Harold R Solbrig1, Eric Prud'hommeaux2, Grahame Grieve3, Lloyd McKenzie4, Joshua C Mandel5, Deepak K Sharma1, Guoqian Jiang6. 1. Department of Health Sciences Research, Mayo Clinic, Rochester, MN, USA. 2. Computer Science and Artificial Intelligence Laboratory, Massachusetts Institutes of Technology, Cambridge, MA, USA; World Wide Web Consortium (W3C), Cambridge, MA, USA. 3. Health Intersections Pty Ltd, Victoria, Australia. 4. Gevity Consulting Inc, Vancouver, BC, Canada. 5. Harvard Medical School Department of Biomedical Informatics, Boston, MA, USA. 6. Department of Health Sciences Research, Mayo Clinic, Rochester, MN, USA. Electronic address: Jiang.Guoqian@mayo.edu.
Abstract
BACKGROUND: HL7 Fast Healthcare Interoperability Resources (FHIR) is an emerging open standard for the exchange of electronic healthcare information. FHIR resources are defined in a specialized modeling language. FHIR instances can currently be represented in either XML or JSON. The FHIR and Semantic Web communities are developing a third FHIR instance representation format in Resource Description Framework (RDF). Shape Expressions (ShEx), a formal RDF data constraint language, is a candidate for describing and validating the FHIR RDF representation. OBJECTIVE: Create a FHIR to ShEx model transformation and assess its ability to describe and validate FHIR RDF data. METHODS: We created the methods and tools that generate the ShEx schemas modeling the FHIR to RDF specification being developed by HL7 ITS/W3C RDF Task Force, and evaluated the applicability of ShEx in the description and validation of FHIR to RDF transformations. RESULTS: The ShEx models contributed significantly to workgroup consensus. Algorithmic transformations from the FHIR model to ShEx schemas and FHIR example data to RDF transformations were incorporated into the FHIR build process. ShEx schemas representing 109 FHIR resources were used to validate 511 FHIR RDF data examples from the Standards for Trial Use (STU 3) Ballot version. We were able to uncover unresolved issues in the FHIR to RDF specification and detect 10 types of errors and root causes in the actual implementation. The FHIR ShEx representations have been included in the official FHIR web pages for the STU 3 Ballot version since September 2016. DISCUSSION: ShEx can be used to define and validate the syntax of a FHIR resource, which is complementary to the use of RDF Schema (RDFS) and Web Ontology Language (OWL) for semantic validation. CONCLUSION: ShEx proved useful for describing a standard model of FHIR RDF data. The combination of a formal model and a succinct format enabled comprehensive review and automated validation.
BACKGROUND:HL7 Fast Healthcare Interoperability Resources (FHIR) is an emerging open standard for the exchange of electronic healthcare information. FHIR resources are defined in a specialized modeling language. FHIR instances can currently be represented in either XML or JSON. The FHIR and Semantic Web communities are developing a third FHIR instance representation format in Resource Description Framework (RDF). Shape Expressions (ShEx), a formal RDF data constraint language, is a candidate for describing and validating the FHIR RDF representation. OBJECTIVE: Create a FHIR to ShEx model transformation and assess its ability to describe and validate FHIR RDF data. METHODS: We created the methods and tools that generate the ShEx schemas modeling the FHIR to RDF specification being developed by HL7 ITS/W3C RDF Task Force, and evaluated the applicability of ShEx in the description and validation of FHIR to RDF transformations. RESULTS: The ShEx models contributed significantly to workgroup consensus. Algorithmic transformations from the FHIR model to ShEx schemas and FHIR example data to RDF transformations were incorporated into the FHIR build process. ShEx schemas representing 109 FHIR resources were used to validate 511 FHIR RDF data examples from the Standards for Trial Use (STU 3) Ballot version. We were able to uncover unresolved issues in the FHIR to RDF specification and detect 10 types of errors and root causes in the actual implementation. The FHIR ShEx representations have been included in the official FHIR web pages for the STU 3 Ballot version since September 2016. DISCUSSION: ShEx can be used to define and validate the syntax of a FHIR resource, which is complementary to the use of RDF Schema (RDFS) and Web Ontology Language (OWL) for semantic validation. CONCLUSION:ShEx proved useful for describing a standard model of FHIR RDF data. The combination of a formal model and a succinct format enabled comprehensive review and automated validation.
Authors: Gil Alterovitz; Jeremy Warner; Peijin Zhang; Yishen Chen; Mollie Ullman-Cullere; David Kreda; Isaac S Kohane Journal: J Am Med Inform Assoc Date: 2015-07-21 Impact factor: 4.497
Authors: Jeremy L Warner; Matthew J Rioth; Kenneth D Mandl; Joshua C Mandel; David A Kreda; Isaac S Kohane; Daniel Carbone; Ross Oreto; Lucy Wang; Shilin Zhu; Heming Yao; Gil Alterovitz Journal: J Am Med Inform Assoc Date: 2016-03-27 Impact factor: 7.942
Authors: Harold R Solbrig; Harold R Solbrig; Eric Prud'hommeaux; David Booth; Cory M Endle; Daniel J Stone; Guoqian Jiang Journal: AMIA Annu Symp Proc Date: 2021-01-25
Authors: Stephany N Duda; Nan Kennedy; Douglas Conway; Alex C Cheng; Viet Nguyen; Teresa Zayas-Cabán; Paul A Harris Journal: J Am Med Inform Assoc Date: 2022-08-16 Impact factor: 7.942
Authors: Eric Prud'hommeaux; Josh Collins; David Booth; Kevin J Peterson; Harold R Solbrig; Guoqian Jiang Journal: J Biomed Inform Date: 2021-03-26 Impact factor: 6.317
Authors: M A Olivero; F J Domínguez-Mayo; C L Parra-Calderón; M J Escalona; A Martínez-García Journal: BMC Med Inform Decis Mak Date: 2020-05-25 Impact factor: 2.796