The number everyone quotes about Medicaid and obesity drugs is thirteen. Thirteen state Medicaid programmes cover GLP-1s for the treatment of obesity under fee-for-service. It comes from a KFF brief, it is the figure that turns up in trade press and advocacy letters alike, and it is correct as far as it goes.
What is rarely quoted with it is the date. The brief carries the dateline “as of January 2026”. It was published on 16 January 2026 and, when we read it again on 21 September 2026, the page still carried that same date as its last modification. The map is eight months old.
That would matter less if the underlying number were stable. It is not. We checked four states at their own agencies, and in three of them the position has moved since January. This article is about how unstable the number is, and about what we could and could not verify ourselves. It is not about the drugs, and it says nothing about whether anyone should take them.
Sixteen, then thirteen, in three months
KFF’s own brief contains the movement inside it. Its annual survey of state Medicaid officials, carried out by Health Management Associates in November 2025, found sixteen state programmes covering GLP-1s for obesity treatment as of October 2025.
By January 2026 the count was thirteen. Four states had dropped coverage in the interval: California, New Hampshire, Pennsylvania and South Carolina. One state, North Carolina, had dropped it and put it back. Sixteen minus four plus one is thirteen.
A snapshot that moves by three states in a quarter is not a map. It is a still frame of something in motion, and its shelf life is shorter than the interval at which it gets refreshed.
The rule underneath
The reason states can move at all is a single clause of federal law. Under the Medicaid Drug Rebate Program, state programmes must cover nearly all FDA-approved drugs of a participating manufacturer. Weight-loss agents sit in a short list of exceptions that a state may leave out, at 42 U.S.C. § 1396r-8(d)(2)(A). So coverage for obesity is optional, while coverage of the same molecules for diabetes is mandatory, and has been mandatory for cardiovascular disease since Wegovy’s March 2024 approval and for moderate to severe obstructive sleep apnoea since Zepbound’s approval in December 2024. For children, coverage is required where it is medically necessary, under Medicaid’s EPSDT benefit.
The consequence is structural rather than medical. The same injection in the same body is a required benefit or an optional one depending on which diagnosis is written next to it. Every state decision described below happens inside that one sentence of statute.
Pennsylvania: what we read ourselves
We did not take the four cuts on trust. Pennsylvania’s is documented in a Medical Assistance Bulletin issued by the Office of Medical Assistance Programs on 24 November 2025, effective 1 January 2026, signed by Deputy Secretary Sally Kozak. We read it in full.
The bulletin is more specific than the summary. Pennsylvania’s Medical Assistance programme began covering GLP-1s for overweight and obesity in January 2023, so the coverage that ended on 1 January 2026 was three years old. From that date the drugs “will not be covered for the treatment of overweight and obesity” and remain covered “for all other medically accepted indications with a prior authorization”. Saxenda (liraglutide) is dropped for every indication. Obesity treatment agents that are not GLP-1s stay on the preferred drug list, phentermine among them, without prior authorisation.
One procedural line is easy to miss and affects everyone, not only people seeking the drug for obesity: every existing prior authorisation for a GLP-1 became invalid after 31 December 2025, “regardless of the condition being treated”. Someone who had been taking one for diabetes for years had to be re-authorised from scratch.
The part that surprised us
Here is what reading the handbook pages rather than the headline produced.
Pennsylvania no longer covers these drugs for obesity. But the renewal criteria for the indications it does still cover are, in several places, written in the language of weight loss and diet.
To renew a prior authorisation for any of the remaining indications, the beneficiary must be “continuing lifestyle changes and behavioral modification (e.g., healthy diet and increased physical activity)”. For obstructive sleep apnoea specifically, a renewal after six months requires, where dose titration is complete and the maximum tolerated dose has been taken for at least three consecutive months, “5% total body weight loss and documentation of dietary changes”.
Read the two together. The state has formally removed body size as a reason to pay. It has kept body size as a condition of continuing to pay for something else. A person prescribed the drug for sleep apnoea must lose five per cent of their body weight to keep the prescription that treats their sleep apnoea.
We are describing how the rule is built, not what it does to anyone’s health. But the construction is worth naming, because it is the quiet version of a pattern this magazine writes about often: the criterion is removed from the front door and reappears as a condition of staying in the room.
North Carolina: ten weeks without
North Carolina is the state KFF counts as having gone out and come back, and it is the one worth reading in the original, because both ends are documented by the agency itself.
On 5 September 2025, NC Medicaid announced that coverage of GLP-1s for the treatment of obesity would be discontinued effective 1 October 2025. The stated reason is in the first line of the discussion: “Given shortfalls in state funding”. Wegovy, Zepbound and Saxenda came off the preferred drug list as an off-cycle change, Saxenda for every indication, and existing prior authorisations stopped being valid after 30 September. The bulletin adds a sentence that is unusual for this kind of document: “NC Medicaid remains committed to the potential of GLP-1s for the treatment of obesity; however, at this time the lack of funding for the program prohibits continued coverage.”
On 19 December 2025 the agency published the reverse. Effective 12 December 2025, in accordance with a directive from the Governor, coverage reverted to the criteria in place as of 30 September 2025, which were themselves the criteria that had taken effect on 1 August 2024. Wegovy returned as a preferred product, Zepbound and Saxenda as non-preferred, which means a beneficiary must try and fail Wegovy, or document why they cannot take it, before either is approved.
The gap is ten weeks and change: 1 October to 11 December 2025. Both bulletins say the same thing about scale, that more than 3.1 million North Carolinians have coverage through Medicaid, and the December one adds an instruction to prescribers that describes the human side of a ten-week gap more plainly than any commentary could: providers “should consider whether members have experienced a lapse in treatment, change in therapy and whether a lower starting dose is necessary”.
That is what a policy reversal looks like from the pharmacy counter. Not a number changing on a map, but a restart at a lower dose.
The money, with the caveat attached
The KFF brief’s second half is about volume, and the figures are worth repeating precisely because they are so often repeated imprecisely.
Medicaid prescriptions for GLP-1s rose from about 1 million in 2019 to over 8 million in 2024, a sevenfold increase. Gross spending rose from about 1 billion dollars to almost 9 billion over the same period, a ninefold increase, and gross spending per prescription reached 1,000 dollars in 2024. In 2024 these drugs were about 1 per cent of all Medicaid prescriptions and over 8 per cent of Medicaid prescription drug spending.
The caveat is KFF’s own and it is load-bearing: gross means before rebates, and states typically receive substantial rebates on brand drugs. KFF notes that Novo Nordisk reported rebates and fees across all payers accounting for about 40 per cent of the cost of Ozempic and Wegovy. Anyone quoting the 9 billion without the word “gross” is quoting a number the source does not claim.
A second limitation, also KFF’s: the public data cannot separate how much of this use is for diabetes, cardiovascular disease, sleep apnoea or obesity. The spending figure is not an obesity-coverage figure, and it is used as one constantly.
Since January: four states, read at the agency
Rather than argue in the abstract that a January snapshot ages, we picked four of the states involved and read what each one’s own agency publishes. Three had moved. The fourth had not moved, but is described wrongly in the lists circulating online.
Massachusetts. MassHealth announced the end of coverage in Pharmacy Facts number 271, dated 12 March 2026. The wording is flat: MassHealth “will no longer cover drugs used for the treatment of obesity or overweight”. The named drugs include Wegovy and Zepbound, as well as Saxenda, Xenical and several older agents. Coverage continues past the cut-off for people with an approved prior authorisation for other accepted indications, listed as established cardiovascular disease, metabolic dysfunction-associated steatohepatitis, and moderate to severe obstructive sleep apnoea, with documentation of severity required. Requests for members under 21 are reviewed under EPSDT.
One detail is worth pausing on, because it is the subject of this article in miniature. The date circulating in secondary summaries is 1 July 2026. The state’s own bulletin says the drugs are no longer covered effective 3 July 2026, with existing authorisations end-dated for 2 July. It is two days, and nobody’s care turns on it. It is also evidence that the summaries were not written from the document.
Rhode Island. Rhode Island has not cut yet. It is about to, and the date is the nearest one in this article. EOHHS publishes its FY2027 Medicaid budget initiatives on a page last updated on 14 September 2026, and the section headed “GLP-1 Benefit Change” reads: “beginning October 1, 2026, Rhode Island Medicaid will no longer cover certain GLP-1 receptor agonists when they are prescribed solely for weight loss.” The page adds two administrative facts that matter for how quickly this happens: the change was directed by the enacted FY2027 budget, and “There is no federal authority required to implement this change.” No waiver, no state plan amendment, no CMS clock. A state that needs no federal permission can move faster than any tracker that describes it.
Michigan. Michigan is the case a map cannot show at all, because on a map it is still coloured in. MDHHS Numbered Letter L 25-73, issued 8 December 2025 and effective for dates of service on or after 1 January 2026, does not end coverage for obesity. It narrows it. Following Public Act 22 of 2025, coverage of a GLP-1 prescribed solely to treat obesity “will be reduced”, and approval now requires that the patient be classified as morbidly obese, that there be documented failure of all other clinically appropriate weight-loss interventions including a trial and failure of preferred anti-obesity agents, and that coverage “be considered only as a measure to avert the need for higher-cost bariatric surgery”. Saxenda and Wegovy become non-preferred for their other indications, which raises the copayment from one dollar to three for beneficiaries who are not exempt. Authorisations approved before 1 January 2026 were honoured for their remaining six months.
Michigan therefore counts as a covering state and is one, in the sense that a claim can be paid. Whether it is one in the sense a reader assumes when they look at a coloured map is a different question, and the map cannot answer it.
Kansas. Kansas is the state we checked because the secondary lists disagree with each other about it. Several overview pages assert that Kansas excludes GLP-1s when prescribed for weight loss. The Kansas Medical Assistance Program’s own Anti-Obesity Medications prior authorisation form, revised June 2025, lists liraglutide (Saxenda), semaglutide (Wegovy) and tirzepatide (Zepbound) as the medications the form exists to authorise, with sections for comorbidities, renewal criteria and switching between agents. Whatever else is true of Kansas, the state publishes a form for a benefit that the lists say it does not have.
That is the finding underneath the date problem. The lists are not merely stale. At least one of them is, on at least one state, the opposite of what the state itself publishes.
What we could not verify
Three things, said plainly.
We could not list the thirteen states from the brief’s text. The count appears in prose; the state-by-state detail is in a map graphic. What the text does name is the movement: the four states that cut, the one that restored, seven states (Connecticut, Louisiana, New Hampshire, New Mexico, North Dakota, Pennsylvania, Texas) that cover at least one weight-loss drug but no GLP-1 at all, and Utah, whose funding the brief records as limited to fiscal year 2026. A list of thirteen names circulating elsewhere is not something we can trace to this source, so we do not print one.
We verified two of the four cuts at the state’s own agency, Pennsylvania and North Carolina. California, New Hampshire and South Carolina we did not verify ourselves, and we are not treating them as our own reporting.
And we do not know what the count is today. What we can say is narrower and firmer than a guess: two of the thirteen, Massachusetts and Rhode Island, are documented by their own agencies as leaving, one in July and one on 1 October 2026, and a third, Michigan, has tightened its criteria to a point most readers would not recognise as coverage. A fourth, Utah, is recorded in KFF’s own footnote as having reported funding limited to fiscal year 2026, a year that ended for Utah on 30 June 2026; we were not able to open Utah’s prior authorisation document ourselves, so we report the footnote and not the outcome. Whether any state has added coverage in the same eight months we do not know, and we are not going to guess. The direction we can document points one way; the total we cannot.
Why the date is the story
There is a wider point here, and it is not about pharmacology.
Access for fat people in the United States is currently described to the public through a small number of periodically refreshed snapshots, and those snapshots are refreshed more slowly than the thing they describe changes. A state can cut in November, restore in December, and be counted in January in a way that is accurate on the day and misleading by March. Coverage becomes a matter of where you live and when someone last looked.
We have written before about how an access rule sorts people by body size and diagnosis in the Medicare GLP-1 Bridge, and about what to do when the world is not built for you in our guide to weight discrimination. This is the same subject at the level of the state line. The rule is not the only thing that decides who gets care. So does the calendar.
One note on framing, since we link to the source. KFF’s brief treats obesity as a treatable disease and coverage as access to treatment. That framing is not ours, and this magazine does not adopt it. We link to the brief because its numbers on coverage decisions and spending are the best documented ones available, not because we share its starting assumption about bodies.
Sources
- Elizabeth Williams, “Medicaid Coverage of and Spending on GLP-1s”, KFF, published 16 January 2026, page last modified 16 January 2026 (checked again and unchanged on 21 September 2026), read in full on 19 and 21 September 2026. https://www.kff.org/medicaid/medicaid-coverage-of-and-spending-on-glp-1s/ (count of thirteen states, the October 2025 survey figure of sixteen, the four eliminations and the North Carolina restoration, the seven-state and Utah footnotes, prescription and gross spending figures, the rebate caveat, the BALANCE model)
- Commonwealth of Pennsylvania, Department of Human Services, Office of Medical Assistance Programs, Medical Assistance Bulletin, “Coverage Change and Prior Authorization of GLP-1 Receptor Agonists”, issue date 24 November 2025, effective date 1 January 2026, including the attached Prior Authorization of Pharmaceutical Services handbook pages dated 1 January 2026. Read in full on 19 September 2026. https://www.pa.gov/content/dam/copapwp-pagov/en/dhs/documents/docs/publications/documents/forms-and-pubs-omap/mab2025112403.pdf
- NC Medicaid, “NC Medicaid to Change Coverage for GLP-1 Weight Management Medications”, 5 September 2025. https://medicaid.ncdhhs.gov/blog/2025/09/05/nc-medicaid-change-coverage-glp-1-weight-management-medications
- NC Medicaid, “NC Medicaid to Reinstitute Coverage of GLP-1s for Weight Management”, 19 December 2025. https://medicaid.ncdhhs.gov/blog/2025/12/19/nc-medicaid-reinstitute-coverage-glp-1s-weight-management
- MassHealth, Pharmacy Facts, Number 271, 12 March 2026 (corrected), “MassHealth Changes to Management of Anti-Obesity Medications”. Read in full on 21 September 2026. https://www.mass.gov/doc/pharmacy-facts-271-march-12-2026-corrected-0/download (end of coverage for obesity and overweight effective 3 July 2026, end-dating of prior authorisations for 2 July 2026, the continuing indications and the EPSDT provision)
- Rhode Island Executive Office of Health and Human Services, “Fiscal Year 2027 Medicaid Budget Initiatives”, section “GLP-1 Benefit Change”, page last updated 14 September 2026, read in full on 21 September 2026. eohhs.ri.gov/initiatives/medicaid-initiatives/fy27-budget-initiatives (coverage ending 1 October 2026 for prescriptions solely for weight loss; no federal authority required)
- Michigan Department of Health and Human Services, Numbered Letter L 25-73, “Update of Pharmacy Drug Coverage for Treatment of Obesity”, 8 December 2025, effective for dates of service on or after 1 January 2026, signed Meghan E. Groen. Read in full on 21 September 2026. https://www.michigan.gov/mdhhs/-/media/Project/Websites/mdhhs/Assistance-Programs/Medicaid-BPHASA/2025-L-Letters/Numbered-Letter-L-25-73-Pharmacy-Final.pdf (the three additional criteria, the non-preferred reclassification and copayment, Public Act 22 of 2025, six-month honouring of existing authorisations)
- Kansas Department of Health and Environment, “Anti-Obesity Medications (AOMs) Prior Authorization Form”, Kansas Medical Assistance Program, revised 06/2025. Read in full on 21 September 2026. https://www.kdhe.ks.gov/DocumentCenter/View/15513/Anti-Obesity-Medications-PA-Form-PDF (liraglutide, semaglutide and tirzepatide listed as authorisable anti-obesity medications)
- 42 U.S.C. § 1396r-8(d)(2)(A), the statutory exception permitting states to exclude agents used for weight loss, cited in both the KFF brief and the Pennsylvania bulletin.