Shape Expressions (ShEx) Schemas for HL7 FHIR R5 Specification

Abstract

The Shape Expressions (ShEx), a W3C standard, is a human-readable and machineprocessable language for describing RDF data. ShEx is an effective way to describe how data should be represented in RDF, and to define structural constraints. ShEx can both be used to define and validate RDF data. The Fast Healthcare Information Resources (FHIR) standard, which forms a foundation for healthcare related data interchange, has added ShEx as an official representational form alongside XML and JSON. Our initial efforts of modeling FHIR profiles into ShEx Schemas proved that ShEx is very useful for describing a standard model of FHIR RDF data by translating FHIR resources into ShEx Schemas. After demonstrating the success and importance of the results of initial algorithmictransformations from FHIR DSTU3 model to ShEx schemas, we continued our study and expanded the prototypical implementation to the official HL7 FHIR R5 version We resolved about dozen issues in initial implementation with consensus from semantic web community - the HL7 ITS/W3C RDF Task Force. We extended our implementation to transform the FHIR R5 specification into ShEx schema. We also developed an enhanced validation mechanism (includes constraints,terminology bindings along with structural properties) of FHIR RDF data using ShEx schema language. First FHIR resource examples were transformed into FHIR RDF instance data. These ShEx schemas and FHIR RDF examples are now part of published FHIR R5. The validation mechanism evaluated these ShEx schemas for completeness and their application on instance data. First, we made sure that schemas have valid structure, cardinality, invariants, value domains and term-bindings. Second, we applied these ShEx schemas to validate the official FHIR RDF examples. The enhanced validation tooling proved critical to ascertain their structural and semantic completeness. We demonstrated that ShEx provides nearly complete coverage of all the FHIR resources.