AMP is WAC -- 07/31/26

Enjoy the Ride.

Chicago this week for NCSL was wild. Best exhibit hall of any conference I’ve ever been to. Grimace and the Chick-fil-a cow are a booth over from the nudists. Who are next to the anti-child brides. Even more wild, four inches of rain in an hour while trying to take public transit was ill advised. Changing my flight to avoid East coast storms and hitting up some of Chicago’s cocktail bars, genius. Meadowlark found a way to make Malort less horrible. IFKYK.

Big Swings. This week Senate Finance Committee Ranking Member Ron Wyden and a group of Senate Democrats dropped an 86-page request for information on private health insurance reform. If you’re a regular reader, you know that this isn’t important for today, next month or even next year – it’s the long game.

This is the 2029 (maybe) preview of a full reimagining of how private insurance works in America: premiums, deductibles, prior auth, MLR gaming, vertical integration, the whole thing.

Three sections. First, reversing Republican cuts and reimagining affordability. Second, making health care simpler for families. Third, taking on corporate greed. Comments are due October 2. Staff told reporters specific policies will start getting drafted in 2027.

Average marketplace deductibles grew by about $1,000 per person in 2026. Family premiums hit $37,840 this year. Individual out-of-pocket maxes are already $10,000, heading to $15,600 under new Centers for Medicare & Medicaid (CMS) rules. And somehow the seven largest insurers generated $54 billion in profit last year. So, yeah.

What caught my attention: the employer market is being tackled. Over 160 million Americans get coverage through an employer, and meaningful reform has basically left that system alone since the Affordable Care Act.

Also interesting: the section on third-party administrators (TPAs). The document explicitly draws a parallel to PBMs - TPAs using “spread pricing, hidden rebates, and patient steering” to extract profit from self-insured employers. Anyone who followed the PBM fight will recognize this framing immediately. My hunch is the same political energy is about to land on TPA practices.

For pharma, the commercial story right now is coverage instability - ACA enrollment chaos, deductible shock, patients abandoning therapy because they can’t afford cost-sharing. This white paper doesn’t fix any of that immediately. But it signals what the policy environment looks like if Democrats win in 2028, if you’re into planning ahead.

Vote Now, Pay Later. On Tuesday, CMS released the 2027 Part D national average monthly bid amount and set itself up for some potential madness right before the election.

The numbers first: the 2027 national average monthly bid amount (NAMBA) is $296.05, and the base beneficiary premium is $41.33. The NAMBA is the enrollment-weighted average of all Part D plan bids; the base beneficiary premium is the statutory starting point for calculating plan-specific premiums. This is a standard annual release.

What’s news: CMS is ending the Part D Premium Stabilization Demonstration at the close of this year. The Demo launched in 2025 as a Biden-era intervention; the IRA’s benefit redesign (the $2,000 out-of-pocket cap and related changes) shifted costs between payers in ways that would have sharply raised standalone prescription drug plan premiums. The Biden administration gave insurers billions in subsidies, through a demo, to cushion the landing. The Demo kept average standalone premiums about $16 a month lower than they would have been otherwise, at a federal cost of $3.6 billion in 2026, per KFF.

CMS’s position now: plan sponsors have had enough experience under the redesigned benefit to accurately price their bids. The program ends; traditional market conditions return. Dr. Oz posted that “premiums will go up by less than $10 for most Medicare recipients, with many even seeing LOWER premiums.” AHIP said it was “reviewing” the announcement. (That’s trade association for “we have concerns we’re not ready to articulate publicly.”)

About 25 million Americans have standalone Part D plans, currently paying $36 a month on average. “Less than $10 for most” is doing a lot of work. The people whose premiums go up more than $10 aren’t a rounding error, and they’ll be making enrollment decisions this fall when open enrollment begins. CMS says individual plan premiums will be public in September. For context: the number of standalone Part D plans available to beneficiaries has dropped from 766 in 2022 to 360 in 2026, per KFF.

I think the timing is interesting. Beneficiaries will be seeing the new premiums just about the time they go to the polls for mid-term elections. I’m not a political strategist, but this seems like a preventable mess.

Fast Track, Same Timeline. A new study in JAMA Health Forum looked at how long it takes for a new drug to go from the start of a clinical trial to Medicare reimbursement. And, the answer, across a full decade of data and four countries, is: about as long as it has always taken.

Researchers analyzed 519 drugs approved by the Food and Drug Administration (FDA) from 2014 to 2024, plus comparable cohorts in France, Germany, and Switzerland. Median total duration from clinical trial initiation to reimbursement: 9.3 years in the US, 8.9 years in Germany, 9.8 years in Switzerland, 11.1 years in France.

Over time, we’ve seen barely any progress. The US ticked up slightly - from 9.0 years in the 2014-2017 cohort to 9.35 years in the 2021-2024 cohort. France got meaningfully worse (9.82 to 12.18 years). Germany improved (9.27 to 8.7 years). Switzerland was roughly flat.

This matters because there’s been enormous policy pressure to speed drugs to market. FDA has expanded accelerated approval, created breakthrough therapy designation, and most recently announced a one-trial default standard. And yet the total timeline isn’t shrinking. Because the bottleneck isn’t primarily regulatory review - it’s clinical development, the time from starting human testing to submitting for approval. That phase isn’t getting faster, and speeding up what happens at the end doesn’t change how long the middle takes.

Cancer drugs are the exception. They have shorter timelines across all four countries, most pronounced in the US and Germany, likely because oncology drugs disproportionately use surrogate endpoints in clinical trials (shorter development), move through accelerated regulatory pathways, and in the US, Medicare’s protected class coverage for antineoplastics essentially eliminates the reimbursement review phase entirely.

The authors make the right call here: before we can target policy to accelerate access, we need to understand where we can cut. Show me the data on where the time goes. Regulatory review? Clinical development design? Reimbursement delays after approval? The answer shapes very different policy responses.

I’d note this study uses Medicare as the US reimbursement benchmark, which misses commercial coverage patterns.

The Generic Escape Clause. On Monday, CMS notified Medicare plan sponsors that Xeljanz and Xeljanz XR will be removed from the Inflation Reduction Act (IRA) Selected Drug List for Medicare negotiated drugs effective January 1, 2029.

A generic tofacitinib (Xeljanz’s active ingredient) has been approved and is being bona fide marketed. Under the IRA’s drug price negotiation program, a brand drug gets removed from the negotiation list when a generic enters the market.

Xeljanz (Pfizer’s JAK inhibitor for rheumatoid arthritis, psoriatic arthritis, and ulcerative colitis) has been on the Selected Drug List for the 2028 price applicability year. Despite already having a generic, it will face negotiated prices next year because of the timing of the generic release.

This is the IRA’s built-in off-ramp functioning as designed. Negotiation fills the gap during patent exclusivity; generic competition takes over when exclusivity ends. That’s the theory, anyway.

I am curious to see what happens in this long gap where generic tofacitinib will compete with the brand on formularies, which should drive costs down. But formulary placement, tier assignment, and prior authorization criteria are still decisions plan sponsors make - and those decisions don’t automatically follow the logic of “generic exists, so access is easy.” We’ve seen plenty of cases where generic entry doesn’t translate cleanly into affordable access for all patients.

The IRA negotiation program covered Xeljanz through 2028. What CMS negotiated as the maximum fair price applied there. Starting in 2029, it’s back to market dynamics with all the opacity and variability that implies.

Worth watching how Pfizer positions the brand against generic competition, and whether plan sponsors route patients to the lower-cost option or find ways to maintain utilization of the branded product.

All That Spending, All Those Gaps. In July, the Congressional Budget Office (CBO) released its annual look at federal health insurance subsidies - and the number at the top is worth sitting with for a moment.

From 2026 to 2036, the federal government is projected to spend $33.6 trillion subsidizing health insurance. Medicare accounts for nearly half - $16.1 trillion, or 48%. Medicaid and CHIP: $8.4 trillion (25%). Employment-based coverage tax exclusions: $7.2 trillion (21%). Premium tax credits: $1.2 trillion (4%).

In 2026, federal health insurance subsidies are $2.4 trillion, or 7.4% of GDP. By 2036, CBO projects that grows to $3.9 trillion. Most of that growth comes from Medicare because of higher prices, more utilization, and enrollment growth as the population ages. That’s demographics doing what demographics do.

The uninsured count is expected to grow from 30 million in 2026 to 37 million in 2036. The primary driver is the 2025 reconciliation act, which CBO projects will reduce Medicaid and CHIP enrollment. And here’s the number that sticks with me - about 60% of the people projected to be uninsured in 2036 will be eligible for subsidized coverage. They won’t lack access because coverage isn’t available. They’ll be uninsured for other reasons: complexity, cost of premiums and cost-sharing even with subsidies, administrative barriers, or simple unawareness.

That’s a solvable enrollment and implementation problem - which makes it more frustrating that CBO is projecting it to get worse.

The pharmaceutical policy angle: 36 million uninsured people is also 36 million people without reliable access to covered prescription drugs.

So Smooth, It’s Invisible. On Monday, Avalere Health released an analysis of 2025 Medicare Part D claims data with a finding that’s either unsurprising or alarming, depending on how optimistic you were about IRA implementation.

Only 1% of Part D beneficiaries, about 333,000 people, enrolled in the Medicare Prescription Payment Plan (MPPP) in 2025. The MPPP lets beneficiaries spread their out-of-pocket drug costs across the year in monthly installments rather than paying it all upfront when they fill a prescription.

It gets more striking when you look at who had the most to gain. About 5.4 million beneficiaries (17% of the study population) reached the catastrophic phase in 2025, with average annual OOP costs of $1,182. Among that group, only 3.8% enrolled. Another 3.6 million had moderately high OOP costs without hitting catastrophic. Only 1.7% of them enrolled.

Avalere identifies three design problems. First, the MPPP is opt-in only. Second, there’s no point-of-sale enrollment (POS)- a pharmacist can tell you the program exists, but you can’t sign up at the counter. You must call your plan separately, before the benefit can apply. Third, monthly payments aren’t fixed - they recalculate based on remaining OOP liability and months left in the year, so your payment can jump around in ways that are hard to anticipate.

These are behavioral economics problems wrapped in a program design problem. Opt-in enrollment consistently underperforms opt-out. Variable monthly payments create confusion and distrust. The fix isn’t complicated: auto-enrollment with opt-out, POS enrollment capability, fixed or more predictable monthly amounts. Whether Congress or CMS will act on it is a different question.

The $2,000 OOP cap is genuinely meaningful. The smoothing tool that was supposed to help people get there is sitting mostly unused.

Reimbursement Fundamentals – Follow the Dollar, UPL Edition

As I prepared for this week’s panel at the NCSL conference talking about upper payment limits (UPLs), I had to update my talking points to include that a federal court blocked one. The timing felt almost instructive.

On July 1, a federal district court in Colorado issued a preliminary injunction preventing the state from implementing its upper payment limit on Enbrel while litigation continues. Amgen (Enbrel’s manufacturer) had sued Colorado’s prescription drug affordability board (PDAB), arguing the UPL interfered with its patent rights under federal law. The court agreed, finding the cap likely “preempted” under the Constitution’s Supremacy Clause.

The ruling is narrow. It applies only to Enbrel. Colorado’s PDAB can still conduct reviews and set UPLs for other drugs, and it’s actively working on one for Cosentyx. The court’s reasoning is also legally shaky in ways these Health Affairs authors lay out clearly: the district court leaned heavily on a 2007 Federal Circuit decision about DC’s price-gouging law without engaging seriously with how pharmaceutical supply chains work. Colorado is expected to appeal. The Federal Circuit will likely have the last word on whether state drug price laws can coexist with federal patent protections.

But here’s what I tried to emphasize from that panel stage this week: even if the court had ruled the other way, UPLs are not the way to improve patient affordability.

I co-authored a paper released this week with the Rare Access Action Project, “Follow the Upper Payment Limit,” that walks through exactly why. The core problem is definitional, and it matters a lot: a UPL is a reimbursement cap. It sets a ceiling on what payers can pay pharmacies or providers for dispensing a drug. It does not set the price a manufacturer can charge. It does not require a manufacturer to sell at the capped price. And it does not guarantee that any savings reach patients at the pharmacy counter.

That last part is what gets lost in most PDAB debates.

A manufacturer sets the Wholesale Acquisition Cost (WAC). Wholesalers buy at WAC, sell to pharmacies at a small discount, and collect a distribution fee tied to WAC. Pharmacies dispense the drug and get reimbursed by a PBM working on behalf of the patient’s health plan. PBMs separately collect rebates from manufacturers in exchange for formulary placement. Those rebates flow back to health plans and are used, in theory, to offset premiums or cost-sharing.

When a state sets a UPL, it caps the reimbursement at the payer level. What the paper models out in detail is what happens to each actor in that chain when the ceiling drops.

If the manufacturer agrees to lower the price to the UPL: the rebate the PBM was collecting largely disappears, because it was calculated off WAC and WAC hasn’t changed (the discount is now built into the acquisition price). That rebate historically flowed back to health plans and offset costs. Under a UPL, it’s gone from the rebate stream. The health plan captures the savings. And unless the state has specifically mandated that savings flow through to patients at the point of sale, and created a mechanism to monitor that, patients see none of it. Colorado is the only state that has attempted to mandate pass-through savings. Its enforcement mechanism is not clearly established.

(And to be perfectly frank, I don’t think they have said HOW the UPL will work. Or how the PDAB envisions the dollar flow working and with what systems. Which is great for something they want to have working in 5 months.)

If the manufacturer doesn’t agree to lower the price: the pharmacy gets reimbursed at the UPL ceiling even though it acquired the drug at full WAC. In our illustrative example (a specialty drug at $4,000 WAC with a UPL set at $3,000), the pharmacy goes from retaining about $146 to retaining negative $838. The pharmacy loses nearly $1,000 per dispense. That’s not a rounding error. For independent community pharmacies already operating on negative margins for some high-cost drugs, it’s an exit decision.

There are two other structural problems worth naming. The chargeback system, where a wholesaler buys at national WAC, supplies at the state-mandated limit, and submits a state-specific chargeback to the manufacturer to recoup the difference, creates real operational complexity that scales badly across hundreds of state-specific commercial transactions. And when legitimate reimbursement rates are compressed below acquisition costs, pharmacies may turn to grey-market sourcing. No bueno.

For patients with rare diseases, the stakes are categorically different. Rare disease drugs cost what they cost because of the economics of developing treatments for conditions affecting a few thousand people nationally. If a UPL disrupts supply chain economics enough that a wholesaler stops stocking the drug in a state or a Center of Excellence can’t absorb reimbursement below acquisition cost, the rare disease patient doesn’t just pay more. They may lose access to a therapy that has no substitute anywhere.

Colorado’s own experience with Trikafta (the cystic fibrosis therapy that can run upwards of $200,000 a year) illustrates the concern. The PDAB considered it. Families mobilized. The board ultimately found it wasn’t unaffordable. The state is now debating whether to create a statutory exemption for rare disease drugs altogether.

And here’s the design trap the paper surfaces on rare disease exemptions: Washington State’s PDAB exemption only protects a drug if it’s approved for a single rare disease indication. If a manufacturer seeks FDA approval for a second rare disease indication on an already-approved product, the drug immediately loses its protection, even if the second indication is also for a rare disease. So, the exemptions designed to protect rare disease patients simultaneously create a financial disincentive to expand access to additional rare disease populations. That’s exactly the kind of second-order consequence that doesn’t come up in floor debates but absolutely shapes research and development decisions over time. This was fixed in the IRA as part of the One Big Beautiful Bill — it remains an issue for PDABs.

As of mid-2026, nine states have active PDABs. Colorado has reviewed five drugs, found three unaffordable, and is completing rulemaking on the first. Maryland has its UPL for Jardiance pending. Washington and Minnesota are still building administrative infrastructure. Oregon has reviewed 23 drugs and set no UPLs.

After all that work, no patient has seen a lower copayment because of a PDAB-set UPL.

States that want to reduce what patients pay have better options: risk pooling and reinsurance, capped copay programs that put a direct ceiling on patient out-of-pocket costs, restrictions on copay accumulator programs that ensure manufacturer assistance counts toward deductibles. These approaches share a design logic: the savings follow the dollar all the way to the patient.

Under current UPL design, the dollar stops at the health plan.

The Amgen ruling may or may not hold up on appeal. The Federal Circuit might clarify the patent preemption question in a way that opens the door for state UPL authority. But the more important conversation, the one I was trying to have this week and am trying to have here, is about what happens even when UPLs aren’t blocked. Because that’s the scenario PDAB advocates are still planning toward, and the supply chain math doesn’t change regardless of what the Federal Circuit does.

Follow the dollar all the way to the pharmacy counter. Right now, it stops well short.