Cosmos Study

Self-Pay Encounters Rose and Medicaid Fell Across All Care Settings from 2022 to 2026

August 31, 2026

Team A:Kersten Bartelt, RNEric Barkley

Team B:Louis Kazaglis, MDGrant KeaneJoe Deckert, PhD

Key Findings

The self-pay share of encounters rose in all four care settings studied between early 2022 and early 2026, but unevenly: the increase was steepest in the emergency department (5.5% to 7.6%), more modest for inpatient admissions (1.9% to 2.6%) and births (0.8% to 1.3%), and moved just a little in primary care (1.8% to 1.9%).
Medicaid declined in all four settings over the same window: emergency (18.2% to 16.1%), inpatient (13.2% to 11.7%), births (23.1% to 21.5%), and primary care (11.4% to 10.3%).

Medicaid is the joint federal-state program that pays for health care for many low-income children and adults. During the COVID-19 pandemic, a federal continuous-enrollment provision kept nearly all Medicaid enrollees covered without the usual periodic eligibility checks. That provision ended on March 31, 2023, after which states began “unwinding” it by redetermining eligibility and ending coverage for people who no longer qualified or did not complete renewal.¹ Because the federal rules left the exact timing to each state, the unwinding was staggered across the country, with states starting termination at different points from spring 2023 onward.¹ We previously found that self-pay encounters (visits where a patient rather than an insurer is the primary expected payer, often used as a proxy for being uninsured) began rising across emergency, hospital, and primary care settings as those terminations resumed,² and that self-pay emergency visits climbed in the months that followed the start of termination.³ Because the unwinding has now largely run its course, it is possible to look across a longer window and multiple care settings to see how the coverage mix shifted and whether those early changes persisted.

We studied more than 550 million U.S. health care encounters between January 2022 and June 2026. For every quarter, we calculated the share of encounters in each setting of interest (emergency department, inpatient admission, birth, and primary care) whose primary coverage was self-pay, Medicaid, traditional Medicare, Medicare Advantage, or commercial/other, counting each patient’s first encounter of a given setting per quarter and excluding encounters with unknown coverage.

The self-pay share of encounters rose in all four care settings between 2022 and 2026, but the increase was not uniform. It was the largest and most visible in the emergency department, where self-pay grew from 5.5% of encounters in the first quarter of 2022 to 7.6% in the second quarter of 2026. Inpatient admissions and births saw smaller increases (1.9% to 2.6% and 0.8% to 1.3%, respectively), while the primary care share changed only slightly (1.8% to 1.9%). The direction is consistent with earlier Epic Research studies showing self-pay encounters rising as Medicaid terminations resumed.^2,3

Figure 1

Percentage of Encounters Covered by Self-Pay by Care Setting Over Time

Figure 1. Quarterly percentage of encounters with self-pay as the primary coverage by care setting.

Over the same period, the Medicaid share declined in all four settings, mirroring the rise in self-pay. Births carried the highest Medicaid share throughout, going from 23.1% in early 2022 to 21.5% in early 2026. Emergency and inpatient encounters showed the steepest declines (18.2% to 16.1% and 13.2% to 11.7%), and primary care again moved least (11.4% to 10.3%).

Figure 2

Percentage of Encounters Covered by Medicaid by Care Setting Over Time

Figure 2. Quarterly percentage of encounters with Medicaid as the primary coverage by care setting.

Examining the full coverage mix in the emergency department, where the movement was largest, shows which shares grew and which shrank, though it does not trace where any individual patient’s coverage went. Between 2022 and 2026, self-pay and Medicare Advantage were the only two categories to grow (self-pay 5.5% to 7.6%; Medicare Advantage 10.3% to 13.0%), while traditional Medicare (12.8% to 11.9%), Medicaid (18.2% to 16.1%), and commercial/other coverage (53.1% to 51.3%) all declined.

Figure 3

Emergency Department Coverage Mix Over Time

Figure 3. Quarterly distribution of primary coverage type among emergency department encounters.

These data come from Cosmos, a dataset created in collaboration with a community of Epic health systems representing more than 310 million patient records from 2,200 hospitals and more than 50,000 clinics from all 50 U.S. states, Canada, Lebanon, and Saudi Arabia. This study was completed by two teams that worked independently, each composed of a clinician and research scientist. The two teams came to similar conclusions. Graphics by Brian Olson.

References

Unwinding and returning to regular operations after COVID-19. Medicaid.gov. Accessed July 8, 2026. https://www.medicaid.gov/resources-for-states/coronavirus-disease-2019-covid-19/unwinding-and-returning-regular-operations-after-covid-19/index.html
Alban C, Joyce B, Bartelt K, Sandberg N. Self-Pay Encounters Increase Following Start of Medicaid Coverage Terminations. Epic Research. https://epicresearch.org/articles/self-pay-encounters-increase-following-start-of-medicaid-coverage-terminations. Accessed on July 8, 2026.
Alban C, Joyce B, Bartelt K, Sandberg N. Uninsured Emergency Visits on the Rise Since the End of Medicaid Continuous Enrollment. Epic Research. https://epicresearch.org/articles/uninsured-emergency-visits-on-the-rise-since-the-end-of-medicaid-continuous-enrollment. Accessed on July 8, 2026.

Data Definitions

Study period: Q1 2022 to Q2 2026

Study population: inclusion: Encounters for patients with a current U.S. residence, at a U.S.-based department, that classify as one of four encounter types (birth encounter, primary care, inpatient, emergency).
Restricted to U.S. departments contributing data across the full study period.
Only the first encounter of a given type per quarter is counted per patient.

Study population: exclusion: Encounters with unknown primary coverage

Encounter classification priority: Each encounter is assigned to exactly one type with the exception that an admission from an emergency encounter was considered both emergency and inpatient.
When an encounter qualified for more than one, it was assigned by: birth encounter → primary care → inpatient → emergency.

Birth encounter: An encounter in BirthFact or with a diagnosis with ICD-10-CM code O60*–O82*

Primary care

An encounter with:

A type of “Office Visit” or “Appointment”
No telehealth procedure code (CPT 99441, 99442, 99443)
A department specialty of “Adolescent Medicine,” “Adult Health,” “Adult Medicine,” “Ambulatory Care,” “Chronic Care,” “Clinical Child & Adolescent,” “Developmental,” “Developmental and Behavioral Pediatrics,” “Elder Care Services,” “Family Medicine,” “Federally Qualified Health Center (FQHC),” “General Care,” “General Internal Medicine,” “General Practice,” “Geriatric Medicine,” “Gerontology,” “Internal Medicine,” “Internist,” “Medical,” “Medical Clinic,” “Physician’s Office,” “Preventative Medicine,” “Primary Care,” “Pediatric Internal Medicine,” or “Pediatrics”

Inpatient: An encounter of type “Hospital Outpatient to Inpatient,” “Emergency to Inpatient,” or “Inpatient Admission”

Emergency: An encounter of type “Emergency” or “Emergency to Inpatient”

Exposure: Quarter of encounter start

Outcomes: Quarterly distribution of primary coverage type among like encounters, reported as percentages.

Primary coverage type: Each encounter’s primary coverage was categorized as either self-pay, Medicaid, traditional Medicare, Medicare Advantage, or commercial/other. Self-pay is a proxy for uninsurance and may include cash-pay or not-yet-adjudicated encounters.

Model specifications: Simple quarterly proportions of each coverage type within each encounter type

Limitations: Coverage-mix trends may partly reflect which health systems contributed data in which quarters rather than true population change, so the analysis is restricted to full-period-contributing departments to mitigate this.
Self-pay primary coverage is a proxy for lack of insurance, not a direct measure.