Assistance Publique-Hôpitaux de Paris (AP-HP) is the biggest teaching hospital in Europe, federating 39 different hospitals and groups of hospitals (22,474 beds) mainly located in Paris and its suburban area. It offers advanced treatments to 8M patients per year, including 1.5M hospitalizations (10% of all hospitalizations in France). It is also the largest employer in the Paris region with 92,000 people working for this public institution representing more than 150 different jobs. With more than 19,000 medical staff, 3000 hospital physicians, including 1000 professors and more than 6000 residents and medical students, AP-HP’s staff cover all the major medical specialities.
As a public health institution, AP-HP pursues 3 main missions: healthcare delivery, medical and paramedical teaching and biomedical and clinical research. AP-HP’ teams develop high quality research activities, including European and International Research projects. In particular, AP-HP is involved and/or co-ordinates more than 80 European research projects supported by the Research and Technological Development Research Programme from the European Union. AP-HP is also the first biomedical research center in Europe with > 500 clinical trials per year and > 8,000 publications per year (25% of all medical research publications in France).
WIND team (UG4303) is implementing the AP-HP Big data platform. This high-performance storage and computing infrastructure hosts the AP-HP Clinical Data Repository (CDR), or “AP-HP cohort,” which integrates clinical data of more than 11 million patients treated at AP-HP. It also offers data integration and processing services for national and international research projects.
The AP-HP Big data platform includes a computing cluster composed of both hadoop distributed processors and GPU technologies to enable the processing of massive data (medical images, high frequency physiological data, textual reports). The open source solution i2b2 associated with Jupyter – a portal dedicated to the analysis of Big data – makes it possible to build secure working environments for each research project where an ecosystem of open source tools for exploitation or visualization is made available. Secured on the material, software and organizational levels, the AP-HP Big data platform was authorized by the CNIL (French Data Protection Authority) on January 19, 2017 and is the subject of a Privacy Impact Assessment for compliance with the European General Data Protection Regulation.