A new analysis in the Journal of Managed Care & Specialty Pharmacy ↗ crosses state-level GLP-1 utilization with state-level diabetes and obesity prevalence in 2023, and reaches a conclusion that anyone tracking the GLP-1 rollout should hold up against their assumptions: the strongest predictor of who gets a GLP-1 drug is not whether they need it. It is whether their insurance covers it.
The data is from Komodo Healthcare Map, the same large pharmacy-claims database that has anchored most major real-world GLP-1 analyses this cycle, covering January through December 2023. State-level utilization rates were calculated separately for the diabetes indication and the obesity indication. State-level diabetes and obesity prevalence came from the CDC Behavioral Risk Factor Surveillance System. Medicaid coverage policies for GLP-1s in weight loss and diabetes were classified as "not covered," "restricted," or "unrestricted."
The diabetes story is straightforward. State-level utilization tracks state-level diabetes prevalence in the commercially insured population. Where there are more diabetes patients, there are more GLP-1 prescriptions, in the same ratio. That correlation weakens in Medicare, and weakens further in Medicaid. The reason is the gradient of coverage. Commercial plans cover GLP-1s for diabetes uniformly; Medicare covers them with conditions; Medicaid coverage varies state by state.
The obesity story is the more striking finding. State-level GLP-1 utilization for obesity correlated only weakly, or negligibly, with state-level obesity prevalence. The states with the most obesity are not the states with the most obesity GLP-1 prescriptions. The driver is Medicaid policy. States with restricted or unrestricted Medicaid coverage of obesity GLP-1s show nominally higher utilization than states with no coverage. The disease is everywhere; access is not.
This matters because the obesity-GLP-1 conversation in 2026 has been dominated by the federal-level question (the Medicare Bridge program, BALANCE, CMS funding through 2027). The federal layer is real, but the JMCP analysis shows that beneath federal policy is a state-level patchwork that already creates two-tier access. A patient in Mississippi with severe obesity and Medicaid coverage may get nothing. A patient in California with the same severity and the same payer category may get coverage. The same body, the same disease, different ZIP codes, different drugs.
For the broader peptide industry the read is straightforward. Commercial-channel demand follows disease distribution. Government-channel demand follows policy distribution. The total addressable market for obesity GLP-1s, calculated against disease prevalence, overstates what the system actually buys. The realistic addressable market depends on the state-by-state coverage map, and Medicaid alone has 50 different answers.
Two limitations of the analysis worth flagging. First, the cross-section is one calendar year (2023), which captures the period before the major coverage expansions of 2025 and the BALANCE pivot of 2026. Second, the obesity-utilization correlation is a single Pearson coefficient, which is a blunt instrument for policy comparison. A state that recently expanded coverage may show low utilization simply because the prescribing infrastructure has not yet caught up. Both limitations argue for a follow-up using more recent data, which the underlying database supports.
What is durable about the result is the gap. The disease distribution is a known map. The drug distribution does not match it. Closing that gap is a policy question, not a science question, and it is the question the next eighteen months of GLP-1 reimbursement debate will turn on.