Crash severity modelling using ordinal logistic regression approach

Int J Inj Contr Saf Promot. 2020 Dec;27(4):412-419. doi: 10.1080/17457300.2020.1790615. Epub 2020 Jul 5.

Abstract

Road traffic accident is one of the major problems facing the world. The carnage on Ghana's roads has raised road accidents to the status of a 'public health' threat. The objective of the study is to identify factors that contribute to accident severity using an ordinal regression model to fit a suitable model using the dataset extracted from the database of Motor Traffic and Transport Department, from 1989 to 2019. The results of the ordinal logistic regression analyses show that the nature of cars, National roads, over speeding, and location (urban or rural) are significant indicators of crash severity. Strategies to reduce crash injuries should physical enforcement through greater Police presence on our roads as well as technology. There is also the need to train drivers to be more vigilant in their travels especially on the national roads and in the urban areas. The Recommendation is, a well thought out and contextualised written laws and sanctioned schemes to monitor and enforce strict compliance with road traffic rules should be put in place.

Keywords: Ordinal logistic regression; chi-square test; injury; proportional odds; road traffic accidents.

MeSH terms

  • Accidents, Traffic* / statistics & numerical data
  • Ghana
  • Humans
  • Logistic Models
  • Models, Statistical*