County-level associations between food retailer availability and violent crime rate

Abstract

Abstract Background Violent crime (i.e., homicide, armed robbery, aggravated assault, and rape) continues to be a major public health concern in America. Several studies have linked the availability and density of specific features of the retail food environment, such as convenience stores and liquor stores, to violent crime rates due to the criminal activity that often occurs in and near these retailers. Nevertheless, there continues to be limited understanding of how other features (e.g., grocery stores, supercenters, restaurants, etc.) are associated with violent crime occurrence. This study aimed to fill this gap in knowledge by examining U.S. county-level associations between food retailer availability and violent crime rate. Methods We analyzed 2014 data on 3108 counties from the U.S. Department of Agriculture’s Food Environment Atlas and Department of Justice’s Unified Crime Reporting Program. Per capita food retailer measures represented the number of stores per 10,000 county residents. Violent crime rate represented the number of police reported violent crimes per 10,000 county residents. We used spatial lag regression models to assess associations between per capita retailer availability and violent crime rate after adjusting for potential confounders (e.g., % under 18, % Black, % Hispanic, % poverty, population density, etc.). In addition, we examined stratified OLS regression models to evaluate associations by metropolitan county status. Results Adjusted spatial regression models revealed that greater supercenter availability [β: 2.42; 95% CI: 0.91–3.93; p-value: 0.001] and greater fast food restaurant availability [β: 0.30; 95% CI: 0.18–0.42; p-value: < 0.001] were associated with higher violent crime rate. Greater availability of farmers’ markets [β: -0.42; 95% CI: -0.77 – − 0.07); p-value: 0.02] was associated with lower violent crime rate. Associations varied between metropolitan and non-metropolitan counties. Stratified OLS models revealed that greater grocery store availability was associated with lower violent crime rate among metropolitan counties only. Greater fast food restaurant availability was associated with lower violent crime rate among non-metropolitan counties only. Conclusions Certain features of the retail food environment appear to be associated with county-level violent crime rates in America. These findings highlight the need for additional research on the influence of food retail and food landscape on violent crime occurrence at the community level.