Medical records have been used for a long time with different forms, aims and usages. This heterogeneity is the result of different professions, ways of working and needs. It is but prejudicial to querying and sharing data and documents. Moreover, we consider that the system must be as close as possible to a more classical, non-computerized way of working such as paper-based medical record and should thus manage documents. Medical records are often loosely- or semi-structured documents, impeding easy retrieval. In our approach, a medical record is considered as a set of documents and a set of data. In this paper, we propose a software system useful for extracting data from loosely structured documents coming from different sources and for querying them in a hybrid way. Querying can be done in a navigation space which represents extracted data or entire documents. Two main parts are described: the extraction of data in loosely-structured documents and the navigation in a unified view of documents and data.