Evercore ISI Addresses Some IRA Issues, Unveils List of Potential Negotiated Drugs

On Sept. 1, CMS will publish its list of 10 Medicare Part D drugs that it has chosen for negotiations for 2026, the first year that negotiated prices will be in effect via the Inflation Reduction Act (IRA). But a lot of unknowns exist around the law, including how CMS will determine the drugs up for negotiation. During an Evercore ISI webinar held Jan. 30, analysts broke down this aspect of the IRA and shared the drugs and manufacturers they expect to be impacted.

“We are looking at an escalating number of drugs subject to negotiation starting in 2026,” noted Tobin Marcus, policy analyst at Evercore ISI. “There’s been some misunderstanding about the fact that these numbers are cumulative.” So while 2026 will have 10 drugs up for negotiation, 2027 will have an additional 15 Part D drugs for a total of 25 for the year. The following year will see an additional 15 Part D and Part B drugs, and then starting in 2029 and later years, 20 more Part D and Part B agents will be up for negotiation.

The products to be negotiated initially will be chosen from the 50 drugs with the highest Medicare Part D spending, followed by the 50 drugs with the highest total Medicare Part B spend. The law requires the HHS secretary to negotiate; it’s not an option.

“How much it looks like a real negotiation I think is a little bit of an open question,” remarked Marcus.

Evercore ISI analyst Umer Raffat explained that the list of IRA drugs to be negotiated is not random but rank ordered, “so you have to really construct it across industry. You can’t just say, ‘these are the drugs that I’m seeing because these are the four companies I’m looking at.’ You have to do it across industry to rank order it.”

Price ceilings exist on the maximum fair price, noted Marcus. For drugs nine to 12 years beyond their initial approval, it’s 75% of non-Federal Average Manufacturer Price (non-FAMP). For drugs 12 to 16 years post initial approval, it’s 65% of non-FAMP. Beyond 16 years, it’s 40%.

“I think the big open question here on what we’re actually looking at is how far, if at all, below those minimum discounts is CMS going to go in trying to negotiate these price caps?” he asked. “The negotiation process is very rigidly structured. CMS makes one offer, the manufacturer makes a counteroffer, and then CMS can counter the counteroffer, and then that’s the last word. And so it certainly is possible that they will try and push these much lower, and we’ve been hearing some noises that they are intending to be aggressive.”

That said, “I don’t think that they’re just going to be able to arbitrarily lowball manufacturers,” stated Marcus. “They’re going to be trying to follow some actual framework around value. But ultimately, whatever they do land on, they have this extremely high excise tax to compel compliance, and the law essentially prevents any administrative or judicial review of those prices and of the selection to be negotiated in the first place.”

CMS will send its initial offers to manufacturers on Feb. 1, 2024, making it “a really important day to watch to see exactly how deep discounts they’re seeking, which has not gotten quite as much attention as the” Sept. 1, 2023, date when the drugs up for negotiation are revealed or the Sept. 1, 2024, date when CMS publishes the final prices for those drugs.

Marcus told attendees that he gets a lot of questions on the chances of getting changes made to the IRA, but “for better or for worse, we expect the IRA drug pricing reforms to be implemented largely as written.”

Will CMS Use Net or Gross Prices?

CMS will use drug prices from June 2022 to May 2023 to determine how it is going to rank order drugs that are going to be negotiated, clarified Marcus. On whether CMS will use drugs’ net prices or gross prices, “my understanding is that it should be net prices; it should be actual prices paid by Medicare.” But “that’s something I think we’re going to have to watch spring guidance for because it’s not spelled out in the statute as far as I know.”

The basket of actual drugs to be negotiated is complex to figure out, noted Raffat. First, he said, you must filter out small molecule drugs less than nine years from their first approval and biologics approved less than 13 years, which the IRA excludes. Other exclusions include orphan drugs with a single indication, plasma-derived products, medications that have generics or biosimilars and agents with less than $200 million in Medicare spend — the last of which is “so critical.…So many drugs are going to get excluded out because of that.” And then the drugs must be rank ordered based on Medicare spending on them.

He noted that many people trying to determine Medicare spending have used the Medicare Spending Dashboard, which contains both Part D and Part B drugs. “But there’s a huge problem with that” for a couple of reasons. First, the dashboard contains data from 2019 and 2020 — not the June 2022 through May 2023 data that the IRA requires be used.

Second, “it does not adjust for things like how many years has it been since launch. Is it small molecule? Is it large molecule? All those exclusions have to happen.” In addition, the dashboard uses gross, not net, sales, which is “a huge problem.”

So Raffat constructed a list entirely himself. First, he captured the U.S. sales forecasts for all products across all global biopharmaceutical firms, the launch dates for each agent and the generic/biosimilar entry dates for each medication. Then he estimated products’ Medicare spend; applied exclusions, such as for plasma-derived products and orphan drugs with a single indication; and then filtered out what agents are affected and in what years in order to meet the nine-year exclusion for small molecule drugs and 13-year one for biologics.

To estimate Medicare spend, “you have to do it on a drug-by-drug level,” he contended. To do this, he compiled data from manufacturers on specific drugs and/or drug classes. Raffat then “paired in CMS datasets. It’s gross sales, but it’s helpful, so you can’t ignore that.” The next step was checking age groups enrolled in clinical trials for these drugs to see what percentage of participants were at least 65 years old to determine potential Part D and Part B exposure “and then any other datasets I could possibly find. I tried to construct actual CMS exposure by drug using a variety of these datasets,” he explained.

For orphan drug exclusions, the IRA specifies “a drug for only one rare disease or condition.” Raffat pointed to AstraZeneca’s Tagrisso (osimertinib), which has multiple indications, but all of them are within lung cancer, so “theoretically, Tagrisso would not necessarily be in the IRA negotiation basket.”

In contrast, Bristol Myers Squibb’s Pomalyst (pomalidomide), AbbVie Inc. and Roche’s Venclexta (venetoclax) and AstraZeneca and Merck & Co., Inc.’s Lynparza (olaparib) have approval for multiple diseases, so “they will not be excluded out for orphan reasons,” he explained.

For the exclusion for plasma-derived products, Raffat said that he is assuming this means that cell therapies will not be subject to negotiation, an assumption held by multiple industry sources, so he excluded these agents from his model as well.

For the small biotech exclusion, the IRA refers to “initial price applicability years 2026, 2027, and 2028,” and if a qualifying single-source drug produces less than 1% of Part D expenditures, then it would be excluded from negotiation. According to Raffat, “Part D spend is $100 billion, and 1% of that is $1 billion. Basically, if a drug like Ingrezza [(valbenazine) from Neurocrine Biosciences, Inc.] or Jakafi [(ruxolitinib) from Incyte] is doing $1 billion or less in Part D spend, it’s not eligible for negotiation.” But since this is a short-term exemption, “I think we’re overstating the significance of the small biotech exemption,” he maintained.

Based on all these criteria, Raffat produced the following 2026 rank-ordered negotiation basket of drugs for Part D:

• Eliquis (apixaban) – Bristol Myers Squibb
• Eliquis – Pfizer (profit share)
• Xarelto (rivaroxaban) – Johnson & Johnson
• Invega Sustenna (paliperidone palmitate) – Johnson & Johnson
• Ibrance (palbociclib) – Pfizer
• Entresto (sacubitril/valsartan) – Novartis Pharmaceuticals Corp.
• Imbruvica (ibrutinib) – AbbVie
• Trelegy (fluticasone furoate, umeclidinium and vilanterol) – GSK
• Vraylar (cariprazine) – AbbVie
• Linzess (linaclotide) – AbbVie
• Genvoya (elvitegravir/cobicistat/emtricitabine/tenofovir alafenamide) – Gilead Sciences, Inc.
• Jardiance (empagliflozin) – Eli Lilly and Co.
• Lynparza – AstraZeneca
• Imbruvica – Johnson & Johnson
• Venclexta – AbbVie
• Calquence (acalabrutinib) – AstraZeneca
• Triumeq (abacavir/dolutegravir/lamivudine) – GSK

Around the eight and nine spots, “there is more variability because multiple drugs theoretically could be relevant,” he said. “There’s a real toss-up at the bottom of the list between a few guys because their 2021/2022 CMS sales are too close to each other.” Also, “you might think Eliquis is double counted, but it’s not. Bristol reports top-line sales; Pfizer reports a profit share.”

For the 2027 basket, “there is an important dynamic” related to the exemption for drugs with less than $200 million in Part D spend, he stated. “If I screen out all those drugs that have less than $200 million exposure to CMS in 2023, I can’t even get to 25 drugs in 2027.” This goes back to the question of whether CMS will look at gross or net sales, he said, and “I think for this reason alone, they have to look at net sales. Because if they’re looking at gross sales, even less drugs will make it in, and they’re incentivized to get more drugs” on the list to be negotiated.

For the following year, he calculated that “much of the inclusions in 2028 will be for Part B drugs primarily. You only have 15 more [total drugs]; it’s not like you need a full 40 drugs.…You just need 15 new,” Raffat clarified.

Those Part B drugs include the following:

• Keytruda (pembrolizumab) – Merck
• Opdivo (nivolumab) – Bristol Myers Squibb
• Darzalex (daratumumab) – Johnson & Johnson
• Prolia (denosumab) – Amgen
• Tecentriq (atezolizumab) – Roche
• Orencia (abatacept) – Bristol Myers Squibb
• Entyvio (vedolizumab) – Takeda Pharmaceuticals U.S.A., Inc.
• Nucala (mepolizumab) – GSK
• Repatha (evolocumab) – Amgen
• Kesimpta (ofatumumab) – Novartis
• Xgeva (denosumab) – Amgen
• Cosentyx (secukinumab) – Novartis
• Tysabri (natalizumab) – Biogen Inc.
• Taltz (ixekizumab) – Eli Lilly
• Benlysta (belimumab) – GSK

However, Taltz and Benlysta, as well as a number of Part D drugs, likely will be excluded based on Medicare sales, Raffat said. Still, unlike 2027, the addition of Part B drugs should allow for a full list in 2028.

“There will be a caveat to everything I showed you: If CMS does decide to use their Medicare Spending Dashboard, then the math is very different,” he pointed out. And “if CMS has some different data, I can’t control that.”

On the magnitude of price cuts, “the buzz I’m hearing is 25%, maybe up to 50%, but somewhere in that range,” said Raffat, adding, “I don’t know how true that is. We’ll find out.”

He noted “a couple of other considerations” with regard to the price cuts. For one, “could you end up with volume gains? Nobody’s really thought about that. Or the wild card is would it only apply to the drug that’s being cut or its competitors as well? I don’t know. A bigger wild card” depends on if the cut is on list price or net price. “For example, is the impact limited to pharma, or would PBMs have to pay some out of their rebates?”

No one has the answers to these questions right now, he maintained. “And don’t also forget, people can game the system” via volume-limited agreements so that a drug loses its single-source status. For example, “let’s say Keytruda has long-dated patents, but in 2028, there’s the IRA coming up. [Merck] might say, ‘Teva, you can launch with a 1% gated volume share,’ and all of a sudden, Keytruda is no longer single source.” Raffat said he’s heard anecdotally that a lot of companies are discussing such a strategy. He noted, however, that authorized generics do not take away single-source status; only generics can do that.

If that strategy starts happening across the industry, “I would say that from my conversations,…it’s possible [CMS] would try to address that in guidance,” said Marcus. “There’s probably some level of flexibility to say, ‘we’re going to try and look through those arrangements.’ And again, there’s no judicial review of the drugs that are selected. If CMS makes a decision about which drug is or is not on the list in a way that the manufacturer doesn’t agree with, there’s not really a lot of recourse if they’re not seeing eye to eye about whether this kind of tactic successfully allows them to evade restrictions.”

When asked during the Q&A part of the call whether he expected private payers to follow CMS’s lead in pricing, Raffat responded that “theoretically they may want to, but I think private payers at some level are probably already nervous. How much of this do they have to pay? And they should be paying something, and for that reason, I don’t think it’s going to extend beyond the Medicare component. I could be wrong, but that’s my take.”

Contact Raffat at umer.raffat@evercoreisi.com.