Supporting online health queries by modeling patterns of creation, modification and retrieval of medical knowledge

Abstract

We evaluated properties of knowledge resources that can be used for building new semantically and behaviorally motivated resources of health guidance and clinical decision making by modeling patterns of creation, modification and retrieval of medical knowledge. We evaluated statistical properties of Wikipedia articles of general terminology and medical terminology based on 25 most common diagnosis names emerging in an electronic health records system. We also evaluated statistical properties of general terminology used in everyday life in respect to occurrence and importance to enable adaptive perspectives to medical knowledge. Our experiments exploit a conceptual co-occurrence network that we created based on a set of 93 medical texts about healthcare guidelines provided by The Finnish Medical Society Duodecim containing 57 679 unique conceptual links. We provide supplementing statistics of an extended range of Wikipedia articles and an n-gram analysis about the set of medical texts.