Current Challenges in Health Economic Modeling of Cancer Therapies: A Research Inquiry

Cancer is the second most common cause of death in the United States, currently accounting for nearly 1 of every 4 deaths annually.¹ Not surprising, the development of innovative cancer treatments is an important focus of pharmaceutical drug development. The payoff has been meaningful, with new treatments helping to significantly increase cancer survival among men and women and in nearly all racial and ethnic groups.^2,3

This improvement, however, has come at a considerable expense; the total direct medical spending on cancer care in the United States is estimated to have cost as much as $124.5 billion in 2010, with estimated expenditures of almost $160 billion by 2020.⁴ The cost of new cancer therapies has garnered the attention of a diverse constituency of patients, clinicians, healthcare payers, and public policymakers.⁵ As decision makers struggle for ways to better manage their budgets while providing the best care possible to patients with cancer, there is a pressing need to assess the value of cancer treatments in a rigorous and robust manner.

Of course, the question about what constitutes “value” in cancer care is a topic of ongoing discussion at all levels. Fundamentally, value is defined as the health outcomes achieved per dollar spent, but there is no universally accepted definition of what good value—or even acceptable value—means in cancer care.^6,7 Value means something different to each stakeholder, whether a patient with cancer or a caregiver, a physician treating patients with cancer, a healthcare payer, or another medical decision maker.

As innovative new cancer therapies come to market bearing high price tags, the need for economic modeling to help provide better insight into what constitutes “value” is of paramount importance. Yet, after nearly 2 decades of cultivating and refining techniques for the modeling of cost-effectiveness and budget impact of cancer therapies—the 2 main economic models used for evaluating cancer therapies (Table 1)—serious methodologic and policy challenges have emerged that bring into question the adequacy of economic modeling, as it currently stands, as a sound decision-making tool in oncology.

As an ongoing research endeavor, in this article we are exploring some of the contentious issues in the development and use of oncology economic models as informative tools in current healthcare decision-making. Our goal is to draw attention to these complex pharmacoeconomic concerns and to promote discussion among the various stakeholders regarding the need for improvements in, and the best practices for, oncology economic modeling.

Method
The focus of our research is on the processes and the means by which economic models are used to evaluate healthcare technologies in oncology. We have structured our inquiry around 4 questions:

1. Are economic models adequately addressing questions relevant to oncology decision makers?
2. What are the methodologic limitations of oncology economic models?
3. What guidelines are followed for developing oncology economic models?
4. Is the evolution of oncology economic modeling keeping pace with treatment innovation?

Within the context of each of these questions, we are investigating issues related to the technical limitations of oncology modeling, the availability of adequate data for developing models, and the problems regarding how oncology modeling analyses and results are presented and interpreted. Our investigation is largely based on our years of professional experience with developing oncology economic models on behalf of the industry (ie, pharmaceutical and biotechnology corporations), insurers and other healthcare payers, and government healthcare technology assessment agencies.

Moreover, as part of our professional duties, we continuously monitor venues of published literature and activities of professional and scientific organizations to keep abreast of the latest changes and advances in pharmacoeconomics and health economic evaluation. In addition to developing numerous oncology economic models, we conduct several literature reviews annually on the economic aspects of cancer treatment, and this has afforded us the luxury of broad and deep exposure in the field of oncology economic modeling.

Economic models typically involve the evaluation of clinical, economic, and humanistic (ie, quality of life) outcomes in 1 or more hypothetical patient cohorts defined by demographics, disease history, clinical characteristics or presentation, and other factors. Oncology-­related economic models are used by pharmaceutical companies, health plans, Pharmacy & Therapeutics Committees, hospitals, clinicians, and the government (eg, Medicare, Medicaid), among many others (Table 2). Clinicians and formulary managers in oncology practice often seek a realistic understanding of the potential impacts of treatment on funding decisions about cancer therapies. Yet, given the current state of oncology economic modeling, are these healthcare decision makers receiving the proper information they need?

Our experience and observation over the years is that model developers tend to target everything and everyone possible, with the idea that analyses and results can be sorted out later by the end user (ie, healthcare decision makers). This lack of focus can lead to considerable uncertainty about what is meaningful and actionable about results and findings from modeling studies.

The issue is further compounded by the divergent perspectives of payers and providers: our experience is that the former often care mainly about direct clinical and economic outcomes (eg, cure rates, survival, costs of care), whereas the latter may care as much (or more) about other outcomes, such as the impact of treatment on patient functioning and on quality of life. Whether the current oncology economic models meet decision maker needs remains an open question, but what little has been reported in the literature suggests that, in general, economic models (not only oncology models) consistently fall short of being able to support informed decision-making.^8-12

One of the reasons that oncology models may fall short of decision maker needs is that the outcomes of the models lack practical meaning. Although a variety of outcomes metrics can be generated by oncology economic models, these usually entail combinations of incremental costs and benefits of treatment to create incremental cost-effectiveness ratios (ICERs). Typical ICERs in oncology economic models are the weekly, monthly, or annual costs of survival; the quality-adjusted or progression-free survival; the cost per responder; and the cost per adverse event avoided.^13,14

Based on our professional experience and what we have observed in real-world healthcare settings, there is considerable uncertainty among decision makers on what to make of these metrics—particularly the more familiar (and controversial) cost per quality-adjusted life-year (QALY). The rationale behind this metric is that it explicitly includes a common denominator that facilitates comparisons across different treatments, at various stages, and even across diseases. From a theoretical perspective, the QALY is meant to reflect benefit or utility arising from a particular treatment. In the case of a cancer therapy, improved utility may arise from improved survival or improved quality of life. From our real-world observations and professional experience, few decision makers, however, understand how the QALY is derived or how to interpret it.

The cost per QALY metric has long been an integral component of cost-effectiveness (or cost-utility) modeling of cancer therapies, and it has been regarded in some circles as the “convenient yardstick” for measuring and comparing health effects of interventions across various diseases and conditions.^15,16 This metric, however, has been politically demonized in deliberation about US healthcare reform.¹⁷ Cancer therapies with high cost and small incremental improvement in survival or quality of life may find it difficult to meet societal thresholds for what is considered acceptable value for the money.^18,19

Therefore, the denominator of the cost per QALY metric has been a focus of debate, with the general misgivings for the QALY being well-illustrated in oncology. A patient’s health state and its associated weight may change, depending on, for example, the patient with cancer’s initial health state, or the state to which the patient’s health is raised, all of which are confounded by the rapid fluctuations in health status experienced by most patients with cancer.^20-22 In addition, the practice of developing QALY weights based on valuations from members of the general population can be problematic, because such individuals may misunderstand what it is really like to have cancer.²²

Most people have strong emotional reactions to the idea of having cancer—cancer is dreaded more than other life-threatening conditions, such as Alzheimer’s disease, heart disease, and stroke.^23,24 There are inherent difficulties collecting valid judgments on cancer health states and QALY weights by people who do not have cancer—a priori fear of cancer is not commensurate with actual experience.²⁵ For these reasons, and for many others, the QALY metric shows limitations in its ability to accurately capture the value of the health gains deemed important by patients with cancer.²²

There is good evidence that the cost per QALY metric is not sufficiently accurate or reliable enough to be used by decision makers as a basis for the comparison of the costs of different cancer therapies. But what are the alternatives? One simple and transparent option is for costs to be related to primary health outcomes relevant to the disease in question, with emphasis on clinically meaningful findings rather than on statistically significant results.^8,21

Decisions about what defines “clinically meaningful” will require coordination between various stakeholders, including physicians providing care, managed care organizations, the government, and drug manufacturers designing and conducting clinical trials. Inroads in this area have already been made by the American Society of Clinical Oncology Cancer Research Committee.²⁶ Thus, from the perspective of oncology economic modeling, greater consideration could be given to disease-specific outcomes, such as cancer progression, intensity of cancer pain, or loss of function as a result of cancer.²⁶ Although Neumann and Weinstein assert that “these outcomes do not permit comparisons among diseases and conditions or between treatment and prevention,”¹⁵ these metrics may provide more meaningful results for clinicians and patients.

Another option is to create models that could accommodate variations in survival outcomes beyond median survival. Recent research indicates that patients place greater value on the probability—even a small probability—of a substantial improvement in survival than they do on the greater probability of smaller gain in survival.²⁷ Incorporating the value of these “hopeful gambles” into models may better reflect the value to patients than current modeling strategies.²⁷ This implies that there should be an increased role for behavioral economics in economic modeling of cancer therapies, at least to complement the traditional grounding of health economic models in standard economic theory.^28-31

Finally, we must also ask, is cost-effectiveness even the right metric to model? Given the life-threatening nature of cancer, neither payers nor clinicians seem particularly compelled to worry about whether a given treatment is below a threshold value for cost-effectiveness. What is likely of greater interest is the balance of comparative benefits and costs among competing treatments.

Therefore, an ideal health economic model would help decision makers assess which treatments work best for which patients under which scenarios. This question is very different from whether a given treatment meets an arbitrary threshold for cost-effectiveness.

A variety of techniques can be used for developing health economic models, but most oncology economic models are based on decision analysis.^32-34 Although oncology economic models can be enormously complex, models by definition are supposed to be simplified representations of reality.³⁵ However, making generalizations and simplifications in economic models about the superiority or inferiority of one cancer therapy over another involves accounting for vast complexities and nuances of actual clinical practice and its associated costs. Using economic models to compute overall cost-effectiveness may hide important heterogeneity, leading to the overestimation or underestimation of particular subgroups of patients with cancer or of patients in diverse treatment settings or various healthcare systems.^36-38

All of these factors pose numerous challenges in the design and development of oncology economic models, and in the worst case can impede clear interpretation of the results generated by the models. Unfortunately, in the end, decision makers are frequently left with noninterpretable “noise.”³⁸ Therefore, reminders are needed that these models are merely representations of reality—some representing reality quite well, with others being fictitious at best.

Oncology economic models are also inherently complex and “data hungry.” They attempt to synthesize biologic, epidemiologic, and economic data from diverse sources. The availability of appropriate and robust data to populate model parameters, however, is quite limited. Traditionally, clinical trials and observational studies of healthcare interventions have provided the crucial clinical data inputs for economic modeling. Some examples include efficacy and safety data regarding disease symptoms and progression, adverse events, and death, as well as data regarding patient drug utilization behaviors (eg, therapy discontinuation or switching), all of which are used to derive probability estimates for model parameters.

Although these data may be appropriately collected, analyzed, and incorporated into an economic model, the fact still remains that they are most applicable to the intents of the original study in terms of treatment setting, patient demographics, health and disease status, and the time period over which the study was conducted. Moreover, much greater attention is paid to collecting efficacy and safety data in oncology clinical trials than to the collection of health economic data; for this reason, the quality of economic data collection may not always meet expectations or satisfy the needs of economic modeling.³⁹

Incentives inherent in the clinical research enterprise lead to clinical trials designed for the specific purpose of regulatory approval and maximizing market penetration.⁵ More often than not, economic data for oncology economic models must come from alternative sources not linked in parallel to the clinical data—such as naturalistic, noninterventional studies (including both prospective and retrospective observational studies).⁵ This poses problems on its own, not only from potential selection bias, but also because of the conundrum that such studies can be conducted only after the US Food and Drug Administration (FDA)’s approval of the drug label, and drug utilization and expenditure have already begun.³⁹

Funneling disparate clinical and economic data directly into models to examine economic questions over an extended period and across different geographic units or population subgroups is complex and fraught with opportunity for bias or outright error. This is especially true when clinical trial safety and efficacy end points are collected over short periods and explicitly omit information pertaining to resource utilization, costs, or patient preferences.⁴⁰ Consequently, uncertainty about the clinical-economic benefits and harms are associated with oncology therapies, which preclude decision makers from reliably assessing the value of these therapies.⁵

In many cases, clinical trial data available for use in oncology economic models are derived from small, early-­phase, nonrandomized trials with only 1 study arm. According to 1 study based on data from Clinicaltrials.gov, 98% of oncology-related clinical trials registered from 2007 to 2010 have 1000 participants or fewer, and 75% have 100 or fewer participants.⁴¹ In fact, compared with cardiovascular disease and mental health clinical trials, oncology clinical trials were found to be the smallest, with a median number of 43 patients. Moreover, most (approximately 71%) of the trials were early phase (phase 0-2), and approximately 65% were single-group, nonrandomized (approximately 64%), and nonblinded (approximately 88%).⁴¹ Although these statistics reflect the reality of drug development, which appropriately designs trials for regulatory purposes, they highlight the fact that clinical trials do not reflect the experience of the majority of patients. Based on these data from Clinicaltrials.gov, small oncology trials are unlikely to be informative for establishing the effectiveness of treatments with modest effects or for comparing effective treatments to enable better decisions in practice.⁴¹

A lack of head-to-head clinical-economic data is one of the larger problems in developing economic models for cancer therapies, especially because there tends to be more routine off-label use of cancer drugs.⁵ Reviews of clinical practice suggest that nearly 50% to 75% of cancer care is provided off label.^5,42 Thus, data from sources other than registration-oriented clinical trials (eg, medical rec­ords and patient registries) may prove necessary to evaluate which patients do best under which treatment scenarios, and to appropriately capture the costs of treatment.⁵

Although some sources of real-world data, such as comprehensive databases of healthcare claims, capture aspects of cancer-related care and its associated costs, they usually do not include the cancer disease stage and may lack other important clinical details (eg, an accurate recording of line of therapy, and positive evidence of survival). Smaller electronic medical record data sets typically capture cancer-specific data elements, but may not include non–cancer-related care and may even miss the important aspects of resource utilization, such as cancer-related hospitalizations, altogether. Whether based on clinical trial data sets or on claims databases, even when modeling data approach the point of being adequate, it is difficult to calibrate them to the real world without real-world reference populations.^9,43

Additional methodologic limitations for oncology modeling coincide with the fact that new oncology treatments are increasingly targeted at advanced (ie, metastatic) disease, and that the majority of drugs for the treatment of cancer that are coming to market have late-stage indications.²² This phenomenon results from the frequent need for an overall survival end point and the ethics of not withholding the current standard of care, which leads to trial designs that begin with patients with advanced-stage cancer who have been heavily pretreated and who have exhausted existing standard treatment options.44 Correspondingly, the majority of oncology economic models are models for late-stage disease. Therapies that may significantly improve survival and/or convey other benefits in early-stage disease often fail to differentiate themselves in late-stage, difficult-to-treat patient populations, and corresponding economic models have an intrinsic ceiling on the type and number of economic benefits that can be evaluated.⁴⁴

Rarely are new economic models developed, or the old models updated, to reflect the shifts in treatment paradigms. Manufacturers struggling to justify the price of new therapies that are only applicable to a small number of patients with metastatic disease are deprived of the true real-world value, because the drugs are used in earlier lines of therapy or in larger, earlier-stage populations, where survival benefits are much greater, and the drugs are much more likely to demonstrate an acceptable effectiveness for the cost.

Conceptually, it would be interesting to compare an economic model for a cancer drug at the time of its launch, and then build variations on that model as its indications and target population expand over time (eg, 10 years later), and to compare the cost-effectiveness ratios from the before and after scenarios to observe how they change.

In recent years, published guidelines have been made available for organizations developing and evaluating health economic models to encourage appropriate conduct for decision-making purposes. Notable examples have been published by the International Society for Pharmacoeconomics and Outcomes Research (most recently in collaboration with the Society for Medical Decision Making),^32,45,46 the Academy of Managed Care Pharmacy,⁴⁷ and WellPoint.⁴⁸

Most health economic modeling guidelines are a subset of comprehensive, country-specific guidelines for economic evaluation in healthcare technology assessments (eg, guidelines issued by the Canadian Agency for Drugs and Technologies in Health [CADTH] and the UK National Institute for Health and Care Excellence). Although published guidelines have provided an important foundation for standardizing economic models, they tend to be more dogmatic (ie, locked in the incontrovertible belief of their own truth) than practical, particularly for the unique challenges posed by oncology. Moreover, the speed at which economic modeling methodologies have progressed in the past 2 decades has been outpaced by advances in oncology treatment technologies. Historically, it is unclear how closely researchers have followed the guidelines for economic model development.

For oncology modeling, guideline deviation may be a consequence of accommodating the unique aspects of new treatments for cancer and the populations in which they are being applied. Without practical guidelines for developing oncology models, researchers interested in developing economic models of cancer therapies have carte blanche choice in modeling methods, resulting in inconsistent or substandard results that prove frustrating for the end users.

Although disease-specific economic modeling guidelines have been published (eg, diabetes⁴⁹), only CADTH has published economic modeling guidelines that are specific to cancer, and the motivation for issuing this guidance closely resonates with the issues we discuss here.⁵⁰ According to Mittmann and colleagues in a publication by the Canadian Agency for Drugs and Technologies in Health, “Current general pharmacoeconomic guidelines do not provide sufficient direction to ensure a consistent approach to the conduct of economic analyses in oncology technology assessment. The decision to develop a guidance document was based on the observed heterogeneity and quality of the analyses in oncology submissions to decision-making bodies where some of these economic analyses have been conducted in an inappropriate or misleading fashion.”⁵⁰

By addressing considerations that are fundamentally important to oncology (ie, model scope, structure, and assumptions; model data choices and quality of data; as well as uncertainty, sensitivity, and validation of model results as they all uniquely pertain to cancer), the CADTH guidelines are a small but meaningful step in the right direction for improving best practices for oncology economic modeling.

It is important to note that there is no clear guidance in the United States about how or by whom the information generated by economic models should be interpreted and incorporated into decision-making processes. Moreover, because most drug formulary decision-making activities occur behind closed doors, researchers must fumble their way through the development of economic models that have a weak sense of expectations and only a vague idea about what the true goals and objectives should be. Aside from the achievement (or not) of formulary placement of a modeled cancer drug, feedback or constructive criticism of model design and presentation are rare.

Oncology is one of the fastest growing therapeutic areas in the pharmaceutical industry and is also one of the most innovative, especially with new products focused on the immunologic aspects of cancer and new targets for interrupting cell proliferation. The advent of targeted cancer therapies, or personalized medicine, that allow physicians to tailor treatment to individual patients is profoundly changing the management of many types of cancer and is presenting interesting challenges in economic modeling.

What has been (or will be) the impact of cancer therapy innovation on the way health economic models are developed and perceived? For the most part, questions about the cost implications of personalized therapies are not easily answered by the hypothetical cohort simulations that are so often used in health economic models.^9,51 The new heterogeneous focus in cancer treatment is creating difficulties for economic modeling, which historically evolved on the paradigm of homogeneous patient populations. Models need to incorporate not only the reality of heterogeneous populations, but also the reality of rapidly evolving treatment paradigms in which newer expensive therapies displace former standards of care that become available generically at significantly reduced costs. Models also need to address the off-label nature of treatments for cancer, but it remains unclear how to best incorporate off-label use when pharmaceutical manufacturers face regulatory restrictions that limit the promotion of the value of their products to what has been approved in product labels.⁵ One notable example is the FDA Modernization Act of 1997, which contains a provision (ie, Section 114) stipulating the conditions under which drug companies can promote health economics information to formulary decision makers.⁵²

Finally, the growing use of companion diagnostics in oncology warrants full consideration in economic models. Both diagnostic radiology and genetic testing have their own issues regarding the costs and benefits that precede treatment selection. In some regards, the value of any given treatment is partially dependent on the accuracy of disease staging and genetic testing, which may predict treatment response, and this is remiss in most oncology economic models.

Drug reimbursement authorities and other healthcare decision makers seldom make definitive yes or no verdicts about the adoption of new therapies for cancer; rather, they tend to make nuanced judgments about how the use of medical technologies affects particular sets or subpopulations of patients.²³ In the future, this is going to be increasingly relevant with the advent of personalized medicine that uses biomarker testing and the push toward patient-centered care, where therapies are tailored to the needs of an individual or a subpopulation, all while limiting budget exposure.^23,53 We observe, however, that although we are heading in this direction, economic models continue to be developed on a one-size-fits-all basis, despite the continued reluctance of US payers to use explicit cost-effectiveness considerations to determine the value of cancer therapies.

It is likely that as personalized medicine in oncology rapidly evolves, payers will pay more attention to economic models that evaluate treatment access and healthcare resources that are directed at patients who are the most likely to benefit, and to models that tie economic outcomes more directly to patient outcomes.²³ Although this would be a step in the right direction, genomics and other means for trying to get the right drugs to the right patients will fall flat if arbitrary or capricious thresholds of value persevere. As such, oncology economic modeling will likely be stretched from competing forces: one side pushing for generating economic evidence that will grant patients with cancer greater access to treatments, and the other side attempting to respond to fiscal pressures and the need to balance expenditures. With the drive toward obtaining system-wide efficiencies from directing resources toward patients who will gain the most benefit at the least cost, it seems clear that oncology economic models will need reorientation for predicting budgetary impact, rather than decision cost thresholds.²³

Conclusions
Methodologic and policy challenges in health economic modeling of cancer therapies are considerable, but perhaps not insurmountable. On one level, oncology is similar to the other areas in healthcare that are under pressure to control expenditures while maintaining or improving quality of care and patient outcomes. However, unlike many other areas in healthcare, the practice of oncology and its rapidly evolving technologies present unique challenges that make assessing and demonstrating value especially complex. There is wide latitude for improvement in oncology modeling methodologies, and the way model results are presented and interpreted. We believe that economic models are good, essential tools for helping decision makers assess complex scenarios. However, the technical and data issues associated with oncology economic modeling raise serious questions that now require discussion within the oncology and health economics research communities. It is our hope that this article will provide a framework to guide future discourse on this important topic.

Funding Source
Funding for this study and for the preparation of this article was provided by Truven Health Analytics. An earlier, unpublished version of this study was funded by a grant from the Pharmaceutical Research and Manufacturers of America.

Author Disclosure Statement
Mr Miller and Dr Foley are employees of Truven Health Analytics; Mr Russell is currently an employee of Truven Health Analytics, and was an employee of inVentive Health Clinical and was working under contract with Truven Health Analytics at the time of the writing of this article.

Mr Miller is Director, Economic Modeling, Strategic Consulting, Truven Health Analytics; Dr Foley is Senior Director, Strategic Consulting, Truven Health Analytics; Mr Russell is Vice President, Strategic Consulting, Truven Health Analytics, Cambridge, MA.

References

1. American Cancer Society. Cancer facts and figures 2014. 2014. www.cancer.org/acs/groups/content/@research/documents/webcontent/acspc-042151.pdf. Accessed March 5, 2014.
2. Eheman C, Henley SJ, Ballard-Barbash R, et al. Annual Report to the Nation on the status of cancer, 1975-2008, featuring cancers associated with excess weight and lack of sufficient physical activity. Cancer. 2012;118:2338-2366.
3. Jemal A, Ward E, Thun M. Declining death rates reflect progress against cancer. PLoS One. 2010;5:e9584.
4. Yabroff KR, Lund J, Kepka D, Mariotto A. Economic burden of cancer in the United States: estimates, projections, and future research. Cancer Epidemiol Biomarkers Prev. 2011;20:2006-2014.
5. Mullins CD, Montgomery R, Tunis S. Uncertainty in assessing value of oncology treatments. Oncologist. 2010;15(suppl 1):58-64.
6. Abernethy A, Abrahams E, Barker A, et al. Turning the tide against cancer through sustained medical innovation: the pathway to progress. Clin Cancer Res. 2014; 20:1081-1086.
7. Porter ME. What is value in health care? N Engl J Med. 2010;363:2477-2481.
8. Nichol MB, Knight TK, Epstein J, et al. Opinions regarding the Academy of Managed Care Pharmacy dossier submission guidelines: results of a small survey of managed care organizations and pharmaceutical manufacturers. J Manag Care Pharm. 2007;13:360-371.
9. Ferrusi IL, Leighl NB, Kulin NA, Marshall DA. Do economic evaluations of targeted therapy provide support for decision makers? J Oncol Pract. 2011;7(3 suppl):36s-45s.
10. Hulme C, Browne C, Mansfield J, et al. Determining cost-effectiveness of advanced cancer care: a systematic review of economic models. BMJ Support Palliat Care. 2011;1 (suppl 1):A21-A22.
11. Leung MY, Halpern MT, West ND. Pharmaceutical technology assessment: perspectives from payers. J Manag Care Pharm. 2012;18:256-264.
12. Avalere Health, LLC. Do decision makers really use budget impact models generated by manufacturers? Slides presented at the International Society for Pharmacoeconomics and Outcomes Research Annual International Meeting; June 2-6, 2012; Washington, DC. www.ispor.org/meetings/WashingtonDC0512/releasedpresentations/IP7-Hughes-Only-Slides.pdf. Accessed March 5, 2014.
13. Fryback DG, Craig BM. Measuring economic outcomes of cancer. J Natl Cancer Inst Monogr. 2004:134-141.
14. Tappenden P, Chilcott J, Ward S, et al. Methodological issues in the economic analysis of cancer treatments. Eur J Cancer. 2006;42:2867-2875.
15. Neumann PJ, Weinstein MC. Legislating against use of cost-effectiveness information. N Engl J Med. 2010;363:1495-1497.
16. Neumann PJ. What next for QALYs? JAMA. 2011;305:1806-1807.
17. Johnson FR. Editorial: Moving the QALY forward or just stuck in traffic? Value Health. 2009;12(suppl 1):S38-S39.
18. Grusenmeyer PA, Wong YN. Interpreting the economic literature in oncology. J Clin Oncol. 2007;25:196-202.
19. Greenberg D, Earle C, Fang CH, et al. When is cancer care cost-effective? A systematic overview of cost-utility analyses in oncology. J Natl Cancer Inst. 2010;102:82-88.
20. Smith MD, Drummond M, Brixner D. Moving the QALY forward: rationale for change. Value Health. 2009;12(suppl 1):S1-S4.
21. McGregor M, Caro JJ. QALYs: are they helpful to decision makers? Pharmacoeconomics. 2006;24:947-952.
22. Garau M, Shah KK, Mason AR, et al. Using QALYs in cancer: a review of the methodological limitations. Pharmacoeconomics. 2011;29:673-685.
23. Neumann PJ, Bliss SK, Chambers JD. Therapies for advanced cancers pose a special challenge for health technology assessment organizations in many countries. Health Aff (Millwood). 2012;31:700-708.
24. Zikmund-Fisher BJ, Fagerlin A, Ubel PA. Risky feelings: why a 6% risk of cancer does not always feel like 6%. Patient Educ Couns. 2010;81(suppl):S87-S93.
25. Williams R. Will the QALY survive? ISPOR Connections. 2007;13:13-14.
26. Ellis LM, Bernstein DS, Voest EE, et al. American Society of Clinical Oncology Perspective: raising the bar for clinical trials by defining clinically meaningful outcomes. J Clin Oncol. 2014;32:1277-1280.
27. Lakdawalla DN, Romley JA, Sanchez Y, et al. How cancer patients value hope and the implications for cost-effectiveness assessments of high-cost cancer therapies. Health Aff (Millwood). 2012;31:676-682.
28. Determinants of health economic decisions in actual practice: the role of behavioral economics. Summary of the presentation given by Professor Daniel Kahneman at the ISPOR 10th Annual International Meeting First Plenary Session; May 16, 2005; Washington, DC, USA. Value Health. 2006;9:65-67.
29. Robinson LA, Hammitt JK. Behavioral economics and the conduct of benefit-cost analysis: towards principles and standards. J Benefit-Cost Anal. 2011;2. Article 5.
30. Frank RG. Behavioral economics and health economics. Working paper 10881, October 2004. www.nber.org/papers/w10881.pdf. Accessed April 12, 2014.
31. Liebman J, Zeckhauser R. Simple humans, complex insurance, subtle subsidies. Working paper 14330, September 2008. www.nber.org/papers/w14330.pdf. Accessed April 12, 2014.
32. Weinstein MC, O’Brien B, Hornberger J, et al; for the ISPOR Task Force on Good Research Practices—Modeling Studies. Principles of good practice for decision analytic modeling in health-care evaluation: report of the ISPOR Task Force on Good Research Practices—Modeling Studies. Value Health. 2003;6:9-17.
33. Nuijten MJ, Starzewski J. Applications of modelling studies. Pharmacoeconomics. 1998;13:289-291.
34. Reeder CE, Gordon D. Managing oncology costs. Am J Manag Care. 2006;12(1 suppl):S3-S16; quiz S17-S19.
35. Roberts M, Russell LB, Paltiel AD, et al; for the ISPOR-SMDM Modeling Good Research Practices Task Force. Conceptualizing a model: a report of the ISPOR-SMDM Modeling Good Research Practices Task Force-2. Med Decis Making. 2012;32:678-689.
36. Hall PS, Rautenberg TA, McCabe C. Cost analysis for cancer subgroups. Br J Cancer. 2009;100:1513.
37. Kreif N, Grieve R, Radice R, et al. Methods for estimating subgroup effects in cost-effectiveness analyses that use observational data. Med Decis Making. 2012;32: 750-763.
38. Sculpher M. Subgroups and heterogeneity in cost-effectiveness analysis. Pharmacoeconomics. 2008;26:799-806.
39. Bodrogi J, Kaló Z. Principles of pharmacoeconomics and their impact on strategic imperatives of pharmaceutical research and development. Br J Pharmacol. 2010;159: 1367-1373.
40. Meropol NJ, Schrag D, Smith TJ, et al; for the American Society of Clinical Oncology. American Society of Clinical Oncology guidance statement: the cost of cancer care. J Clin Oncol. 2009;27:3868-3874.
41. Califf RM, Zarin DA, Kramer JM, et al. Characteristics of clinical trials registered in ClinicalTrials.gov, 2007-2010. JAMA. 2012;307:1838-1847.
42. “Off-label” indications for oncology drug use and drug compendia: history and current status. J Oncol Pract. 2005;1:102-105.
43. Stout NK, Knudsen AB, Kong CY, et al. Calibration methods used in cancer simulation models and suggested reporting guidelines. Pharmacoeconomics. 2009;27: 533-545.
44. Goss TF, Picard EH, Tarab A. Recognizing value in oncology innovation. White paper. June 2012. www.phrma.org/sites/default/files/flash/phrma_innovation_oncology.pdf. Accessed March 5, 2014.
45. Caro JJ, Briggs AH, Siebert U, Kuntz KM; for the ISPOR-SMDM Modeling Good Research Practices Task Force. Modeling good research practices—overview: a report of the ISPOR-SMDM Modeling Good Research Practices Task Force-1. Value Health. 2012;15:796-803.
46. Sullivan SD, Mauskopf JA, Augustovski F, et al. Budget impact analysis-principles of good practice: report of the ISPOR 2012 Budget Impact Analysis Good Practice II Task Force. Value Health. 2014;17:5-14.
47. AMCP Format Executive Committee. The AMCP format for formulary submissions version 3.1. December 2012. http://amcp.org/WorkArea/DownloadAsset.aspx?id=16209. Accessed March 5, 2014.
48. WellPoint. Health technology assessment guidelines: drug submission guidelines for new products, new indications, and new formulations. Updated September 2008. www.anthem.com/shared/noapplication/f0/s0/t0/pw_b155849.pdf. Accessed March 5, 2014.
49. American Diabetes Association Consensus Panel. Guidelines for computer modeling of diabetes and its complications. Diabetes Care. 2004;27:2262-2265.
50. Mittmann N, Evans WK, Rocchi A, et al. Addendum to CADTH’s guidelines for the economic evaluation of health technologies: specific guidance for oncology products. Ottawa, Ontario: Canadian Agency for Drugs and Technologies in Health. December 2009. www.cadth.ca/media/pdf/H0405_Guidance_for_Oncology_Prodcuts_gr_e.pdf. Accessed March 5, 2014.
51. Elkin EB, Marshall DA, Kulin NA, et al. Economic evaluation of targeted cancer interventions: critical review and recommendations. Genet Med. 2011;13:853-860.
52. US Food and Drug Administration Modernization Act of 1997, §114 (21 USC 352[a]). www.fda.gov/downloads/RegulatoryInformation/Legislation/FederalFood DrugandCosmeticActFDCAct/SignificantAmendmentstotheFDCAct/FDAMA/FullTextofFDAMAlaw/UCM089145.pdf. Accessed March 5, 2014.
53. American Society of Clinical Oncology. Accelerating progress against cancer: ASCO’s blueprint for transforming clinical and translational cancer research. November 2011. www.asco.org/sites/default/files/blueprint.pdf. Accessed March 5, 2014.