Background The specific features of crashes involving an alcohol-intoxicated driver have been extensively characterized, but no such data are available for crashes involving a driver who has used a prescription medicine, which could help to plan and target prevention and control initiatives. This study aimed to describe the characteristics of crashes involving drivers under the influence of prescription medicines. Methods We took advantage of CESIR, a French record linkage study for which data were extracted and matched from three French national databases: police reports, the national police database of injurious crashes and the national health care insurance database (HCI database). The drivers included in the study were those involved in an injurious road crash in France from July 1, 2005 to December 31, 2015, whose national identity number, date of birth and sex allowed matching. Prescription medicines considered were those with the two highest levels of warning. Results Similar crash profiles were found when drivers used alcohol or medicines, particularly with respect to injury severity, type of vehicle, type of collision, type of road and cross-track profile. Alcohol-related crashes were over-represented during weekends and in low-density areas and medicine-related crashes were over-represented during weekdays and in cities of fewer than 300 000 inhabitants. While a much higher strength of association with responsibility was found for alcohol than for medicines, the proportion of crashes with drivers using medicines was twice as high as crashes with drivers using alcohol. Conclusion The lower risk carried by medicines is therefore in part compensated by a higher prevalence of use, making medicines one of the hidden factors of road risk. Characterizing these crashes will therefore be useful to better focus road safety intervention on the prevention of driving under the influence of psychotropic medicines.< Réduire