The Inflation Reduction Act (IRA), enacted in August 2022, empowers the Centers for Medicare and Medicaid Services (CMS) to negotiate directly with pharmaceutical companies to set the prices for a limited set of high-cost drugs covered by Medicare. However, the law does not detail a process for determining fair prices, other than a consideration for how long the drugs have been on the market.
Historically, prices for drugs in the United States have been disconnected from the clinical benefit they provide. Without this information, health insurers and pharmacy benefit managers have no way of knowing which investments are the most beneficial — which ones improve health the most for the money spent.
What measures can be used to guide pricing negotiations?
There are myriad approaches to assessing a drug’s value, each with its advantages and disadvantages. The options we summarize below all have one thing in common: their intent is not to bring down prices for important drugs to their lowest possible level, but rather to show how the prices paid relate to the benefits the drugs deliver. By using any one, or a combination, of these measures, CMS would be able to pay drug prices that correspond more closely to the health gains that patients can expect. This would not only save purchasers and patients money but also send the right signals to the pharmaceutical industry about what future innovation should look like.
Below we discuss the available options and set context for how CMS might use these metrics.
Summary measures of health
One type of measurement assesses how well a drug extends life or affects quality of life. Such measures provide a single number that represents the health “gained” through the use of that drug. There are a number of such summary measures. Let’s look at each of these in more detail.
Life year (LY). LY measures the additional survival that might be gained from treatment. LY is a straightforward calculation that allows for comparisons across diseases and treatments. It is particularly useful when survival is the most important outcome — as in the case of vaccines for fatal diseases, or for cancer drugs.
Drawbacks: Estimating LY requires robust information on life expectancy with a given disease, and the LY metric does not capture any improvements in quality of life from treatment.
Quality-adjusted life year (QALY). In contrast to LY, QALY adjusts the additional years of life gained with treatment by the quality of life experienced during those years. Quality-of-life measurement is based on survey data estimating the impact of disease on an individual’s ability to function, expressed as a number between 0 (death) and 1 (perfect health). This value is multiplied by LY estimates to generate a QALY. Like LY, QALY can be compared across diseases and treatments.
Drawbacks: The QALY metric has been criticized for undervaluing health gains in severe and disabling diseases. In addition, its straightforward calculation assumes that a QALY carries the same value for every patient. But is a QALY for an elderly individual with multiple diseases the same as a QALY for a child with asthma?
Equal value life years (evLY). A new measure, evLY attempts to address criticisms of QALY by measuring any gains in the length of life equally, according to a single, societal, quality-of-life estimate. For example, an extra year of life for a cancer patient is weighted the same as an extra year for someone recovering from a broken leg, thereby removing any “penalty” from a lower quality-of-life estimate. Questions remain about how to handle evaluations where evLY and QALY estimates differ substantially.
Drawback: With evLY, there is no reward for treatments that improve quality of life more than the competition.
Health years in total (HYT). HYT is calculated as the sum of life years and a modified version of QALY. One of the advantages of HYT is it allows for quality of life and life extension to be weighted according to the values society places on each.
Drawback: There are concerns about whether HYT is biased toward life-extending treatments over those that improve quality of life. As with evLY, further research is needed to better understand what constitutes good value under the HYT construct.
Disability-adjusted life year (DALY). DALY is the result of an effort by the global health community to measure the burden of disease. It’s the sum of years of life lost from premature mortality (for example, early death from vaccine-preventable disease) and years lost to disability (such as the decline in function caused by multiple sclerosis). DALY is measured on an inverted scale relative to the measures above: 0 represents full health and 1 represents death, so the goal is to avoid a DALY. The measure also allows comparisons across diseases and treatments.
Drawback: Like QALY, DALY may underestimate the benefits of treatments for severe and disabling disease, given the greater disability weights associated with life in those disease states.
Value of a statistical life (VSL). An estimate of how much society is willing to pay for small reductions in mortality risk, VSL is used in both health and nonhealth applications such as transportation and environmental regulations.
Drawbacks: VSL has been criticized as potentially overvaluing drugs, and estimates are sensitive to how VSL is calculated.
Measures of clinical benefit
Other measurement approaches focus more narrowly on the comparative benefits and risks of treatments as observed in clinical studies. Let’s look at these measures of clinical benefit in more detail.
Added benefit. Several countries use a measurement system that judges the level of added benefit that may be obtained from a new treatment. In addition to improved survival, this could include intermediate measures like better control of diabetes or reduced blood pressure. The French system categorizes clinically added value on a five-level scale of improvement: major, important, moderate, minor, or no improvement. Each level is tied to a specific reimbursement policy. Scales like this are simple and clear for policy audiences, though they require subjective judgment.
Drawback: These scales do not include any information on costs, which may complicate funding decisions about treatment alternatives with the same level of added benefit. (See the next section for more about the importance of considering costs along with clinical benefit.)
Clinical effectiveness rating. A clinical effectiveness rating extends the concept of added benefit by also considering the strength of evidence for the net health benefit. The U.S. Preventive Services Task Force and the nonprofit Institute for Clinical and Economic Review use letter grades to assess a treatment’s net benefit in relation to an alternative therapy and assign a level of certainty for the finding.
Drawbacks: As with ratings of added benefit, clinical effectiveness ratings are subjective and do not consider information on treatment cost.
Some experts are concerned that summary health measures are too neat and tidy: wrapping up all the effects of a new treatment in one measure is never as simple as it seems. Similarly, measures based on clinical judgment are sometimes seen as too “loose,” lacking sufficient quantification to make truly comparative decisions. Some have argued for measures that combine these approaches, examples of which are described below.
Multicriteria decision analysis (MCDA). Using quantitative and qualitative scores, MCDA considers a variety of health, societal, and contextual criteria. Some of the factors it considers are not captured by the summary measures discussed above, including racial health disparities, treatment novelty, ease of access, and other dimensions important to patients, clinicians, payers, and other stakeholders. Quantitative scoring allows for weighting these factors according to their relative importance.
Drawbacks: The weights used in MCDA may seem arbitrary in some circumstances. Also, the more complex the analytical approach, the more resource-intensive and difficult it becomes to interpret the results.
Social return on investment (SROI). SROI weighs the total present social value of a treatment’s benefits against the investment. Because both social value and investment are expressed in monetary terms (for example, $4 of social value for every $1 invested), it’s necessary to use proxies to measure social value. For instance, reduced health system spending on emergency-room or hospital care might act as a proxy for improvement in a patient’s ability to manage his or her illness. SROI has the advantage of being a simple and easy-to-understand calculation expressed entirely in monetary terms.
Drawbacks: The SROI measure has historically been used for large programs in public health and other sectors, so its utility in drug pricing negotiation is unknown. In addition, the selection of financial proxies is a subjective exercise.
How are measures used to understand value, and how can CMS use them?
We’ve described a number of measures that can be used in drug price negotiation and discussed their advantages and disadvantages. (You might also find this side-by-side comparison helpful; it contains the various formulas used to calculate each measure.)
But how would CMS implement these measures? Let’s first consider a hypothetical scenario. CMS flags the fictional drug Outstandavil for negotiation based on its contribution to Medicare Part D (nonhospital) drug spending. As stipulated in the Inflation Reduction Act, the drug’s manufacturer must submit information on Outstandavil’s current pricing, nonfederal average manufacturer price, and other considerations. CMS uses one of the measures above to determine that a fair price is $78,000 per year, versus the $195,000 the government currently pays.
The manufacturer has 30 days to accept the proposal or counter the offer, describing any mitigating circumstances, such as a new indication for the drug. In this scenario, CMS agrees with some of the new data submitted by the manufacturer, and a new fair price is calculated to be $110,000. That is still a 44 percent discount from the current price. If 50,000 Medicare beneficiaries were receiving this drug, CMS would realize $4.25 billion in savings in a single year.
But how does CMS actually arrive at a fair price? The answer depends on the measure used. Summary measures are typically employed in cost-effectiveness analysis, which is expressed as a ratio of the difference in costs to the difference in health between the new drug and a comparative drug. Specifically, the analysis might look at the cost per quality-adjusted life year (QALY) gained with the new drug. Alternatively, the analysis might look at the cost per disability-adjusted life year (DALY) averted with the new drug or use one of the other summary measures described above.
CMS could examine the drug’s cost-effectiveness from available reports, compare the current price to the price that meets a common benchmark for cost-effectiveness (such as $100,000 per QALY gained), consider the ceiling prices stipulated in the Inflation Reduction Act, and pick the lowest price of the three.
The value of a statistical life (VSL) measure is used somewhat differently in another type of analysis known as benefit-cost analysis. Both benefits and costs are expressed in monetary terms, so any ratio of benefits to costs greater than 1 is considered a good deal. In this case, CMS could consider a set of drugs that have comparable benefit-cost ratios and pay the lowest available price or the ceiling price, whichever is lower.
The clinical measures would require a slightly different approach. Using the added-benefit approach, CMS might agree to pay a premium price for a drug with a demonstrated “major” or “important” health gain, pay the reference price of a comparable drug for a moderate gain, and pay the lowest available price for a drug with a minor gain or no improvement.
A clinical effectiveness rating system could work in a similar fashion, but additional protections could be put in place when the level of certainty is low or moderate. For example, a “coverage with evidence development” approach, starting with a reference price while new evidence is being generated, might be implemented. As additional evidence is generated, payment levels could be increased, assuming there’s stronger confidence in the level of benefit a drug offers.
Implementing the multiple-measures approach is somewhat less certain, if only because the threshold for what is considered a “good” MCDA score or substantial SROI is not well understood. Still, there is an opportunity for CMS to explore how these results compare across its list of target drugs.
What’s a feasible approach for CMS to implement?
Regardless of the approach taken, one thing is certain: assessing drug value requires a skill set that the Centers for Medicare and Medicaid Services currently lacks, so new staff will need to be brought on board. Some of the approaches we’ve described, like MCDA, would require even more resources and interpretation. And while the summary health measures are attractive for their relative simplicity, the politics of their implementation may be too challenging to navigate.
The measures of clinical benefit are likely the most feasible to implement in the short term. The categories are intuitive and closer to the “medical necessity” rules that public and private payers have long used to make coverage determinations.
Regardless of the approach taken, the government will have to make subjective judgments to a certain extent. Care will be required to ensure the process is transparent and understandable for all stakeholders.