Abstract

Starting from a use case scenario related to the Electronic Medical Record (EMR) information retrieval, this paper proposes several design principles for an ontology-based EMR system, which will allow the physician to perform a semantic search about a specific body structure or zone/space (affected, explored or treated). The research starts from analyzing the information requirements of the family doctors related to the medical history of their patients. By interviewing a group of family doctors, a set of common questions have been identified and modeled as one generic question. The key terms from this generic question then became the building blocks of an ontology, which was further built incrementally and constantly adjusted depending on the query response (design pattern by querying). Elements from other terminologies/ontologies have been reused thus facilitating interoperability with similar constructions. As a result of this research, we tested and proposed for adoption several basic modeling principles for medical ontologies, among them, the particularities of part-whole relations for body regions/spaces and anatomical components, the requirements for propagation of the medical activities from one class to an upper or subsequent class. These modeling principles have a good degree of generality and could be taken into consideration for building any medical ontology.

Keywords

Medical ontology, Semantic search, Electronic Medical Record (EMR), Architecture