Semantic Web Health Care and Life Sciences Interest Group Teleconference

27 Jan 2015

See also: IRC log


Claude_Nanjo, Kerstin_Forsberg, David_Booth, EricP, Brian_Pech, Paul_Knapp, Rob_Hausam, Lloyd_McKenzie, Tony_Mallia, Ingeborg, Guoqian_Jiang, Peter_Hendler, Darrell_Woelk
David Booth and Paul Knapp


<trackbot> Date: 27 January 2015

<GJiang> my area code is 507


Approval of Minutes


that's for Jan 13

David motions to approve Jan 13 minutes. Seconded by Brian.

APPROVED: Jan 13 Minutes

San Antonio minutes: http://wiki.hl7.org/index.php?title=ITS_RDF_Concall_Minutes_20150119

(wait for next week to approve that)

FHIR Ontology Requirements


#8. User Friendly

General agreement to drop this as a requirement.

#9. Datatype IRIs

Paul: Either have a single datatype or a complex type with multiple discrete components, and need to deal with them in different ways.

Guoqian: Question is whether to represent complex types in OWL.

Lloyd: We'll represent instances in RDF.
... Even simple types in FHIR can have extensions.
... Even primitives.
... FHIR type called boolean in the schema is a complex element that includes pure xsd:boolean plus extensions.

<GJiang> http://hl7.org/implement/standards/FHIR-Develop/datatypes.html

Here's an example: http://www.hl7.org/implement/standards/fhir/json.html

Make that a SHOULD.

#10. Articulate Value

Make this a separate goal of the group.

#11. Enable Inference

Peter: this should be a MUST

Claude: Enable of inference might be relevant to use cases that would be supported by inference.

shorten and make it a MUST.

#12. Common Model

Lloyd: Out of scope

Paul: Would help

David: Make it a COULD?

Guoqian: Not sure FHIR can cover all metadata and all vocabularies.
... Common model would be very helpful.
... We're using ISO11179 to manage that. Should be complementary to this FHIR ontology.

Lloyd: Separate project.

David: Complementary to this FHIR ontology, not a part of this FHIR ontology.

Darrell: Agree.

<Claude> What I meant is that having a FHIR ontology would enable a number of use cases where one could develop other ontologies where inferencing is possible. Depending on how the FHIR ontology is created, it might pose a constraint on inferencing at that layer.

Drop as requirement, but make a as separate project.

#13. Valid Against Schemas and Profiles

Roll that into #1.

#14. RDF Quality

Make this a MUST.

#4. Monotonic with Modifier Exensions

Claude: Might need to represent only the physical structure of FHIR in RDF, and then let other ontologies interpret the modifier extensions.
... E.g., status=refuted. RDF would just say that the statement is there, and then another ont coudl interpret the meaning of that.
... Maybe add a requirement that we capture only the physical representation.

Tony: I've been trying to interpret the type from the expression. The modifier will change the logic of the assignment of types to that instance.
... Refuted would be a type in the tbox. How do you have enough clues to link it to that type?

Claude: Example would be status, refuted or entered-in-error.
... Can also say proposed, planned, etc.

Agreement on MUST

Guoqian: Suggestion: are we going to cover versioning of the FHIR ontology? FHIR model is updated frequently? Will we produce a single version of FHIR ontology, or track the FHIR models?

Paul: Given you've been handed an instance, how would you know which version to model into?

ISSUE: How to handle FHIR versioning? Should FHIR ontology track the FHIR versions? How do we know which version the instance data corresponds to?

<trackbot> Created ISSUE-2 - How to handle fhir versioning? should fhir ontology track the fhir versions? how do we know which version the instance data corresponds to?. Please complete additional details at <http://www.w3.org/2014/HCLS/track/issues/2/edit>.

Paul: Need to be able to express different versions of the RDF model.

David: If the FHIR instances themselves do not distinguish versions, then I don't see how the RDF can.

<Claude> I need to drop off but should we have a requirement that the ontology be auto-generatable from the FHIR specification?

Paul: I've suggested requiring that instance data indicate its intended profiles.

#5. Vocabulary Bindings

Leave as is


Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.140 (CVS log)
$Date: 2015-01-27 18:12:04 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.140  of Date: 2014-11-06 18:16:30  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

No ScribeNick specified.  Guessing ScribeNick: dbooth
Inferring Scribes: dbooth
Default Present: +1.678.999.aaaa, Kerstin_Forsberg, DBooth, ericP, +1.604.250.aabb, bpech, pknapp, +1.801.949.aacc, rhausam, Lloyd, Tony, Ingeborg, +1.510.418.aadd, PeterHendler, +1.512.680.aaee, Darrell, [IPcaller], +
Present: Claude_Nanjo Kerstin_Forsberg David_Booth EricP Brian_Pech Paul_Knapp Rob_Hausam Lloyd_McKenzie Tony_Mallia Ingeborg Guoqian_Jiang Peter_Hendler Darrell_Woelk
Found Date: 27 Jan 2015
Guessing minutes URL: http://www.w3.org/2015/01/27-hcls-minutes.html
People with action items: 

[End of scribe.perl diagnostic output]