FAIRness for HL7 FHIR: supporting interoperability of health data sets

Catherine Chronaki1, Giorgio Cangioli2, Alicia Martinez Garcia3, Carlos Calderón3, Philip Damme4
1HL7 Foundation, 2HL7 Europe, 3Andalusian Health Service, 43Amsterdam UMC, University of Amsterdam, Department of Medical Informatics, Amsterdam Public Health Research Institute


Abstract

The FAIR (Finable Accessible Interoperable Reusable) principles are established best practice in the generation of health datasets for open science, research and innovation. HL7 FHIR (Fast Healthcare Interoperability Resources) offer standard APIs (Application Programming Interfaces) to advance interoperability. Research Data Alliance (RDA) is a community that promotes a set of FAIR maturity indicators for datasets. HL7 FAIRness for FHIR (FAIR4FHIR) develops an implementation guide to explain how to deliver FAIR health datasets with HL7 FHIR and how HL7 FHIR supports RDA FAIR Maturity indicators. FAIR4FHIR is connected to the FAIR4Health project which developed a platform to FAIRify health datasets within healthcare institutions and make them accessible to distributed data mining algorithms.

The objective of this work is to assess the FAIRness of health datasets in terms of data and metadata and propose ways to improve their FAIRness using HL7 FHIR.

Using the FAIRness Assessment criteria of RDA we studied health datasets including selected from PhysioNet and associated publications that describe the production, license, and intended use of these datasets. Then, explored how FHIR resources such as library, citation, health study describe the (meta)data of the health dataset. Selected information available in the description of the datasets and related scientific publications were mapped to HL7 FHIR resources. The effort and the cost of providing rich metadata in HL7 FHIR were evaluated.

The studied health data sets fulfil most of the FAIRness criteria. (I)nteroperability and (R)eusability criteria work with accepted community standards. HL7 FHIR resources can improve FAIRness, and particularly interoperability and provenance, as well as simplify combining data sets from different sources.

HL7 FHIR can improve FAIRness and advance interoperability of health datasets. This is important even when only the meta data of the dataset is publicly available and the data is protected in health institutions, allowing only selected queries.