This report presents the results of an exploratory study aimed at assessing the feasibility of establishing an epidemiological surveillance system in prisons based on data collected from inmates’ medical records. The study relied on two complementary data collection methods: interviews conducted with the physicians in charge of the Health Units (HU) at eleven correctional facilities selected in a reasoned manner to reflect the diversity of situations in terms of facility types and inmate populations; the analysis of a sample of approximately three hundred medical records of inmates selected at random from these facilities. The results show that it is not feasible, at this stage, to establish a surveillance system based on inmates’ medical records, due in particular to: i) scattered medical information (the coexistence of multiple records located in separate facilities); ii) the lack of digitization of medical records; iii) the lack of standardization of medical records, with content that varies from one facility to another, and the consequent impossibility of developing the epidemiological indicators necessary for monitoring the health of inmates. The report concludes by proposing avenues for improving epidemiological knowledge regarding the health of inmates in France. (R.A.)