The objective of this study was to compare differences in mortality and nonhome discharge in pediatric patients with firearm and stab injuries, while minimizing bias. Our secondary objective was to assess the influence of insurance on these same outcomes. Methods: Patients aged 0 to 17 years included in the National Trauma Data Bank (2007–2015) with firearm and stabbing injury were matched by propensity score. Logistic regression was used to assess associations of injury type and insurance with long-term care discharge and death. Results: The average age was 14.8 years, 19.2% were female, 48% were African American, 58.4% had an injury severity score ≤8, and assaults accounted for 73.1% of cases. Firearm injuries were associated with a higher risk of discharge to long-term care (adjusted odds ratio [aOR], 2.07) compared with propensity-matched patients who were stabbed.
Similarly, we found a higher risk of mortality in those with firearm injuries compared with stabbing injuries (aOR, 1.85). Regardless of mechanism, self-pay insurance status was associated with a higher risk of mortality (aOR, 2.41). When compared with stab wound patients with commercial insurance, self-pay firearm-injured patients were found to have an increased risk of mortality (aOR, 5.25). Conclusions: Pediatric victims of firearm violence were more likely to die or need additional care outside the home than victims of other types of penetrating injury when accounting for confounding characteristics to minimize bias. Key Words: firearm, stabbing, trauma, mortality (Pediatr Emer Care 2022;38: 147–152) Penetrating injuries account for 7.62% of pediatric trauma admissions to trauma centers in the United States.1 More than half of those admissions (4.4%) are firearm-related traumatic injury.1 Firearm-related injuries are the second leading cause of death among children in the United States.2,3 There is a 10-fold increase in mortality with victims of firearm injury compared with other types of penetrating injury.1 Importantly, many studies are working to identify trends in pediatric firearm violence and intent.4–8
The intent leading to penetrating injuries are most commonly assault, but unintentional injuries are not uncommon.1,6,9–12 Recent studies characterize the sex, race, insurance coverage, and cost of hospitalization for children involved in firearm violence.4,9,10,13–15 Limitations on funding firearm violence research have left gaps in our understanding of outcomes associated with firearm injuries, especially in the pediatric population.16 Factors associated with poor outcomes in pediatric firearm injuries are similar to those with other mechanisms of traumatic injury (low initial pH, high injury severity score [ISS], and traumatic brain injury [TBI]).12,17,18 In the era of near daily headlines of school shootings, there remains vocal public commentary against gun control policies, namely, that knives can cause the same damage as guns, yet policy makers are not proposing knife control policies. Limitations in the literature include a direct comparison of outcomes comparing firearm and stab injuries. The primary aim of this study was to compare differences in mortality and nonhome discharge in pediatric patients with firearm and stab injuries, while minimizing bias. Our secondary aim is to assess the influence of insurance on these same outcomes. To accomplish this, we utilized propensitymatching to control for confounders in a cohort of pediatric penetrating injuries.19
We used the National Trauma Data Bank (NTDB) to identify a cohort of pediatric patients with penetrating firearm or stab injuries.1 The NTDB is a national incident based database sourced from participating trauma centers and produced by the American College of Surgeons. Patients younger than 18 years in the NTDB from years January 1, 2007, to December 31, 2015, with E-codes defined as firearm or cut/pierce were included in the analysis (Fig. 1). Specifically, we defined a firearm injury according to Fowler et al as a weapon that requires a powder charge to fire a projectile.4 Patients discharged from the emergency department were not included in the analysis. If the data set was missing discharge status, ISS, sex, blood pressure on arrival, or hospital teaching status, it was excluded from the inclusion in propensity matching (Fig. 1).
Data collected included age, race/ethnicity, sex, mechanism of injury, hospital type, hypotension, pulse at admission, total ISS, total Glascow Coma Scale score (GCS), requirement for mechanical ventilation, head Abbreviated Injury Score (AIS), intent, insurance, and discharge disposition. We then used the clinical characteristics of the patients to match each firearm patient to a stabbing patient with similar characteristics. The cohort was then processed by propensity matching each firearm injury patient 1 to 1 with the most similar stabbing patient, using a greedy matching algorithm with a caliper of 0.25 SDs of logit of the propensity score, until every gunshot patient had been considered.20 Once the matching occurs, neither patient can be matched again. Patients without a match from either injury type were then excluded from analysis. Specifically, 15 demographic, clinical, and comorbidity covariates were used to build the propensity-matched cohort (Fig. 1). The discharge disposition included death, home, or nonhome discharge. We defined a nonhome discharge as a need for admission to long-term care facility, skilled nursing facility, intermediate care facility, or inpatient rehabilitation facility. Death outcome was defined as in-hospital death after time of admission but before the patient was discharged. Deaths before hospital admission were not included