Impacts of confounding roadway characteristics on estimates of associations between alcohol outlet densities and alcohol-related motor vehicle crashes

Abstract

Introduction and aims: Previous research on alcohol-related motor vehicle crashes (AMVC) share a substantial limitation: sources of geographic variations in background crash risks may confound estimated spatial relationships between alcohol outlets and AMVCs. The aim of this study was to address this concern by examining, spatial-temporally, relationships between alcohol outlets and AMVCs adjusting for a set of six roadway characteristics that may be, independently, related to crash risks. While most similar studies focus on one metropolitan area, we use a unique sample of 50 cities. Design and methods: The spatial sample for this study consisted of 8726 Census 2000 block groups representing 50 mid-sized California cities. Dependent measures were counts of crashes located within Census block groups. Independent measures included socio-demographics, social disadvantage, alcohol outlets and roadway characteristics. We assessed relationships of crashes to independent measures using hierarchical generalised linear models. Results: Greater roadway length, greater percentage of highways, greater average speeds, fewer T-intersections, greater curviness and less fragmentation were related to greater numbers of crashes as was alcohol outlet density. Discussion: Above and beyond alcohol outlet type and density, we found that roadway characteristics were related to AMVC risks across a sample of 50 mid-sized cities. Measures of roadway characteristics are an essential component of any model of motor vehicle crashes that attempts to assess impacts of alcohol outlets on motor vehicle crashes risks.