Health-connect: An ontology-based model-driven information integration framework and its application to integrating clinical databases

Information integration, interoperability and adaptation to evolving data are the key components of clinical information system with high potential impact. Most existing systems accomplish integration and interoperability by employing specific data model standards with static mapping rules that makes system inflexible and inextensible. In response to this drawback, we present Health-Connect system that exploits ontology-based model-driven code generation process to provide integration among heterogeneous databases and adaptation capability for frequent requirements changes. Given an ER model over each source-ontology, our tools are able to generate integrated database schema, dynamically create mapping between data sources using UMLS, translates user queries to access the data sources. In addition, this approach provides domain experts a higher level of operation abstraction. Our positive experience shows ontology-based model-driven code generation is a promising approach for clinical information systems that integrate heterogeneous sources and evolve as the application and data sources changes.