Semantic Challenges in Database Federation: Lessons Learned

In this project an integrated analysis of data from disparate surgery and anaesthesiology departmental information systems was carried out. Due to the lack of shared primary keys, a multi-stage "soft" matching method was implemented. Results of the matching steps are described in detail. Inconsistencies were shown to exist for identifying data, semantic definition of documentation content and documented data itself. Minimum requirements for interdisciplinary documentation in autonomous systems should include shared semantic definitions of documentation content as well as robust and regularly validated interfaces for identifying data.