A single molecule now clears the US market at four different prices at once, and two of them never touch a pharmacy claim. Milliman’s August 2026 white paper collects all four for semaglutide and pairs them with real-world discontinuation reported between 37% and 81% within a year (Milliman, August 2026).
Unit cost and persistency are moving in opposite directions at the same time. The numerator is being reset by federal price negotiation on a schedule that carries a published end date, while the denominator a plan can actually measure erodes as members buy the same drug through cash-pay and telehealth channels that generate no claim.
Key Takeaways
- $274 for a 30-day supply is the Medicare maximum fair price for semaglutide from January 2027, a 71% cut from the calendar 2024 list price, and the only GLP-1 among the 15 drugs in the second negotiation cycle.
- 37% to 81% is the reported spread of real-world discontinuation within twelve months, wide enough that annualised cost per starter cannot be approximated as twelve times any of the four monthly prices.
- 64.8% against 46.5% splits that spread by indication: patients without type 2 diabetes discontinued within a year at 64.8%, those with it at 46.5%, across 125,474 adults (JAMA Network Open, January 2025).
- $675 per month is Novo Nordisk’s reduced list price from January 2027, down from over $1,000. Because rebates are struck as a percentage of list, the cut removes rebate dollars plans currently book against premium.
- 43% of firms with 5,000 or more workers covered GLP-1s for weight loss in 2025, up from 28% a year earlier, so the repricing lands on an exposed base that is itself moving (KFF, October 2025).
Four Prices, One Molecule
The white paper, by Cherie Dodge and Winston Fopalan, is a coverage and cost survey rather than a forecast. Its value to a pricing actuary is the collection itself: four live prices for semaglutide, each attaching to a different population through a different payment path.
The maximum fair price is $274 for a 30-day supply, a 71% reduction from the calendar 2024 list price, and from 2027 all Medicare Part D formularies must include the drug. Alongside it, eligible medications furnished through the Medicare GLP-1 Bridge are reimbursed at $245 per monthly supply against a $50 beneficiary copayment, with CMS covering the remainder.
| Price basis | Amount (30-day) | In force | Population reached |
|---|---|---|---|
| Medicare maximum fair price | $274 | From January 1, 2027 | Part D, mandatory formulary inclusion |
| Medicare GLP-1 Bridge net price | $245, with a $50 copay | July 1, 2026 to December 31, 2027 | Eligible Part D beneficiaries, weight loss |
| Direct-to-consumer cash price | $199 for two months, then $349 | Current | Self-pay patients, no coverage required |
| Manufacturer list price (WAC) | $675, cut from over $1,000 | From January 1, 2027 | Basis for coinsurance and rebate math |
The direct-to-consumer figures come from the TrumpRx platform and Novo Nordisk’s own channel, which advertise Wegovy injection at $199 for the first two months and $349 a month thereafter. Novo Nordisk announced in February 2026 that it would cut the list price for Wegovy, Ozempic and Rybelsus to $675 a month effective January 1, 2027. Eli Lilly had not reduced the list price for Zepbound or Mounjaro as of the report.
The scale behind these prices is why the collection matters. US expenditure on the class rose from $13.7 billion in 2018 to $71.7 billion in 2023, and GLP-1s accounted for close to half of all drug spending growth in 2024, when total US prescription drug spending reached $805.9 billion (ASHP, April 2025). One commercial plan reported that its five most frequently used GLP-1s made up nearly 20% of all drug spending in 2024.
Milliman’s own framing is that the phase has changed: "The GLP-1 landscape is transitioning from rapid gross-spend escalation to a more complex phase shaped by federal price negotiation, MFN pricing, evolving coverage pilots, and adherence challenges" (Milliman, August 2026). Two of the four prices in the table, the cash channel and the demonstration, sit outside ordinary pharmacy benefit adjudication entirely.
What Persistency Does to Cost Per Starter
The paper pairs its price collection with a blunt adherence caveat, noting that "reported discontinuation rates ranging from 37% to 81%" mean payers have to re-examine utilization management against real-world behaviour (Milliman, August 2026). That range is doing more work than it appears to.
The underlying study decomposes it. Across 125,474 adults who newly initiated a dual-labeled GLP-1 between 2018 and 2023, 46.5% of patients with type 2 diabetes and 64.8% without it discontinued within one year (JAMA Network Open, January 2025). The indication, not the drug, is the dividing line.
The same study carries the figure that breaks the simple arithmetic. Among those who discontinued, 47.3% of patients with type 2 diabetes and 36.3% without it reinitiated therapy within a year. Discontinuation is not absorbing. A member who stops at month five and restarts at month eleven is neither a full-year cost nor a dropped one, so the effective utilisation factor is not one minus the discontinuation rate.
At the $274 negotiated price, an uninterrupted year runs $3,288. Apply a 64.8% one-year discontinuation rate and cost per starter falls well below that; add back a 36.3% reinitiation rate and a meaningful share returns, at a timing the claims history cannot yet support. The spread between those two anchors is wider than the 71% price cut that produced the $274 figure in the first place.
Channel mix compounds it in a direction that is easy to get backwards. Steering members toward a cheaper unit price can raise cost per completed course, because the cheapest channels are the ones with the weakest persistency signal and no medical policy review. Online prescribing platforms rely primarily on patient-reported questionnaires rather than comprehensive clinician assessment (JAMA, July 2026), and a plan sees none of that utilisation.
The list-price cut carries its own reversal. Manufacturer rebates have historically been calculated as a percentage of list, so dropping wholesale acquisition cost from above $1,000 to $675 reduces the rebate dollars available to health plans and pharmacy benefit managers. A lower unit price can worsen a plan’s net position while improving a member’s.
All of this lands on a base that is growing rather than static. The Milliman paper anchors employer coverage at 18% of large firms in 2024, but KFF’s 2025 survey puts coverage at 43% of firms with 5,000 or more workers, up from 28%, with 59% of those firms reporting utilisation above expectations and 66% reporting a significant effect on drug spend (KFF, October 2025). The 2026 Milliman Medical Index puts pharmacy trend at 14.8%, with pharmacy and outpatient together supplying 69% of the year-over-year increase.
A Price Basis With a Stated Expiry Date
The cheapest of the four prices is the least durable. The Medicare GLP-1 Bridge is a time-limited demonstration running July 1, 2026 through December 31, 2027, and Part D sponsors do not opt into it (CMS). They carry no risk on the drugs it covers.
That design keeps the experience it generates outside Part D. The $50 copay sits outside the deductible, does not count toward true out-of-pocket costs, and attracts no low-income subsidy. CMS Administrator Mehmet Oz framed the intent plainly: "These treatments are a major medical advancement, but too many seniors are currently unable to access them due to high cost" (CMS, May 2026).
The consequence for a projection is specific. A plan pricing 2028 will hold roughly eighteen months of GLP-1 utilisation generated under a cost-sharing structure that no longer exists, at a net price that was never its own, for a population selected by demonstration eligibility rather than by its formulary. The $274 maximum fair price is statutory and persists; the $245 price stops on a fixed date, and the two apply to overlapping members.
No successor is defined. The BALANCE model was originally proposed for Part D but failed to meet its minimum participation threshold and is not moving forward there; on the Medicaid side it began in May 2026, and as of July 2026 CMS had not announced which state programs elected to participate. The Medicaid baseline it would attach to is itself contracting: only 13 state fee-for-service programs covered GLP-1s for weight loss as of January 2026, with California, New Hampshire, Pennsylvania and South Carolina having dropped coverage in the preceding year, even as Medicaid spending on the category rose roughly ninefold between 2019 and 2024.
The fiscal arithmetic that closed the simpler route is still in place. CMS proposed reinterpreting the statute to allow anti-obesity medication coverage in November 2024, then declined to finalise it in April 2025 after the Congressional Budget Office projected $35.5 billion of added federal spending over eight years. A demonstration that expires inside the projection horizon is what stands in place of that coverage decision, and the discontinuity it creates at the end of 2027 cannot be sized from the demonstration’s own experience.
Further Reading
- Medicare GLP-1 Bridge Launches July 1: What Part D Plan Actuaries Need to Know - The operational mechanics of the demonstration behind the $245 net price, including its central processor design and the 2028 bid problem its December 2027 expiry creates.
- CMS GLP-1 Bridge Sets $50 Copay, Rewriting Part D Actuarial Math - How the copay structure sits outside the deductible, TrOOP and the low-income subsidy, and why that removes the experience from Part D reporting.
- The 2027 Employer GLP-1 Coverage Split and What It Does to Trend - The commercial-market counterpart to the Medicare repricing, where coverage decisions rather than negotiated prices set the exposed base.
- Oral GLP-1 Launch Forces 2027 Pharmacy Trend Repricing - Why adherence to oral formulations is the open variable in the persistency assumption, and how a new dosage form reshapes step therapy.
- Decomposing the 2026 Milliman Medical Index Pharmacy Trend - The 14.8% pharmacy trend figure broken into its drivers, and where GLP-1 utilisation sits within it.
- Catastrophic Claims and Stop-Loss Trend Into 2027 - The attachment-point pressure that GLP-1 utilisation transmits into the stop-loss layer for self-funded plans.
Sources
- Milliman: GLP-1 Medications in the United States, The Status of Coverage, Utilization Management, and Cost in a Rapidly Evolving Market (Dodge and Fopalan, August 2026)
- Milliman: GLP-1 Medications Research Hub
- CMS: Medicare GLP-1 Bridge (Official Program Page)
- CMS: Coming Soon, CMS to Provide $50 Monthly Access to GLP-1 Medications for Medicare Beneficiaries (May 6, 2026)
- CMS: Medicare Drug Price Negotiation Program, Selected Drugs and Negotiated Prices (May 2026)
- JAMA Network Open: Discontinuation and Reinitiation of Dual-Labeled GLP-1 Receptor Agonists Among US Adults With Overweight or Obesity (Rodriguez et al., January 2025)
- JAMA: Online Prescribing of GLP-1 Receptor Agonists (Chetty et al., July 2026)
- KFF: 2025 Employer Health Benefits Survey (October 2025)
- KFF: Medicaid Coverage of and Spending on GLP-1s (January 2026)
- Milliman: 2026 Milliman Medical Index (May 2026)
- Novo Nordisk: Significant Reduction in US List Price for Wegovy, Ozempic and Rybelsus (February 2026)
- ASHP: US Drug Spending Up 10% in 2024, With Weight Loss Drugs Remaining Top Driver (April 2025)
- Congressional Budget Office: How Would Authorizing Medicare to Cover Anti-Obesity Medications Affect the Federal Budget? (October 2024)
- CMS Innovation Center: BALANCE Model (June 2026)