UC Berkeley

Firearm violence is responsible for a substantial amount of morbidity and mortality in the United States, particularly among young Black men (via assault) and older white men (via self-harm). Successful prevention efforts rely on the identification of modifiable risk factors, which can inform new interventions, as well as the evaluation of existent interventions. Individual-level and ecologic studies have found firearm ownership to be a strong risk factor for firearm mortality, but ownership itself can be difficult to intervene on. However, firearm access may be subject to community-level interventions through zoning regulations on firearm dealers and to individual-level intervention via targeted prohibitions among those at high risk of harming others or themselves. Despite general acceptance of the social-ecological model in violence and injury epidemiology, the cross-level risks associated with firearm availability have rarely been explored.

The objective of this dissertation was to identify modifiable risk factors and to evaluate a firearm violence intervention, focusing particularly on cross-level impacts of firearm availability (i.e., how community firearm availability impacts individual risk and vice versa). To do this, we first examined whether community firearm and alcohol availability were independently or jointly associated with individual-level risk of (fatal and nonfatal) firearm assault or self-harm. We conducted two population-based case-control studies of California residents, 2005-2015, drawing on statewide hospitalization, emergency department, firearm transfer, and alcohol license data. The first of these studies focused on firearm assault and homicide and the second focused on firearm self-harm and suicide. We use case-control-weighted g-computation to assess marginal risk differences under various scenarios of interest and to assess additive interactions between measures of firearm and alcohol availability.

Second, we explored whether removal of firearms from high-risk individuals via gun violence restraining orders (GVROs) was associated with the rate of firearm violence in San Diego County, which issued a plurality of orders in the state. California’s GVRO law, enacted in 2016, permits the temporary removal of firearms from people at high risk of committing firearm violence. Using statewide hospitalization and emergency department data, we conducted a quasi-experimental study using the synthetic control method to determine whether GVRO implementation was associated with decreased rates of (fatal and nonfatal) overall firearm violence, firearm assault, or firearm self-harm in San Diego County during the first four years of implementation.

Overall, we found weak cross-level associations between firearm availability and firearm violence. Non-pawn firearm dealer and off-premise alcohol outlet density were associated with modest increases in the individual-level risk of firearm assault, but pawn dealer and off-premise alcohol outlet density were null. Firearm sales density (but not firearm dealer density) was associated with elevated risk of firearm self-harm, but alcohol availability was null or of negligible magnitude. Finally, we found that GVRO implementation was not associated with a reduction in firearm violence of any kind in San Diego County from 2016-2019.

This project will be of interest to policymakers and researchers interested in understanding and preventing firearm violence. Despite our modest findings, this study could inform local interventions relating to community firearm availability. It also provides the first quantitative evaluation of the GVRO law in California, which served as a template for similar laws in 16 other states. Risk factors for firearm violence come from multiple social-ecological levels, including the surrounding social, physical, and policy environments. Successful strategies for prevention will therefore include individual-level interventions, such as GVROs (which may be effective at the individual-level), as well as broad measures that reduce environmental risk. This work adds to the limited literature on cross-level risk associated with firearm availability, but additional research into modifiable risk factors and interventions to prevent firearm violence is needed.