Exploring the Predictors of Accident Severity in Urban Ghana

Abstract

Road traffic accidents involving adverse severe outcomes, injuries and fatalities in particular are among the major challenges facing Ghana. The present study investigates the underlying predictors of fatal and injury accidents in Kumasi, Ghana. A sample of 608 accident cases was systematically sampled from a large traffic police database. Data analysis was performed using the logit model with stepwise variable selection at 0.05 and deletion at 0.10. Of the entire sample, 56.6% were injury and 43.4% involved in at least one fatality. The model revealed that the significant predictors of severe accidents were mainly over­crowding, driver indiscipline on roads, driver fatigue, and design and conditions of roads with estimated odds ratio of 2.42, 3.83, 10.51 and 12.06 respectively. Other variables including speed, drunk driving, not using helmet, mechanical fault, over loading and indiscriminate use of roads by pedestrians were not signifi­cant predictors of severe accidents. Prevention strategies should target effective law enforcement, traffic regulations and safe road engineering.

Keywords: Road Traffic Accident, Fatality, Injury, Stepwise Logit Model