Page tree
Skip to end of metadata
Go to start of metadata

Submitting WG/Project/Implementer Group

W3C/HCLS FHIR RDF Working group


Justification and Objectives

The current FHIR RDF ITS Specification was "hand crafted" and requires the development of FHIR specific software to implement.  The W3C JSON LD working group released a new Committee Draft (CD) of the JSON LD 1.1 specification, which will allow a significant portion of the FHIR RDF specification to be represented as JSON-LD 1.1 contexts. There are still portions of the FHIR specification, however, that cannot be represented using JSON-LD.  This gives us an opportunity to revisit some of the earlier decisions wrt. RDF and to take into account some of the feedback and issues that have been received with respect to usability.  These decisions include:

  • Extensibility – the current RDF specification makes extensibility easy and straightforward, but at the cost of making commonplace RDF/SPARQL operations difficult and non-intuitive.
  • Concept URI's – the RDF specification allows concept URI's but the transformation between the FHIR Coding to the URI's is non-trivial
  • List ordering – the current RDF specification requires the addition of an "index" attribute to maintain list ordering
  • Node root – JSON and XML represent trees, where the "root" is the outermost element.  RDF requires an additional identifier to identify the starting point within an arbitrary directed (potentially cyclic) graph


The objective of this track is twofold:

1) To work the newly created JSON LD 1.1 tooling to transform FHIR JSON representation into RDF and to experiment with the pros and cons of different approaches to solving the issues above.

2) To start with bespoke RDF representations of FHIR  to determine how the JSON-LD Framing Language, ShEx and other tools can be used to transform and validate the RDF as a collection of FHIR resources.


This track will use R5 (latest) version of FHIR.

Proposed Track Lead

Harold Solbrig, solbrig@jhu.edu


Expected participants

TBD – we are interested in involving anyone who expects to use FHIR in an RDF or Linked Data environment as well as other interested parties.


Track Orientation

A webinar was held on Tuesday, December 17 10:00 AM EST to share further participation information about this track.

You can review a recording of this webinar here.

System Roles

Any system that needs to consume or emit FHIR as RDF.  This includes FHIR Servers, DL and NLP modules.

Scenario 1

Goal: Transform FHIR JSON representation into the current FHIR R4 RDF target using proposed FHIR JSON-LD 1.1 contexts.

Material: JSON-LD 1.1 specification, JSON-LD 1.1 FHIR contexts (repository to be supplied) and JSON-LD Playground and JSON-LD 1.1 transformation software.  (As of this writing, the only relatively mature implementation is in Javascript, but others should be forthcoming relatively soon).

Task:

1) Create a transformation process that takes FHIR JSON, the supplied contexts and produces a "minimal" R4 output

2) Develop strategies to extend the input JSON and/or the generated RDF to produce a fully conformant R5 output.

Test scripts:  https://github.com/fhircat/fhir_rdf_validator can be used to compare RDF output to expected and/or run ShEx evaluation


Security and Privacy Considerations

None

  • No labels