Meeting minutes
R6
detlef: We found some missing parts in the fhircat fhir RDF parser
ericp: Tim was working on getting me a dump of the shex and examples, then i was going to run validation and fix what is needed. Make sure they round-trip.
… into HAPI. Then validate HAPI output.
… Tim was going to gen that from an R5 release.
Extending our ont for more semantic interop
ken: COuld be a huge effort. Should validate business cases.
… Want cross-product domain with AI applications.
… I have standard product messages coming across. Some places I've seen in public health have to do w FHIR, CCDA, and US x12 and v2. All are involved in that process. Also E-prescription.
… The other is NLP. When you take unstructured data and you need to structure it to harmonize, having an ont is beneficial.
… Most of this is on a volunteer basis, and if does not align with people's day jobs, it's hard to get cycles.
ericp: Sounds good, but predicated on starting w use cases and queries.
ken: I have a use case abt info transform. But I'm not an SME for query, with inferencing.
ericp: If you can point to "here's what I imagine", we can make sparql query.
ACTION: Ken to write up issue w suggested use case
ken: There can be a mismatch btwn an ont and FHIR data model, and how we navigate those things.
dbooth: this RDF subgroup of ITS is responsible for the RDF format of the FHIR standard.
… The ont is an extra benefit.
… FHIR license is CC0: https://
… And W3C has a clear patent policy that requires free use
ken: Example of ont problem, driven from social services -- housing insecurity in USA.
… It's provided to a household (set of people) who have a set of constraints put on them by fed and local govs.
… We were not allowed to use the concept of a person.
… Also, person relations are important. That's getting some traction, because in FHIR data model only allowed patients, and the only rel they could have was to a related person, and only 1:1. Caused modeling problems.
dbooth: I believe the FHIR RDF group can add the concept of a person to the FHIR ont if we want.
ken: Eric Jahn and I have worked on that. This community group is a joint effort between W3C and HL7.
… If we decompose, there's the RDF level, then there's the FHIR ont. It's independent of the RDF. WE have freedom in the ont.
ericp: The ont w capture what we can get from the structure def. But if you want to add extra connective elements and inference we can.
ericj: HUD is using FHIR RDF, and Temple Univ also, for their DreamKG (NSF-funded) addressing homelessness.
ACTION: Ken to draft an issue of use case for extending the FHIR ont, by Dec 18
Attribution for our work
dbooth: Sebastian did an ack list.
detlef: Do we have some place for ack contributions visibly?
ericp: Suggest we attach messages mentioning product. Eg., we have FHIR R6, and thank the following contributors.
detlef: Need to discuss with our marketing guy. But everytying would be helpful.
ericp: The front page of the CG is writeable by the chairs. Should see what other groups do.
ACTION: DBooth to look into doing that on ouir CG page.
DICOM
erich: I've been on hold. Two months ago I rescanned immaging data commons, and put it into RDF w the extra things we discussed, including CDT. If anyone wants to look it's here: https://
… I'll have more time in Jan.
detlef: Also int in comparing qlever.
erich: I plan to have it loaded by end of Jan.
detlef: Excellent timing.
… Plan to try it with the new DICOM rep.
Holidays
dbooth: We will cancel meetings on Dec 25 and Jan 1.
ADJOURNED