AMP is WAC -- 05/08/26

Tastes Great, Less Filling

I try and embrace the small stuff. First sip of coffee in the morning, a funny text from a friend, and peony season. Peonies are my absolute favorite, and they bloom for a few days a year and then nothing for another 360 days. This week I’m literally stopping to smell the flowers and hope you get a chance to. Unless you’re allergic, then, well, then don’t do that.

Feels Like a TPS Report. On May 1, the Centers for Medicare & Medicaid Services (CMS) issued a clarification memo to all Part D sponsors reminding them, somewhat pointedly, that 340B claim identification in Prescription Drug Event (PDE) reporting has been required since contract year 2025.

Quick reminder for those who do not live in PDE land: the Inflation Reduction Act’s (IRA’s) drug negotiation program gave CMS a reason to know which Part D claims involve 340B-acquired drugs. That matters because the Maximum Fair Price does not fit neatly on top of 340B. If the same claim gets both discounts and nobody catches it, the manufacturer may be left eating both.

CMS updated the PDE layout back in 2023, expanding the Submission Clarification Code field to capture the “20” code for 340B. Two-plus years later, they are sending a memo. The memo tells you sponsors are not universally passing through those values correctly. CMS is now clarifying that if a dispenser supplies the 340B code and it is not on the PDE, the sponsor needs to submit an adjustment. That is not a small ask depending on volume.

Last month, ADVI Health published an analysis of 340B modifier usage in Medicare Part D claims for IPAY 2026 and 2027 selected drugs, including Eliquis, Entresto, Ozempic, Ibrance, and Xtandi. In the first two months of 2026, just 0.5% of IPAY 2026 drug claims and 0.4% of IPAY 2027 drug claims carried a 340B modifier. Prior research estimates 9% to 12% of Part D claims are likely eligible for 340B pricing.

The gap between those numbers is the problem. Voluntary modifier usage is not working as a deduplication mechanism, the HRSA rebate pilot that was supposed to help got blocked in litigation in December 2025, and CMS has declined to take on the deduplication role directly.

For manufacturers with negotiated drugs, this is not an abstract data infrastructure problem. When 340B and MFP stack on the same claim and nobody catches it, the manufacturer eats both discounts. That is the actual stakes of a technical memo that most people would understandably skip.

Election Season Preview. On Wednesday, KFF dropped their latest Health Tracking Poll, and the headline is not surprising: 64% of adults are worried about affording health care costs, tied with gasoline and transportation as the top financial concern.

When insured adults were asked what change to their health insurance would be most important, 46% said paying less out-of-pocket. Only 22% said eliminating prior authorization, which prior KFF polls have consistently flagged as the biggest access pain point other than costs. So even among people who experience prior auth frustration directly, reducing their dollar exposure matters more.

And here is the political angle that is interesting for manufacturers watching Most Favored Nation (MFN) and IRA implementation: voters approve of the Trump administration’s handling of prescription drug costs at 41%, notably higher than their approval on health care costs overall at 33%.

Democrats lead Republicans on trust for addressing drug costs 33% to 26%, but roughly a third of voters do not trust either party. Among independents, half trust nobody on this.

That is the danger zone for industry: not voters who have a coherent drug pricing ideology, but voters who are frustrated enough to support almost anything that sounds like it will lower their costs.

All Work, No Guidance. On Wednesday, KFF released survey results from state Medicaid officials on how states are preparing for the work requirements that take effect January 1, 2027. These are mandatory for Affordable Care Act expansion adults in 43 states under the 2025 reconciliation law.

States are all over the map in terms of progress, federal guidance is late, and the implementation timeline is genuinely tight. Four states—Arkansas, Idaho, Indiana, and New Hampshire—are planning more restrictive verification than the law requires, including longer look-back periods. Iowa, Montana, and Nebraska plan to implement early. Only Iowa and Indiana have declined to adopt any hardship exceptions.

The piece that does not get enough attention in the work requirements debate: the “medically frail” exemption. States want to use Medicaid claims data and allow self-attestation to identify who qualifies. They cannot finalize their systems until CMS defines “medically frail,” and that guidance has not come. Six states are already using AI to assist with implementation. Most others are still figuring it out.

For manufacturers with patient populations concentrated in Medicaid, the coverage disruption risk here is real. Work requirements have a consistent track record of reducing enrollment without meaningfully increasing employment. The administrative churn is where patients lose coverage. And coverage loss is an access problem that shows up in your patient services programs.

Losing More than Weight. Deloitte released the 16th edition of their Measuring the Return from Pharmaceutical Innovation report, and the headline is that pharma R&D returns have improved for the third consecutive year. This sounds like good news until you look at the distribution.

Strip out GLP-1 and GIP assets, and the underlying industry IRR drops to 2.9%, down from 3.8% in 2024. The median barely changed while the mean got pulled way up by a handful of companies with major obesity assets in late-stage development.

For the first time in 16 years of this analysis, obesity has displaced oncology as the largest contributor to pipeline value at roughly 25% of total late-stage forecast sales. Obesity contributed 1% in 2022. The average cost to develop a drug from discovery to launch hit $2,671 million in 2025, up from $2,229 million the year before.

For manufacturers who are not in the GLP-1 race, the bar is higher and the portfolio pressure is real. About 53% of pipeline asset-indications have a forecast peak sales potential below $250 million. At $2.7 billion to develop, the math on that long tail gets uncomfortable fast. The industry is not broadly healthy right now. It is very healthy in one place, and that place is Wegovy-adjacent.

Clip and Save. Cencora dropped their updated U.S. Biosimilar Landscape report this month, and the tl;dr is: the market keeps growing.

As of May 1, 2026, there have been 84 Food and Drug Administration (FDA) approvals and 66 launches. The pipeline is active; Keytruda biosimilars alone have seven in Phase 3, with a first estimated launch in 2028.

For manufacturers on the reference product side, the competitive pressure is real and getting more layered. It is not just about launch timing anymore, it is interchangeability designations, unbranded versions, which payers are driving substitution, and what contracting structures look like in a crowded therapeutic class. Adalimumab is the case study here: nine biosimilars launched in 2023 alone.

Reimbursement Fundamentals — You Know What They Say About Assumptions

On May 7, the White House Council of Economic Advisers released a report projecting that Most Favored Nation (MFN) drug pricing will generate $529 billion in savings over 10 years across all U.S. markets. The voluntary framework requires manufacturers to launch future drugs at prices comparable to what other high-income countries pay, benchmarked to the second-lowest net price among G-7 nations (excluding the U.S.), Denmark, and Switzerland. For existing drugs, manufacturers agreed to offer Medicaid programs MFN pricing, projected to save $64.3 billion over 10 years. The TrumpRx.gov direct-to-consumer channel is already delivering discounted cash prices on GLP-1s and fertility drugs. The math is real. The modeling is detailed. The savings projections are genuinely significant.

The problem is the assumptions.

The entire framework depends on reference country prices rising to meet U.S. prices halfway. The report explicitly states this: prospective MFN will lower U.S. prices and “put upward pressure on prices paid in other wealthy nations.” The policy is designed to work “in tandem with U.S. trade policy efforts” to force foreign governments to pay their fair share. The December 2025 UK pharmaceutical trade deal, which raised NICE’s cost-effectiveness thresholds from £20,000 to £30,000 to £25,000 to £35,000 per QALY and reduced clawback rates, is cited as proof the strategy works.

But one deal with the UK is not a global pricing realignment. France, Germany, Italy, and Canada are not going to raise drug prices just because the U.S. wants them to. They have their own voters, their own budget constraints, and their own incentives to preserve existing price controls. The UK agreed because it needed a trade deal. The rest of Europe does not.

The volume of MFN analysis landing all at once tells you something: the policy is no longer being debated in the abstract. People are starting to model how it actually behaves. There are now three parallel MFN models in motion: GLOBE for Medicare Part B, proposed but not finalized, with a targeted start in October 2026; GUARD for Medicare Part D; and GENEROUS for Medicaid, which is voluntary, open for enrollment, and had its deadlines extended April 29. All three benchmark U.S. drug prices against what peer countries pay. The core logic is simple: U.S. drug prices are over 250% of OECD levels. Import the lower foreign prices. Close the gap.

The savings math, on paper, is genuinely significant. On Thursday, JAMA Health Forum published an analysis from Hwang et al. looking at 76 brand-name cancer drugs that together cost Medicare $49.6 billion in 2024. Applying average international MFN pricing to all of them would have reduced that spend by $27.1 to $35.4 billion. Even for just the nine drugs already in IRA negotiation cycles, MFN would have delivered an additional $3.2 to $4.7 billion beyond what negotiation alone achieved. Cancer drugs are especially exposed because they carry very low average rebates and because international prices run 65% to 81% below U.S. prices, a bigger gap than most other therapeutic areas.

Those estimates assume manufacturers sit still. They will not.

A Health Affairs Forefront piece published this week from Barros, Righetti, and Sá is the clearest articulation I have seen of why MFN pricing is anchoring to a moving target, and it goes beyond the standard gaming argument. Yes, manufacturers will delay market entry in reference countries, shift to confidential discounts to obscure net prices, and in some cases exit markets entirely to eliminate the benchmark. We saw that in real time in February when Amgen pulled Repatha from Denmark rather than let its deeply discounted Danish tender price become reportable under GENEROUS. Amgen’s math: sacrifice roughly $16 million in Danish revenue to protect against exposure in a Medicaid program running roughly $241 million and a Medicare Part D market running roughly $1.75 billion. Rational. Disruptive to 2,000 patients. An early signal of a pattern that will repeat.

The Barros piece adds a second channel that gets underspecified in most of the debate: reference country payers may accommodate U.S. MFN pressure by voluntarily adjusting their own pricing rules upward. The modeling on this is striking. Without accommodation, MFN reduces prices for future drugs launched in the U.S. by about 47.4% versus independent pricing, and global manufacturer revenues fall roughly 17.8%. With accommodation, that price reduction shrinks to 38.0%, and manufacturer revenues actually rise 5.2%, because the reference country’s health system absorbs most of the burden through substantially higher pharmaceutical expenditures. Accommodation is good for manufacturers and bad for UK and German taxpayers. The U.S. savings are smaller than advertised either way.

And then there’s the operational mess. The framework depends on net international prices, not list prices, because list prices don’t reflect what countries actually pay after rebates, clawbacks, and confidential concessions. Manufacturers will voluntarily report net pricing data to CMS following methodological guidance modeled on Germany’s Arbitration Board process. The report says this repeatedly: “voluntarily reported.” There is no global dataset of net prices. The IQVIA MIDAS data used to model the savings estimates explicitly does not include rebates or net pricing. The footnotes say so on page 7. The entire $529 billion projection rests on manufacturers self-reporting the very data that would trigger higher MFN benchmarks and lower their revenues.

For manufacturers, the strategic question is simple: do you believe Europe will actually pay more, or do you plan for a world where they don’t? If the answer is “they don’t,” then MFN is not a rebalancing mechanism. It’s an import of European pricing constraints into the U.S. market with a 10-year hope that trade policy fixes the imbalance later. The report frames this as inevitable: “Failure to fully rebalance the global innovation ecosystem represents an existential risk to the pharmaceutical industry.” That’s true. But the existential risk cuts both ways. Manufacturers can either accept MFN pricing and hope for rebalancing, or they can adjust their portfolios and pipelines to reflect the revenue structure they’re actually going to face.

The baseline assumption, that sovereign health systems will voluntarily raise drug prices to accommodate U.S. trade policy, is the part I can’t get past. Countries that have spent decades building political consensus around price controls are not going to abandon them because the U.S. tied its own prices to theirs. They’re going to delay access, narrow formularies, and find new ways to preserve budget control. And when that happens, the MFN benchmark drifts down, not up.

I remain skeptical.