Lessons learned from the semantic translation of healthcare data

Healthcare data provides a wealth of information that can be used to study and improve patient outcomes. Electronic Medical Records and other sources of healthcare data are often managed in relational database system and archived using modern data warehousing techniques. Contemporary semantic database technology has many advantages over traditional database systems; however, the utility of the semantic data can be limited if the data is not converted properly from a tabular representation. There are a variety of tools which will naively convert tabular data into a Resource Description Format semantic graph. Without proper guidance from the operator, the tools will generate a semantically weak database which doesn't have the necessary richness for semantic analysis. This paper describes the conversion process for two healthcare databases, with the goal of creating a robust dataset for semantic analysis. The 'lessons learned' from this process are detailed in order to serve as a resource for other biomedical researchers and clinicians interested in generating a useful semantic dataset from their own relational databases.