Identifying regions of excess injury risks associated with distracted driving: A case study in Central Ohio, USA

• Published in: SSM - population health Vol. 20; p. 101293
• Main Authors: Lym, Youngbin, Kim, Seunghoon, Kim, Ki-Jung
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.12.2022; Elsevier
• Subjects: Distracted driving; Excessive injury risks; Hierarchical bayesian; Regular; Spatial correlation; Transportation safety; Excessive injury risks; Transportation safety; Distracted driving; Hierarchical bayesian; Spatial correlation
• ISSN: 2352-8273; 2352-8273
• DOI: 10.1016/j.ssmph.2022.101293

This study examines the latent influence of spatial locations on the relative risks of crash injuries associated with distracted driving (DD) and identifies regions of excess risks for policy intervention. Using a sample of aggregated injury and fatal DD crash records for the period 2015–2019 across 1,024 census block groups in Central Ohio (i.e., the Columbus Metropolitan Area) in the United States, we investigate the role of latent effects along with several covariates such as land-use mix, sociodemographic features, and the built environment. To this end, we specifically leverage a full Bayesian hierarchical formulation with conditional autoregressive priors to account for uncertainty (i.e., spatially structured random effects) stemming from adjacent census block groups. Furthermore, we consider uncorrelated random effects from upper-level administrative units within which each block group is nested (i.e., census tracts and counties). Our analysis reveals that (1) addressing spatial correlation improves the model's performance, (2) block-group-level variability substantially explains the residual random fluctuation, and (3) intersection density appears negatively associated with the relative risks of crash injuries, while more diversified land use can increase injury risk. Based on these findings, we present spatial clusters with twice the relative risks compared to other block groups, suggesting that policies be devised to mitigate severe injuries due to DD and therefore enhance public health. •Crash injuries associated with distracted driving are investigated.•Spatial correlation accounts for residual variation in relative injury risks.•Intersection density appears to reduce the risks of crash injuries.•Diversified land use leads to an elevated injury risk.•We identify small areas with excess injury risks.