Research And Implementation Of Medical Data Integration Platform Based On Semantics

In today's medical information system, there exist all kinds of heterogeneous systems such as HIS, LIS, CIS and PACS. For the lack of unified standards among manufacturers, the data structures of these systems are significantly different, which brings a big problem to the data sharing.This paper takes project of Tissue Specimen Database which support by Shanghai committee of science as research background. Problems involved in integration of heterogeneous data in the hospital has been analyzed. Besides, a data integration method based on semantics is proposed. First, this method builds the semantic model of heterogeneous data sources based on Metathesaurus and Semantic Network of Unified Medical Language System (UMLS) . It extracts semantic information from the pattern relationships of heterogeneous data sources to establish local ontology and to establish the global ontology to solve the problem of semantics conflicts between the local ontology ,then use the semantic model and the mapping between data sources to integrate heterogeneous data. Based on this, the paper described the design and implementation of medical data integration platform,include platform architecture,the collaborative ontology editor tool,the semantic model library of medical data based on OWL,the inquiry service of clinical information based on SPARQL and so on.The platform is feasible and effective in actual projects.with other medical data integration systems, the paper has the following characteristics is mainly:1) Introduced Metathesaurus and Semantic Network of Unified Medical Language System (UMLS) into the data integration to solve inconsistencies medical terminology in data source.2) Proposed a semantic based medical data integration method.First,this method establish medical terminology constraints of the heterogeneous medical data sources, and extracts semantic information from the pattern relationships of heterogeneous data sources to establish local ontology and to establish the global ontology to solve the problem of semantics conflicts between the local ontology. Use of the mapping between semantic models and data sources to integrate heterogeneours data.3) The medical data integration platform based on the above method has been in practical projects.The platform provides ontology edit tool to establish the semantic model of heterogeneous data sources and SPARQL-based clinical information inquiry services.