Real-world accident data is used as part of the process by which vehicles and roads are made safer. Typically data is used to identify priorities in injury prevention and to support the development of test procedures. However the provision, nature and integration of accident data with safety policy is in many cases unsystematic and not fully capable of meeting the requirements of it.

This paper examines the existing structures that utilise accident data to improve safety, and compares the existing systems in the US, UK, and EU. The paper concludes that:​

Safety policy needs to take full account of real-world issues – all groups dealing with safety policy need to have a close connection to a strong accident data resource

Data collection needs to be an integrated part of the problem identification->solution generation (technical development of standards)->monitoring solution effectiveness cycle

Data systems must be designed to meet the specific objectives of the main casualty groups

Different levels of data are needed to provide a complete resource including national data, longitudinal studies and focused studies.

The levels of detail in the data gathered must match the detail in the research questions being addressed New technologies, such as event data recorders, have the potential to improve the detail of in-depth data, but there are obstacles from lack of standardisation and privacy regulations.

The paper proposes:​

New system of accident and injury data to integrate with EU structures

Coordinated international approaches to accident data within the framework of EEVC and IHRA

Specific technical areas, including collision severity assessment and injury scaling, where new advances are required to accurately describe injury causation