W3C

- DRAFT -

Healthcare Schema Vocabulary CG Meeting

28 Aug 2015

Agenda

See also: IRC log

Attendees

Present
Marc_Twagirumukiza, RV_Guha, Dan_Brickley, David_Booth, David_Portnoy, Michel_Dumontier, Richard_Wallis, Sajjad_Hussain
Regrets
Chair
Marc Twagirumukiza
Scribe
dbooth

Contents


Introductions

Marc: Agfa, medical doctor. Belgium. prof of clinical pharm.
... Mostly academic, but affiliated w Agfa. Chairing this W3C community group.

<danbri> Guha intro

Guha: Looked at proposal. Meta comments. Learned a few things in the last few years. Appreciate your trying to clean up healthcare vocab.
... Make sure someone who uses it is signed up. Once you have the app you'll have publishers.
... When you start with a smallish vocab (10-15 terms) and get people using it, then it's easy to expand. Starting with 100 terms hasn't worked.
... Anything more than 10-15 terms will make it hard to convince everyone to use it.
... For explaining it to potential users, much easier if you do it in small pieces.

Dan: Google, schema.org.

<danbri> danbri: I work at Google on schema.org

Richard: Google, helping with schema.org extensions.

<danbri> = rjw in irc

Yan: Hospital in toronto. Also involved in mapping and reasoning.

michel: Prof at Stanford, med. Co-chair for W3C HCLS group. Also metadata work for BD2K work. Looking at extending vocab to accommodate data in HCLS.

dbooth: Involved with HL7 work and Yosemite Project using RDF for healthcare information interoperability. Involved with semantic web technology for several years, and applying it to healthcare for the past 5 years.

<Sajjad_> Sajjad Hussain, Clinical Data Architect, RVHS, Toronto, Canada. Involved in EHR4CR project (Semantic Interoperability), HCLS interest group. Research interests: Terminology Mapping and Reasoning

David Portnoy: US Dept of Health and Human svcs. Use cases driven by legislation: provider directories, plan ins coverage and patient data.

Commitments and Requirements

<Marc_Twagirumukiza> https://www.w3.org/community/schemed/wiki/Commitments_and_requirements

<danbri> regarding "(MUST) The working format for this community is RDFa." … I assume this is just practicalities / pragmatics, i.e. using the RDFa/RDFS format for compatibility with the schema.org site infrastructure

David: Why RDFa?

Dan: schema.org site is powered by RDFa files that encode a flavor of RDFS that fits schema.org . Therefore we'll use RDFa to be compatible with the way schema.org .

David: Excellent, thanks.

Standards of Procedures

<Marc_Twagirumukiza> https://www.w3.org/community/schemed/wiki/Domestic_standards_of_procedures

marc: To add a property or class, you need to raise it by an issue on github, then it needs to have 5 supporters (votes), and then it will be discussed to be added.
... comments on procedures?

dan: It's in the W3C style, but not precisely W3C process. We can be faster and more agile.
... Let's not be overly legalistic and slow.
... Sorry it's taken us so long to connect with you. Guha's approach would be healthy: start small and go from there.

<David_Portnoy> Agreed. We should start with the lightest governance and methodology possible, until / unless it needs to be more complex and formal.

Current version of proposal

github repo: https://github.com/twamarc/ScheMed/tree/master/deploys/sdo-schemed

marc: In 2007 HL7 v3 RIM was made, but it was too big and we failed to implement it.
... Then schema.org came. I think it can play a big role between systems.
... When I looked at existing schema.org vocab, it was quite good, but some predicates were missing, and some not placed right.
... Today i'd like to invite you to look at the proposal in github (linked above) in RDFa files.

<rjw> http://health.sdo-schemedex.appspot.com/Diagnosis

marc: The readable version is linked above.

richard: That's a working prototype based on Marc's proposal. If adopted, this is what it would look like.

<danbri> Q: is the health extension considered appropriate for markup of clinical trials datasets / reports? e.g. per http://www.alltrials.net/find-out-more/all-trials/

dan: re Clinical Trials data, do you consider the schemas adequate for them? Looking for use cases of a consumer of this data.

marc: Not yet optimized for clinical trials.
... I think we need to add extra predicates mainly.

michel: I have parsed the clinicaltrials.gov data into RDF, so I have a pretty good idea of that schema. i could assess whehter this is adequate.

dan: We're not sure what schema.org is good for, what's in the center. I don't yet have the data to advocate for this with yahoo and bing, etc.

marc: We have a couple of use case from the first project.
... Modeled around 16 medical entities
... Missing predicates and items. This model is being used in Lombardi Italy hospital. And 3 hospitals in Dresden. Belgium used by Agfa. Switzerland used by ROCHE.

<David_Portnoy> Do we have a location where we're describing the specific use cases that are driving the schema proposals?

marc: But they're not really public.

dan: Use cases are medical informatics, private sharing, and using schema.org as lightweight ont, but not yet using the public web?

marc: Yes.

michel: We're working w much more public data. Eg, daily med from FDA, or adverse events from FDA.
... Could find a smattering of data sets that we can expose using this vocab.

dan: labeling is the topic of GS1 group focus on nutrition, packaged foods of all kinds. If we could bridge those it would be good.

<Sajjad_> For example, the property affectedBy pointed at http://health.sdo-schemedex.appspot.com/affectedBy, where the Canonical URL: http://schema.org/affectedBy. Shouldn't it be more appropriate to have a base URL with namespace http://health.schema.org/ instead of http://schema.org/ ?

dan: You also have a medical term abortion, but probably don't want to put that on a web page.
... Need to draw attention back to the public use cases.

sajjad: the property affectedBy pointed at http://health.sdo-schemedex.appspot.com/affectedBy, where the Canonical URL: http://schema.org/affectedBy. Shouldn't it be more appropriate to have a base URL with namespace http://health.schema.org/ instead of http://schema.org/ ?

dan: Let's take this to github. Several issues: namespaces; general phrases for specific healthcare uses.
... Previsouly we either used one meaning, and other times we made a specific def for a particular extension. 'Recipe' must be usable by anything. Both a strength and weakness.

<danbri> look at http://schema.org/Recipe

sajjad: affectedBy looks like it was made only for med domain.

<danbri> in Recipe we have effectively pseudo-namespaced various otherwise-general terms, e.g. 'recipeCategory', 'recipeIngredient' etc.

<michel> +q

marc: We have some predicates that are generic. Need to discuss one by one. Please read the list of all predicates and mark the ones you think are too generic to stay in core, and what should be in extensions.

<Marc_Twagirumukiza> in Google docs: https://drive.google.com/file/d/0B3XqWCPGZuTeRkRBelJxOVJEcWc/view

michel: Re utility for the vocab, do you have an association between the predicates and the use cases, so we know which are used with which use cases?

<danbri> any actions arising from all this?

Next Steps

marc: meeting frequency?

dan: Monthly

<scribe> ACTION: Michel to assess suitability of vocab for clinicaltrials.gov data [recorded in http://www.w3.org/2015/08/28-schemed-minutes.html#action01]

richard: What would be the minimum set of terms needed?

marc: Let's discuss that next meeting.

<Marc_Twagirumukiza> Next meeting: 25/sep

dan: Please have one issue in the tracker for each thing that needs to be discussed.

<scribe> ACTION: Dan to contact alltrials people [recorded in http://www.w3.org/2015/08/28-schemed-minutes.html#action02]

<David_Portnoy> Proposing meeting approach´: Monthly meetings on general, high-level and meta concepts only. In between, GitHub issue discussions on specific schemas, vocabs and topics. Participants can schedule smaller calls just with interested participants from Github on these topics in between.

<scribe> ACTION: Marc to put all questions as issues on github [recorded in http://www.w3.org/2015/08/28-schemed-minutes.html#action03]

ADJOURNED

Summary of Action Items

[NEW] ACTION: Dan to contact alltrials people [recorded in http://www.w3.org/2015/08/28-schemed-minutes.html#action02]
[NEW] ACTION: Marc to put all questions as issues on github [recorded in http://www.w3.org/2015/08/28-schemed-minutes.html#action03]
[NEW] ACTION: Michel to assess suitability of vocab for clinicaltrials.gov data [recorded in http://www.w3.org/2015/08/28-schemed-minutes.html#action01]
 
[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.140 (CVS log)
$Date: 2015/08/28 18:32:50 $

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/David:/dbooth:/
Succeeded: i/Meeting:/Topic: Introductions
Succeeded: i/Commitments/Topic: Commitments and Requirements
Succeeded: i/Domestic/Topic: Standards of Procedures
Succeeded: s/Bush/ROCHE/
Succeeded: s/Richard ___/Richard Wallis/
Succeeded: s/Yan ___,//
No ScribeNick specified.  Guessing ScribeNick: dbooth
Inferring Scribes: dbooth
Present: Marc_Twagirumukiza RV_Guha Dan_Brickley David_Booth David_Portnoy Michel_Dumontier Richard_Wallis Sajjad_Hussain
Agenda: https://www.w3.org/community/schemed/wiki/August_28th,_2015_Conference_agenda
Got date from IRC log name: 28 Aug 2015
Guessing minutes URL: http://www.w3.org/2015/08/28-schemed-minutes.html
People with action items: dan marc michel

[End of scribe.perl diagnostic output]