GLP-1 Receptor Agonist Coverage for Obesity in Medicaid and Total GLP-1 Receptor Agonist Utilization

Ritikaa Khanna; Michael C Wang; Atheendar S Venkataramani

doi:10.1001/jamahealthforum.2026.3353

. 2026 Sep 25;7(9):e263353. doi: 10.1001/jamahealthforum.2026.3353

GLP-1 Receptor Agonist Coverage for Obesity in Medicaid and Total GLP-1 Receptor Agonist Utilization

Ritikaa Khanna ¹, Michael C Wang ^1,^2,^✉, Atheendar S Venkataramani ^2,³

¹Corporal Michael J. Crescenz Veterans Affairs Medical Center, Philadelphia, Pennsylvania

²Leonard David Institute of Health Economics, University of Pennsylvania, Philadelphia

³Department of Medical Ethics and Health Policy, University of Pennsylvania Perelman School of Medicine, Philadelphia

Accepted for Publication: July 16, 2026.

Published: September 25, 2026. doi:10.1001/jamahealthforum.2026.3353

Open Access: This is an open access article distributed under the terms of the CC-BY-NC-ND License, which does not permit alteration or commercial use, including those for text and data mining, AI training, and similar technologies. © 2026 Khanna R et al. JAMA Health Forum.

^✉

Corresponding Author: Michael C. Wang, MD, Department of Medicine, University of Pennsylvania Perelman School of Medicine, 3600 Civic Center Blvd, 7F, Philadelphia, PA 19104 (michael.wang2@pennmedicine.upenn.edu).

Author Contributions: Drs Wang and Venkataramani had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Ms Khanna and Dr Wang contributed equally as co–first authors.

Concept and design: Wang, Venkataramani.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Khanna, Wang.

Critical review of the manuscript for important intellectual content: Wang, Venkataramani.

Statistical analysis: Khanna, Venkataramani.

Administrative, technical, or material support: Venkataramani.

Supervision: Wang, Venkataramani.

Conflict of Interest Disclosures: Dr Venkataramani reported grants from National Institutes of Health for multiple grants unrelated to the present work, grants from Robert Wood Johnson Foundation for a research grant unrelated to this work, grants from WorkRise Foundation for research unrelated to this work, and grants from Independence Blue Cross (IBX) for research unrelated to this work outside the submitted work; and Dr Venkataramani worked closely with the Eli Lilly Corporation on designing a study related to GLP-1 RA. He was not compensated for this work in any way. This work will culminate in a contract to study the impacts of GLP-1 RA on social and economic well-being. However, the contract is not in place at this time. No other disclosures were reported.

Disclaimer: The article content does not necessarily reflect the views of the US Department of Veterans Affairs.

Data Sharing Statement: See Supplement 2.

^✉

Corresponding author.

PMCID: PMC13615463 PMID: 42789287

Key Points

Question

Is Medicaid coverage of obesity-labeled glucagon-like peptide-1 receptor agonists (GLP-1 RAs) associated with obesity-labeled, diabetes-labeled, and total GLP-1 RA utilization?

Findings

In this cross-sectional ecological analysis of US state-level Medicaid drug utilization data, states covering obesity-labeled GLP-1 RAs had a mean increase of 6.48 prescriptions per 1000 beneficiaries of obesity-labeled GLP-1 RAs. In the subgroup of states with more restrictive prior authorization requirements for obesity-labeled GLP-1 RAs, there was a relative decrease of 3.02 per 1000 diabetes-labeled GLP-1 RAs.

Meaning

Results of this study suggest that substitution effects are possible between obesity-labeled and diabetes-labeled GLP-1 RAs, that could impact coverage decisions.

Abstract

Importance

Glucagon-like peptide-1 receptor agonists (GLP-1 RAs) are effective for treating obesity, but few Medicaid programs cover them due to cost concerns with accelerating utilization. However, potential substitution effects between obesity-labeled GLP-1 RAs and diabetes-labeled GLP-1 RAs, which have identical active ingredients and which all states cover, are not well understood.

Objective

To determine whether Medicaid coverage of obesity-labeled GLP-1 RAs is associated with obesity-labeled, diabetes-labeled, and total GLP-1 RA utilization.

Design, Setting, and Participants

This cross-sectional study used US state-level Medicaid drug utilization data from 2021 to 2024. Synthetic difference-in-differences was used to assess the association of obesity-labeled GLP-1 RA coverage with GLP-1 RA utilization.

Exposure

Medicaid program coverage of obesity-labeled GLP-1 RAs.

Main Outcomes and Measures

The primary outcome was aggregate quarterly prescriptions per 1000 beneficiaries; secondary outcomes included prescription-months (accounting for duration of fills) and Medicaid reimbursement per 1000 beneficiaries.

Results

Of 9 states (n = 89 021 657; mean [SD] of 49.4% [0.5%] were male, 50.6% [0.5%] were female) that began Medicaid coverage of obesity-labeled GLP-1 RAs between 2021 and 2024 and 36 control states (n = 212 786 953; mean [SD] of 49.8% [0.9%] were male, 50.2% [0.9%] were female), overall and diabetes-labeled GLP-1 RA utilization increased substantially in both groups, whereas obesity-labeled GLP-1 RA increased significantly only in treated states. Estimates from synthetic difference-in-differences models demonstrated a mean increase in obesity-labeled GLP-1 RA of 6.48 prescriptions per 1000 beneficiaries (95% CI, 4.00-8.96; P < .001), with estimates increasing nearly each quarter after Medicaid coverage was implemented. By contrast, changes in total GLP-1 RA prescriptions compared with obesity-labeled prescriptions were not statistically significant (4.17 per 1000 beneficiaries [95% CI, −0.33 to 8.67; P = .07]), nor were those for diabetes-labeled GLP-1 RA prescriptions (−2.16 prescriptions per 1000 beneficiaries [95% CI, −5.27 to 0.96; P = .17]). In the subgroup of states with prior authorization requirements more restrictive than the Food and Drug Administration label, there was a significant relative decrease in diabetes-labeled GLP-1 RA utilization of 3.02 prescriptions per 1000 beneficiaries (95% CI, −5.69 to −0.36; P = .03).

Conclusions and Relevance

Results of this study suggest that substitution effects are possible between obesity-labeled and diabetes-labeled GLP-1 RAs, varying by state coverage policies, that could impact coverage decisions.

This cross-sectional study uses US state-level Medicaid drug utilization data to assess whether Medicaid coverage of obesity-labeled glucagon-like peptide-1 receptor agonists (GLP-1 RAs) is associated with obesity-labeled, diabetes-labeled, and total GLP-1 RA utilization.

Introduction

Glucagon-like peptide-1 receptor agonists (GLP-1 RAs) demonstrated significant weight loss of, on average, 12% to 20% of body weight in randomized clinical trials of patients with obesity.^1,2 Due to greater efficacy for treating obesity than prior approved medications, uptake of obesity-labeled GLP-1 RAs has been accelerating.^3,4

The large clinically eligible pool of patients⁵ for these medications has caused wide concerns about escalating drug costs.^6,7 Due to a statutory carve-out for obesity-labeled medications, state Medicaid programs can make individual decisions on coverage instead of following the general policy of covering all US Food and Drug Administration (FDA)–approved drugs from manufacturers participating in the Medicaid Drug Rebate Program.⁸ As a result, less than one-third of states choose to cover GLP-1 RAs for obesity through their Medicaid programs, whereas all cover GLP-1 RAs for diabetes.⁹

Budgetary concerns have led several states to discontinue Medicaid coverage of obesity-labeled GLP-1 RAs as of 2026.¹⁰ However, those states continue to cover diabetes-labeled GLP-1 RAs. The GLP-1 RAs labeled for obesity and diabetes have identical active ingredients. For tirzepatide, the doses are identical between indications, whereas for semaglutide and liraglutide, the 3 lower doses are identical while the obesity-labeled version has higher maximum doses.

The interchangeability of obesity-labeled and diabetes-labeled GLP-1 RAs for therapeutic effects in weight and hemoglobin A_1c sets up potential substitution effects between them. First, patients with comorbid diabetes and obesity could use versions labeled for either indication, and in response to obesity-labeled GLP-1 RA coverage may take an obesity-labeled version rather than a diabetes-labeled GLP-1 RA due to factors such as preference or availability of higher doses. Second, patients taking diabetes-labeled GLP-1 RAs off-label for weight loss could switch to a newly covered formulation labeled for obesity. Third, patients with obesity and prediabetes, who have elevated risk of progression to diabetes, may prevent incident diabetes by using an obesity-labeled GLP-1 RA,¹¹ thus preventing later initiation of a diabetes-labeled GLP-1 RA.

If there are substitution effects between the label indications of GLP-1 RAs, the utilization and cost impacts of decisions to add or remove coverage for obesity-labeled GLP-1 RAs may be less than implied by examining only effects in obesity-labeled GLP-1 RA utilization. In this study, we use the differential coverage decisions of state Medicaid programs to estimate the association of obesity-labeled GLP-1 RA Medicaid coverage with total GLP-1 RA utilization. We hypothesized that coverage of obesity-labeled GLP-1 RAs would increase total GLP-1 RA utilization by less than the increase in obesity-labeled GLP-1 RA utilization through a relative reduction in diabetes-labeled GLP-1 RA utilization.

Methods

Sample and Data

In this cross-sectional ecological analysis, our sample consisted of US states and Washington, DC, from 2021 to 2024. The study period captures the FDA approvals of semaglutide (Wegovy) (June 2021) and tirzepatide (Zepbound) (November 2023), which have driven recent increases in obesity-labeled GLP-1 RA utilization, and excludes data from 2020 to mitigate potential confounding effects of the COVID-19 pandemic. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline. Regulatory approval was not needed per the policy of the University of Pennsylvania as the study does not involve private data or biospecimens.

We obtained data on quarterly drug utilization and cost for each state from the Medicaid State Drug Utilization Data (SDUD).¹² These publicly available data are generated quarterly by state Medicaid programs and submitted to the Centers for Medicare & Medicaid Services (CMS) as part of the Medicaid Drug Rebate Program. Both fee-for-service Medicaid and managed care plans are included. The GLP-1 RAs used for the analysis included the obesity-labeled GLP-1 RAs liraglutide (Saxenda), semaglutide (Wegovy), and tirzepatide, and the diabetes-labeled GLP-1 RAs liraglutide (Victoza), dulaglutide (Trulicity), semaglutide (Ozempic and Rybelsus), and tirzepatide (Mounjaro). State-level population characteristics were obtained from the 2021 American Community Survey and the 2021 Behavioral Risk Factor Surveillance System, and quarterly enrollment data were obtained from CMS. Race and ethnicity were self-reported by individuals and summarized to provide demographic context to the data.

Exposure and Outcomes

The exposure of interest was Medicaid coverage of at least one obesity-labeled GLP-1 RA. For most prescription medications, including diabetes-labeled GLP-1 RAs, Medicaid programs cover all FDA-approved drugs from manufacturers participating in the Medicaid Drug Rebate Program, although they may impose use restrictions.⁸ By contrast, Medicaid coverage of obesity-labeled medications, by statute, is discretionary; thus, some states decided to cover at least one obesity-labeled GLP-1 RA during the study period, whereas most states lacked coverage.

The states that covered obesity-labeled GLP-1 RAs under Medicaid during the study period were identified using prior data.^13,14,15 We then ascertained the quarter of coverage start by obtaining and manually verifying archival records from state Preferred Drug Lists, official announcements and documents on state Medicaid websites, and relevant news articles (eTable 1 in Supplement 1). We further verified empirically that the start of coverage ascertained from primary sources coincided with an expected increase in utilization of obesity-labeled GLP-1 RAs within the data. A total of 14 states reported Medicaid coverage of GLP-1 RA for obesity during the study period. Three states were excluded due to unclear start dates of coverage (Delaware, Kansas, Wisconsin), one for restricted coverage only for patients with cardiovascular disease (North Carolina), and one for coverage only in the last quarter of the study (South Carolina) (eTable 1 in Supplement 1). Nine states were included in the treatment group. Of 37 control states, Kentucky was excluded due to an unexplained discontinuous spike in GLP-1 RA utilization in the second quarter of 2024 within the data; 36 states were included in the control group (eFigure 1 in Supplement 1).

The primary outcome was GLP-1 RA utilization, defined as the number of prescriptions per 1000 Medicaid beneficiaries. Secondary outcomes included prescription-months and Medicaid reimbursement per 1000 beneficiaries. Since prescriptions may be written for different supplies of medication, we constructed a prescription-months variable by multiplying the number of units reimbursed (available in the dataset) by the months’ supply per unit (which we derived from National Drug Codes matched to package label information)¹⁶ (eTable 2 in Supplement 1).

The SDUD only reports, within each grouping of national drug code, quarter, state, and fee-for-service vs managed care, observations for which a state recorded actual prescription fills. Thus, data absent from the dataset reflect a lack of filled prescriptions of that drug and do not contribute to our outcomes. The SDUD also suppresses the number of prescriptions for a given drug, quarter, state, and plan if it is less than 11 for privacy protection. We excluded these observations from the main analysis and tested the implications of this assumption in sensitivity analyses.

Statistical Analysis

We summarized unadjusted trends of GLP-1 RA utilization, overall and separately for obesity-labeled and diabetes-labeled versions. Trends were stratified by obesity-labeled GLP-1 RA coverage status, state, and individual medications.

We then estimated the association of obesity-labeled GLP-1 RA coverage with GLP-1 RA utilization using a synthetic difference-in-differences (SDID) approach.¹⁷ There are likely factors influencing state Medicaid coverage decisions for obesity-labeled GLP-1 RAs, such as the political environment and generosity of formularies, that also affect GLP-1 RA utilization trends. This limits the ability of states without obesity-labeled GLP-1 RA coverage to act as unbiased controls in the traditional difference-in-differences framework. We also observed nonparallel precoverage trends in the data, making the parallel trends assumption unlikely to hold. SDID addresses this concern by constructing weights to control states and preintervention time periods, which create a weighted control group that has parallel precoverage trends, assessed during time periods that are most similar to the postcoverage periods. Since states began obesity-labeled GLP-1 RA coverage during several different quarters, we estimated event study SDID models.¹⁸ We used up to 8 quarters of follow-up, as only 3 treated states had longer than 2 years of follow-up. As these models contain state-level fixed effects, and state-level characteristics are likely to be minimally time varying over the study period, we estimated models without additional covariates.

We estimated separate models for total GLP-1 RAs, obesity-labeled GLP-1 RAs, and diabetes-labeled GLP-1 RAs, which allowed us to assess changes in utilization and cost between versions as factors in overall outcomes. Because specific coverage policies varied by state, we conducted a subgroup analysis based on coverage criteria as of June 2024 in fee-for-service Medicaid programs, as documented by Klebanoff et al.¹⁵ We divided the treated states into coverage with no prior authorization or with prior authorization that matched the FDA label (California, Pennsylvania, Minnesota) or coverage with prior authorization that was more restrictive than the FDA label (Massachusetts, Michigan, Mississippi, New Hampshire, Virginia). Rhode Island was excluded from this analysis due to inability to determine exact coverage criteria. We also estimated models for each treated state individually. Models used bootstrap SEs for inference, except for state-specific models, which used placebo SEs due to having a single treated unit.

We conducted several sensitivity analyses. We examined the impact of excluding suppressed prescription values by computing precoverage averages for obesity-labeled GLP-1 RAs in treated states under assumptions of 10 prescriptions, the maximum possible value, rather than 0 as in the main analysis. We also evaluated whether results were robust to assuming suppressed values are 10 instead of 0, restricting the sample to states that expanded Medicaid prior to 2021, excluding liraglutide formulations (Saxenda and Victoza), and limiting the treatment group to a balanced panel of 4 states with full follow-up to 8 quarters. We also compared our results with models using the Callaway and Sant’Anna estimator.¹⁹ Callaway and Sant’Anna difference-in-differences is designed for settings with multiple time periods and staggered treatment adoption but was not chosen for the primary analysis due to its reliance on a stronger parallel trends assumption. To investigate the potential for bias related to the Medicaid unwinding, which reduced Medicaid enrollment in 2023 and 2024,²⁰ we used 2 approaches: first, we conducted a sensitivity analysis fixing each state’s denominator to its preunwinding average Medicaid enrollment (ie, 2021-2022); second, we investigated whether there was a significant association between the log of the Medicaid enrollee population and total GLP-1 RA prescriptions per 1000 beneficiaries within the SDID model.

All statistical tests were 2-tailed, with a P value <.05 denoting statistical significance. Statistical analysis was conducted using Stata/MP version 18 (StataCorp LLC).

Results

Baseline demographic, economic, and health characteristics were largely similar between the 9 states covering obesity-labeled GLP-1 RAs (n = 89 021 657) and 36 control states (n = 212 786 953). For example, in the control group, a mean (SD) of 49.8% (0.9%) were male compared with 49.4% (0.5%) in the obesity-labeled GLP-1 RA coverage group, and a mean (SD) of 50.2% (0.9%) were female in the control group compared with 50.6% (0.5%) in the obesity-labeled GLP-1 RA coverage group (Table 1).

Table 1. Baseline State-Level Characteristics, 2021.

Characteristic	Mean (SD)
Characteristic	No obesity-labeled GLP-1 RA coverage	Obesity-labeled GLP-1 RA coverage
No. of states	36	9
No. of Medicaid expansion states	27	8
Total population	212 786 953	89 021 657
Income per capita, $	37 720.8 (6952.9)	40 445.2 (6519.4)
Median income (IQR), $	36 108 (34 248.5-39 651.5)	41 753 (38 315-43 756)
Median age (IQR), y	38.9 (37.6-40.0)	39.9 (38.8-40.6)
Unemployment, %	5.7 (1.8)	6.2 (1.5)
Medicaid beneficiaries, %	25.4 (7.6)	25.4 (5.7)
Sex, %
Male	49.8 (0.9)	49.4 (0.5)
Female	50.2 (0.9)	50.6 (0.5)
Race and ethnicity, %^a
Hispanic	13.5 (10.4)	11.9 (10.6)
Non-Hispanic Asian	4.3 (6.1)	5.2 (4.2)
Non-Hispanic Black	9.8 (10.4)	11.3 (10.6)
Non-Hispanic White	64.8 (17.3)	66.2 (15.2)
Obesity, %	33.4 (4.3)	32.1 (3.7)
Diabetes, %	10.8 (2.1)	10.8 (2.0)

Open in a new tab

Abbreviation: GLP-1 RA, glucagon-like peptide-1 receptor agonist.

^{^a}

Race and ethnicity were self-reported by individuals and summarized to provide demographic context to the data.

Total and diabetes-labeled GLP-1 RA utilization increased substantially in both groups from 2021 to 2024 (Figure 1). Prescriptions per 1000 beneficiaries for diabetes-labeled GLP-1 RAs increased from 8.71 in 2021 to 25.28 in 2024 in treated states and from 8.21 in 2021 to 20.74 in 2024 in control states. Total GLP-1 RA utilization per 1000 beneficiaries increased from 8.82 in 2021 to 38.58 in 2024 in treated states and from 8.22 in 2021 to 20.81 in 2024 in control states. Obesity-labeled GLP-1 RA utilization increased significantly in treated states but not in control states, as expected.

Line chart of prescriptions per 1000 beneficiaries by quarter, treated vs control. Single-panel line graph with six time-series. Vertical axis label: Prescriptions per 1000 beneficiaries, ranging from 0 to 50 with gridlines at 10-unit intervals. Horizontal axis labeled by quarter with tick labels 2021 Q1, 2022 Q1, 2023 Q1, 2024 Q1, and 2025 Q1. Legend in the upper right lists six series: Total, treated as a solid gray-brown line with circular markers; Total, control as a dashed gray-brown line; Diabetes labeled, treated as a solid orange line with circular markers; Diabetes labeled, control as a dashed orange line; Obesity labeled, treated as a solid blue-gray line with circular markers; Obesity labeled, control as a dashed blue-gray line near zero. Total, treated rises from about 7 in 2021 Q1 to about 10 by 2022 Q1, about 20 at 2023 Q1, about 28 to 29 around 2024 Q1, then increases to about 33 and peaks near 47 in late 2024, ending near 46. Total, control increases from about 7 to about 10 by 2022 Q1, to about 14 to 15 by 2023 Q1, then to about 19 to 20 around 2024 Q1, reaching about 22 in late 2024 and ending near 21. Diabetes labeled, treated increases from about 7 to 10 by 2022 Q1, to about 17 at 2023 Q1, to about 22 to 23 around 2024 Q1, then reaches about 29 in late 2024 and ends near 26. Diabetes labeled, control increases from about 7 to 10 by 2022 Q1, to about 15 at 2023 Q1, to about 19 around 2024 Q1, then about 22 in late 2024 and ends near 21. Obesity labeled, treated stays near 0 through 2022, rises to about 3 at 2023 Q1, about 5 to 6 mid 2023, dips near 4, then increases to about 6 at 2024 Q1, about 10 mid 2024, about 18 to 19 late 2024, and ends near 20. Obesity labeled, control remains approximately 0 across all quarters. — Descriptive trends in GLP-1 RA prescriptions per 1000 beneficiaries are stratified by obesity-labeled GLP-1 RA coverage status (treated = covered obesity-labeled GLP-1 RAs at some point during the study period; control = did not cover obesity-labeled GLP-1 RAs during the study period) and GLP-1 RA category (diabetes-labeled, obesity-labeled, and total). GLP-1 RA indicates glucagon-like peptide-1 receptor agonist; Q, quarter.

Trends stratified by treated state show that Medicaid coverage was associated with a pronounced increase in obesity-labeled GLP-1 RA prescriptions (eFigure 2 in Supplement 1). Diabetes-labeled GLP-1 RA prescriptions increased steadily over most of the study period, although some states exhibited a deceleration in these trends after obesity-labeled GLP-1 RA coverage began. Increases in obesity-labeled GLP-1 RA utilization primarily came from Wegovy and Zepbound (eFigure 3 in Supplement 1).

For the primary outcome, Medicaid coverage of obesity-labeled GLP-1 RAs was associated with an additional 6.48 (95% CI, 3.98-8.98; P < .001) prescriptions of obesity-labeled GLP-1 RAs per 1000 in treated states compared with control states (Table 2). In comparison, estimates for total GLP-1 RAs were 4.17 (95% CI, −0.33 to 8.67; P = .07) additional prescriptions per 1000, and estimates for diabetes-labeled GLP-1 RAs were −2.16 (95% CI, −5.24 to 0.92; P = .17) fewer prescriptions per 1000, which were not statistically significant. Event study plots suggested increasing treatment effects over time, although quarter-specific coefficients were estimated imprecisely, with 95% CIs that crossed 0 for total and diabetes-labeled GLP-1 RAs (Figure 2). Weights assigned to control states (which sum to 1) were specific to each model and quarter of coverage adoption and, for the primary outcome, ranged from 0 to 0.166 (eTable 3 in Supplement 1).

Table 2. Association of Obesity-Labeled GLP-1 RA Coverage With GLP-1 RA Utilization in Medicaid.

Variable	Baseline average	ATT (95% CI)	P value
Diabetes labeled per 1000 beneficiaries
Prescriptions	14.66^a	−2.16 (−5.27 to 0.96)	.17
Prescription-months	83.09^a	−15.92 (−32.70 to 0.86)	.06
Medicaid reimbursement, $	13 768.36^a	−2276.50 (−5188.38 to 635.39)	.13
Obesity labeled per 1000 beneficiaries
Prescriptions	0.36	6.48 (4.00 to 8.96)	<.001
Prescription-months	6.22	27.07 (14.65 to 39.48)	<.001
Medicaid reimbursement, $	458.63	7162.11 (4683.95 to 9640.27)	<.001
Total per 1000 beneficiaries
Prescriptions	15.03	4.17 (−0.33 to 8.67)	.07
Prescription-months	89.31	6.67 (−14.58 to 27.92)	.54
Medicaid reimbursement, $	14 226.98	4620.59 (−16.38 to 9257.56)	.05

Open in a new tab

Abbreviations: ATT, average treatment effect in the treated; GLP-1 RA, glucagon-like peptide-1 receptor agonist.

^{^a}

Average of the treated states over the 2 preceding quarters before coverage begins.

Three-panel event-study plots of prescriptions per 1000 vs relative time in quarters. Three vertically stacked panels labeled A, B, and C at the upper left of each panel. All panels share a horizontal axis labeled Relative time to treatment in quarters, with tick marks from minus 7 through 8. Each panel includes a vertical dashed line at minus 1 and a solid horizontal reference line at 0. Each panel contains dark teal diamond markers at each quarter and a light gray shaded band surrounding the markers, corresponding to uncertainty intervals. Panel A title at top left: Diabetes-labeled G L P-1 R A s. Vertical axis labeled Prescriptions per 1000 beneficiaries, ranging from minus 10 to 10. Diamonds from minus 7 to minus 1 lie close to 0, roughly between minus zero point five and plus zero point five. From 0 through 8, diamonds fall below 0, approximately minus one at 0, near minus two at 1, around minus one to minus three from 2 to 8, with the lowest near minus three to minus four around 7. The gray band widens after 0, extending upward to roughly plus 6 by quarter 8 and downward to roughly minus 10 by quarter 8. Panel B title: Obesity-labeled G L P-1 R A s. Vertical axis labeled Prescriptions per 1000 beneficiaries, ranging from 0 to 15. Diamonds before minus 1 cluster near 0, slightly below 0 at earlier quarters. After 0, diamonds rise: about 1 at 0, about 3 at 1, about 4 to 6 from 2 to 5, about 9 to 10 at 6 and 7, and about 7 at 8. The gray band expands after 0, reaching near 15 at its upper edge around quarter 6 and remaining above 0 at the lower edge. Panel C title: Total G L P-1 R A s. Vertical axis labeled Prescriptions per 1000 beneficiaries, ranging from minus 5 to 20. Diamonds before minus 1 remain near 0. After 0, diamonds increase to about 1 at 1, about 3 at 2 and 3, about 2 at 4, about 3 to 4 at 5, about 8 at 6, about 5 to 6 at 7, and about 4 to 5 at 8. The gray band widens markedly after 2, with an upper edge near 18 around quarter 6 and a lower edge dropping to around minus 4 by quarter 8. — Event study estimates were derived from synthetic difference-in-differences models. The event study plots reflect the synthetically weighted control group, demonstrating parallel preintervention trends. The shaded areas in each panel indicate the 95% CIs. GLP-1 RA indicates glucagon-like peptide-1 receptor agonist

For the secondary outcome of prescription-months, coverage was associated with an increase in obesity-labeled GLP-1 RAs of 27.07 (95% CI, 14.65-39.48; P < .001) prescription-months per 1000 beneficiaries and no significant change in total GLP-1 RAs (6.67 [95% CI, −14.58 to 27.92; P = .54]) or diabetes-labeled GLP-1 RAs (−15.92 [95% CI, −32.70 to 0.86; P = .06]) (Table 2). For Medicaid reimbursement, we observed a similar pattern.

In the subgroup analysis by coverage policies, there was a statistically significant decrease in diabetes-labeled GLP-1 RAs of 3.02 prescriptions per 1000 beneficiaries (95% CI, −5.69 to −0.36; P = .03) in states with coverage more restrictive than the FDA label but not in states with no prior authorization or coverage that matched the FDA label (Table 3). Increases in obesity-labeled GLP-1 RAs were similar in magnitude and statistical significance in both subgroups. State-specific models were consistent with substantial heterogeneity, although coefficients were estimated imprecisely (eTable 4 in Supplement 1).

Table 3. Association of Obesity-Labeled GLP-1 RA Coverage With GLP-1 RA Utilization in Medicaid by Coverage Restrictions.

Variable	Less restrictive or same as FDA label^a		More restrictive than FDA label^b
Variable	ATT (95% CI)	P value	ATT (95% CI)	P value
Diabetes labeled per 1000 beneficiaries
Prescriptions	−1.27 (−5.08 to 2.54)	.51	−3.02 (−5.69 to −0.36)	.03
Prescription-months	−24.21 (−45.86 to −2.60)	.03	−15.32 (−24.24 to −6.41)	.001
Medicaid reimbursement, $	−1878.13 (−5718.07 to 1961.81)	.34	−2819.97 (−4818.23 to −821.72)	.01
Obesity labeled per 1000 beneficiaries
Prescriptions	7.52 (5.88 to 9.14)	<.001	6.48 (3.47 to 9.49)	<.001
Prescription-months	29.78 (25.77 to 33.77)	<.001	28.39 (10.97 to 45.81)	<.001
Medicaid reimbursement, $	9523.73 (7522.40 to 11525.06)	<.001	6343.40 (4221.98 to 8464.83)	<.001
Total per 1000 beneficiaries
Prescriptions	6.14 (1.21 to 11.08)	.01	3.46 (−0.83 to 7.76)	.11
Prescription-months	−1.02 (−23.56 to 21.52)	.93	12.30 (−9.33 to 33.9)	.27
Medicaid reimbursement, $	7340.01 (1950.13 to 12729.88)	.01	3525.45 (407.52 to 6643.38)	.03

Open in a new tab

Abbreviations: ATT, average treatment effect in the treated; FDA, Food and Drug Administration; GLP-1 RA, glucagon-like peptide-1 receptor agonist.

^{^a}

Includes California (with no prior authorization requirement), Pennsylvania, and Minnesota (with a prior authorization that matched the FDA label).

^{^b}

Includes Michigan, Massachusetts, Mississippi, New Hampshire, and Virginia with prior authorization that was more restrictive than the FDA label.

In sensitivity analyses, patterns of estimated increases in total GLP-1 RA utilization that were less than increases in obesity-labeled GLP-1 RA utilization were generally similar to those of the primary analysis (eTable 5 in Supplement 1). The estimated diabetes-labeled GLP-1 RA coefficient was significantly attenuated when using the Callaway and Sant’Anna estimator. Assumptions regarding suppressed prescription numbers had a negligible impact on precoverage averages of obesity-labeled GLP-1 RA prescriptions (eTable 6 in Supplement 1). The log of the Medicaid enrollee population was not associated with total GLP-1 RA prescriptions per 1000 beneficiaries within the SDID model (eTable 7 in Supplement 1).

Discussion

In this cross-sectional study using synthetic difference-in-differences analysis of 9 states covering obesity-labeled GLP-1 RAs through Medicaid and 36 control states from 2021 to 2024, coverage of obesity-labeled GLP-1 RAs was associated with slower growth in total GLP-1 RAs (4.2 more prescriptions per 1000 beneficiaries) than in obesity-labeled GLP-1 RAs (6.5 more prescriptions per 1000 beneficiaries) and 2.2 fewer prescriptions per 1000 beneficiaries of diabetes-labeled GLP-1 RAs relative to control states. The point estimates are suggestive of substitution effects, which could occur if patients using diabetes-labeled GLP-1 RAs switch to obesity-labeled GLP-1 RAs or if new users of obesity-labeled GLP-1 RAs prevent the onset of diabetes. However, estimates for total and diabetes-labeled GLP-1 RAs were not statistically significant. When considering only states with coverage policies more restrictive than the FDA label, the relative decline in diabetes-labeled GLP-1 RA estimates was economically and statistically significant.

We identified one working paper by Wing et al,²¹ also using Medicaid State Drug Utilization Data, which did not find similar substitution effects to our analysis. Their analysis used a stacked differences-in-differences design, which relies on a stronger parallel trends assumption than synthetic difference-in-differences, similar to the Callaway and Sant’Anna estimator that showed significantly attenuated estimates in our analysis. Another potential factor in differences in findings is the use in Wing et al²¹ of the number of enrollees as weights in their regressions. We did not weight our analyses by beneficiary population to study state-level policy without outsized influence from the most populous states.²² Our differential findings relative to Wing et al²¹ are likely explained by the fact that California and Pennsylvania, the first and fifth largest states in our sample, had opposite-signed effects from the remainder of the states in state-level analyses.²³ Of note, California was the only state that covered obesity-labeled GLP-1 RAs without need for prior authorization, and Pennsylvania’s prior authorization policy matched the FDA label.¹⁵

The finding that states with more restrictive coverage policies had stronger statistical evidence suggesting substitution effects than states with less restrictive coverage policies merits further study. One possible explanation is that restrictive prior authorization may preferentially direct obesity-labeled GLP-1 RAs to patients who otherwise would have received the diabetes-labeled drug (ie, patients with obesity and diabetes and patients with recalcitrant or severe obesity and/or comorbidities associated with elevated risk of incident diabetes), resulting in greater substitution between therapies. In contrast, less restrictive prior authorization may direct obesity-labeled drugs to more previously untreated patients with low diabetes risk, which dominates over potential substitution from diabetes-labeled drugs.

Our findings could have implications for the cost impacts of GLP-1 RAs that are used in Medicaid policy decisions.^24,25 Current projections rely on assumptions regarding uptake, adherence, and reductions in utilization through improvements in health,²⁶ which are being actively tested and debated using empirical data.²⁷ Our results complement these findings by suggesting that substitution effects between obesity-labeled and diabetes-labeled GLP-1 RAs are possible. When making projections on utilization and cost and evaluating policy, payers may need to consider impacts on total GLP-1 RAs, which can vary based on the state setting and coverage environment, rather than focusing on obesity-labeled GLP-1 RAs alone.

Our results in Medicaid are consistent with prior work documenting increasing trends in GLP-1 RA use for diabetes in other settings.^4,28,29 Importantly, diabetes-labeled GLP-1 RA utilization increased in all states regardless of whether they covered obesity-labeled GLP-1 RAs; the negative point estimates for states that covered obesity-labeled GLP-1 RAs reflect slower growth in diabetes-labeled GLP-1 RA utilization relative to control states, not an absolute reduction.

Limitations

This study has several limitations. First, there are data missing not at random due to censoring for small cell sizes. We imputed values at both extremes of possible values (0 and 10 prescriptions) and demonstrated that results were insensitive to these assumptions. Second, the reimbursement estimates in this analysis do not account for rebates through the Medicaid program, which affect the net price states are exposed to but are not publicly available. Third, documentation of Medicaid drug coverage policies and timing from publicly available sources was incomplete, so we were not able to include all potentially eligible states for analysis. Fourth, this study is ecological in nature and does not directly demonstrate whether patients are switching between label indications of GLP-1 RAs. Fifth, cash purchases of GLP-1 RAs outside of insurance (eg, compounded GLP-1 RAs) are not captured in our analysis; however, the liquidity constraints of most Medicaid beneficiaries lower this concern. Sixth, shortages of semaglutide and tirzepatide for both label indications were active during the study period, which could cause confounding substitution effects; because both diabetes-labeled and obesity-labeled formulations were affected, the direction of potential bias is uncertain. Seventh, given we analyzed a policy with staggered adoption, the earliest states to adopt the policy contribute the most follow-up observations and have greater influence on the average treatment effect in the treated estimates, limiting generalizability to late-adopting states.

Conclusions

Results of this cross-sectional ecological analysis suggest that substitution effects are possible between obesity-labeled and diabetes-labeled GLP-1 RAs in the Medicaid program, varying by state coverage policies, that could affect the impact of states’ coverage decisions regarding obesity-labeled GLP-1 RAs on total GLP-1 RA utilization and cost.

Supplement 1.

eTable 1. References to support coverage status determined for each state

eTable 2. Crosswalk between national drug code and months-equivalent prescription

eTable 3. Weights assigned to control states

eTable 4. Association of obesity-labeled GLP-1RA coverage with total GLP-1RA utilization in Medicaid, by state

eTable 5. Sensitivity analyses for the primary outcome

eTable 6. Average prescriptions per 1000 beneficiaries of obesity-labeled GLP-1RAs in treatment states prior to coverage start under varying assumptions of suppressed data

eTable 7. Association of Medicaid enrollment with total GLP-1RA prescriptions per 1000 beneficiaries in the SDID model

eFigure 1. Map of states by coverage status

eFigure 2. Trends stratified by drug category for each treatment state

eFigure 3. Trends stratified by control vs coverage states for individual obesity drugs

jamahealthforum-e263353-s001.pdf^{(645.2KB, pdf)}

Supplement 2.

Data Sharing Statement

jamahealthforum-e263353-s002.pdf^{(15.1KB, pdf)}

References

1.Jastreboff AM, Aronne LJ, Ahmad NN, et al. ; SURMOUNT-1 Investigators . Tirzepatide once weekly for the treatment of obesity. N Engl J Med. 2022;387(3):205-216. doi: 10.1056/NEJMoa2206038 [DOI] [PubMed] [Google Scholar]
2.Wilding JPH, Batterham RL, Calanna S, et al. ; STEP 1 Study Group . Once-weekly semaglutide in adults with overweight or obesity. N Engl J Med. 2021;384(11):989-1002. doi: 10.1056/NEJMoa2032183 [DOI] [PubMed] [Google Scholar]
3.Bozick R, Donofry SD, Rancaño KM. New weight loss drugs: glp-1 agonist use and side effects in the United States. 2025. Accessed February 1, 2026. https://www.rand.org/pubs/research_reports/RRA4153-1.html [DOI] [PMC free article] [PubMed]
4.Ukhanova M, Wozny JS, Truong CN, Ghosh L, Krause TM. Trends in glucagon-like peptide 1 receptor agonist prescribing patterns. Am J Manag Care. 2025;31(8):e228-e234. doi: 10.37765/ajmc.2025.89778 [DOI] [PubMed] [Google Scholar]
5.Emmerich SD, Fryar CD, Stierman B, Ogden CL. Obesity and severe obesity prevalence in adults: United States, August 2021-August 2023. NCHS Data Brief. 2024;(508). doi: 10.15620/cdc/159281 [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Tsipas S, Khan T, Loustalot F, Myftari K, Wozniak G. Spending on glucagon-like peptide-1 receptor agonists among US adults. JAMA Netw Open. 2025;8(4):e252964. doi: 10.1001/jamanetworkopen.2025.2964 [DOI] [PMC free article] [PubMed] [Google Scholar]
7.Baig K, Dusetzina SB, Kim DD, Leech AA. Medicare Part D coverage of antiobesity medications - challenges and uncertainty ahead. N Engl J Med. 2023;388(11):961-963. doi: 10.1056/NEJMp2300516 [DOI] [PubMed] [Google Scholar]
8.42 USC 1396r-8: Payment for covered outpatient drugs. Accessed February 1, 2026. https://uscode.house.gov/view.xhtml?hl=false&edition=prelim&req=granuleid%3AUSC-1999-title42-section1396r-8&num=0
9.Liu BY, Rome BN. State coverage and reimbursement of antiobesity medications in Medicaid. JAMA. 2024;331(14):1230-1232. doi: 10.1001/jama.2024.3073 [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Chatlani S. States retreat from covering drugs for weight loss. Stateline. November 28, 2025. Accessed February 1, 2026. https://stateline.org/2025/11/28/states-retreat-from-covering-drugs-for-weight-loss/
11.Salamah HM, Marey A, Abugdida M, et al. Efficacy and safety of glucagon-like peptide-1 receptor agonists on prediabetes: a systematic review and meta-analysis of randomized controlled trials. Diabetol Metab Syndr. 2024;16(1):129. doi: 10.1186/s13098-024-01371-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
12.State drug utilization data. Medicaid. Accessed March 21, 2026. https://www.medicaid.gov/medicaid/prescription-drugs/state-drug-utilization-data
13.Williams E. Medicaid Coverage of and spending on glp-1s. January 26, 2026. https://www.kff.org/medicaid/medicaid-coverage-of-and-spending-on-glp-1s/
14.Raphael J. A view of Medicaid today and a look ahead: balancing access, budgets and upcoming changes. KFF. November 13, 2025. Accessed March 21, 2026. https://www.kff.org/medicaid/50-state-medicaid-budget-survey-fy-2025-2026/
15.Klebanoff MJ, Chetty AK, Doshi JA. Medicaid coverage and prior authorization for antiobesity glucagon-like peptide-1 receptor agonists: a cross-sectional study of state policies. J Gen Intern Med. 2026;41(7):2003-2006. doi: 10.1007/s11606-025-10012-y [DOI] [PMC free article] [PubMed] [Google Scholar]
16.DailyMed . Accessed March 21, 2026. https://dailymed.nlm.nih.gov/dailymed/
17.Arkhangelsky D, Athey S, Hirshberg DA, Imbens GW, Wager S. Synthetic difference-in-differences. Am Econ Rev. 2021;111(12):4088-4118. doi: 10.1257/aer.20190159 [DOI] [Google Scholar]
18.Ciccia D. A short note on event-study synthetic difference-in-differences estimators. arXiv. Preprint posted online July 5, 2024. https://arxiv.org/html/2407.09565v2
19.Callaway B, Sant’Anna PHC. Difference-in-differences with multiple time periods. J Econom. 2021;225(2):200-230. doi: 10.1016/j.jeconom.2020.12.001 [DOI] [Google Scholar]
20.Tolbert J, Corallo B. An examination of Medicaid renewal outcomes and enrollment changes at the end of the unwinding. KFF. September 18, 2024. Accessed August 21, 2026. https://www.kff.org/medicaid/an-examination-of-medicaid-renewal-outcomes-and-enrollment-changes-at-the-end-of-the-unwinding/
21.Wing C, Lo WL, Potter M, et al. Medicaid coverage for obesity medications: utilization and net-of-rebate spending. National Bureau of Economic Research. March 2026. Accessed August 21, 2026. https://www.nber.org/system/files/working_papers/w34949/w34949.pdf
22.Hoehn-Velasco L, Penglase J, Pesko M, Shahid H. When policies overlap: the fragility of effects amidst multiple treatments. Social Science Research Network. Preprint posted online April 22, 2024:4802701. doi: 10.2139/ssrn.4802701 [DOI]
23.US Census Bureau . State Population Totals and Components of Change: 2020-2025. Census.gov. Accessed April 17, 2026. https://www.census.gov/data/tables/time-series/demo/popest/2020s-state-total.html
24.Pennsylvania restricts Medicaid coverage for GLP 1 drugs. Spotlight PA. December 9, 2025. Accessed February 4, 2026. https://www.spotlightpa.org/news/2025/12/ozempic-glp1-weight-loss-medicaid-pennsylvania-cuts-health/
25.Thompson D. California Ends Medicaid Coverage of Weight Loss Drugs Despite TrumpRx Plan. KFF Health News. January 9, 2026. Accessed February 4, 2026. https://kffhealthnews.org/news/article/california-medicaid-medi-cal-glp1-weight-loss-drugs-ends-coverage-cost/
26.Hwang JH, Laiteerapong N, Huang ES, Mozaffarian D, Fendrick AM, Kim DD. Fiscal Impact of Expanded Medicare Coverage for GLP-1 Receptor Agonists to Treat Obesity. JAMA Health Forum. 2025;6(4):e250905. doi: 10.1001/jamahealthforum.2025.0905 [DOI] [PMC free article] [PubMed] [Google Scholar]
27.Wing C, Cai ST, Sacks DW, Simon KI. Do GLP-1 Medications Pay for Themselves? National Bureau of Economic Research. January 2026. doi: 10.3386/w34678 [DOI]
28.Santoro C. Medicare Part D Spending Trends for Type 2 Diabetes Medications. Am J Manage Care. February 4, 2026. Accessed February 4, 2026. https://www.ajmc.com/view/medicare-part-d-spending-trends-for-type-2-diabetes-medications
29.Li P, Varghese JS, Shah MK, et al. Prescribing trends of glucagon-like peptide 1 receptor agonists for type 2 diabetes or obesity. JAMA Netw Open. 2025;8(10):e2540890. doi: 10.1001/jamanetworkopen.2025.40890 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplement 1.

eTable 1. References to support coverage status determined for each state

eTable 2. Crosswalk between national drug code and months-equivalent prescription

eTable 3. Weights assigned to control states

eTable 4. Association of obesity-labeled GLP-1RA coverage with total GLP-1RA utilization in Medicaid, by state

eTable 5. Sensitivity analyses for the primary outcome

eTable 6. Average prescriptions per 1000 beneficiaries of obesity-labeled GLP-1RAs in treatment states prior to coverage start under varying assumptions of suppressed data

eTable 7. Association of Medicaid enrollment with total GLP-1RA prescriptions per 1000 beneficiaries in the SDID model

eFigure 1. Map of states by coverage status

eFigure 2. Trends stratified by drug category for each treatment state

eFigure 3. Trends stratified by control vs coverage states for individual obesity drugs

jamahealthforum-e263353-s001.pdf^{(645.2KB, pdf)}

Supplement 2.

Data Sharing Statement

jamahealthforum-e263353-s002.pdf^{(15.1KB, pdf)}

[aoi260060r1] 1.Jastreboff AM, Aronne LJ, Ahmad NN, et al. ; SURMOUNT-1 Investigators . Tirzepatide once weekly for the treatment of obesity. N Engl J Med. 2022;387(3):205-216. doi: 10.1056/NEJMoa2206038 [DOI] [PubMed] [Google Scholar]

[aoi260060r2] 2.Wilding JPH, Batterham RL, Calanna S, et al. ; STEP 1 Study Group . Once-weekly semaglutide in adults with overweight or obesity. N Engl J Med. 2021;384(11):989-1002. doi: 10.1056/NEJMoa2032183 [DOI] [PubMed] [Google Scholar]

[aoi260060r3] 3.Bozick R, Donofry SD, Rancaño KM. New weight loss drugs: glp-1 agonist use and side effects in the United States. 2025. Accessed February 1, 2026. https://www.rand.org/pubs/research_reports/RRA4153-1.html [DOI] [PMC free article] [PubMed]

[aoi260060r4] 4.Ukhanova M, Wozny JS, Truong CN, Ghosh L, Krause TM. Trends in glucagon-like peptide 1 receptor agonist prescribing patterns. Am J Manag Care. 2025;31(8):e228-e234. doi: 10.37765/ajmc.2025.89778 [DOI] [PubMed] [Google Scholar]

[aoi260060r5] 5.Emmerich SD, Fryar CD, Stierman B, Ogden CL. Obesity and severe obesity prevalence in adults: United States, August 2021-August 2023. NCHS Data Brief. 2024;(508). doi: 10.15620/cdc/159281 [DOI] [PMC free article] [PubMed] [Google Scholar]

[aoi260060r6] 6.Tsipas S, Khan T, Loustalot F, Myftari K, Wozniak G. Spending on glucagon-like peptide-1 receptor agonists among US adults. JAMA Netw Open. 2025;8(4):e252964. doi: 10.1001/jamanetworkopen.2025.2964 [DOI] [PMC free article] [PubMed] [Google Scholar]

[aoi260060r7] 7.Baig K, Dusetzina SB, Kim DD, Leech AA. Medicare Part D coverage of antiobesity medications - challenges and uncertainty ahead. N Engl J Med. 2023;388(11):961-963. doi: 10.1056/NEJMp2300516 [DOI] [PubMed] [Google Scholar]

[aoi260060r8] 8.42 USC 1396r-8: Payment for covered outpatient drugs. Accessed February 1, 2026. https://uscode.house.gov/view.xhtml?hl=false&edition=prelim&req=granuleid%3AUSC-1999-title42-section1396r-8&num=0

[aoi260060r9] 9.Liu BY, Rome BN. State coverage and reimbursement of antiobesity medications in Medicaid. JAMA. 2024;331(14):1230-1232. doi: 10.1001/jama.2024.3073 [DOI] [PMC free article] [PubMed] [Google Scholar]

[aoi260060r10] 10.Chatlani S. States retreat from covering drugs for weight loss. Stateline. November 28, 2025. Accessed February 1, 2026. https://stateline.org/2025/11/28/states-retreat-from-covering-drugs-for-weight-loss/

[aoi260060r11] 11.Salamah HM, Marey A, Abugdida M, et al. Efficacy and safety of glucagon-like peptide-1 receptor agonists on prediabetes: a systematic review and meta-analysis of randomized controlled trials. Diabetol Metab Syndr. 2024;16(1):129. doi: 10.1186/s13098-024-01371-3 [DOI] [PMC free article] [PubMed] [Google Scholar]

[aoi260060r12] 12.State drug utilization data. Medicaid. Accessed March 21, 2026. https://www.medicaid.gov/medicaid/prescription-drugs/state-drug-utilization-data

[aoi260060r13] 13.Williams E. Medicaid Coverage of and spending on glp-1s. January 26, 2026. https://www.kff.org/medicaid/medicaid-coverage-of-and-spending-on-glp-1s/

[aoi260060r14] 14.Raphael J. A view of Medicaid today and a look ahead: balancing access, budgets and upcoming changes. KFF. November 13, 2025. Accessed March 21, 2026. https://www.kff.org/medicaid/50-state-medicaid-budget-survey-fy-2025-2026/

[aoi260060r15] 15.Klebanoff MJ, Chetty AK, Doshi JA. Medicaid coverage and prior authorization for antiobesity glucagon-like peptide-1 receptor agonists: a cross-sectional study of state policies. J Gen Intern Med. 2026;41(7):2003-2006. doi: 10.1007/s11606-025-10012-y [DOI] [PMC free article] [PubMed] [Google Scholar]

[aoi260060r16] 16.DailyMed . Accessed March 21, 2026. https://dailymed.nlm.nih.gov/dailymed/

[aoi260060r17] 17.Arkhangelsky D, Athey S, Hirshberg DA, Imbens GW, Wager S. Synthetic difference-in-differences. Am Econ Rev. 2021;111(12):4088-4118. doi: 10.1257/aer.20190159 [DOI] [Google Scholar]

[aoi260060r18] 18.Ciccia D. A short note on event-study synthetic difference-in-differences estimators. arXiv. Preprint posted online July 5, 2024. https://arxiv.org/html/2407.09565v2

[aoi260060r19] 19.Callaway B, Sant’Anna PHC. Difference-in-differences with multiple time periods. J Econom. 2021;225(2):200-230. doi: 10.1016/j.jeconom.2020.12.001 [DOI] [Google Scholar]

[aoi260060r20] 20.Tolbert J, Corallo B. An examination of Medicaid renewal outcomes and enrollment changes at the end of the unwinding. KFF. September 18, 2024. Accessed August 21, 2026. https://www.kff.org/medicaid/an-examination-of-medicaid-renewal-outcomes-and-enrollment-changes-at-the-end-of-the-unwinding/

[aoi260060r21] 21.Wing C, Lo WL, Potter M, et al. Medicaid coverage for obesity medications: utilization and net-of-rebate spending. National Bureau of Economic Research. March 2026. Accessed August 21, 2026. https://www.nber.org/system/files/working_papers/w34949/w34949.pdf

[aoi260060r22] 22.Hoehn-Velasco L, Penglase J, Pesko M, Shahid H. When policies overlap: the fragility of effects amidst multiple treatments. Social Science Research Network. Preprint posted online April 22, 2024:4802701. doi: 10.2139/ssrn.4802701 [DOI]

[aoi260060r23] 23.US Census Bureau . State Population Totals and Components of Change: 2020-2025. Census.gov. Accessed April 17, 2026. https://www.census.gov/data/tables/time-series/demo/popest/2020s-state-total.html

[aoi260060r24] 24.Pennsylvania restricts Medicaid coverage for GLP 1 drugs. Spotlight PA. December 9, 2025. Accessed February 4, 2026. https://www.spotlightpa.org/news/2025/12/ozempic-glp1-weight-loss-medicaid-pennsylvania-cuts-health/

[aoi260060r25] 25.Thompson D. California Ends Medicaid Coverage of Weight Loss Drugs Despite TrumpRx Plan. KFF Health News. January 9, 2026. Accessed February 4, 2026. https://kffhealthnews.org/news/article/california-medicaid-medi-cal-glp1-weight-loss-drugs-ends-coverage-cost/

[aoi260060r26] 26.Hwang JH, Laiteerapong N, Huang ES, Mozaffarian D, Fendrick AM, Kim DD. Fiscal Impact of Expanded Medicare Coverage for GLP-1 Receptor Agonists to Treat Obesity. JAMA Health Forum. 2025;6(4):e250905. doi: 10.1001/jamahealthforum.2025.0905 [DOI] [PMC free article] [PubMed] [Google Scholar]

[aoi260060r27] 27.Wing C, Cai ST, Sacks DW, Simon KI. Do GLP-1 Medications Pay for Themselves? National Bureau of Economic Research. January 2026. doi: 10.3386/w34678 [DOI]

[aoi260060r28] 28.Santoro C. Medicare Part D Spending Trends for Type 2 Diabetes Medications. Am J Manage Care. February 4, 2026. Accessed February 4, 2026. https://www.ajmc.com/view/medicare-part-d-spending-trends-for-type-2-diabetes-medications

[aoi260060r29] 29.Li P, Varghese JS, Shah MK, et al. Prescribing trends of glucagon-like peptide 1 receptor agonists for type 2 diabetes or obesity. JAMA Netw Open. 2025;8(10):e2540890. doi: 10.1001/jamanetworkopen.2025.40890 [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

GLP-1 Receptor Agonist Coverage for Obesity in Medicaid and Total GLP-1 Receptor Agonist Utilization

Ritikaa Khanna, MA

Michael C Wang, MD

Atheendar S Venkataramani, MD, PhD

Key Points

Question

Findings

Meaning

Abstract

Importance

Objective

Design, Setting, and Participants

Exposure

Main Outcomes and Measures

Results

Conclusions and Relevance

Introduction

Methods

Sample and Data

Exposure and Outcomes

Statistical Analysis

Results

Table 1. Baseline State-Level Characteristics, 2021.

Figure 1. Line Graph of Unadjusted Trends of GLP-1 RA Utilization for Treated and Control States.

Table 2. Association of Obesity-Labeled GLP-1 RA Coverage With GLP-1 RA Utilization in Medicaid.

Figure 2. Event Study Plots for Association of Obesity-Labeled GLP-1 RA Coverage With Utilization in Medicaid.

Table 3. Association of Obesity-Labeled GLP-1 RA Coverage With GLP-1 RA Utilization in Medicaid by Coverage Restrictions.

Discussion

Limitations

Conclusions

References

Associated Data

Supplementary Materials

ACTIONS

PERMALINK

RESOURCES

Cite

Add to Collections

PERMALINK

GLP-1 Receptor Agonist Coverage for Obesity in Medicaid and Total GLP-1 Receptor Agonist Utilization

Ritikaa Khanna, MA

Michael C Wang, MD

Atheendar S Venkataramani, MD, PhD

Key Points

Question

Findings

Meaning

Abstract

Importance

Objective

Design, Setting, and Participants

Exposure

Main Outcomes and Measures

Results

Conclusions and Relevance

Introduction

Methods

Sample and Data

Exposure and Outcomes

Statistical Analysis

Results

Table 1. Baseline State-Level Characteristics, 2021.

Figure 1. Line Graph of Unadjusted Trends of GLP-1 RA Utilization for Treated and Control States.

Table 2. Association of Obesity-Labeled GLP-1 RA Coverage With GLP-1 RA Utilization in Medicaid.

Figure 2. Event Study Plots for Association of Obesity-Labeled GLP-1 RA Coverage With Utilization in Medicaid.

Table 3. Association of Obesity-Labeled GLP-1 RA Coverage With GLP-1 RA Utilization in Medicaid by Coverage Restrictions.

Discussion

Limitations

Conclusions

References

Associated Data

Supplementary Materials

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases