Technical precisions
Technical precisions
This dashboard provides information on the risk of suffering serious or fatal injuries in road accidents in Belgium between 2005 and 2023. The risk is calculated as the probability of a given type of road user being seriously or fatally injured if they were involved in an accident with at least one injured person. Logistic regression analysis is applied in order to derive an analytical expression for the risk as a function of heaviest opponent in multi-vehicle crashes. Furthermore, variables that can be considered as factors contributing to the risk are also included in the model, such as the mass of the heaviest adverse vehicle and/or the type of road user. The risk is expressed as a percentage. A risk of 10% therefore means that a road user has a 10% chance of being seriously injured or killed when involved in an accident. In the first sheet, via the filters on top, the risk can be displayed for different subgroups. The filtered variables are road user type (pedestrian, cyclist, moped rider, motorcyclist, car occupant, truck occupant) of victims, heaviest opponent (e-scooter, cyclist, moped rider, motorcyclist, car, van, bus/coach, lorry), speed regime (30 km/h, 50 km/h, 70 km/h, 90 km/h, 120 km/h), lighting conditions (day, night, twilight), road conditions (dry, humid, snow/ice, normal conditions), peak hours as traffic density (evening peak, morning peak, off-peak hours). In the second sheet, we provide the risk depending on the mass of vehicles of the heaviest opponent from 2017 until 2022. In the third sheet, we provide risk-matrix depending on the mass of vehicles of the road-user types and the heaviest opponent. It is important to note that only moped drivers, motorcyclists and car occupants are included in this sheet. The accident database is based on crash reports filled in by the Federal Police/DGR/DRI/BIPOL after an accident, and provided by Statbel (Directorate-general Statistics - Statistics Belgium). The information on the mass of vehicles of road user types and the heaviest opponent is retrieved from DIV database.
Â
Terminology
Pedestrian
: Road user going on foot, using a wheelchair or pushing a bicycle or a moped.
Cyclist
: Road user riding a (non-electric) bicycle, an e-bike (<=250W and <=25km/h) or a motorized bicycle (<=1000W and <=25km/h).
Moped
: Class A or B moped, with 3 or 4 wheels, or speed pedelec.
Motorcycle
: Two-wheel motorized vehicle with or without sidecar, fitted with an engine having a cylinder capacity of more than 50 cc and/or capable of exceeding a speed of 45km/h.
Lorry
: Motorized vehicle used for transporting goods, with a maximum gross vehicle mass of more than 3,5T, or tractor with or without semi-trailer.
Heaviest opponent
: In a multi-vehicle accident, heaviest opponent is defined as the road user type that is the heaviest vehicle in the accident. For example, in case of an accident where 3 different road user types such as pedestrian, car and lorry are involved, lorry is considered as the heaviest opponent in the analysis.
Examples : Taking into account the heaviest opponent in an accident, it can be seen from the first sheet that the risk of being seriously injured or killed of a motorcyclist in a multi-vehicle accident is 6% when the heaviest opponent is a cyclists, whereas it is 34% when the heaviest opponent is a lorry. When it comes to age groups, the risk of being seriously injured or killed is highest for victims aged more than 65 years old, while it is the lowest for victims aged between 18-24 years old. Looking at the sheet 2, it can be seen that the risk varies according to the mass of the vehicle of the heaviest opponent. In the risk-matrix in the third sheet, the results show that the risk moves in opposite direction between the mass of the vehicle of the road user type and the heaviest opponent, meaning that the risk increases when the mass of the vehicle of the heaviest opponent increases.
Limitations of the data
When interpreting the data, some limitations should be kept in mind.
- When the data is filtered to show certain sub-groups, it can sometimes be seen that there is no data in certain visuals. This is the case when several filters are used simultaneously and the figures from which the risk is calculated are too small or non-existent. In some cases, therefore, the results must be interpreted with caution.
- Depending on the DIV dataset, each sheet has a different number of observations. Hence, the comparison between different sheets should be carefully done.
- The risk calculation could be improved by taking into account other exposure data such as the total number of kilometres travelled by each road user, the time spent on the road or the purpose of the journey. However, the absence of these data represents a certain limitation in the model applied.
Â
