Pierre-Antoine Gourraud

Professor Pierre-Antoine Gourraud, Ph.D. is a professor at the Faculty of Medicine of the University of Nantes and former associate professor in the Department of Neurology at the University of California at San Francisco (UCSF, USA). He heads the Data Clinic at CHUN which is in charge of developing the biomedical data warehouse at Nantes University Hospital. He leads a team of 12 trained epidemiologists. The Nantes University Hospital includes 900 doctors, 270 postgraduates, 465 medical interns and 1,000 paramedical students. It is ranked 5th for clinical trials (1200 per year), 4th for national and inter-regional clinical research projects (PHRC), 1st for national economic evaluation of health programmes (PRME) and 9th for scientific publications (1000 per year), and is the 1st national choice of residents. CHUN has a full-fledged biomedical data warehouse operating with an experienced team of epidemiologists and integrates medical and administrative data, cleanses and reformats data, into a secured database infrastructure. It provides access to 2.6 million patients, 130 million structured data and 30 million documents. Data is collected from a number of sources: PMSI (Programs for medicalization of the information systems): diagnosis and procedure codes; Plain-text reports (consultations, hospitalizations, imaging, genetics, biology, surgery, laboratory results); Structured lab work results (numerically coded variables). The CHUN biomedical Data Warehouse is accessible through the query management system “eHOP” developed by LTSI at CHU of Rennes and maintained by Enovacom-OBS. The database is used by only 8 specially-trained Epidemiologists of the “data clinic unit” on behalf of the medical center for both internal and external projects. It currently handles 5-10 new research projects every week (including screening and feasibility queries). This service was among the first approved by the national data protection agency (CNIL) in July 2018. By combining queries on structured and unstructured data, the Data Clinic team is able to accurately identify patients with COVID-19 thanks to a daily update of the CHUN biomedical Data Warehouse to create data marts. In addition to ICD-10 codes, natural language queries can be used to search for false-negatives and exclude false-positives whose reports mention the postponement of care due to COVID-19. Biological records will be used as the gold standard to define COVID-19 (200 tests/day) in order to assess the sensitivity of compound queries.