Using geospatial and statistical methods to evaluate the effect of personal factors and environmental conditions on pedestrian injury risk and identifying high risk road segments

Background: pedestrian injuries occur disproportionately with respect to personal characteristics, environmental conditions, and more importantly vary geographically. Metropolitan Atlanta has been ranked fourth worst in the United States regarding pedestrian injuries.

Objectives: this study aims at evaluating factors of demographics, behaviors, and environmental conditions on injury risks to pedestrians, and identifying high injury-risk road segments in Metropolitan Atlanta.

Data and Methods: comprehensive historical pedestrian crash records from 2000 through 2007 (n = 8,498) were obtained from Georgia Department of Transportation. Descriptive statistics were used to explore the variations in the number of cases by year, month, and day of week. Multivariate logistic regression were used to model the injury risks associated with gender, age, intoxication, maneuver, day of week, light conditions, surface conditions, and weather conditions. Geographic Information Systems and spatial clustering methods were used to detect the road segments where pedestrian injuries statistically increase.

Results: the results show fall season (September to November) and weekends have experienced apparently higher incidence of pedestrian crashes and injuries. Injury risk significantly increases (á = 0.05) in male, children (aged 14 or less), intoxication, crossing not at crosswalk, and darkness. Accounting for the space-time interactions, the spatiotemporal clustering detected a road segment along the Buford Highway Corridor that injury risk significantly increased (á = 0.05).

Conclusions: findings from this research are important to identify high-risk populations for heightened education and preventive efforts, as well as to provide insights into hazardous zones for intervention and redevelopment of the area.