Cosmos Study - Collaboration

ED Visits for Motor Vehicle Crashes Associated with Elevated Risk of Subsequent Firearm Injury for Young Male Patients

January 15, 2026

Team A:Kersten Bartelt, RNMichael Levas, MDMaria Beyer, MPHJoe Deckert, PhD

Team B:Dave Little, MDEric Barkley

Key Findings

Male patients aged 10 to 15 who had an ED visit for a motor vehicle crash (MVC) were as much as four times as likely to experience a subsequent firearm injury compared with those whose ED visit was for a fall, in some parts of the United States.
Male patients aged 16 to 25 who had an MVC-related ED visit were up to 81% more likely to sustain a subsequent firearm injury compared with those whose ED visit was for a fall, in some parts of the United States.
Female patients in all age groups and adults over age 25 did not experience a consistent change in firearm-injury risk after an ED MVC visit.

Motor vehicle crashes (MVCs) and firearm injuries are two major contributors to injury-related morbidity in the United States.¹ Prior clinical and public health research has shown that patterns of traumatic injury, risk-taking behavior, and exposure to violence frequently cluster in adolescence and early adulthood.^1,2 While both MVCs and firearm injuries disproportionately affect adolescents and young adults, limited evidence exists on whether an ED presentation for one type of traumatic exposure may identify patients at elevated risk for another. Identifying ED-based markers of elevated future risk could inform targeted prevention strategies, resource allocation, and community-based violence-intervention programs. We aimed to understand the relationship between ED visits for MVCs and the likelihood of future firearm injury.

We studied more than 2 million U.S. patients with an ED visit for an MVC and another 2 million U.S. patients with an ED visit for a fall between 2017 and November 11, 2025. Patients were excluded if they had a documented history of firearm injury, MVC, or fall before the ED visit of interest. Patients with an ED visit for an MVC were matched with patients whose ED visit was for a fall based on age, sex, race and ethnicity, census region, and month of the ED encounter. We additionally accounted for rurality based on residence, social vulnerability based on residence, ED acuity score, smoking history, prior health care utilization, and history of comorbidities including mental health conditions, traumatic brain injury, pervasive developmental disorders, and lead poisoning.

Across most census regions, male patients aged 10 to 15 experienced the clearest pattern of elevated firearm-injury risk following an ED visit for an MVC, with as much as four times as likely as those whose ED visit was for a fall, as seen in Figure 1. For male patients aged 16 to 25, the association persisted across many census regions but was smaller in magnitude; depending on region, patients were 81% more likely to have a firearm injury following an MVC compared to a fall.

Figure 1

Firearm Injury Risk Following an ED Visit Among Males by Age and Region

Figure 1. The risk of a firearm injury following an ED visit for an MVC compared to after a fall among males by age and geographic region.

Of note, significance was not observed for female patients or adults over age 25 in nearly all stratifications.

These data come from Cosmos, a dataset created in collaboration with a community of Epic health systems representing more than 300 million patient records from 1,800 hospitals and more than 41,000 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 scientists, in collaboration with researchers from Children’s Wisconsin. The two teams came to similar conclusions. Graphics by Brian Olson.

References

Cunningham RM, Walton MA, Carter PM. The major causes of death in children and adolescents in the United States. N Engl J Med. 2018;379(25):2468-2475. doi:10.1056/NEJMsr1804754
Sims DW, Bivins BA, Obeid FN, Horst HM, Sorensen VJ, Fath JJ. Urban trauma: a chronic recurrent disease. J Trauma. 1989;29(7):940-947.

Data Definitions

Study period: 1/1/2017 to 11/11/2025

Study population: inclusion

Patients with:

An emergency department (ED) visit
An evaluated sex of female or male
A state of residence in the U.S. documented
A face-to-face visit in the two years prior to their ED visit

Study population: exclusion: History of firearm injury, MVC, or fall before index ED visit

Censoring: Last face-to-face visit

Exposures: ED visit for MVC: Billing or encounter diagnosis with ICD-10-CM code V2[0-9].**XA, V3[0-9].*XXA, V4[0-9].*XXA, V5[0-9].*XXA, V87.*XXA, V88.*XXA, V89.[02]XXA, Y03.8XXA, or SNOMED code 407153006

Outcomes

Firearm injury from an ED, urgent care, or office visit encounter
Negative control outcomes:

Type 1 diabetes: Billing or encounter diagnosis with ICD-10-CM code E10*
Eczema: Billing or encounter diagnosis with ICD-10-CM code L20*
Crohn’s disease: Billing or encounter diagnosis with ICD-10-CM code K50*

Reference population: ED visit for a fall: Billing or encounter diagnosis with ICD-10-CM code W00*-W19*

Face-to-face visit: An encounter with a type of Emergency, Office Visit, Well Child, Follow-up, Telemedicine, Urgent Care, Walk-in, Routine Prenatal, Postpartum Visit, Fetal Care Consult, Emergency to Inpatient, Inpatient Admission, Hospital Outpatient Visit, or Hospital Outpatient Visit to Inpatient

Firearm Injury: Accidental: Billing or encounter diagnosis with ICD-10-CM code W32.[01]XX[AD], W33.[01][01239]X[AD], W34.00X[AD], W34.01[AD], W34.09X[AD], W34.1**[AD]
Intentional: Billing or encounter diagnosis with ICD-10-CM code X72.XXX[AD], X73.[01289]XX[AD], X74.0XA, X74.[89]XX[AD]
Assault: Billing or encounter diagnosis with ICD-10-CM code X93.XXA, X94.[01289]XX[AD], X95.0[129]X[AD], X95.[89]XX[AD]
Undetermined: Billing or encounter diagnosis with ICD-10-CM code Y22.XXX[AD], Y23.[012389]XX[AD], Y24.XX[AD]
Legal: Billing or encounter diagnosis with ICD-10-CM code Y35.0[012349][123][AD]
Terrorism: ICD-10-CM code Y38.4X[AD]
SNOMED code 219257002, 219338008, or 219334005

Confounders

Age:

<10
10-15
16-25
26-35
36-54
55+

Race and ethnicity
RUCA
Census region
Smoking history
ED acuity score
Legal sex
Social Vulnerability Index quintile
Comorbidities prior to index date (including problem list)

ADHD: ICD-10-CM code F90*
Depression/PTSD: ICD-10-CM code F32*, F33*, F06.31*, F06.32*, or F34.1*
Substance use disorder: (diagnosis or ever positive urine drug screen) ICD-10-CM code F10*-F19*
History of suicidal ideation/attempt: ICD-10-CM code R45.851, T14.91
Lead poisoning: ICD-10-CM code R78.71, T56.0
Pervasive developmental disorders: ICD-10-CM code F84*
Conduct disorder: ICD-10-CM code F91*
TBI: ICD-10-CM code S06* or SNOMED code 127295002

Race and ethnicity: Non-exclusive flags for Black/African American patients and Hispanic patients

Census region: West north central: Iowa, Kansas, Minnesota, Missouri, Nebraska, North Dakota, South Dakota
East north central: Illinois, Indiana, Michigan, Ohio, Wisconsin
New England: Connecticut, Maine, Massachusetts, New Hampshire, Rhode Island, Vermont
Middle Atlantic: New Jersey, New York, Pennsylvania
South Atlantic: Delaware, District of Columbia, Florida, Georgia, Maryland, North Carolina, South Carolina, Virginia, West Virginia
East south central: Alabama, Mississippi, Kentucky, Tennessee
West south central: Arkansas, Louisiana, Oklahoma, Texas
Mountain: Arizona, Colorado, Idaho, Montana, New Mexico, Nevada, Utah, Wyoming
Pacific: Alaska, California, Hawaii, Oregon, Washington

Matching

Exposures matched 1:1 to controls
Index event for exposure is first ED visit for MVC in the study period
Index event for controls is first ED visit for fall in the study period
Matched on:

Age
Year and month of ED visit
Census region
Race and ethnicity
Legal sex

Model specifications: Cox Proportional Hazard model

Sensitivity analysis: Negative-control outcomes including type 1 diabetes, eczema, and Crohn’s disease showed no meaningful associations, supporting specificity of the main findings.