Improving EHR semantic interoperability: future vision and challenges

We propose a semantic-driven architecture to improve EHR semantic interoperability. This architecture is constituted by five layers: structured heterogeneous data (i), as found in (un-)standardised clinical information models, which are consumed by a semantic mapping layer (ii), which links the data items to clinical ontologies via user-friendly content patterns. A semantic mediator (iii) then translates these content patterns into ontology-based annotations, which populate a virtual homogeneous data store (iv), which serves the application layer (v).