Firearm injury is the 4th leading cause of violence-related injury in ages 15 to 44 in the U.S. For Americans, firearm homicide and firearm suicide is also the leading cause of violence-related injury deaths in ages 5 to 65+ years. Firearm injury incidences are not randomly distributed across areas. Geographic information systems can be used to understand the geospatial relationship between two variables at the census block level. It was hypothesized that density of tangible community resource locations (TCRL) could play a role in the density of firearm injury incidence in census block groups as prior research has found that geospatial access to care can be associated with reduced deaths from gun violence.
Froedtert Hospital’s Trauma Registry was used to calculate density of firearm injury locations at the census block group level from 2015-2022. Data from the IRS provided non-profit names and addresses of TCRL in Milwaukee County. Criteria for inclusion of TCRLs included those providing health care or social service outreach such as health clinics, community outreach organizations, and childcare entities. A geographically weighted regression (GWR) was performed to explore the relationship between TCRL density and firearm injury density per census block group while controlling for Area Deprivation Index (ADI), a measure of neighborhood disadvantage, along with population and race derived from the U.S. Census. To reduce the risk of multicollinearity within models, only variables representing the population of one race were used within the model.
Between 2015 and 2020, Froedtert Hospital’s Trauma Registry recorded 861 incidents of firearm injury within Milwaukee County, and 733 TCRL were compiled from the Internal Revenue Services’ open data portal. The proportion of TCRL per census block group did not align well with the population per census block group or firearm injury density per census block group. The GWR yielded results consistent with that of the null hypothesis in many areas. The near north/central part of the county showed that greater proportions of TCRL’s was related to lower injury density. However, given the fact that TCRL’s and firearm injury densities are not distributed comparably, this negative relationship of TCRL and firearm injury was not displayed in all of the expected areas of the county.
Geospatial mapping showed inconsistencies in the relationship between TCRL locations and population relative to census block group size. Review of the TCRL coefficients was partially consistent with the null hypothesis in most of the county. Few pockets of the county demonstrated results consistent with the alternative hypothesis. This can be attributed to the disproportionately high rates of firearm injuries within these pockets. The addition of predictors within the GWR could improve future models. For further research, it would be meaningful to investigate the reach of TCRL’s in an effort to increase efficacy within these organizations.
Firearm, Injury, Prevention, Geospatial
Byrne, James P., et al. “Association between Geospatial Access to Care and Firearm Injury Mortality in Philadelphia.” JAMA Surgery, vol. 157, no. 10, 2022, p. 942, https://doi.org/10.1001/jamasurg.2022.3677.
Cook, Alan, et al. “Geospatial Characteristics of Non-Motor Vehicle and Assault-Related Trauma Events in Greater Phoenix, Arizona.” Injury Epidemiology, vol. 7, no. 1, 2020, https://doi.org/10.1186/s40621-020-00258-x.