O3214: Disparities in Pediatric Firearm Injuries During the COVID-19 Pandemic: A Multicenter Retrospective Cohort Study

Background: The COVID-19 pandemic saw a rise in US pediatric firearm injuries. Though disparities in firearm injuries have been described, the differential impact on low-income communities is less well understood. We sought to characterize the change in firearm injuries during the COVID pandemic by socio-economic deprivation index, age, and intent.

Methods: This was a secondary analysis of a multicenter retrospective study of pediatric injuries from two time periods: pre- (January 1, 2019-March 15, 2020) and during (March 16, 2020 – December 31, 2020) the COVID-19 pandemic. Medical record data were obtained from 40 children’s hospitals (39 U.S., 1 Canada). Children < 18 years old with an ICD-10 code for firearm injury at ED presentation were included. Deprivation index (DI) was assigned by home zip code from the US Census based on six American Community Survey measures. Scores range 0-1; higher DI scores reflect higher socio-economic deprivation levels. Scores were divided into quartiles for analysis: Q1, 0.00-0.24; Q2, 0.25-0.49; Q3, 0.50-0.74; Q4, 0.75-1.00 (20 zipcodes only). We compared the change in firearm injury as a proportion of all injuries pre- and during pandemic, by DI quartile overall and by injury intent. We used the Kruskal-Wallis and Chi-square tests, as appropriate, to compare variables of interest pre- vs during pandemic, and Bonferroni correction for multiplicity adjustment for subgroup analyses. We conducted multivariable logistic regression analysis for outcome of firearm injuries adjusting for time and patient characteristics.

Results: There were 1481 firearm injuries: 805 pre- and 676 during pandemic. Mean age of firearm-injured victims was 12.8 years old. Most were male (76.8%). Median DI for firearm-injured patients was 0.35 (IQR 0.26-0.43). Firearm injuries increased with increasing DI quartile (p < 0.0001) except in Q4, which had no firearm injuries. Comparing during to pre-pandemic, for all intents, there was no change in firearm injuries for DI Q1, but significant increases for DI Q2 and Q3 (Figure). When analyzing intent, assault injuries were unchanged during compared to pre-pandemic for DI Q1, but increased for Q2 and Q3 (Table 1); unintentional injuries increased for Q2. There was no difference in self-inflicted firearm injury for any DI quartile. In the multivariable model there were increased odds of firearm injuries during compared to pre-pandemic (OR 1.90, 95% CI 1.72, 2.11), and for DI Q2 (OR 3.04, 95% CI 2.42, 3.82), DI Q3 (OR 4.36, 95% CI 3.40, 5.60) compared to DI Q1.

Conclusion: The pandemic exacerbated existing disparities in firearm injuries, with disproportionate increases in those with higher levels of socio-economic deprivation. Targeted efforts to address upstream causes of disparities, including social determinants, are crucial to addressing the firearm public health epidemic.