Value Set Expansion Minutes

Meetings occur every other week - Tuesdays @ 2pm ET for 60 min. - alternating with VBS call

Telephone uses standard HL7 Vocab number: 1 (770) 657-9270,,,598745# - NOT THE SHARED SCREEN NUMBER FOR MIKOGO OR OTHERS

Meeting will be one of the following web share environments

Normal web share sessions:

Now using Free Conference Call for web share and Audio

Participation Information

Phone Number: +1 605-472-5280
Participant Passcode: 495389
Web Meeting Info

We are now using the Free Conference Call Webinar system. Please join using the following
Online Meeting Link: https://join.freeconferencecall.com/vocab
Online Meeting ID: vocabOnline Meeting Link: https://join.freeconferencecall.com/vocab
Online Meeting ID: vocab

2017-10-24

join.me web share and HL7 call in
Chair - Ted Klein
Attendees - Carol Macumber, Carmela Couderc, Susan Barber, Yanyan Hu
Call commenced at 2:05PM EDT.

Agenda:

1. Continue discussion on general approach to segmenting the binding operations and expansion operation/function.
   1. We examined Frank's diagrams for Derivation and Profiling.
      1. We went round and round on this. Some things are made simpler and other harder if expanding (process) generates everything known about each concept in the expansion. Profiling becomes challenging if only a subset of all the information known about the concepts is included in an expansion, and every expansion must be profiled (process) for use.
      2. We need to think about this further, and get some concrete Use Cases documented so we have a better handle on it.
2. Review definition of what an implementable value set is (a deterministic expansion?)
   1. An "implementable value set" is either an identified Value Set Expansion, or an expansion that can be deterministically recreated from the available arguments (the VSD CLD, the Binding information, the Profile information).
3. Explore notion of what should be in the expanding process vs. what should be in the profiling process, i.e. should the expansion deal only with the codes, and the profiling deal with only the metadata?
   1. It would be nice and clean, but may lead to problems, if the expanding process yields only a set of concepts each one of which has an identifying concept representation in the expansion. All metadata then would be dealt with by the profiling process. This would seem problematical for many folks, who at a minimum feel that a display name should be included in any expansion - but this quickly leads to questions and issues (e.g. code systems with many display names, post-coordinated expressions in the expansion, etc.).
   2. The general sense is that it is appropriate for profiling to deal with tacking on metadata for each concept, and also that it is OK for profiling to include a subset of the concepts in the expansion.
   3. There is general agreement that profiling should be prohibited from resulting in additional concepts in an expansion from the set generated by processing a VSD CLD.
   4. Carol provided a nice argument for why profiling reducing the number of concepts (e.g. when used for concept maps) is also a bad idea.
   5. Can we discuss on the next call if we can assert a rule that profiling deals only with metadata and perhaps a very select explicit few subsetting operations (e.g. ActiveOnly) rather than any subsetting.
   6. Good discussion, but we are out of time. Continue this discussion on the next call.
4. Call was adjourned at 3:04PM EDT. Next call scheduled in two weeks on November 7, Rob McClure to chair (Ted will be traveling and unable to attend).

2017-10-10

join.me web share and HL7 call in
Chair - Ted Klein
Attendees - Frank Oemig, Carmela Courderc, Carol Macumber, Susan Barber
Call commenced at 2:06PM EDT.

Agenda:

1. Continue discussion on general approach to segmenting the binding operations and expansion operation/function.
2. Review definition of what an implementable value set is (a deterministic expansion?)
3. Review thinking on extension notion, i.e. additional codes not in bound value set vs. permission of uncoded data. Note that most HL7 coded datatypes permit both or either code and text as a value.
   1. Frank noted that extensions as permitted by our current thinking about Binding are not technically part of the value set, and probably have no place for discussion relative to Value Set Expansion control. These are simply different notions.
4. Review current state of expansion process diagrams, and make some notes on the profiling vs. expanding operations.
   1. Frank agrees that we need to keep the notions of expanding and profiling separate.
5. Frank would like to lay out his notion of what a value set actually is. Four parts:
   1. First part is a definition
   2. Second part is the control parameters to bind this VSD to a data element.
   3. Third part is constraining of the value set; there are different ways: you can constrain the definition or can constrain the expansion.
   4. Fourth part is usage, how do I deal with the codes related to this.
6. Ted laid out the issue of all of the other material besides just a single primary code and a preferred display string in the value set.
   1. Frank opined that if a code system has more than one identifying concept representation (code) for each concept, this should be specified in the VSD. We can illustrate with with the ISO3166-1 country codes where someone wants the 2-character codes vs. three character codes.
   2. Carmela asked for clarification if we wanted 2 expansions that differed only by for instance the LOINC long common name vs. the LOINC short common names would this be two different definitions?
   3. Frank says that the VSD should have another control information set which will specify what metadata from the code system should go into the unprofiled VSE; this is so that when you profile, you do not need access to the code system and profiling becomes only a constraining/subsetting operation. Ted disagrees.
   4. Ted is concerned that if we make everything dependent upon total granularity against any Use Case, then we may get a proliferation where for every data element in every IG in every jurisdiction for every different business purpose you need a different value set. Ted sees this as imposing intractable difficulty on the community. We are now out of time.
7. Call was adjourned at 3:02PM EDT. Note that Ted's call dropped almost at the top of the hour. Next call scheduled in two weeks on October 24, Ted to chair.

2017-09-26

Mikogo web share and HL7 call in
Chair - Rob McClure, Rob Hausam, Rob Snelick, Carmela Couderc, Yanyan Hu (TJC)

Agenda:

1. Continue discussion on general approach to segmenting the binding operations and expansion operation/function.
   1. Discussion with RS on changing the "Extensabilities" section to be called "Tolerance" because this essentially describes the allowed tolerance for codes not in the expansion. Changes made in the Tolerance section of VBS.
   2. Clarified that this is distinct from "strength of the binding" that would allow changes to a future value set version or even a different value set that is still within the original scope.
2. RS noted that control over sending an "extension code" as allowed by a "tolerant" binding should be different than allowing use of text alone with no code/code system. He thinks this should not be managed using tolerance, but instead should be managed using the data type. Some may feel this would make things too complicated (thinking FHIR) so this should be discussed in next call.

2017-09-05

Chair - Ted K so JOIN.ME will be used for web share
Attendees:  Ted Klein, Rob Snelick, Carmela Couderc, Susan Barber, Rebecca Baker, Rob McClure
Note that the last call was cancelled due to lack of quorum, continuing with the same planned agenda
Call commenced at 2:06PM EDT, quorum reached
Agenda:

1. Continue discussion on Expansion process as illustrated by the Current proposed process and Frank's new slides.
   1. This is a continuation of the discussion last week on the Binding call
   2. From the binding call last week:
      1. We noted that it would cleaner if we could separate Binding (controlling the set of concepts) and Profiling (controlling the metadata for each concept, and which subset is used for a particular implementation instance). Both of these control the content of the Expansion.
      2. We looked at our definitions again. It would probably be good that in this VSE project we explicitly document our understanding of the etiology of the Expansion and the stakeholders who produce/publish/curate any persistent expansions
      3. continue to explore having two objects: a Binding resource, and a Profiling resource (in FHIR-speak). This might involve removing some of the things that are specified in Binding, and certainly removing some of the things currently specified in the FHIR Expansion Profile
   3. We had a wide-ranging discussion on binding and profiling, we examined Frank's slides in further detail.
   4. There seems to be support with the notion of focusing on operations rather than the individual data items in the artifact that controls the binding ("binding parameter specification"?) and the profile (FHIR expansion profile elements). But there may be some issues with applying these operations anywhere.
   5. We need to examine what we have called 'binding strength' and what these operations are, and what the defaults may be.
   6. What approach should we take in order to cover all these pieces and line up FHRI, V3, and V2?
2. Next call plans:
   1. Continue discussions in this area on next week's Binding Semantics call.
3. Call ran over time, and was adjourned at 3:06PM EDT. Next call will be at the same time on September 5, 2017 and will be the final VSE call this cycle. We will continue the discussion in San Diego at the WGM next week.

2017-08-22

Chair - Ted K so JOIN.ME will be used for web share
Attendees:  Ted Klein, Rob McClure
call was cancelled at 2:27PM EDT with quorum unattainable
Next call we will resume with our work
Agenda:

1. Continue discussion on Expansion process as illustrated by the Current proposed process and Frank's new slides.
   1. This is a continuation of the discussion last week on the Binding call
   2. From the binding call last week:
      1. We noted that it would cleaner if we could separate Binding (controlling the set of concepts) and Profiling (controlling the metadata for each concept, and which subset is used for a particular implementation instance). Both of these control the content of the Expansion.
      2. continue to explore having two objects: a Binding resource, and a Profiling resource (in FHIR-speak).

1. Next call plans:
   1. Continue discussions in this area on next week's Binding Semantics call.
2. Call adjourned at 3:00PM EDT. Next call will be at the same time on September 5, 2017 and will be the final VSE call this cycle.

2017-08-08

Chair - Ted K so JOIN.ME will be used for web share
Attendees:  Ted Klein, Rob McClure, Rob Hausam, Richard Esmond, Carmela Couderc, Susan Barber
call commenced at 2:05PM EDT with quorum reached
Agenda:

1. Continue discussion on Expansion process as illustrated by the Current proposed process and Frank's new slides.
   1. This is a continuation of the discussion last week on the Binding call
   2. We will continue the discussion on the clear boundaries between the effects on the expansion of Binding vs. Expansion Profile
   3. We discussed the overlaps and conflicts between the FHIR Expansion Profile and the current definition of the Binding.
   4. Possible new approach: unify into two things
      1. Binding is all in one place and controls the collection of concepts in an expansion. A Binding may be done at an IG and then a derivative binding can be done 'downstream' where we currently think of doing Profiling. Such derivative bindings would follow the rules we are putting together in the Binding Semantics.
      2. Profiling does not affect the concept collection but only the metadata for each concept in the expansion, e.g. language of the designations, etc.
2. Need to also consider interaction with Sharing Value Set (SVS-IHE) specification SVS wiki page and SVS Spec
   1. We did not look at this.
3. Next call plans:
   1. As we ran out of time on today's call, we need to continue and refine/explore the discussions on unifying the notions.
   2. We should explore having two objects: a Binding resource, and a Profiling resource (in FHIR-speak).
   3. Continue discussions in this area on next week's Binding Semantics call.
4. Call adjourned at 3:00PM EDT. Next call will be at the same time on August 22, 2017

2017-07-25

Chair - Ted K so JOIN.ME will be used for web share
Attendees:  Ted K, Richard Esmond, Carmela Couderc, Jeff Danford, Susan Barber
call commenced at 2:05PM EDT with quorum reached

Agenda:

1. Continue discussion on Expansion process as illustrated by the Current proposed process and Frank's new slides.
   1. This is a continuation of the discussion last week on the Binding call
   2. Frank distributed updated slides to the list this morning
2. We had a very good discussion on the diagram and looked at both Frank's slides and the process and the profile. The following items are our notes from the discussion.
3. We updated the VSE generation slides as well with further thoughts on the intermediary object in the expanding and profiling process.
4. Need to also consider interaction with Sharing Value Set (SVS-IHE) specification SVS wiki page and SVS Spec
   1. We did not look at this.
5. Next call plans:
   1. Continue discussions.
6. Call adjourned at 3:00PM EDT. Next call will be at the same time on August 8, 2017

2017-07-11

This meeting conflicted with Vocb use of the call in for harmonization. Also most regular attendees were on other calls so quorum was not met. Meeting ended at 15 min after hour.
Next call is scheduled for July 25 but May need Ted to run the call as Rob will be on travel.

2017-06-27

Chair - Ted K so JOIN.ME will be used for web share
Attendees:  Lisa Anderson, Russ Hamm, Susan Barber, Carmela Couderc
call commenced at 2:05PM EDT with quorum reached

Agenda:

1. Continue discussion on Expansion process as illustrated by the Current proposed process and Frank's new slides.
2. We had a very good discussion on the diagram and looked at both Frank's slides and the process and the profile. The following items are our notes from the discussion.
   1. We reviewed the binding diagram. Carmela suggested that we put the text in the diagram "The connections represent the inputs to the expand process from each of the identified vocabulary objects."
   2. The expansion profile can specify a fixed code system version; if this differs from one specified in the VSD or the Binding, what are precedence rules?
   3. It would simplify things if we could agree to make the statement: "The code systems library with the VSD and the Binding parameters will expand to a superset of concepts for the VSE. The Expansion profile is intended to 1) possibly constrain this superset to a proper subset of concepts; 2) define the specific metadata to be included in the VSE for each concept."
   4. It would help if we make explicit names or labels for the four parts of the binding (because only the part that determines the expansion is used by the expand process in our diagram).
   5. Input to the Binding project: the first part is for the definition and production of a VSE, whereas the other three parts are for the use of the VSE in an implementation (both current use, and future derived uses).
   6. There seem to be numerous elements in the FHIR Expansion Profile which may be in conflict with or certainly duplicative of, items in the VSD and the Binding parameters. Even within the profile itself, there appear to be many ways of constructing mutually exclusive controls that will yield a null VSE, or null metadata for each of the concepts. Is this what is desired? We (vocabulary and FHIR-I) need to go over this is more detail to see how all these moving parts will operate together.
   7. It almost appears that the FHIR Expansion Profile was built because we had no binding syntax, and there exists no means to specify the data missing from an incomplete 'intensional' VSD in order to profile a published and curated VSE in the absence of it being explicitly bound to a data element. Is this a requirement? Do we need to support this Use Case?
   8. It appears that the marketplace wants to have persisted and curated expansions derived from only the VSD and code system libraries, and are free of both Binding statements and Profiles. We need to explore this use case as it has direct bearing on expansion and how they are produced and how the expand process is parameterized.
3. Perhaps look at FHIR Expansion Profile
   1. We did examine it and some of the comments above surfaced.
4. Need to also consider interaction with Sharing Value Set (SVS-IHE) specification SVS wiki page and SVS Spec
   1. We did not look at this.
5. Next call plans:
   1. Everyone please review Frank's slides between now and the next call, as we will review them in some detail on the next call.
   2. Continue discussion on these topics.
6. Call adjourned at 3:00PM EDT. Next call will be at the same time on July 11,2017

2017-06-13

Chair: Carmela Couderc (see GTM above), Attendees:  Richard Esmond, quorum not reached

Agenda

1. Announcement: From the TSC report this week: The Project Approval Request by the Vocabulary WG of the FTSD for Specification for Value Set Expansion at Project Insight 1325 and TSC Tracker 13434 was approved. Expansions are the implementable artefacts from a value set definition. The VSD STU does not define the details of the content of a value set expansion, nor the standard layout of that content. This project will define this and any additional artifacts necessary, and produce an STU for Value Set Expansions which is intended to harmonize the diverse approaches the existing product lines have taken to define the content and layout of implementable value set expansions. It will also address the migration strategy for the product lines.
2. Reviewed the latest version of the expansion process diagram
3. Briefly discussed the relationship between binding, VSD and VSE
   1. Confirmed that a binding is always to a VSD, however conformance is tested against VSE - more on that when Frank's slide deck is discussed
4. Adjourned @2:17 pm Eastern
5. Planned Agenda:
   1. Continue discussion on Expansion process as illustrated by the Current proposed process and Frank's new slides
   2. Perhaps look at FHIR Expansion Profile
   3. Need to also consider interaction with Sharing Value Set (SVS-IHE) specification SVS wiki page and SVS Spec

2017-05-30

Chair: Rob M., Attendees: Ted K, Richard Esmond, Frank O, Carol M, Susan B., Lisa Anderson, Carmela Couderc

Agenda:

1. Confirm Approval of PSS
   1. Not confirmed - emails sent
2. Review updated ppt and look at FHIR expansion profile resource.
   1. RM noted that if "profiling" can change the code system version (as FHIR allows) this makes an "unprofiled expansion" impractical because it would have to have all code systems.
   2. Sequencing of expansion steps is important because a code system version needs to be defined to make an expansion that can then be constrained or enhanced.
   3. Group started to like a modification of the first ppt diagram better: Current proposed process. The "Sequencing" must be clarified - code system version first at least.
3. Next Agenda
   1. RM will not be available - Will confirm that Ted can run call.
   2. continue discussion on process. Perhaps look at FHIR Expansion Profile

2017-05-16

First post-WGM session. Decided to meet week after WGM so attendance may be low
Chair: Rob M., Attendees: Rob H - Quorum NOT met. Call ended after 30min.

Agenda

1. Further discussion on Expand Process ppt to further clarify steps in expansion and should there be a "un-profiled" expansion as a distinct artifact.
   1. Rob and Rob discussed clarification in what the "un-profiled" would contain. To be reviewed by group.
2. Next meeting
   1. Review updated ppt and look at FHIR expansion profile resource.

2017-05-02

Chair - Rob McClure. Attendees: Ted, Rob Snelick, Frank Oemig, Susan Barber, Carmela, Rob McClure, Lisa Anderson (joint commission)
We are using Rob's Mikogo shared screen for today's call
Call began 14:05 EDT
Agenda

1. Status
2. Confirm approval state and draft state of PSS
   1. Added Conformance as co-sponsor
   2. Will confirm no re-votes necessary
3. Madrid Agenda
   1. Monday Q3 time
   2. Timing for future bi-weekly meetings: Need to consider co-sponsors
4. Next Items
   1. Review PowerPoint diagram in DropBox
   2. Work on explicit list of fundamentals
5. Call adjourned at xxx EDT. Next call will be setup in the Madrid planning session next week Thursday Q4.

2017-04-18

Call was cancelled. Next call May 2.

2017-04-04

Chair - Ted, Rob Snelick, Frank Oemig, Susan Barber, Carmela, Rob McClure, Lisa Anderson (joint commission), Rob Hausam
We are using Ted's join.me shared screen for today's call
Call began 14:05 EDT
Agenda

1. Status
2. Confirm PSS Approvals
   1. Ted reviewed the PSS and where we are, and the CIMI comments relative to the fact that some Normative standards are Logical Models, not just implementation models
   2. PSS with all approvals was forwarded to the FHIR-I chairs today with a request to approve
3. Next items to be discussed
   1. Agree on Fundamental Principles
      1. A VSE (Value Set Expansion) is, at its most basic, a list of concepts from one or more code systems, which are signified in some known way;
      2. In addition, a VSE might contain additional information associated with the contained concepts.
      3. An instance of a VSE (Value Set Expansion) is an end-product of the expansion process, whose input consists of Code System(s) content, Value Set Definition(s), Binding Information, and Expansion Profile(s).
   2. Outline Approach
      1. Out of time
4. Actions
   1. Ted will put the powerpoint diagrams he was working on during the call in DropBox.
   2. Ted will cancel the call in the HL7 call calendar
5. Next call
   1. Continue until we agree on our fundamentals
   2. Next call is scheduled for Tuesday April 18, but neither Ted nor Rob McClue will be available. We will have to cancel this call and resume on May 2.
6. Call adjourned at 15:04 EDT.

2017-03-21

Chair - Rob, Susan, Carmela, Richard, Ted
Call 1402 EDT Agenda

1. Finalize PSS approvals
   1. FHIR (LM) has raised concerns about making a normative track spec that does not fully define implementable constructs for impacted product lines. He thinks VSE should be Informative. GG noted that this could not be aligned with CTS2 and this needs to be clarified.
   2. TK notes this is an infrastructure specification as normative out of alignment product family specs could later align.
   3. Changes made to scope.
   4. Will present to CIMI later this week. Contact SD and FHIR to review. Then Vocab in 10 days to finalize.

Next Agenda

1. Confirm PSS approvals
2. Begin to outline Approach for project

2017-02-21 Minutes

Rob - chair; Ted, Frank, Carmela
Call commenced at 2:05PM ET after quorum reached.
Agenda:

1. Continue work on crafting PSS
   1. We began with the version in Dropbox from last call
   2. Saved v0.3

Next meeting March 7 Neither Rob nor Ted will not be available - The meeting will be cancelled. Next meeting March 21.
Next Agenda

1. Finalize PSS approvals
2. Begin to outline Approach for project

2017-02-07 Minutes

Ted - chair; Susan Barber, Carmela Couderc
Call commenced at 2:09PM ET after quorum reached.
Agenda:

1. Continue work on crafting PSS
   1. We began with the version in Dropbox from last call (labeled V0.1, and edited it for V0.2
   2. We made enhancements to the Project Scope and Project Need sections 3a and 3b
   3. We worked extensively on the project deliverable dates; it is unclear if the end date should be for the STU, or include the STU-to-Normative period; Ted to ask Dave H.
   4. We need to decide if Dropbox location is consistent with the TSC guidelines for project repository, and if so, the link and necessary signup information must be put in the PSS. If not, we need to determine an alternate project repository location.
   5. We are almost done, but a little more work is needed.
   6. Ted to update the new draft in Dropbox.

Next meeting Feb. 21 - Rob will run the call, Using Mikogo

2017-01-24 Minutes

Rob - chair, Ted, Richard, Susan, Carol Macumber, Rob H.
Agenda:
Work on crafting PSS
See Draft PSS via link on main page
Next meeting Feb. 7 - Ted will run the call, Using Join Me