Congrès sous l’égide de la Société Française de Génie Biologique et Médical (SFGBM). ; National audience ; Information Retrieval (IR) in Electronic Health Records (EHRs) should provide healthcare professionals with accurate information to the right person at the right time and place. It should also reduce the difficult tasks of manual information retrieval from records in a paper format or computers. In this paper, a flexible, scalable and multi-layer object-oriented query language is described. It is designed for retrieving and viewing data which support any conceptual schema. The search engine deals with structured and unstructured data. Several types of queries on test databases containing some 200,000 anonymized records from 2,000 patients were tested. These queries were able to accuratly treat with symbolic, textual, numerical and chronological data.