Ontology-Based Clinical Heterogeneous Data Integration Research

With the widely use of medical information system, many hospitalsand institutions have their own medical information systems, the medicaldata stored are increasing at index level. Integrating and analyzing themedical data of patient can effectively improve the diagnosis andtreatment of disease, the public health service, the basic research andmedical management efficiency and quality. However, the differentformat that each system adopts and the different meanings of the dataexpressed in the system cause the autonomy and heterogeneous ofmedical data, which cannot be solves by using simple mapping. In ourcountry, making full use of existing resources to realize informationcommunication between hospitals, between superior and grassrootsmedical institutions, between hospitals and medical insurance agenciesbased on the establishment of residents electronic health records canprevent the "information island", which has great significance onimproving the medical service and medical management decision level.So, to solve the problem above, ontology of electronic health record isproposed in this paper. From the perspective of individual clinical dataanalysis, a semi-automatic method is proposed to transform relational databases to a clinical ontology, in order to build the residents’ electronichealthcare record.Firstly, classes and attributes in ontology are extracted from theentities and attributes of the relational databases, ontology axioms aretransformed from the entity integrities, triggers and so on. After allthese, initial ontology is completed. Secondly, ontology alignment isproposed through instance comparison, and electronic healthcare recordontology is fully completed. At last, a prototype is implemented to provethe method and the result prove the feasibility of this method.