Semantic Web Health Care and Life Sciences Interest Group Teleconference

10 Feb 2015

See also: IRC log


David_Booth, Darrell, Marc_Twagirumukiza, Tony_Mallia, EricP, Kerstin_Forsberg, Sajjad_Hussain, Peter_Hendler, Ingeborg, Lloyd_McKenzie, Rob_Hausam, Brian_Pech, Alejandra, Guoqian_Jiang, Claude_Nanjo
David Booth


<trackbot> Date: 10 February 2015

FHIR Ontology Requirements


<inserted> David: Discussing requirement #11.

Who cannot live with these options?

option A: Sajjad

option B: Lloyd, Rob, Tony

option C: Tony

Option D: marc

scribe: sajjad

Option E: (nobody)

Option F: (nobody)

Option Z: Sajjad, Marc, Tony, Rob

Which option would you prefer between E and F?

Option E? (nobody)

<Kerstin> F

Option F? Lloyd, Tony, Marc, Rob, Sajjad, Kerstin, DBooth

Peter: FHIR doesn't say what is Entities, Roles, Participants and Acts. Is there a rule that says you cannot have anything in the RDF that is not in FHIR?

Lloyd: RDF instances cannot have anything extra that is not round-trippable. Need to be able to generate the RDF purely from what's in the XML or JSON.

Peter: But then you don't know the difference between an entity and an act.

<ericP> 41#

<ericP> (where 4 is 'h')

<ericP> you can also do 40# to put your hand down

Peter: Need to be able to round trip, but you can still round trip. In FHIR there's no concept of backbone classes that we're using heavily for data analysis. Should have an upper level of RDF that can be mapped for reasoner use. Then on the way out you can give back what you got from FHIR.
... But without Entities, Roles, etc. you are extremely limited.

Lloyd: Two things we're producing. RDF instances, and RDF/OWL ontology. Instances cannot have extra info in them that is not present in the JSON, to enable lossless round tripping.
... But the ontology could in principle include the RIM mappings where they're relevant, and they could include a formal representation of those mappings to enable inference with RIM, but that would be generated from the profiles not from the instances.
... and part of the OWL ontology -- not the instances. E.g., a BP measurement would say "I am a FHIR observation", and the profile could say "Observataion is a subclass of the RIM observation class, etc"
... Doing that formal RIM mapping is a large amount of work, not in the scope of this work. The mappings we have right now are hand wavy. Good to do, but not part of this effort.

Claude: I agree. Better to do separately.

Peter, does that address your concerns?

Peter: Yes.

Common model for data element management and value set binding


<ericP> 11179 stuff


<Alejandra_> could you please share the link again?

<Alejandra_> I lost connection and missed the link


<Alejandra_> Thanks!

(Eric discusses above slides)

<Marc_Twagirumukiza> excuse us

<Kerstin> Also the EU project SALUS uses ISO11179 - SemamticMDR http://www.srdc.com.tr/projects/salus/blog/?p=181

<Kerstin> PhUSE uses only the lower part of he core four-model, that is not the conceptual top only the representaional to represent CDISC standard data elements

Eric: 11179 does not have a serialization. Need to parse it out, extract the value domains, see if the concepts are in the same conceptual space.
... Concepts are tied back to UMLS Semantic Network.

<Alejandra_> Enterprise Vocabulary Services: http://evs.nci.nih.gov/

Eric: We'll be debugging the things in the CSR, testing.
... Ideally we'll be able to find things in the caDSR that are the same in 11179 and use that to transfer data back and forth.
... We'll be using Shape Expressions (ShEx) -- kind of a higher level way to write sparql. Can also do transformations.

Alejandra: Are you using an ontology for 11179 concepts?

Eric: That's one of the steps. Need to see if it is good enough.

Alejandra: I did work relying on caDSR and created some ontologies to enable queries, but recently came back to this and found an ont for 11179.

Eric: There's a bunch of work I showed, transforming CCDA to FHIR. Enough people to work on that (validation and transformation task force)?

Who is interested in a validation and transformation task force (~1 hr/wk)?

Ingeborg: I am. ONC health IT strategic plan talked about collecting CT data through caDSR.
... Will that do the job?

Eric: My guess is no, but probably it will help.

<Kerstin> Please contact anil@srdc.com.tr working for the SALUS project creating the ontology representing ISO11179. I'm on the advisory board for SALUS and know there are interactions between the folks behind caDSR/EVS and Anil.

Others interested in val/trans TF?

<Kerstin> I'm also interested

Alejandra: I'm interested

Kerstin: interested.

Alejandra: How is this related to the Validation WG?

Eric: SPARQL is verbose. ShEx allows semantic actions to be specified, so that during validation, another form of the data can be emitted.

Alejandra: These things are relevant to the big data to knowledge project Im' working on.

Guoqian: These are initial slides. We may have a common semantic model to enable harmonization between different clinical models. Starting point might be consider how to create a standard representation of ISO 11179 in RDF/OWL.

<Alejandra_> Kerstin, I was reading about the SALUS project

Guoqian: Also a DOD.

<Kerstin> An earlier project between AstraZeneca and IBM to build a ISO11179 based "Clinical Reference Library" http://www.slideshare.net/kerfors/designing-and-launching-the-clinical-reference-library

Tony: Isn't the work going on in ISO about this?
... re 11179. Anyone tracking it?

Guoqian: We're working closely with the people in that group.


<revenge-of-ericP> type type type

ericP, rrsagent seems to be stuck. Not generating minutes.

<ericP> yeah, i see http://www.w3.org/2015/02/10-hcls-irc but not http://www.w3.org/2015/02/10-hcls-minutes

<ericP> maybe try again in an hour and see if you get lucky. then fallback to running the script by hand.


Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.140 (CVS log)
$Date: 2015-02-10 19:38:31 $

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)

Succeeded: s/option:/option B:/
Succeeded: s/CA DSR/caDSR/g
Succeeded: i/Who cannot live with these options?/David: Discussing requirement #11.
No ScribeNick specified.  Guessing ScribeNick: dbooth
Inferring Scribes: dbooth
Default Present: +1.512.680.aaaa, DBooth, Darrell, Marc_Twagirumukiza, Tony, ericP, Kerstin_Forsberg, Sajjad_Hussain, peterhendler, Ingeborg, Lloyd, rhausam, bpech, Alejandra, +1.507.261.aabb, Guoqian, Claude
Present: David_Booth Darrell Marc_Twagirumukiza Tony_Mallia EricP Kerstin_Forsberg Sajjad_Hussain Peter_Hendler Ingeborg Lloyd_McKenzie Rob_Hausam Brian_Pech Alejandra Guoqian_Jiang Claude_Nanjo
Found Date: 10 Feb 2015
Guessing minutes URL: http://www.w3.org/2015/02/10-hcls-minutes.html
People with action items: 

[End of scribe.perl diagnostic output]