Missed Opportunities in Medicare’s Drug Price Negotiations
By using a flawed measure of prescription drug costs, CMS failed to tackle high-priced medicines.
When the Inflation Reduction Act (IRA) became law in September 2022, President Biden promised that long-awaited prescription drug reforms to Medicare would save more than $200 billion over ten years. Backing up his claim, the Congressional Budget Office projected IRA’s prescription drug provisions among the largest deficit reduction measures in the bill at $230 billion. Through 2031, two provisions — a cap on drug price increases for all Medicare-reimbursed drugs to inflation, and the drug price negotiation program — were expected to save the federal government more than $56.3 billion and $98.5 billion, respectively.
While the Centers for Medicare and Medicaid Services (CMS) began collecting inflation rebates this year, the drug price negotiation program begins in 2026. In late August, CMS announced the first ten drugs selected for negotiation. Already, most companies directly affected by drug selection have sued, under the chief argument that lower prices from the program represent an unconstitutional seizure of the companies’ property in violation of the Fifth Amendment.
As indicated in statute, CMS selected among the 50 most expensive drugs paid by Medicare in the Part D prescription drug program, ordered by gross expenditures, from June 2022 to May 2023. CMS touted that the ten drugs represent over $50 billion in Medicare drug spending, implying a large impact on federal spending and company revenue alike.
Due to the complexities of drug supply chains and existing federal law, insurance plans and government payers ultimately pay less than the gross — or “list”— price for a drug. But rather than paying a discounted, negotiated price up front, payers pay the gross price and negotiate to receive rebates from drug companies afterward. After accounting for all rebates and discounts, the total amount paid represents the net price. Depending on how high the company sets its list price and the size of rebates negotiated, a payer may end up paying a net price that is significantly lower than the drug’s list price.
If the goal is to maximize deficit reduction, it is surprising that CMS selected drugs for negotiation based on the gross rather than the net amount Medicare pays for such drugs. But the decision is even more surprising, given that the IRA never intended gross prices to be used at all.
Small change, big effect
Had Congress intended to use gross drug spending as a means to rank the top drugs eligible for negotiation, it could have explicitly said so by adding a definition for gross spending in the IRA. Instead, the IRA references the term “total gross covered prescription drug costs” in Section 1860D of the Social Security Act to describe how the highest spend drugs will be ordered for selection. While it might seem like the IRA is using gross spending by referencing “total gross covered prescription drug costs,” regulation further clarified this term to mean costs “actually paid” by a Medicare Part D prescription drug plan; that is, net costs actually incurred by the Part D plan after all discounts, rebates, or other price concessions are received and applied after the point of sale. Therefore, at the time the IRA was enacted, Congress intended for CMS to choose among Medicare’s highest spend drugs based on net spending.
However, on December 27, 2022, CMS promulgated a rule that, among other things, sought to change the interpretation of “gross covered prescription drug costs” for the purposes of the Medicare drug price negotiation program. In it, CMS removed the phrase “actually paid” from regulation. As a result, CMS changed the IRA’s drug selection process by listing eligible drugs for negotiation based on gross, rather than net, spending. In so doing, the process became easier for CMS officials to select drugs for negotiation. Though public data on gross drug expenditures is only available through 2021, the change also makes it easier for outside analysts to predict which drugs will be selected.
In addition, the change is beneficial if drug companies respond by lowering list prices. Drug manufacturers commonly set high list prices, along with large rebates, in order to get better placement on drug formularies used by insurance companies.
With better formulary placement, drug companies gain more customers because patients pay lower copays and therefore are less price-sensitive to expensive drugs. Lower list prices brought on by the IRA’s implementation would short-circuit this strategy, leaving drug companies with less ability to manipulate formulary placement to increase market share.
At the patient level, lower list prices lead to lower drug spending by insured individuals generally, and specifically for patients who have plans with higher drug cost sharing or for those who are uninsured.
Unfortunately, evidence suggests drug companies continued to raise list prices even when a company’s drug was likely to be selected for the Medicare drug price negotiation program. For the ten drugs selected for 2026, the companies raised list prices by an average of five percent for 2023 compared to the previous year.
Rather than achieving the desired effect of reining in list prices, CMS instead reduced the savings achievable through the program, a result Congress did not intend nor the CBO could have predicted.
The regulatory change could result in less drug savings, both for patients and taxpayers, in two ways. First, CMS may fail to select drugs that have smaller rebates and discounts, since these may fall farther down the list of drugs ordered by gross expenditures but would rank higher based on net prices.
To illustrate the first scenario, consider CMS’ selection of Novolog (insulin aspart), the short-acting insulin marketed by Novo Nordisk. Its selection was based on gross expenditures of $2.6 billion during the 12-month window, a similar amount as other drugs selected including Enbrel (etanercept), Imbruvica (ibrutinib), and Stelara (ustekinumab). However, we estimate Medicare spent only $310 million net on Novolog in the 12-month period that CMS used to select drugs for negotiation. If the selection process was based on net spending, CMS would have selected Xtandi (enzalutamide), a cancer drug marketed jointly by Astellas and Pfizer. (Medicare spent a estimated net amount of almost $2 billion on the drug.) Had this occurred, CMS would likely have saved $300 million or more in 2026 alone—roughly the same amount CMS spent on Novolog in total.
The second way ranking eligible drugs by gross spending could result in less savings is related to the way drug companies continually raise list prices over time. Higher gross spending is often a reflection that a drug has enjoyed monopoly protection for many years as the drug manufacturer raised the list price over time. Thus, a drug with high gross spending is more likely to encounter generic competition sooner because it has been on the market longer. And once generic competition arrives, the drug is dropped from the negotiation program and is not replaced by another, resulting in forgone savings on brand-name drugs.
The second scenario is obvious when one considers the selected drug Stelara, the autoimmune treatment made by Johnson & Johnson. Analysts expect Stelara to remain a negotiated drug for 2026 only. By contrast, if Xtandi had been a selected drug rather than Stelara, the savings on Xtandi would be expected to reach $1 billion more than Stelara from 2026–2029; not because the savings per year would be significantly greater than Stelara, but because Xtandi is expected to have lower negotiated prices three years longer before a marketed generic forces the drug to be dropped from the negotiation program.
Potentially billions in savings lost
In the aggregate, we estimate savings in 2026 from the drug price negotiation program to reach $4.1 billion, with nearly one-third of the savings on Bristol Myers Squibb’s blood thinner Eliquis (apixaban) alone. This figure is more than $700 million less than the CBO originally projected, a difference of 15 percent.
On the other hand, if CMS ranked negotiation-eligible drugs based on net spending, we believe that, among the ten drugs selected for 2026, Novolog and AstraZeneca’s Farxiga (dapagliflozin) would be replaced with two drugs with the highest estimated net spending that were not included for 2026: Xtandi and Bristol Myers Squibb’s Pomalyst (pomalidomide). From this change, CMS would save an additional $480 million in net drug spending for 2026 alone. Furthermore, based on the law’s requirements that a drug is removed from the negotiated list following a generic version’s launch, Xtandi and Pomalyst would remain on the negotiated drug list at least through 2029 and 2027, respectively. At the same time, sales for Novolog and Farxiga are expected to decline further as generic alternatives hit the market and patients switch to newer insulin treatments.
The reduced savings in the drug negotiation program’s first year reveals two important facts. First, Medicare beneficiaries and taxpayers can expect lower savings than the CBO projected, both near- and long-term. By ranking drugs based on gross spending, CMS will likely select drugs with high rebates that limit additional savings and experience generic competition within one or two years of negotiated prices going into effect.
The second fact is drug companies exaggerate when they claim the IRA is the death knell for the industry. According to Bloomberg analysts, negotiated prices for nine out of the ten drugs selected for 2026 will result in revenue losses of less than or equal to one percent for the companies involved. Only discounted prices negotiated for Eliquis are expected to impact drug maker Bristol Myers Squibb more, though negotiations are still expected to reduce the company’s revenues by only three percent.
Furthermore, our research indicates that revenue losses among large pharmaceutical companies often results in negligible impact to the number of innovative drugs developed. For example, in a scenario resulting in 2.8 percent revenue loss at Bristol Myers Squibb — similar to projected losses on Eliquis under the IRA — we estimate a resulting decline of only 0.22–0.33 new drugs developed.
Drug companies argue that attempts to curtail high drug prices will negatively affect the development of new cures. But given the structure of the IRA’s Medicare drug price negotiation program — and the regulatory decision to choose drugs based on gross spending — the evidence suggests such claims are overblown.
