Outcomes of Children With Firearm Injuries Admitted to the PICU in the United States*

Supplemental Digital Content is available in the text.

107 sites provided data from which 101 were trauma sites (level 1: 71, level 2: 27, level 3: 3, and not applicable: 6). The external cause of injury codes (E-codes) from the International Classification of Diseases (ICD), 9th Edition, and ICD, 10th Edition, listings were used to identify the cases, and only the primary admission data were obtained. Once identified, the cases were grouped according to the intent of injury: assault, suicide, unintentional, and undetermined cause. The rationale to classify them in such a manner was to analyze if the outcomes in the PICU vary based on the intent of injury.
We obtained patient demographic data on gender, race, and age. Mortality was calculated when the death occurred in the PICU. Mortality was the primary outcome measured, and Pediatric Overall Performance Category (POPC) and Pediatric Cerebral Performance Category (PCPC) scores were the secondary outcomes. These validated scores are used to describe the shortterm outcomes of children by quantifying the overall functional morbidity and cognitive impairment (8)(9)(10).
The study was approved by the Institutional Review Board of the University of Maryland, Baltimore. The study data included children 1 month to 18 years old admitted with firearm injuries between January 1, 2009, and December 31, 2017.
All data were entered into IBM SPSS Statistics for Windows, Version 24.0 (Armonk, NY: IBM Corp. Released 2016). Table 1 describes the demographic data for our cohort, and of the 1,447 cases, 1,272 (88%) survived, and 175 (12%) died. Higher mortality was observed in White children compared with Black children and Hispanic children (41.1%, 30.9%, 6.9%, respectively; p < 0.05). Mortality was highest in the suicide category as compared to the other types of injuries with firearms (p < 0.05), with head/neck injury as the most common site (30.2%). One third of the children (33.5%) with head/neck category died in the PICU. Table 2 shows the most common cause of injury was unintentional (67.7%). Assault by firearm was disproportionally higher in Black children (65%), whereas suicide attempt was predominantly in White children (63%) (p < 0.05). Table 2 shows the POPC and PCPC scores. Increases in mild, moderate, or severe overall disability show an impact on overall functional outcome. When using the POPC score, most discharged children had mild disability. Children injured due to undetermined cause had the largest decline in POPC scores. Only 10.3% of the total patients had good overall performance at the time of discharge (Supplemental Table, http://links.lww.com/PCC/B807).

DISCUSSION
Important findings of our study include that almost two thirds of the cases in the PICU were due to unintentional firearm injury, admission to the PICU with firearm injury carried high mortality, and head/neck injury with a firearm was a risk factor for death.
Overall, the mortality rate of children admitted to the PICU for all causes is low. In comparison, the mortality rate due to firearm injuries in our study was 12% (n = 175), more than five times the all cause ICU mortality (11). There are very few PICU disease processes that carry such a high mortality rate. Interstage single ventricle physiology (12%, 30 d mortality) (12,13), oncology patients with sepsis (17%) (14), or severe pediatric acute respiratory distress syndrome (33%) (15) are few notable exceptions.
Unintentional injury was common and carried high mortality, with 50% of the children being between 13 and 18 years old. As most of the patients with firearm injury are due unintentional injuries, prevention strategies can significantly decrease this burden in the PICU. Reducing the burden of unintentional injuries should provide a target for societal (policy and advocacy) interventions. It is known that males disproportionately bear the burden of firearm injury and that there are significant differences between older and younger children in terms of intent (intentional vs unintentional) and mortality, with the highest mortality in those greater than 13 years old who attempted suicide (1).
Children attempting suicide with a firearm were more likely to die in the PICU, compared with other categories of firearm injury. Although Black children had the highest rates of firearm injury, White children had the highest rate of suicide and as a subgroup had a significantly higher mortality. We also find that head/ neck injury, common in suicide attempts, had the highest mortality. These studies need wider dissemination to design effective preventive strategies to combat the high mortality in children attempting suicide (16  In addition to the high mortality, survivors had a wide range of morbidities and neurologic disabilities as demonstrated by the POPC and PCPC (Table 2). A significant decline was found upon discharge, and only 10.3% of patients had good overall performance. The POPC and PCPC scores compared with baseline at discharge reveal significant morbidity among victims of firearm injuries indicating that even after survival, a significant number of patients are likely to require rehabilitation services. The disposition of the survivors (1,272) was as follows: 982 (77%) to the floor, 81 (6.3%) home, 145 (11.3%) to a step down unit/telemetry/technology dependent unit, and 33 (2.5%) to a rehabilitation facility.
Our study's limitations include errors in coding, which is inherent to any study based on databases. VPS maintains the accuracy of data by quality check measures and appropriate training of the staff. The data from VPS cannot be generalized as it has not been standardized to national estimates. Pediatric deaths due to firearms outside of the hospital or in the emergency rooms are not included in this database.
Gani et al (17) studied the nationwide emergency department (ED) sample finding about 35% of all ED visits with firearm injury result in inpatient admission. The overall mortality for their cohort was 6%, and interestingly, the ED mortality was lower (3.6%) than the  inpatient mortality (6.6%). Similar to a recent study (6), the higher mortality in our cohort could be because of one of the following reasons. Academic centers, which are more likely to contribute data to the VPS database, are the tertiary centers for pediatric patients and tend to have more acuity and complexity.
Children with firearm injuries admitted to the PICU represent a very small cohort of the whole PICU population. Most of the firearm injuries are nonfatal and do not require PICU admission. Nevertheless, morbidity and mortality in the PICU give a measure of the impact on families and healthcare.

CONCLUSIONS
Our study using data from VPS reports that firearm injury is associated with increased morbidity including neurologic injury evidenced by POPC and PCPC scores and with five times the mortality as the comparable general PICU population. With advances in health informatics and increasing collaboration to form national registries and databases, it has become feasible to better understand the impact of health conditions, like firearm injuries on the society. Accurate epidemiologic data can aid in creating policies that can be used to advocate for pediatric victims of firearm violence.