OBJECTIVES: The use of medico-administrative databases (MADB) for carrying out “real life” studies on health products is now widespread. Linkage with data from other sources (e.g., registry, medical records) is a major lever for enriching the MADBs and broadening research perspectives. The role and use of studies on linked databases have already been described for drugs in several therapeutic areas, but are still poorly understood in the evaluation of medical devices (MD). Our objective was therefore to clarify the role and use of these methods in studies that aim at evaluating MDs through a systematic literature review (SLR).

METHODS: A SLR was performed from the PUBMED database according to the Cochrane recommendations without restriction on the years of publication.

RESULTS: Of the 56 identified studies, 18 were included. The studied MDs were mainly implantable (hip prostheses, breast implants and cardiovascular MDs);national or regional; and mostly conducted by academic teams. Studies were mainly comparative, focusing on MD safety. Less than half of linkages involved MADB . The other linkages involved registries (linked to each other or to medical records). The linkage was direct for 11 studies (using a common unique identifier between the linked databases); and indirect for the other linkages (based on common variables).

CONCLUSIONS: Our study only focused on references published on PUBMED, which is perhaps the limit of our SLR. Most of the linkages were made directly, showing the strength of the use of a unique identification number to facilitate data enrichment. These results suggest that the logistical, regulatory, and analytical difficulties make it a solution that is still little used to date, and the lack of standardization of the methods and their communication make their validation and interpretation difficult. This lack of standardization may also explain why these studies are not listed in PUBMED.