International data about the number of road deaths are fairly reliable and comparable since, mostly, the same definition is used and road deaths are registered fairly appropriately. Information about the number of serious road injuries is, however, still hard to compare and less reliable since road injuries are registered quite differently, and since the common European definition [3] advised in 2013, is still not applied in the same way everywhere.
Definition of a road crash
The international definition of a road crash is: a crash on a public road in which at least one moving vehicle is involved [14]. A crash of a pedestrian stumbling over a parked bike is therefore not considered a road crash, nor is a collision of two cars in a private parking lot. There are minor differences between countries about in- or excluding crashes on private roads, suicide and death by natural causes just prior to a crash. In their road crash registration, the United States and Canada only register crashes in which a motor vehicle was involved: bicycle crashes are therefore not included.
Definition of a road death
The international definition of a road death is: a casualty who, in or after a crash on a public road in which at least one moving vehicle is involved, dies within thirty days from the consequences of that crash, with the exception of suicides [15]. Nowadays, almost all European countries use this definition. France (up to 2015), Spain (up to 2010) and Portugal (up to 2009) were the exceptions. Portugal and Spain only considered casualties who died at the scene of the crash to be road deaths, and France used a time frame of 7 instead of 30 days [16].
Definition of a serious road injury
In 2013, the European Commission chose a Europe-wide definition of a serious road injury: a road injury with a Maximum Abbreviated Injury Score (MAIS) of 3 of higher [3]. MAIS is an international medical measure to classify injury severity. This score may be derived from the various injuries coded for a patient (see SWOV fact sheet Serious road injuries in the Netherlands). Many countries now use this definition, even though the practical application may differ [4]; also see the question How does the Dutch number of road injuries compare to that elsewhere? . The Netherlands still use a MAIS score of 2 or higher as the criterion for serious road injuries. For the sake of uniformity within the EU, and to align the definition to what is customary in the medical world, it stands to reason that the Netherlands will eventually adopt the MAIS3+ criterion as well [17].
Data related to exposure
In order to compare road safety in different countries, data of the number of casualties need to be supplemented with data about population size, road length, vehicle fleet and number of kilometres travelled. In Europe, statistics about numbers of inhabitants and vehicles, and length of the road network are fairly easy to compare. This is not the case for the more direct exposure measures, such as the number of kilometres travelled by road users or vehicles, since for these measures different research methods are used [18].
Data reliability
The registration of road deaths in most Western countries is believed to be in fairly good order and the data to be therefore reliable. In the European Union, 1 to 2% of road deaths are estimated to be left unregistered [19]. The combination with the by now common definition makes international comparisons fairly reliable. It does, however, remain to be seen whether all countries are equally aware of the degree of underregistration, and whether this degree is not underestimated. In the Netherlands, where underregistration is determined using a comprehensive and advanced method, the police fail to register about 15% of road deaths as such [16]. A second explanation of the high degree of underregistration in the Netherlands is that bicycle-only crashes are plentiful here. These are not always registered since police are not usually present at the scene of the crash.
The registration of serious road injuries is much less exhaustive. On the basis of a somewhat older study by Elvik and Mysen [20], the European Commission concluded that, in Europe, around 70% of serious road injuries are registered [5]. This would still be considerably more than the estimated Dutch 30% that are registered by police [17]. Considering the differences in use of the definition and the generally considerable degree of underregistration, international comparisons of serious road injuries are therefore not very reliable. Yannis et al. [13] do, however, conclude that data of OECD countries are generally more reliable than data of non-OECD countries.
