The 9-0 Loss That Came Back Deeper: CMS Now Proposes Paying 340B Drugs 37% Less, and the Survey Is Why

On 2 July, the Centers for Medicare and Medicaid Services proposed a startling change to what hospitals are paid for the specialty drugs they buy through a federal discount program called 340B340BA federal program that lets hospitals serving many low-income patients buy outpatient drugs at steep mandatory discounts from manufacturers, then bill insurers at normal rates and keep the difference.: the drug's average sales price minus 33.4 percent. For a drug with a $100 sticker price, Medicare would hand the hospital about $66.60 instead of today's $106, a cut of about 37 percent from the ASP-plus-6-percent benchmark that has governed outpatient drug payment for roughly two decades. Measured on the ASP scale itself, payment falls 39.4 points, from plus 6 to minus 33.4.

What makes the proposal remarkable is not its size. It is that a shallower version of exactly this cut already went to the Supreme Court, and lost, 9 to 0. The obvious read is that CMSCMSThe Centers for Medicare and Medicaid Services, the federal agency that runs Medicare and sets how much it pays hospitals and doctors. is defying the Court. The closer read is nearly the opposite: CMS was able to cut deeper because it finally did the one thing the Court told it to do first. That inversion, more depth, less legal risk, is the story the headlines miss.

In American Hospital Association v. Becerra (596 U.S. 724), decided 15 June 2022, a unanimous Court, opinion by Justice Kavanaugh, threw out CMS's 2017 decision, effective in 2018, to pay 340B hospitals ASP minus 22.5 percent instead of ASP plus 6.

The part that matters: the Court did not hold that a lower 340B rate was illegal. It held that this particular rate was illegal because of a procedural shortcut. The Medicare statute gives CMS two paths. If it surveys hospitals' actual acquisition costs, it may set a tailored rate that differs by hospital group, a 340B-specific cut included. If it skips the survey, it must pay everyone the same benchmark: ASP plus 6. In 2018, CMS wanted the tailored cut without doing the survey. That, and only that, is what fell.

Between 1 January and 7 April 2026, the response deadline was extended from an original 31 March, CMS ran the acquisition-cost surveyacquisition-cost surveyA government survey asking hospitals what they actually paid for their drugs, used to justify a tailored payment rate. the Court had demanded, asking hospitals what they actually paid for each separately payable drugseparately payable drugAn outpatient drug expensive enough that Medicare pays for it on its own line, rather than folding its cost into the payment for the procedure.. And here is the mechanism of the whole story. The 2018 rate leaned on a MedPACMedPACThe Medicare Payment Advisory Commission, an independent body that advises Congress on Medicare payment policy. estimate that 340B hospitals get at least a 22.5 percent discount off ASP, a floor used as a proxy. The survey of what hospitals really paid found the average 340B discount is deeper than that floor: about 33.4 percent. So the number went up not because CMS chose to punish hospitals, but because the surveyed acquisition cost came in lower than the old proxy assumed. The deeper cut is the survey's finding, not a policy preference bolted onto it.

CMS also reports the survey turned up cases where a hospital was reimbursed more than it paid for the drug, the exact overpayment the tailored-rate authority exists to correct. That became the minus-33.4 proposal, and the political fuel for going deeper.

2018: shallower cut, no survey, lost 9-0. 2027: deeper cut, survey in hand. Depth up; legal exposure down.

The legal posture has flipped end for end. The challenge that won last time, you skipped the survey, is the one argument hospitals cannot make this time.

CMS advertises the proposal as roughly $5.7B in year-one savings: about $1.15B in reduced beneficiarybeneficiaryA person enrolled in Medicare who receives its benefits. out-of-pocket drug costs and about $4.55B on the Medicare program side. Both numbers are real reductions in what gets paid for drugs. But neither one lands as a net saving, because of a piece of plumbing called budget neutralitybudget neutralityA rule requiring that money cut from one area of outpatient payment be added back elsewhere, so total outpatient spending stays the same..

By law, the drug cut cannot shrink total outpatient payments. Every dollar pulled from drug reimbursement has to be added back elsewhere, here, as a roughly 8.44 percent bump to the payments for non-drug services (clinic visits, imaging, infusion administration fees), delivered through the outpatient conversion factorconversion factorThe base dollar figure Medicare multiplies by each service's weight to set its outpatient payment; nudging it up is how CMS recycles the drug cut into non-drug services.. The whole $5.7B is recycled, not removed.

That symmetry is the point the trade coverage skips. On the drug side, beneficiaries paid 20 percent coinsurancecoinsuranceThe share of a medical bill a patient pays out of pocket, here, generally 20% of the Medicare-approved amount. and the program 80 percent. On the non-drug bump they pay roughly the same split, about 20 percent beneficiary, 80 percent program. So beneficiaries get back on the non-drug side roughly what they lose on the drug side, and the same is true for the program. Netted out, both ledgers land near zero. The advertised 'saving' is mostly an accounting artifact of where the money is measured.

If total spending is unchanged, the real question is who gains and who loses inside the fixed pot. Two shifts stand out.

• Across hospitals: drug-heavy 340B safety-net hospitals, which lean on the drug spread, lose, while service-heavy hospitals (including many that are not 340B) gain from the 8.44 percent non-drug bump. A hospital that infuses a lot of expensive 340B drugs but does relatively few imaging and clinic visits comes out behind; one with the opposite mix comes out ahead.
• Across patients: people who receive 340B-acquired Part B drugs pay less in coinsurance, while patients using non-drug outpatient services pay a bit more. The relief is real for the first group and is paid for by the second, not by the program as a whole.

The spread also does work that does not show up on a claim. Under today's ASP plus 6, a 340B hospital buying a $100-ASP drug for about $66.60 keeps roughly $39.40 over cost, money many safety-net systems use to cross-subsidize charity care and thin-margin services. At ASP minus 33.4, the payment lands right at the surveyed acquisition cost, and that margin goes to about zero.

Whether that spread is a windfall to be recovered or a lifeline for vulnerable communities is the live policy fight, and it is a genuine judgment call, not a factual one. CMS frames the spread as Medicare overpaying for drugs. Hospitals frame it as the funding that keeps their doors open. The survey settles what hospitals paid for the drugs; it does not settle what the extra dollars were doing.

The winning move from 2022 is gone. What remains is harder and more fact-bound. Hospitals can attack the survey itself, its design, response rate, or representativeness, or argue the resulting rate is 'arbitrary and capriciousarbitrary and capriciousA legal standard for striking down an agency rule when the agency failed to give a reasoned, evidence-based justification.,' the legal test for a rule an agency failed to justify with reasoned evidence. Both are steeper climbs than the clean procedural win they scored last time, because now there is a survey in the record for a court to defer to.

There is also the collateral machinery. The 2023 remedy ruleremedy ruleThe 2023 regulation in which CMS repaid hospitals about $9 billion for the underpayments the Supreme Court found unlawful, then began clawing the money back through lower non-drug payments. already repaid affected hospitals about $9 billion for the 2018-2022 underpayments and began clawing it back through a 0.5 percent annual cut to non-drug payments starting in 2026. The CY2027 proposal would accelerate that recoupment, lifting the annual reduction from 0.5 to 3 percent, a separate squeeze layered on top of the new rate.

This is a proposed ruleproposed ruleA draft regulation the government publishes for public comment before it becomes final; it can still change., not a final one. Comments are due 31 August 2026, and CMS can move the number before the effective date of 1 January 2027. But the strategic picture is set: by doing the homework the Court assigned, CMS converted a cut it lost unanimously into a deeper cut that is far harder to challenge, while budget neutrality quietly ensures that, program-wide, no dollars are actually saved. The fight from here is over who inside a fixed pot keeps the money, and whether a survey of drug prices can stand in for a judgment about what the drug spread was funding.