Towards Model-Driven Semantic Interfaces for Electronic Health Records on Multiple Platforms Using Notation3

Electronic Health Record (EHR) systems that aim to achieve data interoperability need to conform to established EHR standards, such as HL7 FHIR or openEHR. Manually developing EHR user interfaces (UIs) for the input of standards-compliant health data is not scalable and unsustainable, since this data no longer only originates from the health practitioner's desktop-given an increased focus on illness self-management, health data is being increasingly generated from patient-focused web or mobile applications. This has led to the collection of non-standards-compliant data from a variety of health apps, i.e., vitals, symptoms, treatments, and so on-this data is not interoperable and requires significant processing to load into an standards-compliant EHR system. A solution for the collection of interoperable health data is the (semi-)automatic generation of UIs, guided by standards-compliant descriptions of the required health data and their constraints. We present a framework to automatically generate UIs for the collection of interoperable health data, supporting different platforms (i.e., UI formats) and EHR standards. The generated UIs perform input validation and submit a self-contained, semantically annotated EHR record. Currently, we support HTML+RDFa (web) and Yail (mobile) as UI formats, and HL7 FHIR and openEHR as EHR standards. We utilize Notation3 (N3) to implement our UI generation pipeline, a Semantic Web language for decision-making in an open Web environment.