Given the lack of an overall picture of the extent of workplace accidents, commuting accidents, and occupational diseases—due in particular to the fragmentation of compensation statistics across the various social security schemes—the Public Health Policy Act of August 9, 2004, tasked the French Institute for Public Health Surveillance (InVS) with developing a centralization tool. After describing the data available in the main schemes, the InVS conducted a feasibility study based on information recorded by three schemes: the general Social Security scheme, the agricultural scheme, and the special scheme for local government and hospital employees. Indicators were defined to highlight concerning situations. For workplace accidents, the following were selected: the number of accidents, the frequency rate, the percentage of serious accidents, and the percentage of multiple-injury cases, broken down by gender according to age, sector of activity, the victim’s occupation, and the mechanism of the accident... The classifications used for sector of activity and occupation are those of INSEE. The accidents studied are those that occurred in 2004/2005 in order to assess all their consequences, including long-term effects. The feasibility study shows that, for workplace accidents, the centralization of data from the three Social Security schemes is possible and yields unprecedented results. For the first time, it has been possible to estimate the annual number of these accidents (approximately 914,000 for men and 370,000 for women) independently of the Social Security scheme. Analysis by industry sector reveals new findings, such as the high risk in the food processing industry for both genders (specifically the meat-cutting subsector), as well as more traditional findings: high risk in the construction and transportation sectors for men, and in the healthcare sector for women. The use of multiple indicators and gender-based analysis provide a deeper understanding of the nature of this occupational risk. The feasibility study led to the development of two types of recommendations: - improvements regarding the data (particularly regarding the size of the populations covered), - proposals concerning the implementation of the centralizing tool. (R.A.)