified from administrative registries. The data collection, case ascertainment procedure, and ethical permits have been described in detail elsewhere. 2,3 Incident MS cases diagnosed in the study region that fulfilled the McDonald 4 criteria were included. The classification of disease course to RRMS was performed using standardized definitions. 5 A total of 1242 patients had RRMS, and these patients with RRMS had altogether 2299 EDSS measurements between August 27, 1986, and December 31, 2010. Women accounted for 69.8% of the patients. In all, 62.2% of the EDSS assessments were carried out at the beginning of a DMT episode with an EDSS score of 0–7. In 2010, EDSS values were assessed for all patients alive (July 1, 2010, assumed, if no specific day shown in the data). EDSS Transitions in RRMS Figure A.1 shows all EDSS measurements over time for descriptive purposes. As can be seen, most EDSS measurements were performed for patients with RRMS (green colored dots). The figure also shows that there was censoring in the EDSS measurements in EDSS classes 6.5–9.5. For descriptive purposes, combined Figure A.2 shows the development from one EDSS measurement to the next among patients with RRMS, conditional on particular EDSS scores. EDSS development over time needs to be modeled in order to estimate the progression of MS. MS progression for the model was estimated using integer RRMS EDSS scores (halves rounded up; 9.5 assumed to be 9.0 because the patient is alive when EDSS is 9.5). The JAGS software V3.3.0, 6 which is a statistical program capable of analyzing Bayesian hierarchical models by Markov Chain Monte Carlo (MCMC) 0 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8 8.5 9 10 1990 1995 2000 2005 2010 Calendar time EDSS (jittered) Subtype PPMS RRMS SPMS All EDSS measurements over time Figure A.1. All EDSS measurements (jittered to prevent over-plotting) over time, by MS type (RRMS, green; PPMS, red; SPMS, blue). EDSS ¼Expanded Disability Status Scale; PPMS ¼primary progressive multiple sclerosis; RRMS ¼relapsing-remitting multiple sclerosis; SPMS ¼secondary progressive multiple sclerosis. Clinical Therapeutics 557.e1 Volume 39 Number 3 RRMS: next EDSS, current EDSS is 0 RRMS: next EDSS, current EDSS is 1.5 or 2 RRMS: next EDSS, current EDSS is 3.5 or 4 RRMS: next EDSS, current EDSS is above 5.0 RRMS: next EDSS, current EDSS is 4.5 or 5 RRMS: next EDSS, current EDSS is 2.5 or 3 RRMS: next EDSS, current EDSS is 1 10.0 7.5 5.0 2.5 0.0 10.0 7.5 5.0 2.5 0.0 10.0 7.5 5.0 2.5 10.0 7.5 5.0 2.5 0.0 05Yrs 10 0510 15 Yrs 10.0 7.5 5.0 2.5 0.0 0510 15 10.0 7.5 5.0 2.5 0.0 0 5 Yrs 10 15 Yrs Yrs Yrs Yrs 0510 15 0510 15 0510 15 20 10.0 7.5 5.0 2.5 0.0 EDSS(jittered) EDSS(jittered) EDSS(jittered) EDSS(jittered) EDSS(jittered)EDSS(jittered) EDSS(jittered) Figure A.2. EDSS development in RRMS population over time showing next EDSS scores. The blue line gives the average expected EDSS over time and the shaded area is the 95% CI obtained by unadjusted local polynomial smoothing of the raw data. CI ¼confidence interval; EDSS ¼Expanded Disability Status Scale; RRMS ¼relapsing-remitting multiple sclerosis. E. Soini et al. March 2017 557.e2 Table A.I. Annual transition probability matrix by EDSS for patients with RRMS based on the Finnish data. From/To RRMS EDSS 0 RRMS EDSS 1 RRMS EDSS 2 RRMS EDSS 3 RRMS EDSS 4 RRMS EDSS 5 RRMS EDSS 6 RRMS EDSS 7 RRMS EDSS 8 RRMS EDSS 9 RRMS EDSS 10 RRMS EDSS 0 0.67822 0.26314 0.04275 0.01136 0.00364 0.00077 0.00003 0.00003 0.00003 0.00003 0.00000 RRMS EDSS 1 0.11299 0.60711 0.17922 0.06484 0.02725 0.00711 0.00037 0.00037 0.00037 0.00037 0.00000 RRMS EDSS 2 0.01770 0.17312 0.37521 0.22712 0.14263 0.04960 0.00365 0.00365 0.00365 0.00365 0.00001 RRMS EDSS 3 0.00547 0.07282 0.26542 0.25690 0.24007 0.11065 0.01216 0.01216 0.01216 0.01216 0.00002 RRMS EDSS 4 0.00155 0.02710 0.14772 0.21289 0.30097 0.18210 0.03189 0.03189 0.03189 0.03189 0.00009 RRMS EDSS 5 0.00045 0.00981 0.07124 0.13607 0.25204 0.20969 0.08010 0.08010 0.08010 0.08010 0.00031 RRMS EDSS 6 0.00001 0.00027 0.00276 0.00786 0.02314 0.04173 0.23071 0.23071 0.23071 0.23071 0.00141 RRMS EDSS 7 0.00001 0.00027 0.00276 0.00786 0.02314 0.04173 0.23071 0.23071 0.23071 0.23071 0.00141 RRMS EDSS 8 0.00001 0.00027 0.00276 0.00786 0.02314 0.04173 0.23071 0.23071 0.23071 0.23071 0.00141 RRMS EDSS 9 0.00001 0.00027 0.00276 0.00786 0.02314 0.04173 0.23071 0.23071 0.23071 0.23071 0.00141 RRMS EDSS 10 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 1.00000 Clinical Therapeutics 557.e3 Volume 39 Number 3 simulation methods, was used to estimate the EDSS transition probabilities. When estimating the RRMS EDSS 0–9 transitions, uniform priors were assumed because no earlier Finnish transition probabilities data were available. Based on a prior knowledge of the data in question, 60% of the mortality was assumed to be MS-related. 7 This estimate was conservative in comparison to other estimates, which have a higher proportion of MSrelated mortality (eg, 78.3% in Goodin et al 8 ). The results shown in Table A.I are well in line with the recent British Columbia results. 9 RRMS to SPMS Transition The hazard rate (HR, λ) for conversion from RRMS EDSS 1 to SPMS was calculated assuming an exponential survival function (ie, a constant hazard of converting to SPMS over time): SðtÞ¼expðλtÞ λfor an exponential distribution could be estimated from the median time of conversion to SPMS, reported to be 15 years based on London Ontario data, 10,13 ie: λ¼lnð2Þ=15 This gives an annual HR of 0.0462 for SPMSconversion of patients in EDSS 1. The Finnish dataset includes only a few observations of conversion to SPMS, and an EDSS-specific rate could not be estimated from these. Based on the London Ontario data, the Cox proportional hazards model was: HðtÞ¼HðtÞEDSS1:expðβXÞ where H(t) is the HR of conversion for any EDSS state; H(t) EDDS1 , the HR of conversion for EDDS 1; and β, the coefficient (0.25270) of the relationship between EDSS and the HR of progression between the base case EDSS 1 and all other EDSS states. 10,13 Using Bender et al, 11 the relationship was reformulated as: ln HðtÞ HðtÞEDSS1  ¼β:X Thus: HðtÞ¼λ:eβ:X This was used to derive the HR of conversion from EDSS 1 through each successive stage to EDSS 8 (Table A.II). All estimated HRs were then subsequently converted into probabilities 12 : p¼1expðrtÞ EDSS Transitions in SPMS For SPMS transitions, data from the London Ontario MS registry 10,13 were available and used (Table A.III), because the Finnish register data had too few EDSS measurements for patients with SPMS. Table A.II. Annual probabilities of conversion to SPMS from RRMS by EDSS score. EDSS score Calculation Hazard rate of conversion Calculation Probability 1 ln(2)/15 0.046210 1-exp(-0.046210) 0.045158 2 0.04621*e (0.25270*2) 0.076600 1-exp(-0.076600) 0.073739 3 0.04621*e (0.25270*3) 0.098622 1-exp(-0.098622) 0.093915 4 0.04621*e (0.25270*4) 0.126975 1-exp(-0.126975) 0.119245 5 0.04621*e (0.25270*5) 0.163480 1-exp(-0.163480) 0.150817 6 0.04621*e (0.25270*6) 0.210480 1-exp(-0.210480) 0.189805 7 0.04621*e (0.25270*7) 0.270993 1-exp(-0.270993) 0.237378 8 0.04621*e (0.25270*8) 0.348902 1-exp(-0.348902) 0.294538 9 † 1.000000 10 0.000000 † Information for EDSS 9 was not available from the London Ontario dataset. Thus, a 100% conversion rate for patients with RRMS in EDSS 9 was assumed. E. Soini et al. March 2017 557.e4