Under-reporting of road crash casualties in France
Introduction
In most countries, epidemiological knowledge of road crash casualties is mainly based on data collected by the police. Fatal casualties are quite clearly defined and well reported, but this is not the case for non-fatal casualties. The under-reporting of non-fatal casualties is well known (Hauer and Hakkert, 1988, James, 1991, Hvoslef, 1994, Stutts and Hunter, 1998, Aptel et al., 1999, Elvik and Mysen, 1999, Alsop and Langley, 2001, Cryer et al., 2001, Dhillon et al., 2001, Rosman, 2001, Laumon and Martin, 2002, Langley et al., 2003); the degree of under-reporting can be quite large, and furthermore it varies according to characteristics such as injury severity and road user type.
The validity of data used as a basis for epidemiological knowledge of road casualties should be assessed, and this should be done in each country since the police definition of road casualties and police reporting practice are country-specific. We analyse here the police under-reporting of non-fatal casualties in France: the degree of under-reporting and how it varies according to some crash or casualty characteristics; in other words, it consists of identifying and quantifying selection bias factors.
This can be achieved thanks to the existence of a medical road trauma registry. This registry covers all victims of road crashes in the Rhône county who seek medical attention in health facilities in the county or close surroundings. Record-linkage has been applied to the registry and police files so that casualties reported by both sources are identified. It is therefore possible to analyse the police “filter”, i.e. the filter through which casualties in the registry are “selected” into being police-reported. This is done by modelling the probability of being police-reported among casualties in the registry, as a function of their crash and/or casualty characteristics. In this analysis, police under-reporting covers both under-reporting to the police and under-reporting by the police as we cannot distinguish between the two. Under-reporting to the police corresponds to no-one calling the police; under-reporting by the police corresponds to the police not writing a crash report even though present at the crash scene, or omitting some of the casualties within a reported crash. The unit of the whole analysis is the casualty, not the crash.
Additionally, a survey on casualties identified in the registry but not in the police files has been set up, with the aim of having a better understanding of police under-reporting. In particular, this study aims at identifying and quantifying reasons for not calling the police.
Section snippets
Police traffic crash data
The French police are required by law to write a crash report for every road crash causing at least one casualty. A road crash is officially defined as a crash involving at least one vehicle and occurring on the network open to public traffic. Skateboard or roller skate users are considered as pedestrians by the police, and, as such, are only classified as road casualties if hit by a vehicle. There is no restriction about motorised vehicles, in other words there is no exclusion criteria on
Main analysis
As previously stated, we focus on non-fatal casualties, as they suffer from a large under-reporting compared to fatal casualties, and also because the underlying reasons and characteristics of their under-reporting might be quite different. We compare the police data to the registry, which we restrict to the casualties that fulfil the police definition of road casualties: this implies that roller skate, skateboard and scooter users should be considered as pedestrians and excluded if no third
Under-reporting and selection bias
The registry reports 53988 non-fatal casualties from road crashes in the Rhône county during the 1997–2001 period. Of these we exclude 2374 roller skate, skateboard or scooter users not hit by a vehicle, leading to 51,614 casualties in the registry. The police reports 22,498 non-fatal casualties. Of these, 14,398 are identified in both sources, leading to a total of 59,714 non-fatal casualties (Fig. 1). The police reporting rate is 37.7%.
For further analysis, we have to exclude 678 casualties
Discussion
The strength of this study lies mostly in the Rhône road trauma registry. This registry has been acknowledged by the French National Registry Committee. The study is based on a 5-year period, on a county with a large population, and hence on a large number of subjects.
The underlying assumption of the study is that each police force having a centralised national structure (not regional) and receiving centralised instructions, each one has the same way of reporting crashes and casualties all over
Conclusion
The large extent of police under-reporting of non-fatal casualties and the associated selection biases means that police data give the wrong picture. Slight casualties are under-represented; their consequences in terms of health care, disabilities and associated costs are not as dramatic as serious casualties, but since there are many more of them, they do represent a public health issue. Casualties involved in crashes without a third party are under-represented, and even more so for some
Acknowledgements
We wish to thank the following people for having participated in the data collection and data entry, in the framework of the Association for the Registry of road traffic casualties in the Rhône (ARVAC, president Banssillon V., director Ndiaye A., scientific consultant for the registry, Laumon B) and in the framework of INRETS-UMRESTTE: Ait Idir T., Ait Si Selmi T., Alloatti D., Andrillat M., Artru F., Asencio Y., Assossou I., Auzaneau F., Bagès-Limoges F., Bagou G., Balogh C., Banssillon G.,
References (24)
- et al.
Under-reporting of motor-vehicle traffic crash victims in New-Zealand
Accident Anal. Prevent.
(2001) - et al.
Road accident statistics: discrepancies between police and hospital data in a French island
Accident Anal. Prevent.
(1999) - et al.
Assessment of hospital and police ascertainment of automobile versus childhood pedestrian and bicyclist collisions
Accident Anal. Prevent.
(2001) - et al.
Complementing police road-crash records with trauma registry data- an initial evaluation
Accident Anal. Prevent.
(2000) The Western Australian road injury database (1987–1996): ten years of linked police, hospital and death records of road crashes and injuries
Accident Anal. Prevent.
(2001)- Abdel-Aty, M., Keller, J., Brady, P., 2005. Analysis of the types of crashes at signalized intersections using complete...
- et al.
Le registre des victimes d’accidents de la circulation routière du Rhône [The trauma registry of road casualties in the Rhône county]
(2002) Practical introduction to record linkage for injury research
Injury Prevent.
(2004)- et al.
Investigation of bias after data linkage of hospital admission data to police road traffic crash reports
Injury Prevent.
(2001) - et al.
Incomplete accident reporting; meta-analysis of studies made in 13 countries
Transport. Res. Rec.
(1999)
Model-based estimation of relative risks and other epidemiologic measures in studies of common outcomes and in case-control studies
Am. J. Epidemiol.
Extent and some implications of incomplete accident reporting
Transport. Res. Rec.
Cited by (184)
A systematic review and meta-analysis of data linkage between motor vehicle crash and hospital-based datasets
2024, Accident Analysis and PreventionPolice and hospital data linkage for traffic injury surveillance: A systematic review
2024, Accident Analysis and PreventionAdvances, challenges, and future research needs in machine learning-based crash prediction models: A systematic review
2024, Accident Analysis and PreventionSeeing through eyes of Drivers: Space consideration in investigating visibility of Vulnerable road users involved in crashes from Driver's perspective
2023, Accident Analysis and PreventionLearning about injury severity from no-injury crashes: A random parameters with heterogeneity in means and variances approach
2023, Accident Analysis and PreventionMaritime accidents in the Yangtze River: A time series analysis for 2011–2020
2023, Accident Analysis and Prevention