Nosocomial Infections (NI) and iatrogenic events (IA) affect 5 to 35% of patients admitted to Intensive Care Units (ICUs) in the United States , costing millions of Euros, affecting the well being of its citizens and increasing morbidity. The aim of the RHEA project is to develop a decision support tool based on a data warehouse architecture. The warehouse is fed by a selective extraction from the local database of each Intensive Care Unit. A biostatistical exploitation of the warehouse enlightens correlation between data and enables specifying prediction tools on the evolving of the patient's gravity, including iatrogenic and nosocomial events, with the view to reducing their incidence thus improving health care. The project includes now a network of 30 hospitals.