Report numberRA-2006-80
TitleAccidents in thorougfares
SubtitleA risk analysis of thoroughfares as a whole
AuthorsKurt Van Hout
Published byPolicy Research Centre for Traffic Safety 2002-2006
Number of pages51
Document languageDutch
Work packageOther: Infrastructure and space

This report is for now the last in a series of 3 reports concerning traffic safety in the built-up environment. The first report was a literature review of the different infrastructure characteristics that influence traffic safety. The second report describes the results of a risk analysis of Flemish through roads (regional roads through a built-up area) base don the elements of the cross section of the road. This report does the same, but this time on the level of the through road as a whole.


The data used in this research originate from a inventarisation done by the Flemish Administration of Roads and Traffic as part of a priority analysis concerning through roads. The datafile contains 321 through roads, spread over the Flanders region. Road accidents from 1996 to 2001 are used in the analysis.


The goal of this study is working out the impact of specific characteristics of traffic, road and surroundings on traffic safety in Flemish through roads, based on the data available. This was done by a cross sectional survey in which data from a certain point in time are gathered for several through roads in the research area.


Through generalised linear modelling (by means of the procedure GENMOD in the statistical package SAS) models are derived that give the number of accidents as a function of some parameters available on the level of a through road as a whole. First some models are built that give the overall number of accidents base don the data for all road sections in the database. Special attention is paid to the form of the intensity function (the assumed relationship between traffic intensity and the number of acidents).


Additional models are built for more specific accident types (e.g. accidents wich involve bicyclists, pedestrians, serious accidents, accidents on cross roads or road sections, …) or for more specific through road types (4 types are considered according to speed limit, number of lanes and the situation of the road in relation to the built-up area (approach road, touching or cutting across the built-up area, presence of a ring-road)).


The models explain the variation in the number of accidents on existing through roads. The relationships found aren’t necessary causal. The models are therefore not suitable to predict a change in the number of accidents following a through road redesign. The change in number of accidents can be calculated when the models are combined with the results of a Before-After study on the subject. The models can however be an aid in determining priorities of through road redesign.


Models that want to predict the number of accidents on through road level have to contain the explaining variables traffic intensity en the length of the through road. When the number of accidents involving a specific type of road user has tob e predicted, it is recommended that the number of that specific road user is also added as an explaining variable. Additional variables as the number of cross roads, the number of facilities (shops, community services, …) or the number of inhabitants can also contribute to a smaller improvement of the models.

DownloadPDF icon RA-2006-80.pdf


The Policy Research Centre for Traffic Safety carries out policy relevant scientific research under the authority of the Flemish Government. The Centre is the result of a

cooperation between Hasselt University, KU Leuven and VITO, the Flemish Institute for Technological Research.


Leuven vito