Can you ever feel sad looking at photo of a chimp wearing clothes? Maybe it will make the latest policy news go down just a little smoother.
340B’s Split Screen. February 26 brought two developments in the 340B contract pharmacy fight. They pointed in opposite directions.
First, a judge in the Middle District of Tennessee dismissed AbbVie’s challenge to the Tennessee Hospital Protection Act. The law prohibits manufacturers from restricting 340B contract pharmacy access or requiring hospitals to submit claims data as a condition of participation.
That decision adds to a growing list of losses for manufacturers. At least thirteen district courts have now ruled against them, with three circuit courts affirming. AbbVie has now lost both at the preliminary injunction stage and on the merits in Tennessee, on top of earlier losses in Mississippi, Louisiana, Arkansas, Hawaii, Maine, Colorado, Nebraska, and South Dakota.
Manufacturers argue they need utilization data from hospitals to monitor diversion and duplicate discounts. The Tennessee court said nope. Federal 340B law never required covered entities to submit claims data to manufacturers. If the statute never required it, the court reasoned, states are not interfering with federal law by blocking the practice.
Then, later that same day, the Justice Department (DOJ) stepped in on the other side of the fight.
DOJ filed a brief in AbbVie’s appeal of Colorado’s 340B law and sided with the company. The department argued the state law interferes with the federal 340B program. Manufacturers have been making that federal preemption argument for years with little success in court. Hearing it from DOJ carries a different kind of weight.
Maybe there is finally light at the end of the tunnel for manufacturers? HRSA recently extended the comment deadline on its proposed 340B rebate model after the previous version was blocked in court. Eli Lilly is also moving forward with a policy that requires hospitals to submit detailed utilization data to access 340B pricing. Either way, the next phase of the fight will play out in the courts.
Gophers Digging into 340B. On February 27, the Minnesota Department of Health released its second annual report on the 340B program. It offers the clearest picture yet of how much revenue the program generates inside a single state.
I was going to say dam, but that’s beavers, not gophers. Either way, damn. Covered entities in Minnesota reported at least $1.34 billion in net 340B revenue for 2024 and the state says that figure likely understates the total. Some covered entities struggled to fully capture infusion drug data.
Nationally 340B drug purchases reached about $81 billion in 2024, up from $66 billion the previous year. The program requires manufacturers to sell drugs at discounted prices to eligible safety net providers, but federal law does not require providers to report how they use the resulting revenue.
Minnesota’s transparency effort starts to change the policy conversation. Legislators, payers, and manufacturers now have something that rarely exists in the 340B debate. A number. And a state-based number feels way more real. It feels like something policymakers should be able to dig into and ask: where did the money go and what did we get for it?
No Pill for That. On February 26, GAO released a report examining the Center for Medicare & Medicaid Services’(CMS) Part D Premium Stabilization Demonstration. In July 2024, CMS launched the demonstration under its Section 402 authority. The program allowed the agency to reduce premiums by up to $15 per month and limit annual increases to $35.
The Inflation Reduction Act shifted more catastrophic drug spending onto plan sponsors and away from the federal government. That redesign changed the risk profile for insurers almost overnight. When plans submitted bids for 2025, the variation in expected spending translated directly into premium volatility.
GAO estimates that without intervention, premiums for standalone prescription drug plans would have nearly doubled on average. About 37% of beneficiaries would have faced monthly increases above $40. Total cost: $9.8 billion across 2025 and 2026. On paper, the intervention worked. Average standalone Part D premiums increased only slightly, from $42 to $43, and enrollment grew.
Multiple policy changes hit the Part D market at the same time, making it difficult to isolate how much stability came from the demonstration itself. An ongoing evaluation is expected to sort that out. The demonstration begins winding down in 2026, with smaller subsidies and more premium flexibility for plans.
CMS spent $9.8 billion buying time. The next few bid cycles will show whether the Part D redesign can stand on its own. I remain doubtful.
One Big Beautiful Something. On February 26, RAND released a state by state analysis of the Medicaid provisions in the One Big Beautiful Bill Act, signed last summer. The topline number is large. RAND estimates the law will reduce state Medicaid spending by about $665 billion between 2025 and 2034. Federal savings reach about $714 billion.
RAND projects there will be 7.6 million fewer Medicaid enrollees nationwide by 2034. But the distribution of those cuts is far from uniform. States that rely heavily on provider taxes and state directed payments to finance Medicaid are expected to take the largest hits. Arizona, Iowa, and Nevada face projected reductions exceeding 15% of their Medicaid budgets. California and New York face the largest absolute reductions, at roughly $112 billion and $63 billion respectively.
Florida, which relies less on those financing tools, sees minimal impact in RAND’s modeling. Some expansion states that also avoided heavy reliance on those mechanisms, such as North Dakota and Nebraska, see smaller net effects because a rural health provision offsets some losses.
The political story around this law often frames the fight as red states versus blue states. RAND’s analysis suggests a different dividing line.
For manufacturers, the implication is straightforward: prepare for a growing number of uninsured patients, on top of those already losing Exchange coverage.
Go Ahead and Work It. Speaking of Medicaid changes…States are beginning to estimate what it will cost to implement Medicaid work requirements under the One Big Beautiful Bill Act. And, as you might imagine, the numbers add up quickly.
An Associated Press analysis of state projections suggests implementation will exceed $1 billion before the policy generates any federal savings. The law requires most adults ages 19 through 64 in the Medicaid expansion population to complete 80 hours per month of work, education, or community service. States must implement the requirement by January 1, 2027.
Congress provided $200 million in federal funding to help states build the necessary systems. State projections suggest the gap between federal funding and actual costs will be significant.
- Maryland estimates about $32 million.
- Kentucky estimates $46 million.
- Colorado projects $51 million.
- Arizona estimates $65 million and expects to hire about 150 additional staff.
Medicaid eligibility systems in many states are decades old. They were not designed to verify monthly work activity, track exemptions, or manage compliance periods. Arkansas offers a recent example of how tricky all of this can be. In the first seven months of implementation of their changes, about 18,000 people lost coverage before a federal court halted the policy. Evidence later suggested many of those individuals remained eligible but struggled with the reporting process. Another thing for patient support programs to prepare for. Add it to the list.
Reimbursement Fundamentals: A Look Under the Hood at Hospital Drug Prices
On Thursday, 3 Axis Advisors and Patient Rights Advocate released what is probably the most comprehensive look at hospital drug pricing under the federal transparency rules to date. They pulled chargemaster files from more than 1,300 hospitals, focused on high-expenditure drugs used in cancer and MS treatment, and ran the numbers. Woof.
Here’s the short version: posting a file and providing useful information are not the same thing. 93% of downloaded files met CMS formatting requirements. Only 62% contained usable pricing for the drugs the researchers were looking for.
The pricing data that does exist is jarring. For six drugs — Keytruda, Darzalex, Faspro, Opdivo, Ocrevus, and Tysabri — the researchers looked at what hospitals are charging commercial insurers relative to Medicare’s Average Sales Price (ASP) + 6% benchmark. That benchmark is a reasonable stand-in for what it costs a hospital to acquire the drug. The results: negotiated commercial rates averaged somewhere between 136% and 209% of Medicare reimbursement depending on the drug, with third-quartile rates reaching 363% to 422% of Medicare. For Ocrevus, the average cash price was 1,300% of Medicare. The gross charge, that is the number that is the basis for patient billing for the uninsured, was as high as 2,271% of Medicare for Tysabri. Read that again. 2,271%.
That is before you factor in 340B. Over half of outpatient drug sales run through hospitals participating in the 340B program, where MedPAC says acquisition costs average about 22.5% below ASP. If the benchmark used to calculate those eye-popping markups already overestimates what 340B hospitals actually pay, the real margins are higher. The report flags this explicitly: the analysis likely underestimates hospital markups.
The within-hospital variation is what I can’t get over. For a given drug at a given hospital, negotiated rates can differ by a factor of ten or more depending on which payer contract you’re looking at. The average max-to-min ratio of negotiated prices across the dataset was 2,347x. The median was about 79x. When one insurer pays $1 and another pays $2,347 for the same drug at the same hospital on the same day, there is no real price. There is a number that appears on a document, and it means whatever the contract says it means that day.
For uninsured patients, the numbers are ugly. About one in four times, cash prices offer no discount at all off the inflated gross charge. Half the time, the cash discount is 30% or less. Commercial insurers, by contrast, average around 40% off gross charges, meaning an uninsured patient walking in off the street often gets a worse deal than someone with coverage. But the data cuts both ways. In some cases, the uninsured cash price actually beats the insurer-negotiated rate entirely, the opposite of what most people would expect insurance to be doing.
The margin sitting between acquisition cost and commercial reimbursement is real, large, and almost entirely invisible in a policy debate that has spent a decade focused on manufacturer list prices.
The rebate system compounds this. High WAC means high rebates flowing back to PBMs and payers, calculated off list prices, giving PBMs and plans a structural incentive to keep list prices elevated even as they negotiate them down on the back end. The hospital markup story and the manufacturer rebate story aren’t competing explanations. They’re the same system extracting margin from different pressure points, with patients and employers paying for all of it.
Affordability has more address points than the public debate acknowledges, and hospital margin on drugs has been hiding in plain sight behind data nobody could previously read.
That’s what makes this report worth paying attention to. The tool 3 Axis built at hospitaldrugprices.org is a genuine contribution, but also a demonstration of how much technical work it takes to make publicly mandated disclosures usable. You cannot build a functioning market on prices only experts can interpret.
