Gunshot Injuries in Children Served by Emergency Services
OBJECTIVE: To describe the incidence, injury severity, resource use, mortality, and costs for children with gunshot injuries, compared with other injury mechanisms.
METHODS: This was a population-based, retrospective cohort study (January 1, 2006–December 31, 2008) including all injured children age ≤19 years with a 9-1-1 response from 47 emergency medical services agencies transporting to 93 hospitals in 5 regions of the western United States. Outcomes included population-adjusted incidence, injury severity score ≥16, major surgery, blood transfusion, mortality, and average per-patient acute care costs.
RESULTS: A total of 49 983 injured children had a 9-1-1 emergency medical services response, including 505 (1.0%) with gunshot injuries (83.2% age 15–19 years, 84.5% male). The population-adjusted annual incidence of gunshot injuries was 7.5 cases/100 000 children, which varied 16-fold between regions. Compared with children who had other mechanisms of injury, those injured by gunshot had the highest proportion of serious injuries (23%, 95% confidence interval [CI] 17.6–28.4), major surgery (32%, 95% CI 26.1–38.5), in-hospital mortality (8.0%, 95% CI 4.7–11.4), and costs ($28 510 per patient, 95% CI 22 193–34 827).
CONCLUSIONS: Despite being less common than other injury mechanisms, gunshot injuries cause a disproportionate burden of adverse outcomes in children, particularly among older adolescent males. Public health, injury prevention, and health policy solutions are needed to reduce gunshot injuries in children.
- AIS —
- abbreviated injury scale
- CI —
- confidence interval
- ED —
- emergency department
- EMS —
- emergency medical services
- GCS —
- Glasgow Coma Scale
- ICD-9-CM —
- International Classification of Diseases, Ninth Revision, Clinical Modification
- ISS —
- injury severity score
- WESTRN —
- Western Emergency Services Translational Research Network
What’s Known on This Subject:
Gunshot injuries are an important cause of preventable injury and mortality in children, with emergency services often providing the initial care for patients. However, there is little recent population-based research to guide public health, injury prevention, and health policy efforts.
What This Study Adds:
Gunshot injuries are uncommon in children, but cause greater injury severity, need for major surgery, mortality, and costs compared with other injury mechanisms. There is also large variation in the population-adjusted incidence of pediatric gunshot injuries between regions.
Gunshot-related injuries are a leading cause of death and non-fatal injury among children and adolescents in the United States.1 Gunshot injuries rank second only to motor vehicle crashes as a cause of death for children ages 15 to 19 years.1 From 2001 through 2010, 29 331 children age 0 to 19 years died of gunshot-related injuries.1 Another 155 000 children were injured seriously enough to undergo treatment in emergency departments (EDs).1 After peaking in 1994, the gunshot-related mortality rate for children ages 0 to 19 years fell 54.6% by 2003, but has remained almost unchanged since.1
Pediatric gunshot-related injury and death has been the subject of several population-based epidemiologic studies from the mid-1980s through 2002.2–8 Other studies during this time period grouped gunshot injuries with other forms of intentional injury.9,10 More recent population-based research has been limited and focused exclusively on mortality11,12 or was restricted to hospital-based samples.13–15 Data from the National Violent Death Reporting System provide insight into the problem,16 although only include fatalities. Thus, there remains a lack of recent population-based research on gunshot-related injuries in children, with almost no studies evaluating or integrating the important role of emergency medical services (EMS) in the care of these patients.
To address these important research gaps, we conducted a population-based study of injured children served by EMS agencies in 5 regions of the western United States, matched to ED and hospital records. By using a large sample starting with 9-1-1 EMS contact, we compared children with gunshot injuries to those with other injury mechanisms regarding incidence, injury severity, hospital interventions, mortality, and costs.
This was a multiregion, population-based, retrospective cohort study using the Western Emergency Services Translational Research Network (WESTRN). Eleven institutional review boards in 5 regions approved this protocol and waived the requirement for informed consent.
The study included injured children evaluated by 47 EMS agencies serving more than 2.2 million children17 in 5 regions and transporting to 93 hospitals (11 Level I, 5 Level II, and 77 non-tertiary community and private hospitals) over a 3-year period. The 5 regions included: Portland, OR/Vancouver, WA (4 counties); King County, WA; Sacramento, CA (2 counties); Santa Clara, CA (2 counties); and Denver County, CO. Each of these regions represents a predefined geographic “footprint” consisting of a central metropolitan area and surrounding areas (suburban and some rural), defined by EMS agency service areas.
We included all injured children up to age 19 years for whom the 9-1-1 EMS system was activated within the 5 predefined geographic regions from January 1, 2006 through December 31, 2008. Injured children were identified by provider primary impression in EMS patient care reports. Specifying the sample in this manner allowed us to identify a broad, population-based, out-of-hospital injury cohort of children served by EMS providers in multiple trauma systems. We excluded children transferred between hospitals without an initial presentation involving EMS and 9-1-1 calls without patient contact.
We recorded out-of-hospital variables from EMS electronic health records exported from each EMS agency and transmitted to a central data coordinating center. In regions where multiple EMS agencies (eg, fire departments and private ambulance agencies) care for the same patients, we matched EMS records at the patient level. We have previously detailed and rigorously evaluated the methods for constructing this cohort18 and have validated the all-electronic data collection processes used in this study.19
The primary exposure variable of interest was mechanism of injury (15 categories), as recorded by EMS providers. EMS agencies collected mechanism of injury based on standardized categories outlined in the National EMS Information System20 for “cause of injury.” The guidelines for collecting this variable in electronic patient care reports were based on standard EMS training and charting practices in these agencies. Because of small numbers of children in certain categories, we collapsed the mechanism variable into 6 categories: gunshot (including powder-charged and air guns), cut or piercing (eg, stabbing), occupant motor vehicle crashes, pedestrians struck by motor vehicles, falls, struck by blunt object (eg, assaults), and “other” mechanisms.
To evaluate the mechanism of injury term, we matched EMS chart narratives available from 2 sites (Portland, OR and Denver, CO) by using all available sources of EMS patient care reports (fire agencies, ambulance agencies, and base-hospital phone records transcribed from EMS calls). Chart narratives provide contextual descriptions of the scene, patient, and circumstances surrounding the injury event. We assessed inter-rater reliability by randomly sampling 20% of patients with non-missing values for mechanism and a matched chart narrative (n = 1915). An investigator blinded to the EMS-coded mechanism term abstracted these records for mechanism of injury by using the same categories. We then compared the 2 independent mechanism terms using κ scores, which ranged from 0.64 (“other” category) to 0.93 (motor vehicle occupant). In addition, we abstracted all available chart narratives for patients with gunshot injuries to estimate the proportion of injuries from powder-charged firearms versus air guns (including BB, pellet, and paint ball guns) in the sample.
We also collected out-of-hospital disposition, EMS transport mode, patient demographics (age, gender), field trauma activation status (a dichotomous measure of patients meeting predefined field trauma triage criteria), initial field physiologic measures (systolic blood pressure, Glasgow Coma Scale [GCS] score, respiratory rate), out-of-hospital procedures (intubation attempt, intravenous line placement), and transport destination (including hospital type). We categorized hospital destinations as major trauma centers (Level I and II trauma hospitals) based on their American College of Surgeons accreditation status and state-level designations versus non-tertiary hospitals. We also recorded the occurrence of inter-hospital transfer after initial EMS transport.
Hospital outcome measures included abbreviated injury scale (AIS) score, injury severity score (ISS), surgical interventions, blood transfusion, in-hospital mortality, and average per-patient acute care costs. To determine hospital outcomes, we matched EMS records to hospital records using probabilistic linkage (LinkSolv v8.2; Strategic Matching, Inc, Morrisonville, NY). Record linkage methodology has been used to link large EMS data files to hospital records in previous studies,21 validated for matching ambulance records to trauma registry data22 and rigorously evaluated in this database.18 Sources of electronic hospital records included local trauma registries, state hospital discharge databases, and state ED databases (3 sites did not have ED data available). Because AIS and ISS are not collected in state discharge or ED databases, we used International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes and a mapping function (ICDPIC, Stata v 11; StataCorp, College Station, TX) to convert ICD-9-CM diagnosis codes to AIS and ISS values.23 Previous studies have validated the process of mapping administrative diagnosis codes to generate anatomic injury scores24,25 and we have previously validated ICDPIC-generated injury scores against chart-abstracted scores in this database.26 We defined “serious injury” as an AIS ≥3 in specific body regions and ISS ≥16 for overall injury severity.27 We categorized major surgical interventions by body region: abdominal, thoracic, spine, brain, neck, and vascular (including interventional radiology procedures). The methods for calculating per-patient acute care costs are described in the Supplemental Information; we did not evaluate costs beyond the acute care period.
We used descriptive statistics to characterize the sample and compare children with gunshot injuries versus those injured by other mechanisms. We calculated annual incidence estimates by averaging the number of children injured by each mechanism over the 3-year period (numerator) and by using 2010 census data17 for each region to define the at-risk population of children (denominator).
To preserve the population-based sampling design and minimize bias in the analysis, we used multiple imputation28 to handle missing values for important out-of-hospital variables. These variables included mechanism of injury (13% missing), physiologic measures (16%–29% missing), and EMS procedures (3%–8% missing). Details on the use of multiple imputation in this study are included in the Supplemental Information. For analyses of hospital measures, we restricted the sample to children transported to acute care hospitals with matched hospital records available. As a sensitivity analysis, we recalculated estimates for hospital measures by using all transported children (matched and unmatched to hospital records) with multiple imputation to impute missing hospital outcomes. We managed the database and conducted analyses using SAS (v 9.2, SAS Institute, Cary, NC).
A total of 49 983 injured children age 0 to 19 years were evaluated by EMS providers in the 5 regions over the 3-year period, of whom 505 (1.0%) had gunshot injuries. Among children with gunshot injuries, 421 (83.2%) were between age 15 and 19 years, and 427 (84.5%) were male. For children injured by other mechanisms (n = 49 478), 24 283 (49.1%) were between age 15 and 19 years, and 29 003 (58.6%) were male. Among children with gunshot injuries who had an EMS chart narrative available for review (n = 184, 36.4% of all children injured by gunshot), there were 27/184 (14.7%) air gun injuries, 156/184 (84.8%) firearm injuries, and 1/184 (0.5%) could not be determined. Children with gunshot injuries had greater out-of-hospital physiologic compromise, required more EMS medical procedures, had a higher proportion of deaths declared in the field, and had more field trauma activations, as compared with children injured by other mechanisms (Table 1).
The population-adjusted annual incidence of children with gunshot injuries ranged from 1.9 to 30.9 per 100 000 by region (Fig 1A), a 16-fold difference. Older adolescents (age 15–19 years) had a higher incidence of gunshot-related injuries than younger children at all sites. Compared with other injury mechanisms, gunshot injuries had the lowest incidence (7.5 per 100 000 children) (Fig 1B). When stratified by age group, the incidence of gunshot injuries differed more by age group than any other injury mechanism (24.9 per 100 000 older adolescents versus 1.7 per 100 000 younger children).
Among children transported by EMS with a matched hospital record available (n = 11 593, 31.6% of all children transported to an acute care hospital), there were differences in injury severity and hospital interventions by mechanism of injury (Fig 2). Compared with other injury mechanisms, children with gunshot injuries had the highest proportion of ISS ≥16 (23.0%, 95% confidence interval [CI] 17.6–28.4). When assessed by body region, gunshot injuries caused the greatest proportion of AIS ≥3 abdominal/pelvic injuries (15.4%, 95% CI 10.0–19.9) and extremity injuries (13.9%, 95% CI 9.6–18.3). Serious head injuries were similar between children injured by gunshot (9.0%, 95% CI 5.3–12.7) and those struck by motor vehicles (9.1%, 95% CI 6.8–11.3). The proportion of AIS ≥3 chest injuries was similar between those with gunshot injuries (14.5%, 95% CI 10.1–19.0) and cut/pierce injuries (16.9%, 95% CI 13.0–20.8). There were also differences in hospital interventions by injury mechanism, with children having gunshot injuries incurring the greatest proportion of major surgical interventions (32.3%, 95% CI 26.1–38.5) and blood transfusions (9.1%, 95% CI 5.5–12.7). Children with gunshot injuries commonly required abdominal (20.5%, 95% CI 15.3–25.7), thoracic (8.5%, 95% CI 4.9–12.1), and vascular (7.1%, 95% CI 3.9–10.3) surgery.
Figure 3 compares in-hospital mortality and acute care costs by mechanism of injury. In-hospital mortality for children with gunshot injuries (8.0%, 95% CI 4.7–11.4) was almost 4 times greater than the next highest group (cut/pierce, 2.2%, 95% CI 0.7–3.7). Among all 150 injured children who died (n = 60 declared in the field, n = 90 after transport), 32 (21%, 95% CI 15.1–28.8) died after gunshot injuries. The total number of injury deaths was exceeded only by motor vehicle occupant crashes (n = 53); however, in-hospital mortality for motor vehicle crashes (0.7%, 95% CI 0.5–1.0) was notably lower than for gunshot injuries. Total acute care costs followed a similar pattern, with children injured by gunshot incurring the highest average per-patient costs ($28 510, 95% CI 22 193–34 827). The mechanism group with the next highest cost was struck by a motor vehicle ($15 566, 95% CI 12 764–18 368).
Sensitivity analyses (in which missing hospital variables were imputed), including all children transported by EMS to an acute care hospital (n = 36 644), demonstrated similar comparisons between children with gunshot injuries versus other injury mechanisms (Supplemental Figs 4 and 5). However, point estimates for hospital metrics were slightly lower across all mechanisms, reflecting the inclusion of injured children discharged from the ED after EMS transport for whom no hospital record was available for linkage.
In this study, we demonstrate that gunshot injuries in children are uncommon, but are associated with a disproportionate burden of adverse outcomes compared with other injury mechanisms. Gunshot injuries were more severe, requiring more frequent major therapeutic interventions and resulting in higher mortality and per-patient costs than any other injury mechanism. Although gunshot injuries accounted for only 1% of injured children, they were associated with more than 20% of deaths after injury. The incidence of gunshot injuries also varied considerably between regions and age groups, with older adolescent males accounting for the majority of gunshot-related injuries.
Findings from this study are important for several reasons. There have been few population-based studies of children with gunshot injuries over the past decade. Similarly, gun research integrating the role of 9-1-1 emergency services has been limited. Our study involved both of these important elements. We demonstrate that one-third of children with an EMS response dying from gunshots are declared dead in the field and a substantive number are treated outside of major trauma centers, highlighting the limitations of previous research. Whereas previous studies have demonstrated the lethal nature of gunshot injuries in children,2–4,6–8,10,13–15 few have evaluated both out-of-hospital and in-hospital measures of injury severity, resource use, outcomes, and costs, compared with children injured by other mechanisms. These comparisons illustrate the immense per-patient impact of gunshot injuries in children. Even when compared with children injured by a cut/pierce mechanism (commonly grouped with gunshot injuries as “penetrating injury”), children with gunshot injuries were more severely injured and had notably higher mortality.
The frequency of gunshot injuries in children also differed widely by age group. Older adolescent males accounted for the majority of gunshot injuries and had substantially higher population-adjusted incidence rates compared with younger children. These findings suggest that injury prevention efforts to reduce gunshot injuries in children should begin at an early age, but focus particularly on older adolescent males.
Gunshot-related injury is a major public health issue for children. The low incidence and high burden of pediatric gunshot injuries suggests that more effective means to reduce such injuries must be strategic to have an impact. This pattern is in contrast to other injury mechanisms (eg, motor vehicle crashes) with high incidence and much lower per-patient injury burden. Lessons in improving child motor vehicle safety (eg, booster seats,29 age-appropriate restraint systems and seat position,30–32 air-bags33–36) and the associated reductions in pediatric motor vehicle mortality37 provide important examples of the potential impact of injury prevention efforts. The gains in motor vehicle safety required rigorous research to understand the problems, partnerships with national organizations (eg, National Highway Traffic Safety Administration) and the automotive industry, and evidence-based legislation. Reducing gunshot injuries in children will likely require similarly broad-based, interdisciplinary efforts, in addition to rigorous research. Exploring the regional variation in pediatric gunshot injuries may also help direct injury prevention efforts, particularly in areas with disproportionately high rates of gunshot injuries among children.
We used a retrospective cohort study design, which can be limited by unmeasured confounding and bias. We attempted to minimize potential bias through population-based sampling from multiple EMS agencies, regions, and hospitals. Nevertheless, the robustness of our findings would be enhanced if replicated in larger prospective studies that include additional US cities. We attempted to maximize case ascertainment of injured children served by EMS and minimize bias in our incidence estimates through rigorous electronic data mechanisms18,19 and inclusion of EMS agencies with broad catchment areas. However, it is possible that some children with gunshot injuries did not have contact with EMS (eg, obvious death in the field without EMS involvement) or that certain children in our sample were non-residents of the geographic regions. Both of these factors could have affected our incidence estimates. Also, we did not have information about intent and therefore were unable to determine this important aspect of gunshot injuries to further delineate potential outcome differences.
Our estimates regarding the impact of gunshot-related injuries in children are conservative owing to several factors: (1) inclusion of children injured by powder-charged firearms and air guns in the “gunshot” category; (2) inclusion of children who died during hospitalization in the cost estimates, which could have suppressed costs due to the high mortality rate; (3) lack of long-term outcomes and years of productive life lost to fully quantify the costs borne by society; (4) use of ISS, which can underestimate the severity of injury from penetrating trauma38; and (5) lack of functional outcomes. Inclusion of these additional factors would further magnify the measurable impact of gunshot injuries in children. Also, our results do not include the intense emotional impact and life changes incurred by surviving parents, family members, and affected communities after the death of a child.
Data were missing for a portion of important variables (eg, mechanism of injury). Rather than excluding patients with missing out-of-hospital values (complete case analysis), we used multiple imputation to handle missing values, which has been shown to reduce bias39–44 and preserve the sampling design. Sensitivity analyses that included multiply imputed hospital values and different ways of coding patients’ missing mechanism of injury yielded qualitatively similar findings.
Finally, the categories used for mechanism of injury were completed by EMS providers, rather than hospital-based data abstractors. Although mechanism of injury was completed using standardized categories by on-scene providers who evaluated the patient and injury setting, it is possible that the assigned categories could have differed if completed by trained hospital data abstractors. Also, the electronic health record for most participating EMS agencies had a single field for mechanism of injury and therefore did not allow us to assess scenarios in which a patient may have had more than 1 injury mechanism.
We used a unique population-based cohort of injured children in 5 western US regions to demonstrate the disproportionate burden and adverse outcomes associated with gunshot injuries compared with other pediatric injury mechanisms. Although the relative incidence of gunshot injuries is low, the impact is large, particularly among older adolescent males.
We acknowledge and thank all the participating EMS agencies, EMS medical directors, trauma registrars, and state offices that supported and helped provide data for this project.
The WESTRN investigators are: Craig D. Newgard, MD, MPH, Department of Emergency Medicine, Oregon Health & Science University; Eileen Bulger, MD, Thomas Rea, MD, MPH, Susan Stern, MD, Departments of Surgery and Medicine, Division of Emergency Medicine, Harborview Medical Center/University of Washington; Nathan Kuppermann, MD, MPH, James F. Holmes, MD, MPH, Deborah Diercks, MD, Department of Emergency Medicine, University of California at Davis; Renee Hsia, MD, MSc, Department of Emergency Medicine, San Francisco General Hospital/University of California at San Francisco; Kristan Staudenmayer, MD, N. Ewen Wang, MD, James Quinn, MD, Department of Surgery and Division of Emergency Medicine, Stanford University; Jason Haukoos, MD, MSc, Department of Emergency Medicine, Denver Health Medical Center/University of Colorado; N. Clay Mann, PhD, Erik Barton, MD, MS, MBA, Departments of Pediatrics and Surgery, Division of Emergency Medicine, University of Utah; Mark Langdorf, MD, Department of Emergency Medicine, University of California at Irvine; Sean Henderson, MD, Department of Emergency Medicine, Los Angeles County/University of Southern California; Cameron Crandall, MD, David Sklar, MD, Department of Emergency Medicine, University of New Mexico.
- Accepted August 7, 2013.
- Address correspondence to Craig D. Newgard, MD, MPH, Department of Emergency Medicine, Center for Policy and Research in Emergency Medicine, Oregon Health & Science University, 3181 SW Sam Jackson Park Rd, mailcode CR-114, Portland, OR 97239-3098. E-mail:
Dr Newgard conceptualized and designed the study, orchestrated funding for the project, conducted all database management and statistical analysis, critically evaluated all study results, drafted the initial manuscript, compiled and integrated all manuscript revisions, and takes full responsibility for the study; Dr Kuppermann assisted with study conceptualization and design, secured funding for the project, interpreted study results, provided leadership and guidance for the study group, and reviewed and revised the manuscript; Dr Holmes assisted with data acquisition, funding for the project, interpretation of study results, and critical revision of the manuscript; Dr Haukoos assisted with interpretation of study results and critical revision of the manuscript; Mr Wetzel assisted with data abstraction, interpretation of study results, and critical revision of the manuscript; Dr Hsia assisted with data acquisition, funding for the project, interpretation of study results, and critical revision of the manuscript; Dr Wang assisted with interpretation of study results and critical revision of the manuscript; Dr Bulger assisted with data acquisition, interpretation of study results, and critical revision of the manuscript; Drs Staudenmayer and Mann assisted with data acquisition, funding for the project, interpretation of study results, and critical revision of the manuscript; Dr Barton assisted with interpretation of study results and critical revision of the manuscript; Dr Wintemute assisted with study conceptualization and design, funding for the project, and content expertise, interpreted study results, provided leadership and guidance for the study group, and reviewed and revised the manuscript; and all authors approved the final manuscript as submitted.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: Supported by the Robert Wood Johnson Foundation Physician Faculty Scholars Program; California Wellness Foundation (grant 2010-067); the Oregon Clinical and Translational Research Institute (grant UL1 RR024140); University of California Davis Clinical and Translational Science Center (grant UL1 RR024146); Stanford Center for Clinical and Translational Education and Research (grant 1UL1 RR025744); University of Utah Center for Clinical and Translational Science (grant UL1-RR025764 and C06-RR11234); and University of California San Francisco Clinical and Translational Science Institute (grant UL1 RR024131). All Clinical and Translational Science Awards are from the National Center for Research Resources, a component of the National Institutes of Health (NIH), and NIH Roadmap for Medical Research. The funding period for these grants was July 1, 2007 through June 30, 2010. Funded by the National Institutes of Health (NIH).
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
- ↵Centers for Disease Control and Prevention. Web-Based Injury Statistics Query and Reporting System (WISQARS). Available at: www.cdc.gov/injury/wisqars/index.html. Accessed February 22, 2013
- Eber GB,
- Annest JL,
- Mercy JA,
- Ryan GW
- DiScala C,
- Sege R
- Zavoski RW,
- Lapidus GD,
- Lerer TJ,
- Banco LI
- Dowd MD,
- Knapp JF,
- Fitzmaurice LS
- Nance ML,
- Carr BG,
- Kallan MJ,
- Branas CC,
- Wiebe DJ
- Schecter SC, Betts J, Schecter WP, Victorino GP. Pediatric penetrating trauma: the epidemic continues. J Trauma Acute Care Surg. 2012;73:721–725
- ↵Karch DL, Logan J, McDaniel D, Parks S, Patel N. Surveillance for violent deaths—National Violent Death Reporting System, 16 states, 2009. MMWR. 2012;61(SS 6):1–43
- ↵American fact finder—community facts. United States Census Bureau. Available at: http://factfinder2.census.gov/faces/nav/jsf/pages/community_facts.xhtml. Accessed February 1, 2013
- Newgard CD,
- Malveau S,
- Staudenmayer K,
- et al.,
- WESTRN investigators
- ↵Uniform Pre-Hospital Emergency Medical Services (EMS) Dataset—Final Documentation and Data Dictionary. Version 2.2.1 (2006, last modified 4/9/2012). National Emergency Medical Services Information System. US Department of Transportation, National Highway Traffic Safety Administration
- Clark DE,
- Osler TM,
- Hahn DR
- ↵MacKenzie EJ, Steinwachs DM, Shankar BS, Turney SZ. An ICD-9CM to AIS conversion table: development and application. Proc AAAM. 1986;30:135–151
- Fleischman RJ,
- Mann NC,
- Wang NE,
- et al
- ↵AAAM. The Abbreviated Injury Scale (AIS), 2005 Revision, Update 2008. Des Plaines, IL: Association for the Advancement of Automotive Medicine
- Rubin DB
- Braver ER,
- Whitfield R,
- Ferguson SA
- Berg MD,
- Cook L,
- Corneli HM,
- Vernon DD,
- Dean JM
- ↵Kahane CJ. Fatality Reduction by Air Bags—Analyses of Accident Data Through Early 1996. DOT HS 808 470. Washington, DC: NHTSA; 1996
- National Highway Traffic and Safety Administration Traffic Safety Facts
- van der Heijden GJMG,
- Donders ART,
- Stijnen T,
- Moons KGM
- Greenland S,
- Finkle WD
- Bazzoli GJ,
- Kang R,
- Hasnain-Wynia R,
- Lindrooth RC
- Corso P,
- Finkelstein E,
- Miller T,
- Fiebelkorn I,
- Zaloshnja E
- Consumer Price Index. Bureau of Labor Statistics, United States Department of Labor. Available at: www.bls.gov/cpi/. Accessed February 22, 2013
- Raghunathan T,
- Copyright © 2013 by the American Academy of Pediatrics