Advertisement

How Does Option Value Affect the Potential Cost-Effectiveness of a Treatment? The Case of Ipilimumab for Metastatic Melanoma

Open ArchivePublished:May 16, 2019DOI:https://doi.org/10.1016/j.jval.2019.02.002

      Abstract

      Background

      Innovations that extend life can generate option value and cost of experiencing future technologies.

      Objectives

      To understand how consideration of option value may affect the potential cost-effectiveness of a treatment through a case study of ipilimumab for previously untreated metastatic melanoma.

      Methods

      We estimated the cost-effectiveness of ipilimumab in 2 scenarios: a conventional scenario, for which we constructed the model using the standard methods that rely on efficacy data directly from the phase III trial of ipilimumab, and an option value scenario, where we incorporated future hypothetical improvements in mortality for metastatic melanoma owing to innovations. We developed 2 approaches to incorporate option value. In the first approach, we forecasted mortality trends based on historical trends from the Surveillance, Epidemiology, and End Results (SEER) Program registry. Alternatively, we identified drugs being studied in clinical trials at the time of ipilimumab’s approval on clinicaltrials.gov and estimated their likelihood and timing of approval, potential efficacy, and cost. We accounted for increases in overall cancer treatment cost and unrelated medical cost in the option value scenario.

      Results

      In the option value scenario, using the SEER approach, the incremental quality-adjusted life-years (QALYs) gained and the incremental cost increased by 6.2% and 3.8%, respectively, whereas the incremental cost-effectiveness ratio (ICER) decreased by 2.3% compared with the conventional scenario. Using the clinicaltrials.gov approach, the incremental QALY gained and the incremental cost increased by 7.5% and 7.1%, respectively, whereas the ICER decreased by 0.40%.

      Conclusions

      We developed generalizable approaches to estimating option value in cost-effectiveness analysis.

      Keywords

      Introduction

      The past 2 decades have witnessed a sharp increase in the use of cost-effectiveness analysis (CEA) to assess the value of medical technologies. Many of these analyses were conducted on newly approved medicines based on efficacy and safety data from their pivotal trials, with the goal of influencing decision making on their pricing, reimbursement, and use. Many of these models adopted a time horizon that was longer than the length of the pivotal trial, and one challenge brought by a long time horizon is whether and how to account for future technology advancement, both related and unrelated to the disease and treatments under study. Most CEAs to date have simply ignored this issue, and as a result, they may have omitted the potential “option value” of some life-extending treatments, which is the opportunity to benefit from future innovations during the extended life.
      • Becker B.
      • Murphy K.
      • Philipson T.
      The Value of Life Near Its End And Terminal Care.
      • Lakdawalla D.N.
      • Doshi J.A.
      • Garrison L.P.
      • et al.
      Defining elements of value in health care—a health economics approach: an ISPOR Special Task Force Report [3].
      (See Appendix in Supplemental Materials found at https://doi.org/10.1016/j.jval.2019.02.002 for a brief overview of the origin of option theory in financial economics.)
      In the healthcare value assessment literature, a few studies have estimated the option value of life-extending medical technologies. Philipson et al. examined the case for monotherapy zidovudine (AZT) for patients with human immunodeficiency virus or acquired immunodeficiency syndrome, which can improve life expectancy by a few months to 1.6 years, before the advent of the highly active antiretroviral therapy (HAART), which can improve life expectancy by about 10 years.
      • Philipson T.J.
      • Becker G.
      • Goldman D.
      • Murphy K.M.
      Terminal Care and the Value of Life Near its End.
      They estimated the option value of monotherapy zidovudine to be as much as 400% of the conventional value in cohorts of patients with human immunodeficiency virus or acquired immunodeficiency syndrome diagnosed right before the arrival of HAART.
      • Philipson T.J.
      • Becker G.
      • Goldman D.
      • Murphy K.M.
      Terminal Care and the Value of Life Near its End.
      Using a similar methodology, Thornton Snider et al. estimated the option value of ipilimumab for extending the survival of metastatic melanoma patients to when pembrolizumab and nivolumab became available. Their findings suggested that for those diagnosed in 2013, the option value of ipilimumab was as much as 49% of its conventional survival benefit. These 2 aforementioned studies treated the arrivals of HAART, pembrolizumab, and nivolumab as events with certainty and used their actual approval dates (rather than forecasts) in the calculations. The option value estimates they arrived at using this approach were therefore the “ex post” option value of the treatments. Another line of research on the option value of cancer drugs estimated the “ex ante” option value by forecasting future improvements in survival among cancer patients. Sanchez et al. examined the case of tyrosine kinase inhibitors for chronic myelogenous leukemia and estimated the option value from future medical innovation to be equivalent to 9% of the average survival gains from existing treatments.
      • Sanchez Y.
      • Penrod J.R.
      • Qiu X.L.
      • Romley J.
      The option value of innovative treatments in the context of chronic myloid leukemia.
      Using similar methodologies, Thornton Snider et al. estimated the option value of nivolumab to be 18%, 5%, and 10% of the conventional value for renal cell carcinoma, squamous nonsmall cell lung cancer (NSCLC), and nonsquamous NSCLC, respecitvely.
      • Thornton Snider J.
      • Batt K.
      • Wu Y.
      • Tebeka M.G.
      • Seabury S.
      The option value of innovative treatments for non-small cell lung cancer and renal cell carcinoma.
      This little literature on quantifying the option value of life-extending treatments has demonstrated that in cases where the main goal of treatment is life extension and the speed of innovation is rapid, option value may not be negligible. Nevertheless, none of the existing studies directly addressed the issue of how the option value of a new medical technology can be estimated in a CEA and how considering option value may affect its potential cost-effectiveness. In this study, we used ipilimumab for the treatment of previously untreated metastatic melanoma to demonstrate how option value can be incorporated in a CEA.
      Ipilimumab was the first metastatic melanoma drug that has demonstrated significant survival benefit over placebo in randomized controlled trials.
      • Robert C.
      • Thomas L.
      • Bondarenko I.
      • et al.
      Ipilimumab plus dacarbazine for previously untreated metastatic melanoma.
      • Hodi S.
      • O’Day S.
      • McDermott D.
      • et al.
      Improved survival with ipilimumab in patients with metastatic melanoma.
      Since its approval in 2011, seven additional new molecules have been approved to treat metastatic melanoma either as a monotherapy or as a part of a combination therapy, and nearly all demonstrated significantly superior overall survival benefit over either chemotherapies or ipilimumab in their phase III trials.
      U.S. Food and Drug Administration. [email protected]
      FDA Approved Drug Products. Drug Databases 2016.
      • Chapman P.B.
      • Hauschild A.
      • Robert C.
      • et al.
      Improved survival with vemurafenib in melanoma with BRAF V600E mutation.
      • Andtbacka R.H.I.
      • Kaufman H.L.
      • Collichio F.
      • et al.
      Talimogene laherparepvec improves durable response rate in patients with advanced melanoma.
      • Flaherty K.T.
      • Robert C.
      • Hersey P.
      • et al.
      Improved survival with MEK inhibition in BRAF-mutated melanoma.
      • Hauschild A.
      • Grob J.J.
      • Demidov L.V.
      • et al.
      Dabrafenib in BRAF-mutated metastatic melanoma: a multicentre, open-label, phase 3 randomised controlled trial.
      • Larkin J.
      • Ascierto P.A.
      • Dréno B.
      • et al.
      Combined vemurafenib and cobimetinib in BRAF-mutated melanoma.
      • Robert C.
      • Schachter J.
      • Long G.V.
      • et al.
      Pembrolizumab versus ipilimumab in advanced melanoma.
      • Robert C.
      • Long G.
      • Brady B.
      • et al.
      Nivolumab in previously untreated melanoma without BRAF mutation.
      • Weber J.S.
      • D’Angelo S.P.
      • Minor D.
      • et al.
      Nivolumab versus chemotherapy in patients with advanced melanoma who progressed after anti-CTLA-4 treatment (CheckMate 037): a randomised, controlled, open-label, phase 3 trial.
      For metastatic melanoma, ipilimumab in itself can prolong survival and improve quality of life (before progression), and these constitute the value of ipilimumab captured by a conventional CEA. Nevertheless, longer survival can also give patients the opportunity to benefit from later innovations, and this is the potential option value of ipilimumab.
      In this study, we estimated the cost-effectiveness of ipilimumab in 2 scenarios: the conventional scenario and the option value scenario. In the conventional scenario, we constructed the CEA model using the standard methods that rely on efficacy data directly from the phase III study of ipilimumab. In the option value scenario, we forecasted future mortality and incorporated it in the CEA. We used data available by 2011, when ipilimumab was approved, to do the forecast for post 2011. The goal for this study was to illustrate how the ex ante option value can be estimated for a new drug when the initial decision on pricing and reimbursement is made.

      Method

       Model Structure and Patient Population

      A Markov model consisting of 3 health states—progression-free survival (PFS), progressive disease, and death (Fig. 1)—was constructed in Microsoft Excel to estimate the cost and quality-adjusted life-years (QALYs) on ipilimumab plus dacarbazine versus dacarbazine alone for previously untreated unresectable stage III or stage IV melanoma under the 2 scenarios. The model’s population started in PFS and can transition to progressive disease, directly to death, or stay progression-free in each cycle. Those in the progressive disease state can either transition to death or stay in progressive disease in each cycle. Complete or partial response was modeled as a substate of the PFS in which patients have a different utility but the same cost. The model had a cycle length of 1 month and a lifetime horizon for a typical patient of 57 years of age. The analysis was conducted from the health system’s perspective.

       Dosing

      Dosing and frequency of administration for each arm in the Markov model reflected those in the phase III trial.
      • Robert C.
      • Thomas L.
      • Bondarenko I.
      • et al.
      Ipilimumab plus dacarbazine for previously untreated metastatic melanoma.
      Treatments were discontinued if any of the following occurred: progression of the disease, development of drug-related adverse events, or the end of the study.
      • Robert C.
      • Thomas L.
      • Bondarenko I.
      • et al.
      Ipilimumab plus dacarbazine for previously untreated metastatic melanoma.

       Transition Probabilities: Conventional Scenario

      The monthly transition probabilities for the conventional scenario were derived directly from the phase III trial. As indicated by the Kaplan-Meier overall survival (OS) curves in the publication, a large amount of censoring happened in the fourth (last) year of the trial and approximately 20% of study participants did not have the event of interest (death) by the end of the trial.
      • Robert C.
      • Thomas L.
      • Bondarenko I.
      • et al.
      Ipilimumab plus dacarbazine for previously untreated metastatic melanoma.
      As a result, extrapolation of mortality beyond the trial follow-up was required. Standard parametric methods were first attempted, but they provided poor fit to the actual survival data for both arms. Given the clear plateau of the Kaplan-Meier OS curves at the tails, we assumed, for both arms, that the patients who survived through the end of the trial were “cured” and died at the same rate as the general population after the trial ended. The mortality rates were taken from the period life table in 2010 produced by the Centers for Disease Control and Prevention (CDC). This approach assumed that ipilimumab did not have survival benefit beyond the length of trial (4 years). Extrapolations of PFS curves were not needed because no study participant remained progression free by the end of the trial. The monthly transition probabilities during the trial period were calculated nonparametrically using the published PFS and OS curves for each arm.

       Transition Probabilities: Option Value Scenario

      In this case, the option value of ipilimumab has 2 main components: the option value from potential future reduction in disease-specific mortality, and the option value from potential future reduction in background mortality. In the Markov model, we assumed the transition from progressive disease to death represented disease-specific mortality, whereas the transition from PFS to death represented background mortality. Option value from reduction in background mortality can be roughly accounted for by using the cohort life tables by the Social Security Administration (SSA) to approximate the progression-free mortality. Unlike the Centers for Disease Control and Prevention period life tables, these SSA cohort life tables take into account future trends in mortality, affected by factors including the development and application of new diagnostic, surgical, and life sustaining techniques.
      The other component of option value, which is from future reduction in disease-specific mortality, can be accounted for by lowering the mortality on progressive disease. We used 2 approaches to forecasting—one based on forecasting future trends of melanoma-specific mortality (SEER approach) and the other on forecasting likely future drug approvals for metastatic melanoma (clinicaltrials.gov approach). We conducted this analysis as if we were in 2011, and only used information that was available at that time.
      In the SEER approach, we estimated the past mortality trends for metastatic melanoma using the Surveillance, Epidemiology, and End Results Program (SEER) registry.

      Surveillance, Epidemiology, and End Results (SEER) Program (www.seer.cancer.gov) Research Data (1973-2014), released April 2017, based on the November 2016 submission.

      We selected patients with unresectable stage III/IV melanoma from the registry from 1988 to 2010, and fit a Cox proportional hazard model to all-cause mortality, melanoma-specific mortality, and other-cause mortality in the first 4 years after diagnosis, with the year of diagnosis as the main predictor of interest and adjusted for age, sex, marital status, race, ethnicity, and number of tumors. The hazard ratio (HR) of being diagnosed 1 year later on all-cause mortality, melanoma-specific mortality, and other-cause mortality were 0.964 (P < .001; Appendix Table 4), 0.962 (P < .001; Appendix Table 5), and 0.970 (P = .219; Appendix Table 6), respectively. We then restricted our data to between 1993 and 2010, a period when there was no new FDA approval for metastatic melanoma, and the hazard ratio for diagnosis year was 0.957 (95% CI: 0.943-0.971; P < .001; Appendix Table 7). We further included only patients who survived at least 6 months after diagnosis and patients who survived at least 12 months after diagnosis, and ran the Cox regressions conditional on their survival to 6 and 12 months (assuming that this is when second-line treatment started). The hazard ratios for diagnosis year were 0.965 (95% CI: 0.946-0.984; P < .001; Appendix Table 8) and 0.944 (95% CI: 0.918-0.970; P < .001; Appendix Table 9), respectively. These results demonstrated that mortality for metastatic melanoma in the 2 decades before ipilimumab had been declining similarly across different lines of therapy, and the result was robust in a period when there was no new FDA approval for the disease. Because we adjusted for clinical and demographic characteristics of the patients in the Cox regressions, the observed mortality decline was independent of any change in patient mix and thus may be attributed to the use of new drugs, better use of existing drugs, better management of disease, and so on. We assumed that these trends would continue into the future and adjusted the postprogression mortality from the trial with mortality trends after 6 months postdiagnosis to reflect second-line treatments after 2011.
      In the clinicaltrials.gov approach, we assumed that melanoma-specific mortality would decrease only when a new treatment for the disease becomes available. To forecast new arrivals, we systematically reviewed phase III clinical trials registered on clinicaltrials.gov in 2011. We identified new molecules, biologics, or gene therapies that were in phase III testing for metastatic melanoma, extracted the results (response rate and median OS) of their phase II testing, and estimated their time and likelihood of approval (Appendix Tables 1 and 2) based on published statistics on the R&D time and success rate of oncology drugs.
      • DiMasi J.
      • Hermann J.
      • Twyman K.
      • et al.
      A tool for predicting regulatory approval after phase II testing of new oncology compounds.
      • DiMasi J.A.
      • Reichert J.M.
      • Feldman L.
      • Malins A.
      Clinical approval success rates for investigational cancer drugs.
      • DiMasi J.A.
      • Grabowski H.G.
      • Hansen R.W.
      Innovation in the pharmaceutical industry: new estimates of R&D costs.
      • Sosman J.A.
      • Kim K.B.
      • Schuchter L.
      • et al.
      Survival in BRAF V600-mutant advanced melanoma treated with vemurafenib.
      • Ascierto P.
      • Minor D.
      • Ribas A.
      • et al.
      Phase II trial (BREAK-2) of the BRAF inhibitor dabrafenib (GSK2118436) in patients with metastatic melanoma.
      • Senzer N.N.
      • Kaufman H.L.
      • Amatruda T.
      • et al.
      Phase II clinical trial of a granulocyte-macrophage colony-stimulating factor-encoding, second-generation oncolytic herpesvirus in patients with unresectable metastatic melanoma.
      • Kim K.B.
      • Kefford R.
      • Pavlick A.C.
      • et al.
      Phase II study of the MEK1/MEK2 inhibitor Trametinib in patients with metastatic BRAF-mutant cutaneous melanoma previously treated with or without a BRAF inhibitor.
      Based on these data, the next arrival for metastatic melanoma after ipilimumab was likely in 7 months, with a 77% probability of approval and a median OS of 13 months. In the base case, we assumed that 60% of patients in the progressive disease state initiated another line of therapy after a new treatment arrived.
      • Robert C.
      • Thomas L.
      • Bondarenko I.
      • et al.
      Ipilimumab plus dacarbazine for previously untreated metastatic melanoma.
      We lowered the mortality on progressive disease to the level of the new treatment conditional on its arrival and uptake starting from the eighth cycle.

       Other Clinical Inputs

      In addition to the OS and PFS data, the following clinical parameters from the trial were used in the model: the proportion of patients who had grade III or IV adverse events for each arm, the proportion of patients who had complete or partial response, and the proportion of patients who had stable disease. We included adverse events that affected at least 5% of patients in either arm. A list of clinical, cost, and utility inputs was summarized in Table 1.
      • Robert C.
      • Thomas L.
      • Bondarenko I.
      • et al.
      Ipilimumab plus dacarbazine for previously untreated metastatic melanoma.
      • DiMasi J.
      • Hermann J.
      • Twyman K.
      • et al.
      A tool for predicting regulatory approval after phase II testing of new oncology compounds.
      • Sosman J.A.
      • Kim K.B.
      • Schuchter L.
      • et al.
      Survival in BRAF V600-mutant advanced melanoma treated with vemurafenib.
      • Ascierto P.
      • Minor D.
      • Ribas A.
      • et al.
      Phase II trial (BREAK-2) of the BRAF inhibitor dabrafenib (GSK2118436) in patients with metastatic melanoma.
      • Senzer N.N.
      • Kaufman H.L.
      • Amatruda T.
      • et al.
      Phase II clinical trial of a granulocyte-macrophage colony-stimulating factor-encoding, second-generation oncolytic herpesvirus in patients with unresectable metastatic melanoma.
      • Kim K.B.
      • Kefford R.
      • Pavlick A.C.
      • et al.
      Phase II study of the MEK1/MEK2 inhibitor Trametinib in patients with metastatic BRAF-mutant cutaneous melanoma previously treated with or without a BRAF inhibitor.
      • Sacco J.J.
      • Botten J.
      • Macbeth F.
      • Bagust A.
      • Clark P.
      The average body surface area of adult cancer patients in the UK: a multicentre retrospective study.
      • Barzey V.
      • Atkins M.B.
      • Garrison L.P.
      • et al.
      Ipilimumab in 2nd line treatment of patients with advanced melanoma: a cost-effectiveness analysis.
      • AnalySource
      Premier access to the First Databank drug pricing database.
      • Howard D.H.
      • Bach P.B.
      • Berndt E.R.
      • Conti R.M.
      Pricing in the market for anticancer drugs.
      • Centers for Medicare & Medicaid Services
      Physician Fee Schedule 2018.
      • Guy G.P.
      • Yabroff R.
      • Ekwueme D.U.
      • et al.
      Economic burden of chronic conditions among survivors of cancer in the United States.
      • Tsao H.
      • Rogers G.S.
      • Sober A.J.
      An estimate of the annual direct cost of treating cutaneous melanoma.
      • Seidler A.M.
      • Pennie M.L.
      • Veledar E.
      • Culler S.D.
      • Chen S.C.
      Economic burden of melanoma in the elderly population: population-based analysis of the Surveillance, Epidemiology, and End Results (SEER)—Medicare data.
      • Rashid N.
      • Koh H.A.
      • Baca H.C.
      • et al.
      Economic burden related to chemotherapy-related adverse events in patients with metastatic breast cancer in an integrated health care system.
      • Fitch K.
      • Pelizzari P.M.
      • Pyenson B.
      Cost Driver of Cancer Care: A Retrospective Analysis of Medicare and Commercially Insured Population Claim Data 2004-2014, Milliman.
      • Beusterien K.M.
      • Szabo S.M.
      • Kotapati S.
      • et al.
      Societal preference values for advanced melanoma health states in the United Kingdom and Australia.
      Table 1Clinical, cost, and utility inputs in the Markov model.
      ParameterBase case95% CIDistributionSource
      General
       Age at diagnosis57 years
      • Robert C.
      • Thomas L.
      • Bondarenko I.
      • et al.
      Ipilimumab plus dacarbazine for previously untreated metastatic melanoma.
       Percent female40%
      • Robert C.
      • Thomas L.
      • Bondarenko I.
      • et al.
      Ipilimumab plus dacarbazine for previously untreated metastatic melanoma.
       Average body surface area1.78 m2
      • Sacco J.J.
      • Botten J.
      • Macbeth F.
      • Bagust A.
      • Clark P.
      The average body surface area of adult cancer patients in the UK: a multicentre retrospective study.
       Year of analysis2011
       Months until the next arrival73.4-10.6NormalAnalysis
       Probability of approval77%38%-100%Beta
      • DiMasi J.
      • Hermann J.
      • Twyman K.
      • et al.
      A tool for predicting regulatory approval after phase II testing of new oncology compounds.
       Overall survival on new 2L drug138.2-19.6Normal
      • Sosman J.A.
      • Kim K.B.
      • Schuchter L.
      • et al.
      Survival in BRAF V600-mutant advanced melanoma treated with vemurafenib.
      • Ascierto P.
      • Minor D.
      • Ribas A.
      • et al.
      Phase II trial (BREAK-2) of the BRAF inhibitor dabrafenib (GSK2118436) in patients with metastatic melanoma.
      • Senzer N.N.
      • Kaufman H.L.
      • Amatruda T.
      • et al.
      Phase II clinical trial of a granulocyte-macrophage colony-stimulating factor-encoding, second-generation oncolytic herpesvirus in patients with unresectable metastatic melanoma.
      • Kim K.B.
      • Kefford R.
      • Pavlick A.C.
      • et al.
      Phase II study of the MEK1/MEK2 inhibitor Trametinib in patients with metastatic BRAF-mutant cutaneous melanoma previously treated with or without a BRAF inhibitor.
       Percent of patients on progressive disease initiating new 2L drug60%40%-80%Beta
      • Robert C.
      • Thomas L.
      • Bondarenko I.
      • et al.
      Ipilimumab plus dacarbazine for previously untreated metastatic melanoma.
       Discount rate3%
      Dosing
       Ipilimumab3 mg/kg
      • Robert C.
      • Thomas L.
      • Bondarenko I.
      • et al.
      Ipilimumab plus dacarbazine for previously untreated metastatic melanoma.
       Dacarbazine850 mg/m2
      • Robert C.
      • Thomas L.
      • Bondarenko I.
      • et al.
      Ipilimumab plus dacarbazine for previously untreated metastatic melanoma.
      Efficacy
       Ipilimumab complete response1.6%
      • Robert C.
      • Thomas L.
      • Bondarenko I.
      • et al.
      Ipilimumab plus dacarbazine for previously untreated metastatic melanoma.
       Dacarbazine complete response0.8%
      • Robert C.
      • Thomas L.
      • Bondarenko I.
      • et al.
      Ipilimumab plus dacarbazine for previously untreated metastatic melanoma.
       Ipilimumab partial response13.6%
      • Robert C.
      • Thomas L.
      • Bondarenko I.
      • et al.
      Ipilimumab plus dacarbazine for previously untreated metastatic melanoma.
       Dacarbazine partial response9.5%
      • Robert C.
      • Thomas L.
      • Bondarenko I.
      • et al.
      Ipilimumab plus dacarbazine for previously untreated metastatic melanoma.
       Ipilimumab stable disease18.0%
      • Robert C.
      • Thomas L.
      • Bondarenko I.
      • et al.
      Ipilimumab plus dacarbazine for previously untreated metastatic melanoma.
       Dacarbazine stable disease19.8%
      • Robert C.
      • Thomas L.
      • Bondarenko I.
      • et al.
      Ipilimumab plus dacarbazine for previously untreated metastatic melanoma.
      Trend in melanoma mortality hazard on second-line treatments
       Hazard ratio annually0.9650.955-0.975NormalAnalysis
      Treatment discontinuation due to adverse events
       Ipilimumab, induction34%
      • Robert C.
      • Thomas L.
      • Bondarenko I.
      • et al.
      Ipilimumab plus dacarbazine for previously untreated metastatic melanoma.
       Dacarbazine, induction4%
      • Robert C.
      • Thomas L.
      • Bondarenko I.
      • et al.
      Ipilimumab plus dacarbazine for previously untreated metastatic melanoma.
       Ipilimumab, maintenance9%
      • Robert C.
      • Thomas L.
      • Bondarenko I.
      • et al.
      Ipilimumab plus dacarbazine for previously untreated metastatic melanoma.
       Dacarbazine, maintenance0%
      • Robert C.
      • Thomas L.
      • Bondarenko I.
      • et al.
      Ipilimumab plus dacarbazine for previously untreated metastatic melanoma.
      Grade 3/4 adverse events
       Fatigue, ipilimumab10.9%
      • Robert C.
      • Thomas L.
      • Bondarenko I.
      • et al.
      Ipilimumab plus dacarbazine for previously untreated metastatic melanoma.
       Fatigue, dacarbazine4.8%
      • Robert C.
      • Thomas L.
      • Bondarenko I.
      • et al.
      Ipilimumab plus dacarbazine for previously untreated metastatic melanoma.
       Increase in liver enzymes, ipilimumab21.9%
      • Robert C.
      • Thomas L.
      • Bondarenko I.
      • et al.
      Ipilimumab plus dacarbazine for previously untreated metastatic melanoma.
       Increase in liver enzymes, dacarbazine1.2%
      • Robert C.
      • Thomas L.
      • Bondarenko I.
      • et al.
      Ipilimumab plus dacarbazine for previously untreated metastatic melanoma.
      Costs
       Ipilimumab cost per administration30 00014 712-45 288Gamma
      • Barzey V.
      • Atkins M.B.
      • Garrison L.P.
      • et al.
      Ipilimumab in 2nd line treatment of patients with advanced melanoma: a cost-effectiveness analysis.
       Dacarbazine cost per administration9145-137
      • AnalySource
      Premier access to the First Databank drug pricing database.
       Cost per treatment course, new 2L treatment140 61768 959-212 275Gamma
      • Howard D.H.
      • Bach P.B.
      • Berndt E.R.
      • Conti R.M.
      Pricing in the market for anticancer drugs.
       Chemotherapy administration, first hour117
      • Centers for Medicare & Medicaid Services
      Physician Fee Schedule 2018.
       Chemotherapy administration, additional hour24
      • Centers for Medicare & Medicaid Services
      Physician Fee Schedule 2018.
       Chemotherapy administration, additional drug55
      • Centers for Medicare & Medicaid Services
      Physician Fee Schedule 2018.
       Monthly disease management cost, ipilimumab822
      • Barzey V.
      • Atkins M.B.
      • Garrison L.P.
      • et al.
      Ipilimumab in 2nd line treatment of patients with advanced melanoma: a cost-effectiveness analysis.
       Monthly disease management cost, dacarbazine1518
      • Barzey V.
      • Atkins M.B.
      • Garrison L.P.
      • et al.
      Ipilimumab in 2nd line treatment of patients with advanced melanoma: a cost-effectiveness analysis.
       Monthly cancer cost on progressive disease35782175-4981Gamma
      • Guy G.P.
      • Yabroff R.
      • Ekwueme D.U.
      • et al.
      Economic burden of chronic conditions among survivors of cancer in the United States.
      • Tsao H.
      • Rogers G.S.
      • Sober A.J.
      An estimate of the annual direct cost of treating cutaneous melanoma.
      • Seidler A.M.
      • Pennie M.L.
      • Veledar E.
      • Culler S.D.
      • Chen S.C.
      Economic burden of melanoma in the elderly population: population-based analysis of the Surveillance, Epidemiology, and End Results (SEER)—Medicare data.
       Annual unrelated medical cost23951456-3334Gamma
      • Guy G.P.
      • Yabroff R.
      • Ekwueme D.U.
      • et al.
      Economic burden of chronic conditions among survivors of cancer in the United States.
       Cost of treating severe fatigue2069
      • Barzey V.
      • Atkins M.B.
      • Garrison L.P.
      • et al.
      Ipilimumab in 2nd line treatment of patients with advanced melanoma: a cost-effectiveness analysis.
       Cost treating increase in liver enzymes3540
      • Rashid N.
      • Koh H.A.
      • Baca H.C.
      • et al.
      Economic burden related to chemotherapy-related adverse events in patients with metastatic breast cancer in an integrated health care system.
       Annual increase in cancer treatment cost, inflation-adjusted0.33%0%-1.21%Beta
      • Fitch K.
      • Pelizzari P.M.
      • Pyenson B.
      Cost Driver of Cancer Care: A Retrospective Analysis of Medicare and Commercially Insured Population Claim Data 2004-2014, Milliman.
      Utilities
       Complete/partial response0.88
      • Beusterien K.M.
      • Szabo S.M.
      • Kotapati S.
      • et al.
      Societal preference values for advanced melanoma health states in the United Kingdom and Australia.
       Stable disease0.80
      • Beusterien K.M.
      • Szabo S.M.
      • Kotapati S.
      • et al.
      Societal preference values for advanced melanoma health states in the United Kingdom and Australia.
       Progressive disease0.52
      • Beusterien K.M.
      • Szabo S.M.
      • Kotapati S.
      • et al.
      Societal preference values for advanced melanoma health states in the United Kingdom and Australia.
       Elevated liver enzymes−0.17
      • Beusterien K.M.
      • Szabo S.M.
      • Kotapati S.
      • et al.
      Societal preference values for advanced melanoma health states in the United Kingdom and Australia.
       Severe fatigue−0.17
      • Beusterien K.M.
      • Szabo S.M.
      • Kotapati S.
      • et al.
      Societal preference values for advanced melanoma health states in the United Kingdom and Australia.

       Cost Inputs: Conventional Scenario

      Four categories of costs were included in the conventional scenario: ipilimumab or dacarbazine drug and administration costs, disease management costs in PFS, costs of treating drug-related adverse events, and postprogression treatment cost. Drug unit costs were taken from AnalySource and the literature.
      • Barzey V.
      • Atkins M.B.
      • Garrison L.P.
      • et al.
      Ipilimumab in 2nd line treatment of patients with advanced melanoma: a cost-effectiveness analysis.
      • AnalySource
      Premier access to the First Databank drug pricing database.
      The costs of administration were taken from the CMS Physician Fee Schedule.
      • Centers for Medicare & Medicaid Services
      Physician Fee Schedule 2018.
      Disease management costs in PFS were taken from the literature and were treatment specific.
      • Barzey V.
      • Atkins M.B.
      • Garrison L.P.
      • et al.
      Ipilimumab in 2nd line treatment of patients with advanced melanoma: a cost-effectiveness analysis.
      Costs per adverse event episode and postprogression treatment cost were derived from the literature.
      • Barzey V.
      • Atkins M.B.
      • Garrison L.P.
      • et al.
      Ipilimumab in 2nd line treatment of patients with advanced melanoma: a cost-effectiveness analysis.
      • Rashid N.
      • Koh H.A.
      • Baca H.C.
      • et al.
      Economic burden related to chemotherapy-related adverse events in patients with metastatic breast cancer in an integrated health care system.

       Cost Inputs: Option Value Scenario

      Because we accounted for potential improvement in survival owing to future technology advancement when we calculated the QALYs in the option value scenario, it is important to consider the cost implications of technology advancement as well. Costs in the option value scenario differed from those in the conventional scenario in a number of ways. First, we assumed that the cost of cancer care is rising faster than the cost of medical care in general and that the medical care inflation-adjusted cost of cancer care increased 0.33% annually.
      • Fitch K.
      • Pelizzari P.M.
      • Pyenson B.
      Cost Driver of Cancer Care: A Retrospective Analysis of Medicare and Commercially Insured Population Claim Data 2004-2014, Milliman.
      This was estimated by calculating the increase of cancer care cost above the Consumer Price Index (CPI) Medical Component.
      • Fitch K.
      • Pelizzari P.M.
      • Pyenson B.
      Cost Driver of Cancer Care: A Retrospective Analysis of Medicare and Commercially Insured Population Claim Data 2004-2014, Milliman.
      • Bureau of Labor Statistics
      Consumer Price Index 2018.
      Second, we assumed that annual unrelated medical cost for this patient population was $2395 and did not increase above the inflation of medical care.
      • Guy G.P.
      • Yabroff R.
      • Ekwueme D.U.
      • et al.
      Economic burden of chronic conditions among survivors of cancer in the United States.
      Third, in the clinicaltrials.gov approach, we forecasted the costs of new metastatic melanoma drugs using a model developed by Howard et al. on pricing for anticancer drugs.
      • Howard D.H.
      • Bach P.B.
      • Berndt E.R.
      • Conti R.M.
      Pricing in the market for anticancer drugs.
      All costs were adjusted to 2011 US dollars using the Consumer Price Index medical component.
      • Bureau of Labor Statistics
      Consumer Price Index 2018.

       Utility Inputs

      Utility for each health state and disutility associated with each grade III/IV adverse event were taken from a societal preference elicitation study for advanced melanoma.
      • Beusterien K.M.
      • Szabo S.M.
      • Kotapati S.
      • et al.
      Societal preference values for advanced melanoma health states in the United Kingdom and Australia.
      We assumed that each severe adverse event lowered the patient’s quality of life for 1 month.

       Base Case Analysis

      We calculated the incremental costs, QALYs, and cost-effectiveness ratios (ICERs) of ipilimumab plus dacarbazine versus dacarbazine alone under both the conventional and the option value scenarios. We also calculated the changes in incremental costs, QALYs, and ICERs from the conventional scenario to the option value scenario. Discount rates were set at 3% for both costs and outcomes.

       Scenario Analysis

      In scenario analysis, we varied ipilimumab’s efficacy (HR of OS relative to dacarbazine) by ±20% and examined its impact on the change in incremental QALY gained from the conventional scenario to the option value scenario.

       Sensitivity Analysis

      We conducted one-way and probabilistic (PSA) sensitivity analyses to examine the impact of key parameters on the estimate of option value: the timing, likelihood, efficacy, and cost of the next innovation in melanoma in the clinicaltrials.gov approach, the trends in melanoma-specific mortality approach, and treatment costs in the SEER approach. For inputs that did not have empirically estimated standard errors, we derived the range for sensitivity analysis through varying the base case value by roughly ±50%. We conducted 10 000 simulations in the PSA.

      Results

       Base Case

      The base case results are summarized in Table 2. In the conventional scenario, the incremental QALY gained of ipilimumab plus dacarbazine versus dacarbazine alone was 0.76. In the option value scenario, it increased by 6.2% to 0.81 using the SEER approach and by 7.5% to 0.82 using the clinicaltrials.gov approach.
      Table 2Base case results: costs, QALYs, and ICERs in conventional and option value scenarios.
      QALY gainedCancer cost, $Healthcare cost,
      Healthcare cost included unrelated medical cost.
      $
      Difference between option value scenario and conventional scenario
      QALY gained (%)Cancer cost, $ (%)Healthcare cost,
      Healthcare cost included unrelated medical cost.
      $ (%)
      Conventional scenario
       Ipilimumab+dacarbazine2.29275 188284 816
       Dacarbazine1.53103 366109 728
       Incremental0.76171 822175 087
       ICER, $/QALY224 901229 175
      Option value scenario—SEER approach
       Ipilimumab+dacarbazine2.60307 278318 3330.31 (13.5)32 090 (11.7)33 518 (11.8)
       Dacarbazine1.79129 104136 6770.26 (17.2)25 738 (24.9)26 949 (24.6)
       Incremental0.81178 174181 6560.05 (6.2)6352 (3.7)6569 (3.8)
       ICER, $/QALY
      Negative values indicate that the ICER was smaller in the option value scenario than in the conventional scenario.
      219 651223 944−5250 (−2.33)−5231 (−2.28)
      Option value scenario—clinicaltrials.gov approach
       Ipilimumab+dacarbazine2.71356 921368 4660.42 (18.1)81 733 (29.7)83 650 (29.4)
       Dacarbazine1.89172 991181 0100.36 (23.5)69 625 (67.4)71 281 (65.0)
       Incremental0.82183 930187 4560.06 (7.5)12 108 (7.0)12 369 (7.1)
       ICER, $/QALY
      Negative values indicate that the ICER was smaller in the option value scenario than in the conventional scenario.
      223 964228 258−937 (−0.42)−917 (−0.40)
      ICER indicates incremental cost-effectiveness ratio; QALY, quality-adjusted life year.
      Healthcare cost included unrelated medical cost.
      Negative values indicate that the ICER was smaller in the option value scenario than in the conventional scenario.
      In the conventional scenario, the incremental cost of ipilimumab was $175 087 and $171 822, with and without unrelated medical cost. In the option value scenario, the incremental costs increased by 3.8% and 3.7% to $181 656 and $178 174 using the SEER approach, and increased by 7.1% and 7.0% to $187 456 and $183 930 using the clinicaltrials.gov approach.
      In the conventional scenario, the ICERs of ipilimumab, with and without unrelated medical costs, were $229 175/QALY and $224 901/QALY, respectively. In the option value scenario, the ICERs decreased by 2.28% and 2.33% to $223 944/QALY and $219 651/QALY using the SEER approach and decreased by 0.40% and 0.42% to $228 258/QALY and $223 964/QALY using the clinicaltrials.gov approach.

       Scenario Analysis

      Results of scenario analysis indicated that the option value of ipilimumab increased with its survival benefit (Fig. 2). In the scenario where the HR of ipilimumab versus control was 0.58 (20% lower than in the base case), the incremental QALY gained in the option value scenario was 8.1-8.3% higher than that in the conventional scenario. In the scenario where the HR was 0.86 (20% higher than in the base case), the incremental QALY gained in the option value scenario was only 0.5%-2.5% higher than that in the conventional scenario.
      Figure thumbnail gr2
      Figure 2Scenario analysis of ipilimumab’s efficacy (hazard ratio of all-cause mortality relative to dacarbazine) and the change in incremental QALY gained from the conventional scenario to the option value scenario. In the clinical trial (base case), the hazard ratio of ipilimumab was 0.72.
      QALY indicates quality-adjusted life year; SEER, Surveillance, Epidemiology, and End Results.

       Sensitivity Analysis

      In one-way sensitivity analysis for the clinicaltrials.gov approach, where all inputs were varied by about ±50% of the base case value, the median OS of the next innovation for metastatic melanoma appeared to be the biggest driver of the difference in ICER between the option value scenario and the conventional scenario (Appendix Figs. 8 and 9). Annual increase in cancer care cost also had a big impact because it determined how costly survival will be in the future. The cost of ipilimumab and the cost of the new innovation had the opposite effects on the change in ICER: the more expensive the new innovation is relative to ipilimumab, the more it raises the cost per QALY for ipilimumab. In the SEER approach, annual increase in cancer treatment cost appeared to be the biggest driver (Appendix Figs. 10 and 11). This perhaps largely reflected the greater uncertainty around the estimate of this input. The hazard ratio that measured potential future reduction in mortality had a relatively small impact, possibly owing to its tight confidence interval from the Cox regression. Like in the clinicaltrials.gov approach, the cost of ipilimumab and the cost of the new innovation had the opposite effects on the change in ICER.
      In PSA, in the SEER approach, the ICER on average decreased by $5083/QALY and $5101/QALY in the option value scenario compared with the conventional scenario, with and without unrelated medical cost (Table 3). In the clinicaltrials.gov approach, the ICER on average increased by $2853/QALY and $2839/QALY in the option value scenario compared with the conventional scenario, with and without unrelated medical cost. The 90% credible intervals for the change in ICER in both approaches were tight around zero, indicating a small impact on ICERs after accounting for option value.
      Table 3Simulated values for difference in ICERs between option value and conventional scenarios.
      Base Case ICER
      Base case ICER and base case value are the same as in Table 2.
      Change in ICER from conventional to option value scenario, $/QALY
      Base case value
      Base case ICER and base case value are the same as in Table 2.
      Mean value
      Mean value is the mean of the 10 000 simulations.
      5-95 percentile
      SEER approach
       QALY/cancer cost
      Negative values indicate that the ICER was smaller in the option value scenario than in the conventional scenario.
      219 651−5250−5101−11 440 to 4573
       QALY/healthcare cost
      Healthcare cost included unrelated medical cost.
      ,
      Negative values indicate that the ICER was smaller in the option value scenario than in the conventional scenario.
      223 944−5231−5083−11 420 to 4603
      Clinicaltrials.gov approach
       QALY/cancer cost
      Negative values indicate that the ICER was smaller in the option value scenario than in the conventional scenario.
      223 964−9372839−8178 to 18 137
       QALY/healthcare cost
      Healthcare cost included unrelated medical cost.
      ,
      Negative values indicate that the ICER was smaller in the option value scenario than in the conventional scenario.
      228 258−9172853−8145 to 18 138
      ICER indicates incremental cost-effectiveness ratio; QALY, quality-adjusted life year; SEER, Surveillance, Epidemiology, and End Results.
      Base case ICER and base case value are the same as in Table 2.
      Mean value is the mean of the 10 000 simulations.
      Healthcare cost included unrelated medical cost.
      § Negative values indicate that the ICER was smaller in the option value scenario than in the conventional scenario.

      Discussion

      In this study, we developed 2 approaches for incorporating the option value of life-extending treatment into a CEA and applied them to the case of metastatic melanoma. We found that for the case of ipilimumab, the incremental QALYs gained increased by 6%-8% and the ICER decreased by 0%-2%, after accounting for option value from technology advancement. Although economic theory suggests the presence of option value in the context of medical treatment, this source of value has been ignored in the practice of CEA thus far. This is to our knowledge the first study that incorporated option value in an ex ante CEA.
      Assuming a $150 000/QALY threshold, the results of this study suggest that a “value-based” price of ipilimumab could increase by $7500-$9000 above $114 000 per course of treatment.
      Institute for Clinical and Economic Review
      Final value assessment framework for 2017-2019.
      In disease areas where there are more and bigger breakthroughs, the increase in value-based price could be larger. Different decision-making bodies often use different metrics for value in their assessment, and option value can be included in different forms: increase in value-based price, increase in incremental QALY gained, change in ICER, and pace and scale of technology advancement. Ultimately, whether and how option value should be incorporated depends on the decision-makers’ values and the decision-making process.
      The main challenge with estimating option value is forecasting trends in disease-specific and background mortality, and the costs of treatments in the future. In this study, we developed 2 alternative approaches to forecasting, from which we obtained similar results. For the case of metastatic melanoma, we used the SSA cohort life table to account for improvement in background mortality because noncancer mortality was very low owing to competing risk. For early stage cancers, where cancer-specific mortality is not much higher than noncancer mortality, one could use the methods by the SSA with SEER data to forecast noncancer mortality. In disease areas where there is no such registry on survival, disease-specific mortality can be removed from the overall mortality of the general population, as demonstrated by Rosenberg, and this other-cause mortality of the general population can be used to approximate the other-cause mortality of the disease population.
      • Rosenberg M.A.
      Competing risks to breast cancer mortality.
      Alternatively, death records from the National Center for Health Statistics can be linked to some claims and survey data to estimate and forecast mortality from various causes for different populations.
      For the disease-specific mortality trend, the SEER approach can be easily generalized to other cancers, with adjustments to when on average subsequent line of therapy starts. Some may wish to use administrative claims data or electronic medical records to overcome the SEER registry’s lack of detailed information on treatment history. With such information, mortality trends on a certain line of therapy can be more accurately estimated by analyzing patients who have failed the previous line of therapy. Nevertheless, obtaining these databases can be costly and analyzing them time-consuming. As an alternative, information on clinicaltrials.gov is publicly available, and the review of pivotal trials for a disease can be done quickly. In disease areas where there is no publicly available patient registry, the clinicaltrials.gov approach might be the only practical approach with time and budget constraints. In disease areas where both pipeline information and survival data on survival are readily available, it might be worthwhile to try both approaches if the assumptions for both are likely to be met because using different approaches can help us understand the robustness of the results.
      There are several limitations of this study that are worth noting. First, we did not include potential quality-of-life improvement from technology advancement and assumed that the increase in incremental QALYs came solely from reduced mortality. This likely gave us a conservative estimate of the change in incremental QALYs and a conservative estimate of the change in ICER. Second, in the SEER approach to forecasting future reductions in mortality, we used historical data before 2011 and assumed the historical trend will continue into the future. Nevertheless, the rate of new drug approval for metastatic melanoma has accelerated since the approval of ipilimumab in 2011. Therefore, our approach likely gave us a conservative estimate of the survival improvement after 2011. Third, in the clinicaltrials.gov approach, we limited our search of clinical trials to phase III studies that compared a new molecule entity or novel gene therapy to standard of care for forecasting new arrivals. By doing this, we ignored the possibility of new arrivals being approved without a phase III study. For future practice, the review can be potentially expanded to include phase I and phase II trials and can be combined with an analysis of the FDA’s Fast Track, Breakthrough Therapy, and Accelerated Approval programs. Furthermore, expert opinions can be solicited on likely approvals, their potential use (FDA approved and off-label), and potential efficacy. Fourth, with SEER data we cannot observe when second-line treatment started, so we assumed that second-line treatment started 6 months after diagnosis. To get a more accurate estimate of mortality trend on second- and later-line treatments conditional on first-line treatment, SEER-Medicare data or other administrative claims or clinical data that contain diagnosis, treatments, and survival are required. Last but not least, this research does not address the issue of how to apportion option value. Option value is created by the combination of the current treatment and future innovations. Alhough we are not able to provide an answer to this question, our study is the critical first step to address it.

      Conclusion

      We developed methods for incorporating option value into standard CEAs and provided one example of their application. The study results underscored the need for private and public payers to be aware of treatment attributes that are not reflected in traditional value metrics. Further research is needed on how to reward option value among the emerging new treatments, which are economic complements to prior treatments.

      Acknowledgment

      The authors would like to thank the three anonymous reviewers for their insightful comments.

      Supplemental Materials

      References

        • Becker B.
        • Murphy K.
        • Philipson T.
        The Value of Life Near Its End And Terminal Care.
        National Bureau of Economic Research, Cambridge, MA2007
        • Lakdawalla D.N.
        • Doshi J.A.
        • Garrison L.P.
        • et al.
        Defining elements of value in health care—a health economics approach: an ISPOR Special Task Force Report [3].
        Value Health. 2018; 21: 131-139
        • Philipson T.J.
        • Becker G.
        • Goldman D.
        • Murphy K.M.
        Terminal Care and the Value of Life Near its End.
        Cambridge, MA, 2010
        • Sanchez Y.
        • Penrod J.R.
        • Qiu X.L.
        • Romley J.
        The option value of innovative treatments in the context of chronic myloid leukemia.
        Am J Manag Care. 2012; 18: 265-279
        • Thornton Snider J.
        • Batt K.
        • Wu Y.
        • Tebeka M.G.
        • Seabury S.
        The option value of innovative treatments for non-small cell lung cancer and renal cell carcinoma.
        Am J Manag Care. 2017; 23: e340-e346
        • Robert C.
        • Thomas L.
        • Bondarenko I.
        • et al.
        Ipilimumab plus dacarbazine for previously untreated metastatic melanoma.
        N Engl J Med. 2011; 364: 2517-2526
        • Hodi S.
        • O’Day S.
        • McDermott D.
        • et al.
        Improved survival with ipilimumab in patients with metastatic melanoma.
        N Engl J Med. 2010; 362: 2251-2259
      1. FDA Approved Drug Products. Drug Databases 2016.
        • Chapman P.B.
        • Hauschild A.
        • Robert C.
        • et al.
        Improved survival with vemurafenib in melanoma with BRAF V600E mutation.
        N Engl J Med. 2011; 364: 2507-2516
        • Andtbacka R.H.I.
        • Kaufman H.L.
        • Collichio F.
        • et al.
        Talimogene laherparepvec improves durable response rate in patients with advanced melanoma.
        J Clin Oncol. 2015; 33: 2780-2788
        • Flaherty K.T.
        • Robert C.
        • Hersey P.
        • et al.
        Improved survival with MEK inhibition in BRAF-mutated melanoma.
        N Engl J Med. 2012; 367: 107-114
        • Hauschild A.
        • Grob J.J.
        • Demidov L.V.
        • et al.
        Dabrafenib in BRAF-mutated metastatic melanoma: a multicentre, open-label, phase 3 randomised controlled trial.
        Lancet. 2012; 380: 358-365
        • Larkin J.
        • Ascierto P.A.
        • Dréno B.
        • et al.
        Combined vemurafenib and cobimetinib in BRAF-mutated melanoma.
        N Engl J Med. 2014; 371: 1867-1876
        • Robert C.
        • Schachter J.
        • Long G.V.
        • et al.
        Pembrolizumab versus ipilimumab in advanced melanoma.
        N Engl J Med. 2015; 372: 2521-2532
        • Robert C.
        • Long G.
        • Brady B.
        • et al.
        Nivolumab in previously untreated melanoma without BRAF mutation.
        N Engl J Med. 2015; 372: 320-330
        • Weber J.S.
        • D’Angelo S.P.
        • Minor D.
        • et al.
        Nivolumab versus chemotherapy in patients with advanced melanoma who progressed after anti-CTLA-4 treatment (CheckMate 037): a randomised, controlled, open-label, phase 3 trial.
        Lancet Oncol. 2015; 16: 375-384
      2. Surveillance, Epidemiology, and End Results (SEER) Program (www.seer.cancer.gov) Research Data (1973-2014), released April 2017, based on the November 2016 submission.

        • DiMasi J.
        • Hermann J.
        • Twyman K.
        • et al.
        A tool for predicting regulatory approval after phase II testing of new oncology compounds.
        Clin Pharmacol Ther. 2015; 98: 506-513
        • DiMasi J.A.
        • Reichert J.M.
        • Feldman L.
        • Malins A.
        Clinical approval success rates for investigational cancer drugs.
        Clin Pharmacol Ther. 2013; 94: 329-335
        • DiMasi J.A.
        • Grabowski H.G.
        • Hansen R.W.
        Innovation in the pharmaceutical industry: new estimates of R&D costs.
        J Health Econ. 2016; 47: 20-33
        • Sosman J.A.
        • Kim K.B.
        • Schuchter L.
        • et al.
        Survival in BRAF V600-mutant advanced melanoma treated with vemurafenib.
        N Engl J Med. 2012; 366: 707-714
        • Ascierto P.
        • Minor D.
        • Ribas A.
        • et al.
        Phase II trial (BREAK-2) of the BRAF inhibitor dabrafenib (GSK2118436) in patients with metastatic melanoma.
        J Clin Oncol. 2013; 31: 3205-3211
        • Senzer N.N.
        • Kaufman H.L.
        • Amatruda T.
        • et al.
        Phase II clinical trial of a granulocyte-macrophage colony-stimulating factor-encoding, second-generation oncolytic herpesvirus in patients with unresectable metastatic melanoma.
        J Clin Oncol. 2009; 27: 5763-5771
        • Kim K.B.
        • Kefford R.
        • Pavlick A.C.
        • et al.
        Phase II study of the MEK1/MEK2 inhibitor Trametinib in patients with metastatic BRAF-mutant cutaneous melanoma previously treated with or without a BRAF inhibitor.
        J Clin Oncol. 2013; 31: 482-489
        • Sacco J.J.
        • Botten J.
        • Macbeth F.
        • Bagust A.
        • Clark P.
        The average body surface area of adult cancer patients in the UK: a multicentre retrospective study.
        PLoS One. 2010; 5: e8933
        • Barzey V.
        • Atkins M.B.
        • Garrison L.P.
        • et al.
        Ipilimumab in 2nd line treatment of patients with advanced melanoma: a cost-effectiveness analysis.
        J Med Econ. 2013; 16: 202-212
        • AnalySource
        Premier access to the First Databank drug pricing database.
        https://www.analysource.com
        Date accessed: March 24, 2018
        • Howard D.H.
        • Bach P.B.
        • Berndt E.R.
        • Conti R.M.
        Pricing in the market for anticancer drugs.
        J Econ Perspect. 2015; 29: 139-162
        • Centers for Medicare & Medicaid Services
        Physician Fee Schedule 2018.
        • Guy G.P.
        • Yabroff R.
        • Ekwueme D.U.
        • et al.
        Economic burden of chronic conditions among survivors of cancer in the United States.
        J Clin Oncol. 2017; 35: 2053-2061
        • Tsao H.
        • Rogers G.S.
        • Sober A.J.
        An estimate of the annual direct cost of treating cutaneous melanoma.
        J Am Acad Dermatol. 1998; 38: 669-680
        • Seidler A.M.
        • Pennie M.L.
        • Veledar E.
        • Culler S.D.
        • Chen S.C.
        Economic burden of melanoma in the elderly population: population-based analysis of the Surveillance, Epidemiology, and End Results (SEER)—Medicare data.
        Arch Dermatol. 2010; 146: 249-256
        • Rashid N.
        • Koh H.A.
        • Baca H.C.
        • et al.
        Economic burden related to chemotherapy-related adverse events in patients with metastatic breast cancer in an integrated health care system.
        Breast Cancer. 2016; 8: 173-181
        • Fitch K.
        • Pelizzari P.M.
        • Pyenson B.
        Cost Driver of Cancer Care: A Retrospective Analysis of Medicare and Commercially Insured Population Claim Data 2004-2014, Milliman.
        2016
        • Beusterien K.M.
        • Szabo S.M.
        • Kotapati S.
        • et al.
        Societal preference values for advanced melanoma health states in the United Kingdom and Australia.
        Br J Cancer. 2009; 101: 387-389
        • Bureau of Labor Statistics
        Consumer Price Index 2018.
        https://www.bls.gov/cpi/#
        Date accessed: March 23, 2018
        • Institute for Clinical and Economic Review
        Final value assessment framework for 2017-2019.
        2017
        • Rosenberg M.A.
        Competing risks to breast cancer mortality.
        J Natl Cancer Inst Monogr. 2006; 53706: 15-19