When a state Medicaid programme starts paying for GLP-1 drugs labeled for obesity, the bill for those drugs goes up. A study in JAMA Health Forum suggests the net increase is smaller than that figure implies, because some of the new obesity-labeled prescriptions appear to replace diabetes-labeled versions of the same molecules, which every state already covers. The finding bears directly on the coverage decisions several states have been reversing on cost grounds — the counterfactual to “cover Wegovy” is not “no GLP-1 spending”.
What the study found
Researchers at the University of Pennsylvania and the Philadelphia VA used Medicaid’s public State Drug Utilization Data for 2021 to 2024. Nine states that began covering at least one obesity-labeled GLP-1 drug in that window were compared with 36 that did not, using a synthetic difference-in-differences method that weights the comparison states to match the treated states’ earlier trends. Five other covering states were excluded, mostly because their start dates were unclear.
Coverage added 6.48 obesity-labeled prescriptions per 1,000 beneficiaries per quarter (95% confidence interval 4.00 to 8.96), from a starting point of 0.36. Total GLP-1 prescriptions rose by 4.17 per 1,000 (−0.33 to 8.67), and diabetes-labeled prescriptions fell by 2.16 per 1,000 (−5.27 to 0.96). Put another way, the fall in diabetes-labeled fills offset roughly a third of the obesity-labeled gain (2.16 ÷ 6.48 ≈ 33%). Neither the total nor the diabetes-labeled estimate reached statistical significance, so this is a signal rather than a settled figure. In states whose prior authorisation rules for diabetes-labeled drugs were stricter than the FDA label, the diabetes-labeled drop was significant: 3.02 fewer per 1,000 (0.36 to 5.69).
Medicaid reimbursement for obesity-labeled drugs rose by $7,162 per 1,000 beneficiaries against a baseline of $459; total GLP-1 reimbursement rose by $4,621, with an interval running from a $16 fall to a $9,258 rise.
Why it matters for the price argument
The active ingredients are identical. Tirzepatide is sold as Mounjaro for diabetes and Zepbound for obesity at the same doses; semaglutide as Ozempic and Wegovy, with Wegovy reaching a higher top dose. The authors lay out three routes to substitution: patients with both conditions choosing the obesity version, patients using a diabetes version off-label switching once the obesity version is covered, and people with prediabetes who never progress to needing the diabetes drug.
For a budget office, that means projections built only on new obesity-labeled prescriptions overstate the cost of coverage, and projected savings from dropping coverage may be overstated too if patients drift back to diabetes-labeled fills. For patients outside Medicaid, the study is a reminder that label, not molecule, drives what gets covered; our guide to what these drugs cost without insurance sets out the self-pay side.
What is not yet known
This is an ecological study of state-level totals, not a record of individual patients switching, so it cannot show that any given person moved from Ozempic to Wegovy. The key substitution estimates were not statistically significant overall, and only three treated states had more than eight quarters of follow-up. The data end in 2024 and predate list-price cuts and the coverage withdrawals of 2026. Small counts are suppressed in the source data for privacy. One author disclosed uncompensated work with Eli Lilly on designing a separate GLP-1 study.
Sources: Khanna R, Wang MC, Venkataramani AS. “GLP-1 receptor agonist coverage for obesity in Medicaid and total GLP-1 receptor agonist utilization.” JAMA Health Forum 7(9): e263353, published 4 September 2026. doi:10.1001/jamahealthforum.2026.3353.
