Amgen Comments on ICER’s Proposed Methods Adaptations For Assessments of Potential Cures and Other Transformative Therapies
Amgen appreciates the opportunity to comment on ICER’s Proposed Methods Adaptations for Assessments of Potential Cures and Other Transformative Therapies. Although curative therapies are still quite rare, ICER has identified the need to expand upon traditional valuation methods. It is our hope that the resulting methods will continue to incentivize the development of curative therapies through balanced, science-based assessments of these unprecedented therapies that will likely provide an increasing proportion of health benefits in the future.
More pharmaceuticals are truly transforming outcomes: ICER can lead by putting principles of full and fair valuation from multiple perspectives ahead of uninformed reactions and perceptions attached to such treatments. Health care is a continuum of treatment types, from daily incremental symptom fixes, to one-dose curative therapies for previously incurable diseases. That said, an increasing number of pharmaceutical disease treatments are characterized by durable and meaningful disease improvements that were only imagined 10-20 years ago. This may present a seemingly new analytic challenge of comparing the value of more typical ‘incremental’ treatments that the healthcare system is accustomed to, with the increasing number of new treatments that might produce years or even decades of benefit. Potential cures and other transformative therapies may be more challenging to value for several reasons, driven by issues related to methods, equity and perception, including:
- The greater uncertainty (and accompanying actuarial risk) associated with larger and longer-term benefits compared with short term, incremental and lower cost health care. (Methods question regarding uncertainty).
- Mismatch between when payment might occur (today), with when benefits may accrue (over decades). (Methods question regarding discounting and uncertainty).
- Increasing concentration of benefit for a few people gaining large benefits (e.g., curative therapies for rare cancers) as opposed to many people deriving small benefits (e.g., NSAIDs for a headache). (Equity question regarding distribution).
- Less intuitive familiarity by society in general with these types of treatments and what they are ‘worth’. (Perception issue).
- The high cost of higher value treatments compared with what seems ‘normal’. (Perception issue).
There is an increasing need to help society understand the value of reshaping entire diseases by using metrics that will fully capture the value of curative therapies. This helps in making the inevitable trade-offs between short- and long-term health goals, while at the same time providing incentives to prioritize the higher effort (and much greater long-term rewards) of such progress. This approach should be grounded in robust methods, and once the methods are correct, then adjust for factors such as equity, and finally overcome unfounded perceptions of what the answer should be.
Although there may be a sense of urgency to address these questions, the introduction of curative therapies is still a somewhat rare event in healthcare. It is important that these discussions do not preempt the development and evolution of multiple valuation methodologies, as well as the incentives and funding mechanisms for curative therapies. We appreciate that ICER has taken a collaborative approach to broadly engage stakeholders in this dialogue. We also agree that the evolution of the assessment of cures in ICER’s framework should be an iterative process allowing for change and adaptations as various stakeholders evaluate and interpret the methods. Strategically ICER should also be looking for solutions to help overcome inherent system bias against longer-term, more uncertain, and less equally shared transformative treatments, since most other agents in health care will likely be incentivized to deliver ‘here and now’ health care. Done well, ICER can help promote societally equitable long-term realization of full health for more people, compounded over time, and encompassing multiple generations that more myopic analyses may undervalue.
ICER’s initial problem statement contains some indication that the proposed valuation methods development exercise may be at risk of being driven by perception. ICER’s proposal articulates that the extremely high value of cures requires “a solution to the most egregious prices that would otherwise be recommended by traditional cost-effectiveness methods.”1 This introduces a perspective that the value of curative therapies is already a ‘problem’ that requires a solution, even when that value is supported by established methods. From a value assessment standpoint, it is imperative that the value assessor, in this case ICER, maintain objectivity and ensure impartial scientific methods. In the case of transformative treatments, the methods that HTAs, including ICER, employ to value ‘typical’ treatments should not be discarded based on a matter of perception and instinct for what health care ‘should’ cost because our health care system has not yet invented better ways to share the costs and risks of cures. Any new methodologies for the valuation of a potential cure or transformative therapies should not artificially decrease their high estimated value to fit into a preconceived notion of what the ‘right’ cost should be. We encourage ICER to follow the science of valuation and allow society, stakeholders, and others to debate what might appear to be uncomfortable answers and tradeoffs as a related discussion that is separate from valuation.
It is critical that ICER craft methodologies that accurately capture a curative therapy’s value, separate from affordability: this will ensure optimal investment in healthcare which impacts generations to come. With this goal in mind, Amgen’s guidance on this initiative focuses on three main recommendations: 1). Ensure a broad, flexible and intuitive cure definition, characterized by longevity and quality of life for present and future patients indistinguishable from that of the general population as the ultimate objective (abandoning the term “single or short-term transformative therapies (SSTs)”; 2). Account for areas where modeling and discounting go directly against empirical findings in human behavior and preferences for curative therapies; and 3). Endorse dynamic methods that capture value fairly over time and place contextual and alternative considerations of value quantified on par with the quality-adjusted life-year (QALY). Below we present our key recommendations in more detail in answer to the issues raised in ICER’s proposed adaptations.
1. Defining potentially curative and transformative therapies by durability and outcome achieved
Amgen recommends abandoning the term “single or short-term transformative therapies” (SSTs) and simply referring to cures as “potentially curative or transformative therapies”. We appreciate ICER has broadened their proposed definition of cures to be more inclusive. Recognizing that this is a dynamic space that will evolve in the coming years, this definition needs to be intuitive to patients and stakeholders, and durable as it is refined over time. The term “SSTs” introduces a complex and unnecessarily time-bound definition that is confusing and may impede this quest for a unifying set of methods to help stakeholders appropriately value transformative innovations in all of healthcare. Instead of utilizing the term SSTs, ICER should simply call it what it is – potentially curative or transformative therapies and anchor the definition to relevant transformative domains: 1) marked improvement in outcome achieved, and 2) marked increase in durability of effect, as informed by disease experts which will vary by therapeutic area and impacted population. Having a cures definition with valid conceptual underpinnings will support a stronger foundation for the relevant methods, which can then inform the appropriate valuation. Moreover, appropriate terminology will help frame a more balanced assessment and help avoid prematurely diminishing a curative therapy based on the perception of one-time or all-inclusive high prices, timing of payments vs. benefits, and affordability, which are separate from ‘value.’
2. Cost and benefits time divergence: eliminate or apply lower discount rates to curative therapies
Discounting should begin when the patient gaining the benefit starts their treatment to avoid discounting intrinsic value for future patients. Most health economic analysis employs discounting for treated individuals starting from the present and discounting over time. A more controversial issue when calculating societal costs and benefits is whether we should speak for current and future generations of undiagnosed patients by discounting the value of benefits they will receive when they are diagnosed in the future. Put another way: if a person somehow knew they were going to be diagnosed with cancer in ten years, how much less would they value a cure being developed today, even if they would not get to use it for 10 years? The perspective of present and future patients both warrant consideration.
While ICER’s proposed approach tests both different discount rate values and differential rates, this approach does not address the issue of how to balance the needs of the patients known to us today versus those who will need cures in the longer term. ICER typically models a hypothetical cohort, with limited consideration for a treatment needed in the future. This is a complex area that might not be immediately clear to a lay person or patient, but it is important that patients understand that this approach involving distribution and equity, devalues the curative therapies that patients could need in the next few years. Discounting health is a contentious ethical issue. In fact, the farther in the future this benefit occurs, the more discounting brings the ‘current’ benefit to zero, with a severe impact on those treatments that have the greatest long term benefits. (We recommend ICER refer to other discount rate research such as environmental economics.2,3,4,5)
Non-constant time discounting should be incorporated. Static discount rates developed in 1937 and used by ICER, are out of step with more recent research on discount rates that suggests that individuals apply dynamic discount rates in reality.6 Psychology and behavioral economist field experiments have uncovered strong evidence of human ‘preference reversals’, where individuals prefer x today over y tomorrow, but choose y in a year and a day over x.7 Individuals empirically exhibit preferences for dynamically changing discount rates that are not constant, which might for example, echo a hyperbolic pattern of a 1-3% discount rate initially followed by a far lower rate over time.8 For specific rates, ICER should at minimum use the latest Treasury Green Book guidance of 3.5% for costs and 1.5% for health benefits for curative therapies.9 An accurate discount rate that reflects individual preference is germane to curative therapies to prevent policy that results in disproportionately reduced cure development, especially for younger and pediatric patients.10 Further, recent research supports a hyperbolic discounting effect (even outside of the market failure characteristic of healthcare) as application of static rates could lead to unpredicted collapse in innovative healthcare resources, in this case with curative therapies. Amgen suggests ICER revisit this.11,12
3. Uncertainty: Ensure methods are objectively applied
ICER should apply real-world evidence and clinical data to alleviate uncertainty and ensure there is not an over-reliance on sensitivity analyses. Just as discounting favors short-term over long-term treatments, using higher uncertainty to reduce value will also work against longer-term treatments and future societal benefits. In fact, discounting and uncertainty used together produce a rapidly compounding effect that highly favors short-term and incremental treatments. Consideration should therefore be given as to whether using both discounting and uncertainty to reduce present value is a form of double discounting. Sensitivity analysis must capture the extent to which discounting and uncertainty assumptions lead to big shifts in realized value to avoid reinforcing short term preferences, including analyses where discounting and uncertainty are assumed to be negligible. ICER puts considerable thought into uncertainty and proposes several approaches, and then where feasible, does an excellent job of testing these in existing models of CAR-T, SMA, and Hemophilia A. A few additional concerns:
- For all sensitivity analyses, Amgen recommends ICER simulate only plausible scenarios, not a set of pre-specified analyses. ICER proposes to test varied assumptions on durability, safety and effectiveness as well as provide analysis at different time horizons. Modeling methods that include extreme and implausible scenarios can lead to incorrect conclusions. For example, ICER’s suggestion of varying time periods could lead to incorrect decisions, given the body of economic research that demonstrates that time horizon has an extensive impact on health economic analyses results.13,14,15
- The assessment of value for curative therapies should be separate from policy paradigms on outcomes-based contracting. ICER’s proposal appears to suggest tying outcomes-based arrangements (OBAs) to probabilistic sensitivity analysis (PSA) results. Layering longer term transformative treatments with more assumptions around payment, on top of the potential for discounting and uncertainty, further clouds the intrinsic value of curative therapies. All of these value modifiers are, in effect, mechanisms that penalize any treatment where the benefits are not matched with the costs at every moment in time, which is in effect, an accounting problem rather than a health outcomes value problem. It is important that ICER maintain objectivity and separation in value assessments and allow payers and other stakeholders to evaluate both intrinsic health outcomes value as well as the potential financial value of outcomes-based contracting based on the resulting health economics. Finally, ICER should also acknowledge the fundamental limitations of PSA, even when uncomplicated by multiple and additive forms of discounting noted above.16,17 PSA results even with modest discounting will likely appear to lead to far more uncertain results than an equivalent analysis of a treatment with lower upfront costs and short term returns. So even PSA itself could potentially lead policymakers to incorrect conclusions and poor choices if applied without a high degree of transparency and clarity in communication.
- Different discount rates should be tested in sensitivity analyses. ICER’s proposed adaptations suggest that a test of differing discount rates in the sensitivity analysis is not necessary, however, as ICER’s testing of discount rates has shown,18 discount rates have a disproportionately large impact on the cost-effectiveness results that would be valuable for any decision-maker to see.
In particular for potential rare disease curative therapies, uncertainty must be appropriately balanced with the need for breakthrough therapies which has necessitated the FDA to deem it is in the public’s interest to approve a treatment. Without this consideration, there is greater risk for harm to patients and society. Rather than re-adjudicate the value of trials, new methods for valuing curative therapies should tolerate more uncertainty than might normally be the case. ICER should answer these methods questions and focus on the best ways to extrapolate trial results into the future, by acknowledging signposts of potential future medical value.19
4. Actively incorporate additional dimensions of value
We urge ICER to actively test and refine approaches to incorporate additional aspects of value for the assessment of cures. ICER’s proposed adaptations explore the addition of new elements of value for curative therapies highlighted in the International Society for Pharmacoeconomics and Outcomes Research (ISPOR)’s technical brief towards developing a value-framework, but ICER concludes that these cannot be applied empirically.20 In ICER’s 2017-2019 Value Framework, the QALY (which has significant limitations) is everything, meaning that it has such a disproportionate impact on an assessment that it eclipses other aspects of value. Although the QALY may be an appropriate starting point given a lack of valid alternatives, it needs to be heavily supplemented to account for its limitations. Per ICER’s current framework guidance, products falling above $175,000 cost per QALY would automatically be labeled ‘low value’, hence silencing any role for other elements of value, which are then discussed and considered afterwards by the Panel. This approach confounds the true value of therapies, which would be particularly amplified in the assessment of a cure.
ICER should continue to engage in, and apply findings from its methods research into its assessments, including considering an earlier MCDA type approach with weightings informed by patients/experts. There is a precedent for ICER conducting research to help inform more accurate and appropriate methods for the capture of alternative dimensions of value. ICER should invest in cure value methods related research similar to ICER’s investment in modeling. This should be accompanied by other major changes in both ICER assessments and their engagement with stakeholders, independent panel composition, and voting processes. We recommend ICER re-attempt multi-criteria decision analysis (MCDA), embedding it earlier in the process, with patient / expert input data to inform the relative rankings of the criteria rather than the panel’s implicit vote.21 Thus, ICER should incorporate novel elements of value into the base-case of every cure assessment. Additionally, ICER should encourage manufacturers and academic groups to generate appropriate evidence of novel elements of value for curative therapies prior to an ICER assessment, incorporating them into the base case.
5. Economic Surplus: Focus on curative therapy assessment, leaving surplus to policy makers
Rather than focusing on economic surplus, incorporate the natural reductions in price resulting from competitive entrance and loss of exclusivity (LOE) into models. Full valuation of potential curative therapies may result in prices that seem high to some, but will ensure that we are not potentially mortgaging future cure discovery by succumbing to inappropriate pressure to discount the most transformational aspect of curative therapies: future outcomes. ICER suggests that “Transformative treatments offer the potential for magnitudes of health gain and /or cost offset that raise concerns that traditional cost-effectiveness methods will allocate too much of the economic surplus to innovators and will assign fair prices to transformative treatments that are manifestly unaffordable in the near term”22 This has not been supported by research into consumer surplus nor in empirical research.23,24 As an empirical example, in research analyzing consumer and producer surpluses for HIV/AIDS drug therapies in the late 1980's onwards, innovators appropriated only 5% of the social surplus.25
There are approximately 20,000 diseases today, but most do not have cures and lack treatments.26 Current development of treatments consists of chronic and/or symptomatic therapy but rarely involves a curative aspect, wherein morbidity and mortality are eliminated. Curative therapies, first and foremost, are cures for society, allowing subsequent generations to live free from the threat of morbidity and mortality from disease. These need to be recognized for the life changing value they bring. When a cure goes off patent, the price radically declines with biosimilar and generic competition, as is often the case for non-curative treatments. The correct approach to valuing a cure begins with having correct methods, and then adjusting for matters such as equity and distribution and overcoming issues with perception of what the ‘right’ answer should be. ICER should ensure a broad, intuitive and flexible cure definition. This definition should represent the value of a cure characterized by longevity and quality of life for present and future patients equivalent to the general population as the ultimate objective. It should account for areas where modeling and discounting go directly against empirical findings in human behavior and preferences for curative therapies. It is important that ICER incorporate methodologies that reflect value fairly over time, placing contextual and alternative considerations of quantified value on par with the QALY with the goal of more inclusive and representative decision-making processes.
1. ICER. Value Assessment Methods and Pricing Recommendations for Potential Cures: A Technical Brief. Draft Version. August 6, 2019. p. 26
2. Arrow K, Cropper M, Gollier C, Groom B, Heal G, et al. Determining Benefits and Costs for Future Generations. Science. July 2013, 341: 349-350.
3. Weitzman M L. Why the Far-Distant Future Should Be Discounted at Its Lowest Possible Rate, Journal of Environmental Economics and Management. 1998: 36 (3): 201-208.
4. Ramsey F. Mathematical Theory of Saving. Economic Journal, 1928: 38(152): 543–59.
5. Attema AE, Brouwer WB, Claxton K. Discounting in economic evaluations. PharmacoEconomics. 2018 May 19:1-4. Link
6. Samuelson PA. A note on measurement of utility. Rev Econ Stud. 1937: 4(2):155–161
7. Sterner T, Persson UM. An even Sterner review: introducing relative prices into the discounting debate. Rev Environ Econ Policy. 2008: 2(1):61–76.
8. Hepburn C, Duncan S, Papachristodoulou A. Behavioural economics, hyperbolic discounting and environmental policy. Environmental and Resource Economics. 2010 Jun 1;46(2):189-206.
9. HM Treasury. The green book: Central government guidance on appraisal and evaluation. London: HM Treasury. 2018.Link
10. Sterner T, Persson UM. An even Sterner review: introducing relative prices into the discounting debate. Rev Environ Econ Policy. 2008: 2(1):61–76.
12. ICER states they are not exploring hyperbolic discounting as this approach is “more useful descriptively rather than prescriptively”. However, our experience is that this technique more accurately reflects individual preferences over time than what is currently proposed: ICER. Value Assessment Methods and Pricing Recommendations for Potential Cures: A Technical Brief. Draft Version. August 6, 2019. p. 81.
13. Kim DD, Wilkinson CL, Pope EF, Chambers JD, Cohen JT, Neumann PJ. The influence of time horizon on results of cost-effectiveness analyses. Expert review of pharmacoeconomics & outcomes research. 2017 Nov 2;17(6):615-23.
14. Basu A, Maciejewski ML. Choosing a Time Horizon in Cost and Cost-effectiveness Analyses. JAMA. 2019 Mar 19;321(11):1096-7.
15. Additionally, for questions of durability, theoretically, cured patients will approach the general population’s morbidity/mortality: general population morbidity/mortality already captures the probability to develop a secondary disease, or relapsing from the same disease, requiring a benchmark comparing the risk of relapse to the risk of developing the disease in the first place.
16. Baio G, Dawid AP. Probabilistic sensitivity analysis in health economics. Statistical methods in medical research. 2015 Dec;24(6):615-34. Link
17. Baltussen RM, Hutubessy RC, Evans DB, Murray CJ. Uncertainty in cost-effectiveness analysis: probabilistic uncertainty analysis and stochastic league tables. International Journal of Technology Assessment in Health Care. 2002 Jan;18(1):112-9. Link
18. ICER. Value Assessment Methods and Pricing Recommendations for Potential Cures: A Technical Brief. Draft Version. August 6, 2019. p. 45-49.
19. Jena A, Lakdawalla D. Value frameworks for rare diseases: Should they be different. Health Affairs Blog. 2017 Apr;12. Link
20. Lakdawalla DN, Doshi JA, Garrison Jr LP, Phelps CE, Basu A, Danzon PM. Defining elements of value in health care—a health economics approach: an ISPOR Special Task Force report . Value in Health. 2018 Feb 1;21(2):131-9.
21. Jit M. MCDA from a health economics perspective: opportunities and pitfalls of extending economic evaluation to incorporate broader outcomes. Cost Effectiveness and Resource Allocation. 2018 Nov;16(1):45. Link
22. ICER. Value Assessment Methods and Pricing Recommendations for Potential Cures: A Technical Brief. Draft Version. August 6, 2019. p. 2.
23. In Farrell and Shapiro's 2004 paper, Intellectual Property, Competition and Information Technology, the authors argued that a patent monopoly does not reward the inventor enough. Research would not be undertaken if there is zero producer surplus and consumer surplus is maximized. Farrell J, Shapiro C. Intellectual Property, Competition, and Information Technology. UC Berkeley Competition Policy Center Working Paper No. CPC04-45. Link
24. Philipson TJ, Jena AB. Who benefits from new medical technologies? Estimates of consumer and producer surpluses for HIV/AIDS drugs. Inform for Health Economics & Policy 2006 Jan 2 (Vol. 9, No. 2). De Gruyter.
26. MalaCards website. Link