A study of thousands of Colorado patients who were prescribed GLP-1 drugs – among the most talked-about medications of recent years – reveals that four out of 10 of these prescriptions for diabetes or obesity or both were never filled, even though the patients were insured.
The study also reports sharp differences among racial and ethnic groups in GLP-1 prescription fill rates. It finds that Black and Hispanic patients – groups with higher rates of diabetes and obesity than the general population – were less likely to fill their prescription orders than white patients.
And it shows that patients whose prescription was for an indication of obesity alone were significantly less likely to fill their prescription than if the prescription was for both diabetes and obesity, or for diabetes alone.
The study draws on data on nearly 10,000 Colorado prescriptions of glucagon-like peptide-1 receptor agonist (GLP-1RA) drugs over a 4⅟₂-year period. Foster Goss, DO, MMSc, FACEP, an adjunct professor and site principal investigator for this study at the University of Colorado Anschutz Department of Emergency Medicine, co-authored a research letter on the study, published recently in JAMA Health Forum.
“Seeing that 40% of the orders were never filled was surprising, and so were the disparities by race and ethnicity and the differences by indication,” Goss says. He points to varying levels of insurance coverage, and how drug costs impose barriers that impact some groups more than others, as likely factors in these fill-rate disparities.
Goss is an emergency physician and a clinical informaticist – a relatively new board-certified specialty that applies data, information systems, and technology to the practice of medicine, with the goal of improving the quality, safety, efficiency, and outcomes of health care.
“When I first entered the field, people often asked, ‘What is informatics?’” Goss says. “Now, with the rapid growth of AI, I’m approached by residents and fellows all the time, asking me about clinical informatics. It’s become foundational to how we practice medicine and deliver care.”
GLP-1 drugs were introduced to manage blood sugar in patients with type 2 diabetes, beginning with the approval of Byetta in 2005, followed by Ozempic in 2017 and Mounjaro in 2024. GLP-1s were also found to have significant weight loss benefits. In 2014, Saxenda was the first GLP-1 introduced specifically for chronic weight management, followed by Wegovy in 2021 and Zepbound in 2023.
Use of these drugs has skyrocketed in recent years, including a more than 700% increase in GLP-1 use for weight management between 2019 and 2023. In a November 2025 national survey, about 12% of U.S. adults said they were currently taking a GLP-1 drug either to lose weight or treat a chronic condition.
In their new study, Goss and his research partners analyzed data from UCHealth electronic health records and the Colorado All Payer Claims Database, which includes records on millions of public and private health insurance claims. Individual patients were not identified in the data.
The study focused on 9,848 GLP-1 prescriptions ordered between January 2018 and September 2022 for 6,094 patients who were enrolled either in Medicare or commercial insurance. The average age of patients in the sample was 60.9 years.
Overall, 60.1% of patients in the sample filled their GLP-1 prescription, the study showed. Those rates were lower for Black patients (55.3%) and Hispanic patients (58.4%) than for white patients (60.0%).
The study found that the medical indication leading to the prescription made a big difference in fill rates. Patients with both diabetes and obesity were far more likely to fill their order (64.6%) than were patients with diabetes only (47.5%) or obesity only (37.2%).
When prescriptions were filled, the average out-of-pocket cost for a 30-day supply was $71.90. But those costs varied widely depending on indication, from $70.32 for patients with diabetes and obesity to $134.04 for those with obesity alone. The study suggests the higher costs for obesity likely reflected “less comprehensive insurance coverage for this indication.”
“We definitely identified a cost barrier associated with prescriptions for obesity alone,” Goss says. “These findings likely reflect underlying socioeconomic gradients— shaped by education, the built environment, social influences including social relationships, and chronic exposure to social and environmental stressors — that limit patients’ ability to fill these medications.”
Exactly what is driving these disparities in prescription fill rates remains an open question for future research, Goss says. “It would be important to reassess this pattern now with more widespread adoption” of GLP-1s for obesity alone, since the study period ended in 2022. The study did not include people who paid cash to fill their prescriptions.
The study concludes by saying that “policymakers should explore options to improve equitable access” to GLP-1 drugs. Goss says that further research will be essential to help answer which strategies are most effective.
“There are social determinants of health at play here that are driving health disparities among race and ethnicity,” Goss says. “As we consider policy solutions, we need a deeper understanding of these drivers to design more effective solutions.”
“In the emergency department, he adds, “we care for patients with health-related social needs every day, whether it's access challenges, socio-economic barriers, or factors related to race or ethnicity. We’re often a safety net for these patients.”
The GLP-1 study is part of an broader portfolio of research on prescription drug costs, co-led by Goss and Anna Dour Sinaiko, PhD, at the Harvard T.H. Chan School of Public Health. The work is supported through funding from a major R01 grant from the National Institutes of Health’s National Institute on Minority Health and Health Disparities.
Another recent study in the series examined how clinicians use real-time benefit tools that alert them to out-of-pocket prescription price estimates for their patients and lower-price alternatives. These tools have been integrated into the electronic health record system at UCHealth facilities.