Amgen's Comments on ICER’s Value-Based Price Of A Cure Collaborative: Methodological Considerations

Amgen appreciates the opportunity to comment on ICER’s review of methods for valuing cures in healthcare. Progress in treating diseases with transformative therapies inevitably presents the challenge of how to value these advances. Although cures are still quite rare, recent successes suggest the time to expand upon traditional valuation methods into the cure space is now and we look forward to a productive dialog in this area.

We believe capturing the full value of a cure requires a rethinking of many elements of value that are often inaccurately, incompletely, or not captured in traditional health technology assessment (HTA).  The enduring nature of a cure not only for today’s patients but for future generations, demands that we derive outcomes which more comprehensively and dynamically reflect societal preferences than the Quality-Adjusted Life Year (QALY), with its inherent limitations, as currently applied by many HTAs.  This initiative presents an opportunity to consider expanded societal value and propose new concepts of societal equity, inclusive of all current participants in the healthcare system as well as future generations.  The goal then becomes a long-term, strategic realization of full health for more people, over time: a true generational perspective.

Across all of society, even the best partial treatments for disease pale in comparison to the value of a cure and the elimination of long-term ill health.  Accordingly, there is a need to translate the value of reshaping entire diseases into metrics that will fully capture the value of cures, which in turn will provide incentives to prioritize the higher effort (and much higher long-term rewards) of such progress.  In this capacity, we recommend ICER:

• Capture all types of cures when applying cure-specific methods.  Disease cures come in all shapes and sizes including functional cures (disease held at bay but not “eradicated”), cures that only cure a proportion of a population, cures with very promising but unknown durability (long tail), cures that replace a less desirable cure that came before it, and potentially transformative areas of medicine such as gene therapy where data are particularly sparse, but reward is also potentially extremely high.
• Endorse health economic methods for cures that will capture the dramatic and transformational nature of a cures in three critical ways:
  1. Go beyond “typical” health economic analysis to capture the transformational benefit of a cure on today’s patients.  Such methods should tolerate more uncertainty and go beyond the QALY to capture unique and sometimes hard to quantify benefits, perhaps using multi-criteria decision analysis, or other alternatives to QALYs.  Disease severity, deadliness, and evidence of “curability” (diseases with promising signals and signposts pointing to cure) might also figure into the value of a cure to today’s patient.  Evidence of curability such as empiric, incremental, and painstaking progress attained combining existing non-curative treatments might be used to justify tolerance of more uncertainty in some disease areas when extrapolating beyond the available data.
  2. Assess value from a societal and generational perspective and fairly value cures for future patients: Future patients' needs are not often accurately considered in economic analysis and even if they are, they are discounted away. But even if we disregard equity for future patients, since the primary value of a cure is for people who may not yet have the disease, and a huge beneficiary of a cure is society itself, cures should not be subject to standard discount rate approaches. The health care system already has too many financial disincentives for considering long-term outcomes. The health care system is not a single person looking to maximize utility, but a construct that should be seeking long term sustainability and preservation of resources for future patients. Typical methods that overly discount future illness and costs is counter to what cures should be accomplishing and puts them at a remarkable disadvantage compared to short-term fixes. For instance, some countries such as France and the United Kingdom have adopted discount rates that decline over time to address the shortcomings of standard discounting methodology¹. The methodology used should be thoughtful and appropriate for long-term investment decisions that are likely to impact generations to come.
  3. Endorse more dynamic health economic methods that accurately look at costs vs benefits over long periods of time.  The backbone of innovation is the patent system which plays out over decades.  But too often, the analysis period ignores the recurring benefit to society of the post-patent period where most of the benefit may be obtained.  Diseases cures, which seem to create durable benefits across generations, are particularly disadvantaged by these distortions.
• Maintain value and affordability as separate constructs.  To accurately calculate value, it must be assessed outside of the influence of affordability.  Separate mechanisms are available to address affordability. (Further addressed under ‘Additional Considerations’)  

We provide more detailed input below, structured as general and methodological considerations.  

GENERAL CONSIDERATIONS

Curing disease encompasses a diverse spectrum of treatment types and potential outcomes, therefore, a broad and flexible definition of cure should be considered.  The classic definition of a cure might be a single pill or injection that eliminates a serious disease forever.  But the definition of cure should not be limited to such rare types of cures since this might discourage the search for less “classic” but potentially even more valuable treatments.  A noteworthy example is antibiotics.  These exceptional compounds could be thought of as treatments for infection at the patient level, but in fact, their real value has been (so far) in curing most common infections in most people thereby preventing death, over a nearly 100-year period.  Most HE valuation techniques that look at new antibiotic classes simply value the class as a replacement for a current class, in today’s patients, with stagnant projections of infectious disease virulence.  Reinvention of the valuation techniques for this class would more comprehensively value their ability to maintain known curative benefits.² 

Cures take many forms, have varying impacts on people and society, and evolve over time. It is therefore more important to focus on the potential to attain benefits proportional to costs than on the “shape” of the cure itself.  Also, the definition of a cure should be tolerant of considerable variation on the classic idea of a cure such as (detailed in Appendix):

• Functional Cures. Cures are often not “one and done.”  Cures may need to be taken once, intermittently, or for a lifetime.
• Uncertain (long-tail) Cure. Sometimes the best measures of a disease suggest that a disease is cured, but there is still uncertainty.  Nevertheless, in medicine, long tails tend to get longer.
• New and Improved Cures.  A disease might already be cured.  But cures have variable tolerability.  A cure that is challenging to tolerate may not be accessible to all patients that may need it, making improved cures highly valuable.
• Partial Cure.  A cure may only work for a proportion of the population.  But, improvement in the cure rate is often incrementally improving.  Assumptions about the progress toward cure for rapidly advancing diseases is important to valuing the anticipated future state.
• Reinvented Cures.  Cures are dynamic because disease may be dynamic. Cures may need to be reinvented, refreshed, or updated, as in the antibiotic example.

Cures are the ‘holy grail’ of innovation.  The cost of not curing disease is years or decades of chronic, suboptimal treatments accompanied by both economic and intangible cost burdens to the system for current and future people with the disease and for their caregivers. Consideration of cures therefore demands a generational perspective. A cure for society, allows children of all succeeding generations to grow up without being at risk of severe morbidity or death from disease.  This outcome is a transformational, perpetual, and (often eventually) nearly free societal good.³  Unfortunately, developing a curative treatment is still an exceptional event; the current standard largely consists of chronic and / or symptomatic therapy that incompletely eliminates the morbidity and/or mortality of disease. There are an estimated 20,000 diseases today; most do not have cures and many lack treatments.⁴ It is therefore not surprising that health economic methods have not evolved to address transformation of healthcare (vs simply adding up utility and costs for today’s potential outcomes).⁵ Valuation of cures must at least reflect the base case perspective as generational to avoid the massive discounting of future benefit for future health care recipients, reflected by the unwavering demand by society that the end goal of medicine ultimately be elimination of disease.

Valuing a cure or potential cure requires that health economists and policy makers become more comfortable with the uncertainty of modeling the long-term impact of relatively new curative treatments.  Such uncertainly even extends to whether the new technology in question is even a cure (yet).  As with many things, it may very well be the conflict between short-term certainty and long-term promise that derails future cures.  A good example of this uncertainty is evolving in the field of oncology.  At present, some types of cancer are curable, but the cures have revealed themselves only after painstaking trial and error and continuing experimentation.  The cure for Hodgkin’s disease is one such example where cure for many people was “discovered” over the course of many years with multi-drug treatments.  Such cures as they are happening might be called dynamic cures, and therefore valuation requires a more dynamic cost effectiveness methodology.  Some treatments such as antiretroviral therapy (ART) for HIV become essentially functional cures that also evolved dynamically, with much clinical experience supplementing the body of randomized trials.   Multiple myeloma may be on a similar trajectory today.  In short, for many diseases, there was no definitive RCT or publication in which suddenly a cure was bestowed on society.  Evaluation of diseases that are “In the midst of being cured” may therefore be the most difficult aspect of properly valuing treatments progress is initially and erroneously viewed as incremental, but in fact a tipping point is reached that firmly differentiates a new treatment from an older one.  A promising pathway to cure may therefore be cause for increased future valuation.^{6, 7}

METHODOLOGICAL RECOMMENDATIONS

1. How should value-based prices for potential cures reflect substantial uncertainty regarding clinical safety and effectiveness due to limitations in study design, outcome measures, and the size and duration of clinical trials?

Uncertainty will always be greatest for the most ground-breaking therapies just entering the market, especially for curative therapies with small population sizes and clinical trial limitations.  This is often the case in the rare disease setting, where the very nature of the disease results in limited clinical trial data with many uncertainties. However, the nature of breakthrough therapies necessitates that the FDA and other regulatory agencies deem it is in the public interest to approve the treatment with the consideration that without it, there is greater risk for harm to patients and society. Rather than re-adjudicate the value of trials, new methods for valuing cures should tolerate more uncertainty than might normally be the case, and should focus on the best ways to extrapolate trial results into the future (without arbitrarily underestimating future outcomes for future generations), by acknowledging signposts of potential future medical value, given our introductory remarks.

The value-based price for a cure at launch should fairly reflect the full value-proposition with respect to the given indication for the life of a product, both on and off patent, allowing market mechanisms to potentially discount from a “fully valued” baseline.  Full valuation of potential cures 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 cures: future outcomes.  That said, once a policy group values a cure, it may be that societal tendencies still manifest in the market for products.  Simple market mechanisms like competition among curative multiple treatments will often drive the cost well below the actual value, as evidenced by HCV treatments.

Once a fully valued price for a cure is known, payers can negotiate based on limitations in evidence for their own plans, but the negotiation should start with a level playing field.  Innovative contracting and cost structuring, including increasingly prevalent risk share agreements and outcomes based agreements, could also be used to share risk between healthcare market participants.⁸  Manufacturers could also remove some of the uncertainty in the fully valued price by addressing the durability of potential cures through the collection of real-world evidence to support a therapy’s value proposition and by working with payers in collaboration on coverage with evidence development.  Adjustments to the original value-based price could and should occur over time as the value evolves, reflecting dynamic cost-effectiveness influenced by real-world evidence and greater clinical understanding of a product’s unique place in the care continuum. 

Potentially curative therapies offer, by definition, higher reward.  They should not be penalized due to higher uncertainty which might otherwise lead to undervaluation in a standard cost-effectiveness analysis with conservative assumptions.  For example, in ICER’s recent assessment of Voretigene Neparvovec (VN), ICER assumed, without evidence,  that VN would only have a 10-year treatment effect in the base case analysis; VN’s efficacy was assumed to wane over the subsequent 10 years.⁹  The difference between this efficacy profile versus a lifetime durability was significant. Specifically, in older patients the incremental cost-effectiveness (from a health care system perspective) was $643K/QALY (ICER’s base case) versus $385K/QALY (for lifetime duration). Incorporating assumptions to offset uncertainty, particularly in ultra-rare diseases with very limited data can result in a bias toward investment in therapies which address short-term endpoints, which is opposite to the intent of valuing cures.¹⁰

The incomplete capture of treatment benefits to generations of patients and society is a fundamental limitation of health technology assessment and one that is further amplified with curative therapies.  In curative therapies, the benefits and costs-offsets can be of an order of magnitude that eclipses some chronic therapies, yet they are undervalued by the tools used to assess them.  For example, some HTA’s initially rated sofosbuvir-based cures for Hepatitis C as low value despite the treatment’s step change in outcomes with a greater than 98% cure rate.^{11, 12}  The majority of the value of these agents is likely to accrue to future generations, as HCV is increasingly eradicated from the human population, mandating a dynamic analysis of curative benefit.  A systematic analysis of all published studies in HCV in which such benefits were often more generously applied concluded that second generation DAA’s are cost-effective at any price at or below $227,000 and that they save money in the U.S. at any price less than $79,000.¹³  The unprecedented impact of this cure combined with significant favorable cost-effectiveness using more dynamic and intergenerational analysis has generally not been reflected in the HTA of these molecules.  As a result, these drugs may be being used in a suboptimal and more costly manner than if society recognized the potential additional benefits of disease eradication (value of preemptive elimination of disease).

2. How should value-based prices for potential cures reflect uncertainty regarding inclusion of additional elements of value that may be important for potential cures, but which are not part of standard cost-effectiveness methods?

The valuation of a cure should capture elements that are beyond a traditional cost-effectiveness analysis and most relevant to patients and society.  Stakeholders (including patients, manufacturers, opinion leaders) should be engaged early, before the assessment is framed, to help identify the most important elements specific to that assessment. Potential elements of value include, for example pursuit of equity, value of hope, disease severity and reduction in uncertainty, altruistic value and insurance value, all captured in Figure 1.^{14, 15}  In addition, assessments of cures like other treatments should reflect society’s preferences on value and include the societal perspective with a range of options, ensuring inclusion of cost offsets and difficult to monetize benefits such as reduced patient and caregiver burden, productivity gains and losses and other indirect and non-medical costs.

FIGURE 1: ELEMENTS OF VALUE & THE COST PER QALY

3.  How should value-based prices for potential cures reflect extreme magnitudes of lifetime health gains and cost offsets that are far beyond those generated by traditional therapies?

New tools should allow comparison of cures and non-cures on a level playing field. Some may argue that the generational approach and tolerating future uncertainty will result in very high valuations for cures that may surprise some stakeholders and even "crowd out" some current health care spending, but this is, in fact, a desirable outcome of a comparative exercise. Cures are, in fact, still very rare. Their valuation necessitates some methodological innovation to account for long-term generational effects that may represent monumental step changes over chronic therapies. Ignored or reduced by methodologists, these effects might otherwise not be considered when looking at spending resources on curing vs merely treating disease. Given the rarity of cures, the danger is, in fact, that cures will be crowded out by "medicine as usual" if the full value of a cure is not articulated.

Valuation of curative treatments requires rethinking the QALY as currently used to measure and then usually discount future benefits. Shown in Figure 1, QALYs have some major and well-documented shortcomings as they relate to capturing the full impact of cures. Likewise, the temptation to discount the gains of cures (simply because there is so much more gain) is exactly the opposite direction the field of health economic should be taking. The use of modelling of long-term benefits that take arbitrarily conservative or negative assumptions on areas such as treatment durability render a disservice to the valuation of cures. The effort to value cures should not start with an immediate discount; the resultant systematic distortion compared to "standard" medicine will result in immediate internal bias and an inability to compare a cure to a non-cure.  

The appropriateness of discounting of future outcomes for future patients for curative therapies should be seriously considered. At a minimum, methodologies that are specific to intergenerational investment should be used.   Traditional HTA typically discounts all outcomes of a new treatment occurring after one year: the farther in the future a benefit occurs, the more discounting brings the “current” benefit to zero, with a severe impact on treatments having the greatest promise of long-term and even intergenerational benefits. This treatment of costs and outcomes is standard in economics and may be justified when valuing health care options for today’s “economic rational actors” making choices for themselves.  But blanket discounting is problematic in curative treatments because it minimizes the impact of remarkable step changes and does not capture the increased magnitude of cost savings and improved benefits which extend for decades. In fact, many prominent economists are making the distinction between intragenerational discounting and intergenerational discounting and are applying different discounting methodologies for these intergenerational decisions, such as investment to improve environmental problems and other investments with long time horizons. ^{16, 17, 18} It is recognized that typical discounting techniques (especially of QALY-like gains for future patients) suppresses important societal objectives (cures) by overweighting the preferences of people alive and with disease right now, as shown in the Case Study in Figure 2. A sense of equity suggests that the present generation should not be allowed to discount the gains of future generations who cannot speak for themselves, even if their own benefits are subject to discounting. ¹⁹

Because of the limitations of QALYs, especially regarding the accurate ascertainment of hard to measure benefits to both patients and society, evaluation of cures should be evaluated using various alternatives to the QALY. Many efforts are underway to identify alternative approaches to the QALY, for example, the ISPOR Special Task Force on US Value Assessment Frameworks has catalyzed efforts to better characterize value and the findings of these efforts should be considered as part of this initiative.²³ We note ICER's recent addition of the equal value of life years gained (evLYG) to their assessments as an adjunct to the QALY,²⁴ and would like to understand if this metric may be applied productively in the curative setting.²⁵ In addition, ICER should consider other areas such as further development of multi-criteria decision-analysis (MCDA) and/or augmented or extended cost-effectiveness analysis (CEA) and the Burden Augmented by Deadliness and Impact.^{26, 27} All of these techniques are meant to more comprehensively capture aspects of value that include health and non-health benefits such as wider public health effects, positive net tax flow, distribution of health, stimulation of medical innovation, peace of mind and increased macroeconomic growth.

Structural assumptions in HTA and specifically, the long-term modelling of cures is an important area for innovation in valuation. ICER has worked on innovations with collaborators most recently in the valuation of the SMA gene therapy, which uses interesting assumptions on the impact of this potential breakthrough therapy for SMA Type 1 patients. Here ICER projected Phase 1 trial data with the assumptions that patients remain for a lifetime in the same health state after the completion of the clinical trial.²⁸ Likewise, durability in CAR-T treatment involved methodological innovations in oncology that allowed for cures to be more practically defined at 24 months when the treatment mortality was similar to that of a sex- and age-adjusted general population; likewise, ICER working with collaborators identified new ways to project survival. ²⁹

ADDITIONAL CONSIDERATIONS

Value vs. Affordability

Affordability is a separate and distinct construct from ‘value’, and experience shows that perceptions of affordability based solely on price may need to be managed.   Cures are, in fact, still very rare and paying for them even at “accepted”, undiscounted levels of QALY valuation represents a small part of health care budgets.  Given the scarcity of available curative therapies, affordability should not yet present an issue when viewing cures within the health system holistically.  However, as noted above, the sheer magnitude of potentially extreme gains coming from cures may result in sensible defensible valuations and prices that are nonetheless perceived as very high by society.  Unmanaged and uninformed, the prices themselves may be used to intuit value, discarding the actual evidence of value (see Sovaldi example).

When cures consistently dominate most other health care choices, the system is more likely to pivot from our current “break and fix continual maintenance” paradigm to one that fully values “predict, prevent and cure”.  (Prevention is beyond the scope of this response but is tantamount to pre-emptive cure.  This reframing would reveal the vast underinvestment and undervaluation in preventive medicine, which, like cures, fare poorly using current short-term health economic valuation techniques.)

As noted previously, where there are perceived affordability challenges, they will be addressed by the market with innovative financing mechanisms, whose development must be accelerated ^{30, 31}  In addition, competition drives intrinsic market-based mechanisms for long term affordability. Devaluation of cures should not begin with the scientifically based groups that are trying to empower society to make better decisions.

Value Framework Updates

ICER has noted that the value of cures initiative will inform their broader framework methods updates.  We urge ICER to reconsider the ambitious timeline for this initiative to enable more thoughtful dialogue and productive collaboration. Additional time for consideration will also allow greater clarity into how the valuation of potential cures should align with evolving US and ex-US value frameworks and HTA methodologies. We would also be interested in how such issues will be reflected in National Institute for Health and Care Excellence (NICE)'s Methods Update as well as HTA methods applied by the Canadian Agency for Drugs and Technologies (CADTH).

CONCLUSIONS

Cures, because they are so rare, need to be recognized for the life changing value they bring to patients, caregivers and society. A wide variety of potential therapy types, patient populations, and treatment circumstances contribute to the complexities of defining what constitutes a cure. A cure or potential cure should be defined as a treatment or therapy that is likely to leave individuals today and in the foreseeable future functionally free from disease and/or harmful genetic mutation for a clinically meaningful period. Further, there may be scientific basis for suggesting that such cures will become more attainable for more people in the future. Our input on this initiative focuses on three main recommendations. 1). Ensure a broad, inclusive and flexible cure definition, characterized by mortality and quality of life for present and future patients approaching that of the general population as the ultimate objective. 2). Endorse dynamic methods that capture value fairly over time and for future generations (e.g., not applying standard discount rates), and 3). Assess value outside of affordability. Also, treatments that offer hope to extend or improve life (or allow the opportunity for a cure) are also important and should not necessarily be valued 'less than a cure', especially when one does not exist. Lastly, we emphasize the importance of getting this right to ensure optimal investment in health impacting generations to come.
 

REFERENCES

APPENDIX

A. The Wide Spectrum of Cures Influencing Their Definition

• Cures are often not a “one and done.”  Cures may need to be taken once, intermittently, or for a lifetime.  For example, Gleevec offered a significant breakthrough in the treatment of leukemia but must be taken for a lifetime to be effective.  In contrast, gene therapy or CAR-T is typically given only once.  The duration of the treatment should not be material to the definition of a cure, but a curative state should be achieved even with a continuous treatment.

• If the cure is challenging to tolerate, it’s not really a cure applicable to all patients that may need it.  For example, while beta interferon might have been considered a cure for a subsection of patients with HCV, the side effects of the treatment were poorly tolerated by patients.  Only the introduction of Sovaldi allowed a majority of patients with HCV (and their physicians) to consider themselves candidates to be cures.

• Partial cures are still cures for many people now, and likely more in the future: For example, in oncology, some patients can be cured by a treatment however, this may be a function of tumor pathology, time with disease and other prognostic factors.   The fact that not all people can be cured should not discount the value placed on attainment of some cures as a signpost for eventual cure for more and more patients.  Science has shown that sporadic or partial cure is usually a stage that must be attained before complete cure is possible.

• As noted in the anti-infective and anesthesia examples above, cures may need to be reinvented, refreshed, or updated.  Put simply, valuation of cures cannot assume that diseases will stay cured forever using current techniques.

• Some potential cures are truly “uncharted waters” but the penalty for undervaluation of the most promising science is likely higher than assuming zero future benefits: For example, spinal muscular atrophy (SMA) gene therapy currently under evaluation by the FDA suggests it could potentially be a cure, however, durability for gene therapies are unknown and it is impossible to run a RCT forever until we are sure.  Unfortunately another “black swan” could be the emergence of a terrible side effect for these treatments.  As noted above, the value of cures means we will need to use our best clinical judgement as to how we extrapolate current evidence into the future.  Also, as for the cure for Hodgkin’s Disease, uncertainly also means that continuous data collection and use of empiric evidence is also critically important and will require continuous reevaluation of the disease area under question.
   

B. Anesthetics: A Fundamental Enabler to Surgical Care

An illustrative example of cures is anesthetics accompanying surgery.  Prior to anesthesia, surgery was a last-ditch effort to prevent death, and often resulted in death of the patient from the surgery itself.  Anesthesia is therefore a “cure for near-certain surgical death” that has allowed the evolution of surgery to be the life-saving (and often curative) intervention that it is today.  But unlike anti-infectious agents, there is little evidence that current anesthetics are losing their effectiveness or need to be reinvented to maintain most of their positive effects.  The low investment (and valuation) for new anesthetics could therefore potentially be a rational long-term decision for society, unlike the potential short sightedness of under investment in maintaining the curative power of anti-infectives.  New techniques for valuing cures should be capable of making such comparisons to allow better societal decision making for health, and better health care policy decision making for society.