OBJECTIVES: The aim of this benchmarking study is to provide a detailed analysis of medical devices (MD) price and to identity what drove price dispersion. METHODS: A large panel of MD level price data was collected in 3 French public healthcare institutions and 10 centralized purchasing groups (representing 37% of french public hospitals). MD were selected according to the Pareto law (20% of the MD represent 80% of the expenditure) and expert opinion to ensure that each MD had sufficiently large demand. Several factors were considered such as volume purchased, affiliation to a purchasing group, procurement procedure and contract start date. RESULTS: Finally, 18 MD were retained following up on the provided answers (5 elastic bandages, 2 implants, 8 common MD and 3 captive MD). In terms of pricing, results between hospitals being close for similar quantifies and none can be defined as the benchmark leader. Rebates are a common mechanism and the level of discount depends on the MD considered and type of funding (activity based or additional payments). Open public tenders are the most commonly procurement procedure used, whereas negotiated procedures are more efficient for captive MD. There is practically no relationship between volume purchased and purchased prices. MD price can change over time and the relevance of the contract start date is confirmed: older the contract is, cheaper is the price for some MD, or on the contrary for others. CONCLUSIONS: There is no connection between catalogue prices and purchased prices especially for MD paid by activity based payment (discount rates can reach 50% to 90%). As the volume effect has no evident impact on MD discounts, the advantage of joining a centralized purchasing group has not been confirmed. The crucial factors are mainly the MD considered, the contract start date and the procurement procedure.