HL7 ITS RDF Subgroup / W3C HCLS -- 18 Nov 2014

<Marc_Twagirumukiza> Hi every one this is Marc Twagirumukiza from Agfa Healthcare

Minutes

<dbooth> http://wiki.hl7.org/index.php?title=ITS_RDF_ConCall_Agenda#Agenda

<dbooth> ITS page: http://wiki.hl7.org/index.php?title=ITS_WG

<dbooth> RDF subgroup: http://wiki.hl7.org/index.php?title=ITS_RDF_ConCall_Agenda#Agenda

<dbooth> Meeting minutes: http://www.w3.org/2014/11/04-hcls-minutes

previous minutes

<scribe> scribenick: ericP

pknapp: PROPOSED to accept http://www.w3.org/2014/11/04-hcls-minutes as a record of the last meeting

Logistics

APPROVED: accept http://www.w3.org/2014/11/04-hcls-minutes as a record of the last meeting

(unanimous)

dbooth: two wiki choices or this HL7/W3C joint work
... which shall we use for administative stuff?
... other work, e.g. FHIR ontology, can be on another wiki

<Marc_Twagirumukiza> I support HL7 wiki as well

dbooth: but for the admin stuff i propose that HL7 host the adminstrative wiki

PROPOSED use the HL7 wiki as the administrative wiki

APPROVED: use the HL7 wiki as the administrative wiki

(unanimous)

-> https://www.w3.org/wiki/HCLS/ClinicalObservationsInteroperability COI Task Force

ericP

Followup on last week's discussion with Mark Musen

ericP: note the tech topics, e.g. FDA therapeutic areas, terminfo, etc.

<dbooth> The recording: https://www.dropbox.com/sh/d7a9bb6ppqgo9xn/AAApWheLsxRSyEKFa0Kr4qNna?dl=0

dbooth: any followup on Mark's presentation?

<dbooth> Rough notes: http://www.w3.org/2014/11/11-hcls-minutes

Tony: following up with MarkM on the details hidden by the ICAT interface.

<dbooth> ACTION: Tony to find out more details about how iCat handles ICD-11 ont and report back [recorded in http://www.w3.org/2014/11/18-hcls-minutes.html#action01]

<Zakim> ericP, you wanted to point out the snomed examples in the FDA TA work

-> https://www.w3.org/wiki/File:Snomedct_extract_descriptive.ttl SNOMED example

<dbooth> Eric: If you're digging into this, here's an example of SNOMED that I took, and found all of the examples of everything that they said about it -- level in the cell, test results, in range, out of range. All the entities that SNOMED teases apart. It would be great to take the same slice of ICD-11

<daniel> There is a new way of representing lab in SNOMED CT as of recently

<dbooth> ... Not a legit SNOMED extraction though, because I changed the nubmers to human readable URLs.

Tony: i've got the 110M SNOMED snapshot from earlier this year.

<dbooth> David: Might be a good use case!

ericP: you shouldn't have to dig back into the 110M 'cause i already did that

Tony: still waiting for access details from MarcM
... I've been looking at the bits that i've already dug into, e.g. allergies to bee stings

<daniel> Here's an alpha version of new lab content

<Marc_Twagirumukiza> ...just to disambiguate it's Mark

<daniel> Daniel Karlsson

<daniel> https://csfe.aceworkspace.net/sf/go/doc10215?nav=1

daniel: there's a new style for SNOMED

Tony: is there a style guide you can point to?
... once we dig down into the details of the ontologies, you're sucked in for 2 years

<daniel> daniel.karlsson@liu.se

Proposed work items

dbooth: we have lots of potential work items and limited resources
... we can persue multiple things but it will depend on the individual champions.

FHIR ontology

<Marc_Twagirumukiza> I would volunteer in the FHIR ontology, --- just as related work to what we did for SALUS project

<Marc_Twagirumukiza> http://www.srdc.com.tr/projects/salus/blog/?p=342

dbooth: claude nanjo and ericP have worked on this. we should schedule reviews

Tony: we discussed this in the HCLS COI call
... the issues were what models for what use cases and how to build the ontology

dbooth: on one end of the spectrum there's "what ontology facilitates machine transformation"
... on the other, "what ontology facilitates humen understanding"

Tony: there's also which supports SemWeb inference
... we may want to bring these together or maybe not

Marc_Twagirumukiza: reusable graphs need to be human-readable, machine-readable, etc.
... we did this for SALUS, "entities" were RDF graphs at the domain level. we could map it to other data.
... so hoping to contribute to this task

<dbooth> Eric: I can present what Josh and I did.

dbooth: so you're stressing the convergence

<dbooth> Claude's work attempted to be more human friendly.

ericP: i can present the stuff that Josh and I did

<Marc_Twagirumukiza> Actually what can help there is not to 'just' ontologyze FHIR but to build up a partitioned reusable domain RDF graphs from the complexe information model

<Marc_Twagirumukiza> those will be a key for converging and mapping as well

dbooth: claude did an ontology as well

Tony: i can present what i did at VA

<dbooth> ACTION: Tony and Eric to work with each other and Claude to prepare comparison of FHIR ontology approaches [recorded in http://www.w3.org/2014/11/18-hcls-minutes.html#action02]

<dbooth> My view is that the *most* important is an ontology that works well for machine processing -- instance data translation, inference, etc. That needs to be the basis. Then making it more human friendly -- or conceptually understandable -- would be nice to have.

dbooth: priority should be machine processing

Tony: agree. "reasoner-friendly"
... something we can use with existing tools

ICD-11 and SNOMED

dbooth: we discussed this earlier.

Bridging VA, Intermountain and other models

Tony: depends on what MarkM can give me. i have no access to the bits.

<dbooth> David: Not a current work project. Create a wiki page for project ideas backlog?

dbooth: maybe we should create a backlog page

Tony: rep from VA would be rafael (not on this call)

Marc_Twagirumukiza: sometimes the backlog becomes a brainstorming place with no followup.

<dbooth> ACTION: David to create a wiki page for backlog [recorded in http://www.w3.org/2014/11/18-hcls-minutes.html#action03]

Marc_Twagirumukiza: but very necessary, also needs an issue tracker

Issue tracking

<dbooth> Eric: W3C Tracker is integrated with IRC

ericP: we can start with tracker, which will work with IRC and public-semweb-lifesci

Use Cases

<dbooth> ACTION: ericP to send pointers to existing wiki pages [recorded in http://www.w3.org/2014/11/04-hcls-minutes.html#action01] [DONE]

<dbooth> (Eric already pointed to the W3C COI page)

dbooth: what use cases with what goals?

<dbooth> Eric: Migration of CCDA to FHIR

<dbooth> Eric: Utility on FHIR RDF that would motivate mapping of FHIR XML to FHIR RDF

dbooth: need different levels of use cases
... need one simple enough for a slide, e.g. BP
... it would be nice if they could be used across different purposes.

PhUSE-FDA project (formerly CDISC2RDF)

Kerstin_Forsberg: know the CDISC2RDF. can work on this.

Ingeborg: i know SDTM quite well

<Kerstin_Forsberg> kerstin.l.forsberg@astrazeneca.com

<Ingeborg> ingeborgholt@gmail.com

<dbooth> ACTION: Kerstin and Ingeborg to prepare a status and future state ideas for PhUSE-FDA work [recorded in http://www.w3.org/2014/11/18-hcls-minutes.html#action05]

<dbooth> Kerstin and Ingeborg have volunteered to champion the PhUSE-FDA work.

ericP: our use cases for FDA-TA were to:
... .. make different studies in the same therapeutic area compatible for safety and efficacy analysis
... .. make e.g. drugs in differnt TAs comparable for safety signals

C-CDA RDF representations

<dbooth> ACTION: Eric to establish/make a wiki page for C-CDA RDF representations work [recorded in http://www.w3.org/2014/11/18-hcls-minutes.html#action06]

dbooth: ericP is champion

Joshua_Phillips: i'd like to work on this.

<dbooth> ACTION: Eric and Joshua to report on C-CDA RDF representations work plan [recorded in http://www.w3.org/2014/11/18-hcls-minutes.html#action07]

High-level concept mapping to RDF (AR typeCodes, etc.)

Tony: this is kind of light-weight.
... "how do you represent the terminology?"
... from the top down, identify topics, e.g. transport mechanisms

<Kerstin_Forsberg> FDA/PhUSE Semantic technology project's first delivery was "just" to represenet existing CDISC foundational standards in RDF. Not representing the actual clinical trial data (e.g. blood pressure). Hwoever, they are moving beyond that and I know there was a presentation from FRDA at CDISC Interchannge conference last week on representing CDASH (raw data such as blood pressure) in RDF.

Tony: it's a style guide. a set of patterns.

-> https://www.w3.org/wiki/HCLS/ClinicalObservationsInteroperability/TerminologyExpressivity high-level terminology approaches

<Kerstin_Forsberg> There is also a FDA/PhUSE subproject to reperesent analysis results data, e.g. avg. blood pressure for propulation reported in a study report

dbooth: if there's already a wiki, use it.

<dbooth> David: I would caution that there can be so many different views on high level models, adoption/use would have to be voluntary. I would worry about being perceived as telling people how that they MUST do things. But best practices recommendations sound great.

-> https://www.w3.org/wiki/HCLS/ClinicalObservationsInteroperability/FDATherapeuticAreaOntologies#Coding_specificity terminology coding example

Tony: i'll list the topics. some may be black holes

RobHausam_: i'm interested.

<dbooth> Rob Hausam

<dbooth> ACTION: Tony and Rob to report their plan on High-level concept mapping to RDF work [recorded in http://www.w3.org/2014/11/18-hcls-minutes.html#action08]

Other work proposed work items

RobHausam_: want to look at the interface between terminology and infor model

<dbooth> Rob: Propose Term info work

<dbooth> Eric: Charlie Mead and I created a wiki page on that

<daniel> I'm in

<dbooth> ACTION: Tony and all to decide on a wiki for Term Info work [recorded in http://www.w3.org/2014/11/18-hcls-minutes.html#action09]

<dbooth> ADJOURNED

- DRAFT -

HL7 ITS RDF Subgroup / W3C HCLS

18 Nov 2014

Attendees

Contents