€¦ · Web viewProposal initiated by Canada originally based upon education on using SNOMED in conjunction with HL7 FHIR. There are already projects in this area. This is …

HL7 Vocabulary Minutes – MadridMay 2017

Q3 Sunday - Agenda PlanningHL7 Vocabulary Meeting Minutes

Location: Madrid, Spain(CP co-chair present), P Present, T present on telecon.

Date:2017 May 7 - 12

Facilitator Jim Case Note taker(s) Heather Grain

Attendee Name Affiliation

PC Jim Case NLM

PC Rob Hausam Hausam Consulting

PC Rob McClure MD Partners

PC Heather Grain Global eHealth Collaborative

Quorum Requirements Met: No

All changes to the agenda were logged on the wiki: http://wiki.hl7.org/index.php?title=Baltimore_Sep_2016&action=edit

Sunday Q4 Project review and Training ReviewHL7 Vocabulary Meeting Minutes

Location: Madrid, SpainDate: 2017 May 7th - 12th

Facilitator Ted Klein Note taker(s) Heather Grain

Attendee Name AffiliationPC Jim Case NLM

PC Ted Klein Klein Consulting


PC Robert Hausam Hausam Consulting


Quorum Requirements Met: Yes

Action List Reach out to folks for participation in ISO Binding ballot – in progress. The intention is to

that the ISO work be more inclusive, such as FHIR compatibility. Provide a short list of value sets to be reviewed in Madrid Tuesday Q1 – Grahame. Action:

Rob Hausam will follow up with Grahame. Review Harmonization training material for review – Ted Work with Dave Hamil to close projects 1169 and 1170 – Heather - done

http://wiki.hl7.org/index.php?title=Baltimore_Sep_2016&action=edit

http://wiki.hl7.org/index.php?title=Baltimore_Sep_2016&action=edit

Review UTG document and add bits to cover the closed projects 1169 and 1170 Heather and Ted – Done

Reach out to Lloyd and Grahame and straighten out the vocab joint sessions Add the need for a policy statement on SCT allowed binding strengths in FHIR to the HTA

agenda for clarification – Rob H and OO fell it is sufficiently clear to address their needs – HG Tooling call times – there are difficulties inhibiting advancements in Terminology but these

calls are not progressing and times difficult for attendees – this needs to be addressed with urgency – Russ H. ACTION no action so far – Ted will discuss with Wayne

Send updated UTG document to vocab list (Ted) – in progress. This document tis in development and not yet sufficiently stable for circulation. Removed from the action list.

Review, update and complete the Harmonization Preparation webinar materials (Ted) Send examples of good and bad practices in creating / updating terminology artefacts to

Heather – (All) - none yet received. Action: Heather will prepare words and a) send request to the list, and b) provide words for a section on the wiki – of requests to the Vocab Community. Heather to report back to the WG on whether this assists or not.

Send a few examples of V3 value sets and show them in HRIR specification format (Russ) not yet done, Action: Rob Hausam agreed to take this on.

Put together some ideas for discussion on how a repository of FHIR terminology artefacts can be published (Heather). Removed

Draft new FHIR TermInfo PSS (Rob H) - in progress, delayed due to higher priorities. Seek examples of specific problems around representing multi-part observations in CDA (Gay

Dolin and SD). Action: ask if these examples are available as a follow up and associated with other planned discussions this week.

Set up a shared call with PA and FHIR about quality issues especially around enumerated value sets (Ted). – Ted will speak with Alexander and Lyn about this in Madrid

Log a tracker item about codes in value set expansions being unique DST code NC6”(Rob M) – Action Rob M will look into this

Prepare a draft policy between vocabulary, conformance and InM for best practice for V2 code content in Implementation Guides related to versioning for discussion in Madrid – to be discussed in Q1 Thursday. Versioning best practices for V2 vocabulary.

Remove exceptions marking methodology for CEA write up in Binding Semantics (Rob M).

Project ReviewAll Vocabulary projects on Project Insight were reviewed: (http://www.hl7.org/special/committees/projman/searchableprojectindex.cfm?sortBy=&sortDirection=&FilterKeyword=&FilterProjectNumber=&FilterProjectSponsor=Vocabulary+Work+Group&FilterProjectType=&FilterStatus=notArchived&FilterProductType=&FilterDateField=&FilterStartDate=&FilterEndDate=)

Only projects that are active (not awaiting approval) were discussed.

Vocabulary Tutorial ReviewHeather Suggested the following webinars for development

Intro to TermInfo – needs a more descriptive title for those who don’t know what TermInfo is. – how use snomed CT in HL7, including FHIR.

Prepare a harmonization proposalTed mentioned the ‘hole’ digging problem associated with Harmonization and asked that this be added to HTA agenda.

http://www.hl7.org/special/committees/projman/searchableprojectindex.cfm?sortBy=&sortDirection=&FilterKeyword=&FilterProjectNumber=&FilterProjectSponsor=Vocabulary+Work+Group&FilterProjectType=&FilterStatus=notArchived&FilterProductType=&FilterDateField=&FilterStartDate=&FilterEndDate



Heather reported that the glossary tutorial has been updated with additional examples of what is good and what is not and how to write a definition properly.

Heather has also suggested review of specifications for:Value Set AuthoringExternal Terminology

Intention is to break into smaller webinars

Monday Q2 Terminology Governance Project and Tooling

Facilitator Jim Case Note taker(s) Heather Grain




P Chris Chute Johns Hopkins

P Carol Macumber Apelon


P Monica Jones Animo Consulting, NHS UK


P Sandy Stuart Kaiser Permanente

P Frank Ploeg HL7 Netherlands

P Susan Barber TN Dept. Health

P Oyvind Aasse Norway Directorate of e-Health

P Jane Millar SNOMED International

P Julie James Blue Wave

Quorum Requirements Met: Yes 13 present plus chair.

Objective of this session: Increase the visibility and knowledge of Unified Terminology Governance Project.Ted Klein presented an update on this topic. PowerPoint is circulated and available at http://www.hl7.org/documentcenter/public/wg/vocab/presentations/UTG4TSC.ppt: Issues and objectives of the project were presented and the rationale for a new process was clearly defined. The Vocabulary WG will be the custodian for the terminology used in HL7 standards. This requires coordination of work groups and this is improving but still requires considerable change to support consistency across all HL7 products.

This is now urgent, FHIR maturity is increasing and there is a need for more predictable and stable terminology content. V2 has been included in the Harmonisation process but the process is completely manual.Ballot generation and publication tooling are broken and vocabulary can no longer be rendered without human intervention. The primary scope of this project involves curation of value sets and concept domains. The objective is that HL7 not create our own vocabularies – except for those items which are required for the structure of the information models in HL7. There are near 3000 value sets published in HL7, 900 concept domains, about 500 terminology collections – called code systems for historical reasons. In total about 12,000 codes.

Earlier drafts of the documents have been provided. Updated copy will be provided after this meeting to the Vocab WG for comment.The intent is that this process will replace current Harmonization processes, with new ways to incorporate governed shared vocabulary into published Standards, publish the HL7 Terminology, increased shared review of shared terminology objects.

Using a crowd sourced architecture – so that previous releases will be available to the community and the approval process will be as straight forward and quick as possible.

The envisioned process flow was presented and discussed. Some of the process is automated where that is appropriate but there are consensus and manual components to assist in quality assurance.

It is intended to crowd source good practice and governance, but for things which are non structural as other components are governed through the ballot process. So instance identifiers for model objects (structural components) will not be decided through crowd sourcing.

Monday Q3 Value set expansionFacilitator Rob McClure Note taker(s) Heather Grain






P Peter Jordan HL7 NZ

P Vadim P - note full name not provided Furore

Frank Ploeg HL7 Netherlands

P Sandy Stuart Kaiser Perminente



The Vocabulary Working Group raised glasses in memory of Woody Beeler.Objective of this session

To bring everyone up to date Encourage participation Begin the conversation about what is to be specified and included in this work.

Discussion:Normative specifications should only specify things which are to be implemented. Dates on the PSS were updated and confirmed.Discussion on scope and issues with FHIR:

We can’t make the value set definition be the final source of truth for enumerated content. Information which can be excluded to generate an initial expansion A diagram has been developed which identifies some of the use cases and relationships

between the expand operation and the profiling operation and binding parameters. see the presentation for a flow diagram of the operations.

The intermediary normative expansion is created through the expansion operation and profiling operation which should also procedure the VSE Instance (same value set identifier for all expansions).

It is understood that different products in HL7 use terms such as expansion profile to mean different things. Groups such as VSAC also use that term and it does not mean the same as it means in this work. i.e. VSAC it means the code system and versions

It was suggested that this process might be called expansion process. This name change make it a unique description of the process.

Inputs to expand process are Code system and version, Value Set Definition and version, binding parameters – which spits out the usable value set. The FHIR use case seems to require the definition of an un-profiled concept expansionProfile restricts the rows and extends the columns. Profiling should only constrain available concepts and may add concept metadata. May add languages etc but not change the scope or conceptual content included. Motion: an Expansion profile shall only constrain available concepts from the un-profiled concept expansion and may add concept metadata. Seconded Jim Case. Vote: In favour 9, Abstain 0 Against 0

Action: on the next call this topic needs to be pursued further. Particular regarding identifiers of persistent objects. Ted to include in call agenda for Tuesday 2pm calls.

Monday Q4 TermInfo

Facilitator Rob Hausam Note taker(s) Heather Grain

Attendee Name AffiliationPC Rob Hausam Hausam Consulting


PC Jim Case NLM


Online Daniel Karlson HL7 Germany

P Peter Jordan HL7 NZ

P Frank Ploeg HL7 Netherlands

P Jason Steen HL7 Australia

P Ken Load

Online Piper Ravallo

Online Ron Van Duyne

Online Yongsheng Gao

Online Gay Dolan Trifolia



Quorum Requirements Met: Yes 9 in room (excluding the chair) plus 2 on the phone

The objective of this session is to: Progress planning and complete work on TermInfo FHIR implementation Guide.

Updates from SNOMED Business MeetingProposal initiated by Canada originally based upon education on using SNOMED in conjunction with HL7 FHIR. There are already projects in this area. This is a high priority. There may be some

difficulties in the FHIR profiles which some countries SNOMED International representatives are concerned about. There is a prioritized work item at SNOMED International. The CIO indicated that there is a list of activities within a collaborative agreement between HL7 and SNOMED international. There is a plan established on who is to do what and what deliverables are required. This plan includes a plan for the days of work required. Each organisation is then progressing the activities and have identified the skills and workforce needed to progress the work items. When this was reported at the business meeting, several countries were supportive, including Australia. Jane Millar is coordinating this activity. At a strategic level, they will consider who wants to be involved with this. A call was made for contributors for HL7 in this area.Action: Jason Steen from HL7 Australia will contact NSW and Queensland State Health Departments to see if there are possible contributors from these environments.

The role of the TermInfo project in HL7 is still to be determined. Some planning has already been done and the meeting was concerned to ensure that HL7’s execution commitment is met.

TermInfo FHIR Implementation GuideKen Load indicated his interest which is specific to information exchange. They have found that descriptions are not sufficient for one to understand what a semantic concept really means. This results in potential for inadvertent misinterpretation. There is a desire to extend to include additional metadata. In trying ot move this forward they don’t want to boil the ocean but like to see how LOINC concepts can be linked back to other concepts and terminologies. Creating their small ontology for this purpose to link to ontologies such as SNOMED or LOINC etc at a high conceptual level. The objective is to leverage work done by others and to provide semantic clarity but this may also be able to be used in other work to enable semantic web operations.Links to relevant implementation formats for RDF. These are source files for blending FHIR and RDF.https://github.com/BD2KOnFHIR?BLENDINGFHIRandRDF all of the FHIR resources have definitions in this format. This seeks to have the information model and the terminology in a common representation to support reasoning. The SNOMED owl class information is also provided at this link. This was also demonstrated in Protégé which shows the promising development to support reasoning across models.

There is a presentation on this by Harold Solbrig which has been circulated to the Vocab list and can be found at https://drive.google.com/file/d/0B8TzXCdmLirJZWhweU9RT1ZWRm8/view

Ongoing Call Schedule4pm New York Time (8pm UTC) (6am Australian Eastern Time). It was agreed to retain this call schedule.

Tuesday Q1 Vocab Multiple topicsFacilitator Rob Hausam Note taker(s) Heather Grain

Attendee Name AffiliationP Susan Barber TN Dept Health

P Heather Grain Global eHealth Collaborative

https://github.com/BD2KOnFHIR?BLENDINGFHIRandRDF

P Rob Hausam Hausam Consulting

P Peter Jordan HL7 New Zealand

P Ted Klein Klein Consulting


P Danielle Friend Epic

P Rob McClure MD Partners

P Ron Shapiro Qvera

P Rieu Watker Qvone

P Joel Schneider NMDP

P Isabelle Gibaud HL7 France

P Mark Shafarman Shafarman consulting

P Stefan Lang HL7 Germany

P Daniel Vreeman Regenstrief



Review Value Sets shared between V3 and FHIR and Planning for R4There are issues which are known which have deprecated content included in their enumerated value sets. The extent has not been determined.It was suggested that we focus on the classes before we worry about the instances. The ballot is scheduled for Dec 17th for the January Ballot. There was concern that the content required may not have had suitable oversight by that time.

There was discussion on the implementation capacity. In general value sets and operations are mature but terminology service considerations are not yet (e.g. subsumption) and further testing is required on these.

The FHIR maturity Metric Level Grid was reviewed and though the systems may publish content the full range of expansion profiles or other functionality may not be tested and included. There is a capability statement to which you must comply.

Much of the information in the build.fhir.org/terminology-module.html web site is out of date.

The objective is that the resource has been tested sufficiently and the operations in the resource need to be fully operational or it has not achieved the testing required.The terminology service must fully comply with the functionality required – which in this instance are to read and to search in a RESTful server situation. This is not transactions, messaging, updates, deletes, creation, or ensure correct management of history.It was suggested that there is a need to more clearly indicate that normative is for specific functions – and not have statements such as fully conformant terminology capability – the term fully conformant is misleading – consider Conformant to minimum requirements.

We are currently being asked to - assess the elements in value set, and code system along with their operations (together) are

stable enough to be normative. It is not clear how the capability statement - determine whether the terminology service is mature enough to be normative. To assess this

there is a need to update the capability statement requirements. The current capability statement for terminology service is insufficient to support normative statements in this area. The current requirement is only for read and search and does not include code system.

Action: Ask FHIR to update this requirement. Further discussion this afternoon.The determination of use is that read and search are being used for value set and content map – it is used in production. Vocab may argue that the minimum set should be larger.In an evaluation of a value set resource regarding normative maturity level, the timeframe is reasonable for going normative. However, we suggest that a process of evidence of the full expression of usage in the production systems be part of the judgement. This may or may not be able to be generalised but requires further discussion. This is above and beyond the existing conformance statement of the terminology service. This may require formal testing processes to be developed. And these may be independent for individual resources, concept maps and code systems and operations. This needs to be confirmed and clarified. Closure is not appropriate for concept maps and requires further testing. For ConceptMap all criteria are met to go to level 5 except for closure.

It was suggested that the information currently provided also needs to make clear to users (including non technical users such as decision makers) what the implications of the minimum level of conformance means and how this impacts their functionality.Question:Some terminologies have grammar and there is work in the expand operation but not all terminology services may include this processing. There could be a possibility that an expand could be broken by the terminology grammar – this was considered not to be a problem.

Tuesday Q2 Vocabulary Hosting Structured DocumentsFacilitator Rob McClure Note taker(s) Heather Grain

Attendee Name AffiliationP Swapna Abhyankar Regenstrief

P Beau Bannerman Lantana

P Susan Barber TN Dept Health

P Calvin Beebe Mayo Clinic

P Jeff Brown ASCO’s Cancer Lnq

PC Jim Case NLM

P Steve Fine Cerner

P Benjamin Flesser Redox

P Sarah Gaunt Lantana

P Rick Gemir Lantana


P Dave Hamill HL7 PMO


P Emma Jones Allscripts


P Austin Kreisler Leidos

P Ken Laed JPS



P MaryKay McDarvie Cognosante

P Nancy Orvis Dept. of Defence USA

P Mat Rohn ONC

P Mark Shafarman Shafarman consulting

P May Terry Flatiron Health

P Daniel Vreeman Regenstrief Institute

Quorum Requirements Met: Yes 24 Present (excluding the Chair)

Review of output from "Updating C-CDA value sets" projectValyue /set QA overview as at April 28th

An initial export and comparison of the value sets in VSAC and Trifolia has been finished. Initial review of the 104 value sets listed for C0CDA R2.1 identified:

- 21 are not in VSAC- 1 is in VSAC but doesn’t align because its OID is missing the final digit - 82 exist in Trifolia and VSAC and of these:

o 45 match exactly on contento 37 have mismatches of one or more codes.

Structured docs intend to migrate to VSAC away from Trifolia as the ongoing tooling for this content. The speed of the process is one of the concerns of Structured Docs.

After the QA exercise each component will be resolved and value set content going forward in VSAC is still to be discussed.

The comparison spreadsheet was reviewed and explained. This project includes stakeholder identification. Include source to begin with. The report will include a list of value sets missing from VSAC . l A list of value sets in VSAC that require changes to their scope, definition or expansion. A list of the code systems where there are issues.

Ballot submission comments for CDA_R2.1 ballot. Vocabulary Binding - Should the FHIR binding required and preferred? Attempts have been made to receive guidance on this from Grahame but nothing received at this point. Vocab is struggling to find

a single approach to binding across all profiles. The meeting recommended, preferred, experimental, Open, Closed, Fixed

Missing link – will be updated

Date version issue – the use of a single date to force a single value set expansion to be identified for use is insufficient. A value set definition, of any type (enumerated list or an expansion) will result in a set of concept identifiers. If that definition is set to a particular value set version, the definition will be static. BUT the expansion based on that definition can change over time if the code system version is NOT ALSO set, OR the value set definition is specified to allow inactive concepts to be included in the expansion. One potential solution would be to state that a static binding fixes both the value set definition AND the code system version to a specific version. If both versions can be set by only one date, then a date might work, but this is not always true. Need to validate the location of this note, if in the guidance for IC binding, then it is to be reviewed to indicate both the value set and the cod system require versioning.

Tuesday Q3 Vocabulary Hosting - FHIR infrastructureFull list of attendees not available. Vocab Co-Chairs and total number present only is available.

Facilitator Russ Hamm Note taker(s) Jim Case

Attendee Name AffiliationP Jim Case NLM

P Grahame Grieve




Quorum Requirements Met: Yes 20 plus Chair (full list not available)

Topic: Rework HL7 FHIR terminology tools and processes

Discussed in later section

summary of where we are (Ted/Lloyd/Grahame)

We have vocab in all product families and that is not an acceptable situation. V2 tables are generated out of the database. Maintenance is being done on the database. Changes now go through harmonization and are reviewed. V3 goes through harmonization, but CDA changes do not generally. It is suspected that there are situations when V3 content changes do not go through harmonization. The master CDA terminology is in the PDF guides. The V3 master file is in the Core MIF file. Detailed discussion of the complexities of V3 vocabulary were described.

FHIR imports the V2 database he replaces the V3 MIF when a new version comes out. A few code systems and value sets are managed independently by FHIR, managed in GForge subversion. Ted mentioned the current issues with the V3 tooling so that the validation of the terminology may not be correct. Graham described some of the “messy” issues with the V3 MIF representation of values sets and the work he needs to do when someone asks to use a V3 value set. This is a known tooling issue that has been around for a long time.

FHIR implementation guides define some of their own code systems and value sets. Bound to FHIR general policy, but these are not being followed by developers of implementation guides. New value sets are added to IGs independently, but these are then accepted into the terminology server. While IGs should not be publishing these independent of HL7, but the policy at the HL7 level does not have the processes or policies in place to handle this.

Possible solutions to these issues were proposed, but the difficulties around the diversity of sources identified problems that we do not have solutions yet. Ted mentioned that this is the purpose of the UTG, but it has not been fully defined. We need to define the scope of vocab maintenance. Graham proposed:

- Core business content- HL7 bound terminology- Overlap with content that is not actively maintained externally

RegistryFHIR.org will be up and running soon where code systems, value sets, etc. can be published and distributed for things that they manage. Graham asked for clarification about terminology managed by harmonization and that managed through ballot. Ted explained that there needs to be some validation as to the fitness for purpose of value set, Graham countered that this is the purpose of the ballot process. Ted mentioned that for some areas where ballots move very slowly, updating the terminology needs another process. Static vs. extensible bindings may be the differentiation of which process is used. If the value set evolves faster that the artifact that it is bound to then there needs to be a process for vetting these changes. Dynamic binding is the use case that needs to be addressed.

definition of success (Ted)

Ted outlines a number of criteria for the success of the project (went too quick to get them all), but included. Getting rid of the harmonization process and getting rid of the incompatible value sets across

strawman candidate for going forward (Grahame)Graham proposes that we move all of the terminology to FHIR resources and put them in a github repository and that be the source of truth for the terminology across all product families. All changes will be submitted through the github repository and changes can be proposed through the repository which would be reviewed by the responsible moderators and that this review would replace the harmonization process. Publication would be based on the current version of the terminology as needed at publication time. The format would be family specific and the guthub would be the terminology master for everything except terminology maintained in spreadsheets.

Graham said we can do that now except for generating MIF out of the terminology server. This is in line

with the details outlined in the draft UTG document.

Discussion

Is it possible to have access control? Yes, this can be controlled at three levels; read; read and suggest changes; full access.

Does this include a user interface for suggesting and approving changes? A: There is an infrastructure to build a UI, but there is not one proposed. Ted was concerned that we have had no great success without simple user interfaces.

It looks like we will have access to Michael Lawley’s value set editor so that will assist users. Mapping tool is also available. Everything is terminology service based.

In terms of using github, HL7 terminology is not open source, so is that going to be a problem? Not a technical problem, just a process/policy issue. There are access/copyright statements in every resource that would “require” implementers to adhere to license.

The github implementation would serve internal HL7 development. Question is whether it would also support the publishing needs for value sets for HL7 artifacts? A number of possible solutions to IP restricted content were discussed, but no definitive solution proposed.- Generate V2 data base- Generate MIF- Import MIF and validate content- Integrate tools from Michael Lawley

Question as to how to handle concept domains. Lloyd mentioned that these could be handled by a single code system that contained concept domains. What is not current code system supplement are not currently supported by MIF. This is to be able to define new properties with a code system. Canada has such an extension for LOINC, but this is not handled by the MIF currently. Possible solutions to this deficiency were discussed. Challenges and benefits of this proposal were discussed, such as the operations on the code system assumes working on the base code and not on the supplement, so a merge would be needed to return the proper content. This is a problem if the filtering is based on a supplemental property.

To put all this together will need funding and Graham will determine if and where these would come from. Timeline would be a demonstration in the September.

Motion (Ted): Accept the strawman proposal from Graham, with the goal of a working prototype to be demonstrated at the September meeting representing the universal terminology governance implementation. Second Lloyd.

Lloyd had questions about vetting content through the github process and Graham expected that this would be possible. Ted said that a full review of the technical needs in the UTG document would need to be done to make sure they were all addressed

Vote: 20/0/1 Approve/Opposed/Abstain

Discussion of adding FHIR readable identifier to V3 code systems

Pushed to next joint session. Grhaam will present a set of lists to represent these.

Scheduling VSE future quarters to include FHIR (and CNF)Would want FHIR to participate, so will need to cooperate to get a common quarter where they can meet. Graham is assisting vocab in this area so he suggested that we get info from Lloyd as he runs the FHIR schedule. The VSA meetings are currently Q3 Monday.

Proposal that the VSE discussions are held at the normal VSA quarter and then validate them when we have common quarters with FHIR. Also proposed that the technical discussions be done at the connectathon prior to the WGM. Agreed.

Tuesday Q4 – Vocab/Patient Care (Hosting)

Topic: Negation requirementsSee Patient Care minutes for details of this session.

Wednesday Q1 – Joint with MnM/FHIR (Vocab Hosting)Facilitator Rob Hausam Note taker(s) Heather Grain

Attendee Name Affiliation



PC Jim Case NLM/SNOMED International

P Jeff Danford Allscripts


P Stefan Gang HL7 Germany

P Grahame Grieve Health Intersections


P Simone Heckmann HL7 Germany

P Peter Jordon HL7 NZ



P Lloyd McKenzie Gevity

P Jane Millar SNOMED International/

P Jean Narcissi ADA

P Dave Shaver Corepoint

P Ron Shapiro Qvera

P Elliot Silver Change Healthcare

P May Terry Flatiron Health

Quorum Requirements Met: Yes 17 plus Chair

Topic: Maturity Level associated with terminology

Vocab Q1 Wed – Vocab/FHIR/MnMPlanning for R4FMG has asked for four questions to be answered by each project group:

- Which artifacts in your space are meant to be normative?

o CodeSystem – maturity level 5, but there are some issues with the $compose operation. So how can this be split out? Operations are separate artifacts so while the $compose operation is not ready for normative, the underlying artifact that it depends on may go normative.

o ValueSet – maturity level 5, no issues

o $Expand operation is ready for normative, so should the ExpansionProfile resource be a candidate for normative as well? Rob M. has issues with making it go normative. It has not been tested by a number of organizations. The valueset expansion project needs an implementation and this might be the best candidate. It does not yet meet the maturity criteria. Want to make sure that there is congruence between the valueset expansion project and this resource going normative. The timeline for the valueset expansion project is after the next WGM. Need active participation from FHIR to make the timelines.

o ConceptMap?

o Related operations to these resources

- Which new artifacts do you expect to have in R4?

- Maturity planning for existing artifacts

- Any issue that FMG could offer assistance

Operations – issues raised earlier were largely addressed regarding operations having to go inside the resources to operate appropriately. The relationship to the capability statement was clarified and will be more clearly documented and further consideration will be needed to determine future requirements and implications.If you are exercising a resource to demonstrate additional terminology server capability this needs to be clear. How can we decide what is the minimum requirements and how do we determine conformance?It was suggested that publication proceed with additional considerations to handle negation requirements, and terminology server requirements and profiles to be considered for next release. The key factorIt was noted that Vocabulary need to spend more time and consideration on the maturity levels and their terminology related requirements for operation and profiling.

ConceptMapImplementers don’t necessarily need this to be normative. It is currently at level 3 and closure is still an issue. There is concern about the changes vendors might need to make if this operation was normative. This is not stable – structural changes are still being made. There are 4 or 5 servers implementing ConceptMap. This is something that is useful e.g. grahame has used this for R2 – V2 mapping. If this is balloted as normative in the May ballot. If b allotted as normative there are 3 outcomes:

- Everyone says OK - People say what! And remove that component cause it is so ‘bad’.- Make changes required and proceed with it in the ballot to normative.

There is a task to do:- How do we decide that a resource is ready to go normative – from a terminological

perspective? Usage criteria need to be clear and the specification also needs to consider the use case/s. There is a case to exercise the resource more strongly, particularly consideration of update requirements. A resource should be exercised across all of its capability

- Do we need something in the documentation that indicates the impact of the minimum normative requirement on functionality delivered?

Action: It was agreed that code system does need to be added to the Terminology Service Capability Statement. Ask FMG for space for further clarification on terminology service capabilities and issues discussed here.

- Value Sets- Code Systems – not including compose- ConceptMap - not including closure

Motion: Rob McClure, seconded Ted that - Value Sets- Code Systems – not including compose- ConceptMap – not including closureBe considered normative

Vote: 0 Against, 0 Abstain, 20 In favour

Do we need to have the conversation about what we need to prove safe and stable to use prior to going to maturity level 5?

- There are internal navigation pages which are not normative – they have no status- There are advisory pages – which are also not normative but are informative- Pages with shall or should are aimed at normative.It would be helpful for the web system to be able to indicate if you are pointing to a normative page or not. Grahame indicated that he will have to work out the effort involved.

Action: revisit the Terminology Service Page and using codes in FHIR (include on calls – Ted)

Moved Rob McCllure that we should allow sections of a page to indicate their normative or other status clearly within pages. Allow the same capabilities for a note for each section of a resource in order to override the default status of the page. Seconded Lloyd McKenzieRendering implications were considered and felt to be within the requirements identified in the motion.


We need to beef up the capability statement and processes associated with that statement. Designating production servers as evidence of level 5 maturity – functionality for maturity levels – we need a formal way to determine that capability – this may require test scripts to be established.

It was agreed that Data Types are not an issue in this area

Motion: Rob McClure moved that Vocab and FHIR propose that in order to demonstrate Terminology Service Capability should be based upon some form of formal testing which demonstrates the base resource functionality and operations to which the terminology service is conformant.Seconded: Peter JordanDiscussion: it is recognised that development of such formal test scripts will e time consuming.


Tracker Item: Clarification on ICD modifiersExample: lung metastasis due to breast cancer: if it is left is it left lung cancer or left breast cancer? Recurrence is also a simple problem.

There is a need to improve the ICD-10 page to more clearly indicate how these examples should be handled.

Action: A quarter should be put aside for the next meeting to go through pages and FHIR stuff to confirm content and advance these issues. (Jim), The objective is to

- make pages clear and useful to the user community, not just us. - work on the links.

Specific Questions: 13210 – maintenance of modifiers such as dagger and asterisk in ICDPrevious guidance was to separate the codes by a space. This is insufficient to accurately represent the concepts. This is an expression – which requires appropriate syntax.

Moved by Simone and Seconded Rob McClure. Document this in the “Using ICD-[X] in FHIR page. This will include dagger-asterisk, * | and potentially other identified syntaxes.


Wednesday Q2 – Joint with MnM (Vocab Hosting)

Facilitator Ted Klein Note taker(s) Heather Grain


PC Jim Case NLM/SNOMED International




P Lloyd McKenzie


P Frank Oemig HL7 Germany

P Birtil Reppen

P Ron Shapiro


Topic: Vocabulary Maintenance

Out of the MnM Monday session they decided not to proactively try and resolve known problems until they are raised as issues by users.In the terminology space Vocab is moving on with the vision for unified terminology and there are terminologies for harmonisation published in core MIF for V3:

- structural codes in the RIM need to be maintained to support relevant implementations. Lloyd stated that the RIM in terms of vocabulary will be maintained.

- V3 code systems and value sets which have been imported to FHIR resources. These must be maintained to support FHIR effort. There does not seem to be a reason why FHIR could not hold an absolute set of HL7 terminology.

- Some code systems not in FHIR which we don’t know whether they are used at all.

Ted suggested that having a unified terminology for HL7 with specific statements of which are used in which product lines, with the ability to identify overlaps between product lines. We inform and apply unified sets of rules for quality and vocabulary models on all the terminology used in HL7. The packaging is then an enabler rather than the ‘’issue’’. Ted seeks endorsement of this principle by both Vocabulary and MnM

It was agreed yesterday that a pilot of unified terminology tooling would be – tooling that is ready in the FHIR terminology service and editor. This would be tested prior to the September working group meeting for TSC to review. Vocab will be working with Grahame to undertake this and document what is and is not covered by this process and tools.

Discussion with publishing is also occurring to understand the publishing requirements and management. What are the MnM requirements associated with evolving terminology maintenance processes?

The meeting considered the vocabulary related information on http://wiki.hl7.org/index.php?title=FHIR_Conformance_QA_CriteriaIt is difficult to determine how much information to provide without being too detailed which results in people not using the guidance at all. The FHIR focus is always on what does the implementer need. It was agreed that people are foolish to ignore the information requirements specified and complete the relevant harmonisation paper work the document won’t evolve past a certain maturity level.

A resource will bind the value set knowing that what makes a breaking change – i.e. a code system version change. Then binding should be binding to this version and not a subsequent version or if they bind just to the value set dynamically. Approval from the community can be a blocking point for poor practice. The maturity levels are also used to indicate that the work is not ready to be locked down. Once the content is normative changes become much more difficult and are discouraged.

The question arose whether we want to make the code system part of the normative specification. There may also be a need for the ability to override. Binding syntax have discussed requirements for downstream modifications of value set (more constrained, less constrained etc) some of that gives guidance on what can be added or removed in a value set. There is an overlap between some of these notions in FHIR processes.

As we start to apply FMM levels and think about the notion of ‘normative’ to terminology artifacts, the community needs to understand what sorts of changes will be permitted without consulting the community *level 4/5) or at all (normative). In some cases, these will be quite limiting – perhaps to the level of any change to the expansion. In other cases, it may be open to such a significant change as migrating fromICD9 to ICD 10 or SNOMED CT. Because variability in behaviour needs to be allowed, the explicit rules need to be captured at a ‘per valueset’ level. It may be this should be a choice of code/string so we can define a few common conformance behaviours, but also allow for a particular code system or value set to have it’s own “specific/special” defined behaviour.

http://wiki.hl7.org/index.php?title=FHIR_Conformance_QA_Criteria

http://wiki.hl7.org/index.php?title=FHIR_Conformance_QA_Criteria

Possible code system behaviours:a) Fixed expansionb) Expansion can increase, but no changes to existing concepts (including no

overlapping/ancestor concepts)c) Expansion can increase, but no changes to existing concepts, new overlapping/ancestor

concepts are allowedd) Existing concepts can be increased in scope (non-overlapping with existing concepts) and

new concepts can be added including overlapping / ancestor concepts- ….z) No constrains

We need to identify what behaviours are typical/useful

Example value set behaviours:a) Definition is fixedb) Additional concepts can be introduced from the same code system, but no new code

systemsc) Existing concepts can be replaced with concepts from a different code system

It was decided not to vote now but to get further input, from Grahame G and others prior to deciding.

There is a concern that the more we get into the details we may need different roles for different artefacts. The language codes as defined involve a language tag creation doing post-coordination where the different terms

Wednesday Q3 – Joint with Publishing (Publishing Hosting)Refer to publishing for minutes of this session

Facilitator Beal Thyph Note taker(s) Heather Grain


PC Beal Thyph HL7 Switzerland

P Jim Case NLM / SNOMED International

P Heather Grain Global eHealth Collaborative


P Jean Narcisi American Dental Association

Quorum Reached – Yes Total: 5 plus the chair

Publishing of V2 VocabularyA stop gap measure is needed until this information can be stored on an organisational resource which supports publishing. Frank’s database is the source of truth and the ballot copy is generated from the database. All updates to the database content goes through Ted at harmonisation. Ted structures and validates it. As soon as this process can be replaced with a less intensive option. The source of truth for V2 vocabulary will be the new HL7 Terminology Repository which FHIR are going

to develop as a pilot project in the FHIR terminology service. An extract will either be a delta which cn be applied to the active database or a complete set. The complete set option may not be sufficient as FHIR does not have all the metadata in Frank’s system. The content delta can copy metadata from existing value sets, but for newly created codes it is in the harmonisation data and Frank will update manually.

We need to modify the direction of travel. The Harmonisation process needs to go Frank then when all is good to FHIR. The long term approach will be harmonisation – determine the most current available vocabulary otherwise a snapshot of the current terminology server data will be provided to go into the ballot through a Delta load to Frank’s database.

The intent is to have a single source of truth and the changes will all come from the community at large with a higher level of oversight.

Vocabulary suggests that if a coded concept you need has been added in a more recent version of HL7 that includes that content. A more clearly documented policy on this should be considered.

Publishing of V3 vocabularyThere are a few current problems with the tooling – which cannot publish the vocabulary.In the foundation documents for V3 publishing we need a section of generated data tables for concept domains, value sets etc. which can be clicked on should be structured to links to concept domain content directly to the source of truth. These links should not go to static pages they should go directly to the FHIR resource page for this vocabulary or the underlying tooling pumps the xml for the content through a style sheet so that it looks familiar.The process and requirements need to be spaced out over the next 12 months to support urgent tool development.

These changing processes are one of the reasons why Vocabulary is concentrating on tutorials which cover the updated underpinning requirements of value set definition and good process to prepare people prior to the new process being introduced.

PHIN VACS publishing issues re: URL changesThe requirements include the publishers and distributors of terminologies will assert that something will be a permalink. The link is never changed. It may be a link to information about the code system or directly to specific content. That is not being mandated.

For CDA implementation guides include publication of URLs to on CDC but HL7 didn’t get agreement from CDC that these would be permalinks. Vocabulary didn’t realise that this could be possible – we didn’t ask if these were URLs and if URLs confirm that they are permalinks and that an agreement on that status exists with the provider. We need to make clear that any implementation guide with a published link requires a guarantee from those making it available agree that they are permalinks.

You should take under advisement whether a URL should be published in a normative document if the URL is not under a permalink agreement. Instead publish the OID.

Wednesday Q4 – Vocab / Patient Care (PC Hosting)Topic: AllergySee Patient Care for minutes of this session.

Thursday Q1 – Vocab (host)/CGIT/InMFacilitator Jim Case Note taker(s) Jim Case and Heather Grain

Attendee Name AffiliationP Swapna Adhyankar Regenstrief Institute


P Nathan Bunker AIRA

P Jim Case NLM

P Heather Grain GeHCO

P Tony Julien Mayo Clinic




P Craig Newman CDC

P Frank Oemig HL7 DE

P Dave Shaver Corepoint

P Rob Snelick NIST

Quorum Requirements Met: Yes 12 plus chair

ISO Update ISO TC215 Health Informatics are undertaking activities of relevant to HL7 Vocabulary and associated communities. A report of these activities was provided by Heather Grain (Convenor of TC215, WG3 Semantic Content) and Ted Klein (HL7 Liaison) on upcoming activities requiring Vocabulary input. The link between ISO standards and European legislative implementation requirements was explained. This legislative requirement impacts the use of HL7 content.

Metadata Requirements (11179 related topics)This work item has active engagement from the ISO Information Technology working groups and has the objective to identify:

Additional metadata needed to support the continuum of healthcare (not just data repositories for e.g. national reporting).

Representation forms of information for different situations

Action: Heather will take the MetaReq 11179 related document and send it to the Vocabulary WG for comment along with the comment form. Send comments back to Heather who will collate the comments and submit to ISO.ISO Update – other current work items

ISO Binding

The progress of taking HL7 Binding material into the ISO environment. There is great interest in including FHIR binding and to undertake binding

Heather and Ted explained the European situation which requires the use of ISO standards where they exist.

Versioning – best practice for V2 vocabulary.We have, so far, taken our current practices today, which did not track vocabulary changes, the only thing which was possible was manually implemented scripts. Given the work to integrate HL7 vocabulary for all HL7 products there is a need to be able to manage version control across members of the product family.Though we are not yet ready to put up tooling, we do have a set of V2.9 comments in ballot. We need to establish versioning mechanisms in place now before we go out to normative with V2.9. Ted requested input and ideas on what is needed in HL7 V2 for version management.

What do we do today now that we have cleaned up in 2.9. How do we encourage movement to 2.8. If I am starting a new 2.5 guide should I stick with 5 and adopt components of 2.9.

Vocabulary has said in the context of use with code systems such as SNOMED CT or LOINC you are required to use the latest vocabulary version. The problem is that HL7 Vocabularies include things we call structural – e.g. the table of event codes – table 253 – the codes which identify these structures – you can’t use the latest table as the structures are not able to apply in V2.5. For structural vocabulary, you must use the structures relevant to that version. Ted recommends using the latest table version for all but structural concepts. i.e. everything that is non structural – you should use the latest table. Occasionally we have changed the binding as well as the content. If the binding changes and someone looks at the implementation guide it may not be clear that it is the binding that has changed rather than the intent of the table.

Decision to reduce the workload and not to change things on people. We have only applied changes to version where real changes occur (e.g. typo change from taint to paint). Misspellings were cleaned up. The 2.9 source is the best quality available and meets requirements for use of current terminologies. Purchase database from HL7 is available.

Actions: Vocabulary WG should include the following in the call agenda (Ted) Add a flag to the database to indicate if an item is structural or non-structural Generate content for each table indicating if it is structural or non-structural Provide a short guidance document on best practice

Question – do we need a general document on how to select the relevant code system (version) for use. When something is an HL7 table you have to use it – when it is a user defined table there may be suggested values (which are not necessarily recommended). Such a summary should indicate not only what to do but why it is important. We also have to have sufficient flexibility to recognise that information systems do not change as readily as standards do. It must be clear who is the intended audience for this guidance. FHIR use some V2 tables as a source but have modified the table without any change of name or ID. This is confusing to users.

Guidance Document should address: That Terminology is not yet unified and therefore there are multiple sources for content (e.g.

the ballot or FHIR) and we know that there are inconsistencies between these different sources.

Appropriate terminology management When an organization decides that they need a different binding – they should develop a

profile or implementation guide – that clearly documents the variations from the existing standard. Though we would recommend inclusion of the reason for such variations that information would not be mandatory

The ability to extend an HL7 structural V2 value set – explain how to do this appropriately and in what situations extensions are appropriate.

How to address knowns, others, unknowns and unknowable.

This work should also be used to develop a best practices webinar. Vocab is working on best practice for developing value set content. How many value sets do you want to bind to a data element and at what time in the constraint process do you want to do that. Best practice for binding specification is still emerging. The best practice document may be an evolving one. The objective is to provide guidance on what is known and identify what is not yet clear (under development section).

We will seek use cases to validate or invalidate assumptions in the document.

Nulls in V2 / CDASome content has unknowns and some do not. It was suggested that there is a recommendation to use null flavour in all situations without making it mandatory. The fact that different product families use nulls differently is a known problem. Because there is no explicit way in transforming between V2 and V3 and nulls are contained in the value set in one case and additional supplementary codes in the other. FHIR handles nulls the same way V2 does. They are enumerated in the Value Set. The sense is that we should discourage unless needed explicitly for other business reasons the use of the null value as an external supplementary set of codes. This needs to be discussed with Structured Documents.If you are enumerating codes that represent forms of null – including other and unknown – whether they are part of the code system or if not part of the coding system (such as SNOMED CT) you should use the HL7 flavours of null.

The question is how to handle different types of null flavour in different families Other and unknown are data like any other as they have value which represents something which can be actioned. Flavours of null are often just reasons why you don’t have data in a specific field.

Each question you ask has subtle differences to the information not being available. We have to think about how much of the real world we want to implement. It might be worth considering where an underlying code system has concepts for what we call flavours of null – we assume that in the contexts of the semantic coverage those concepts mean something very specific. In one code system might bean I don’t know the patient s condition – or it might mean it is unknowable. We

should assume that we have null flavours for a reason, that they have value. We include documentation where this would be helpful. Where you don’t have these values in a code system – the primary code system, doesn’t have specific codes for representing the forms of unknown – the hL7 flavours of null are available for inclusion.

Code systems with null flavours in them should have a context associated with them. In SNOMED CT the context is provide din how it is used. Often the context is provided by the information model not the code system itself.

The use of other is completely context dependent and is reliant upon the values in the code system. If you are passing something in the null flavour there is no other – you can’t do another kind of other or a null kind of other.

Statement to take back to PHIR – rather than creating a complex text based representation – if a code system contains values for unknown and other etc those values should be included in the value set from a binding perspective. If it does not they should use an HL7 defined null flavour and include these values in the value set.

Thursday Q2 Vocabulary (host)/ CGITFacilitator Ted Klein Note taker(s) Heather Grain


P Nathan Bunker AIRA

PC Jim Case NLM

PC Heather Grain GeHCO




P Frank Oemig HL7 DE

P Rob Snelick NIST


Binding syntaxRob McClure explain the information on binding on the HL7 wiki which reflect the current state of work in this area.

The spreadsheet of VSBS – use cases was discussed. The objective is to define the main use cases.1. Bind to an entire code system 2. Bind to concept domain, no code system specified3. Bind to Code System but implement only latest version of the code system4. Bind to a specific, entire Code System – implementing one specific code system version only.5. Bind to multiple Code System versions – alternative acceptable value sets6. Bind to a defined subset of a code system

a. From only one specified code system version (static)b. From only the most recent code system versionc. From more than one code system version – to include inactive legacy codes

7. Bind to ad hoc string data elements (such as those used in questionnaires). i.e. content which is not in a published code system. E.g. codes that are not part of a publicly registered code system. Such codes only live in the context of the model element to which they are used in an implementation. This is an ‘’In Line’’ code system in FHIR. People are also doing this in V2 implementation guides. This is only valid for use within the defining model artefact.

8. Bind to value set with multiple code systems a. Where the version of each code system is independently specified.b. Where the version of one or more code system is not specifiedc. Where the version of one or more code systems is specified as most recent

9. Bind to a value set and specify a value set definition versiona. Specify most recent definition versionb. Don’t specify which VSD version

10. Bind to a specific VS expansion – don’t do this one – explain why not.11. Bind to a value set with any other parameterization - extensibility - allow extensibility to be

managed similarly to the way binding is handled in the first case, rather than redeveloping different mechanisms.

a. Extensibility: Define how codes/text that are not in the expansion are handled, Permit extensibility from a priori defined set – from an implementation

perspective, Permit extensibility in an ad hoc situation – from an implementation

perspective Consider for inclusions or exclusions – can use, cannot usea. Pick from a code system – specific version, most recent etc, b. Pick from a Value Set – such as a local code but note that this is not in the

expansion set.c. Textd. Local code

12. Bind to a single codea. From any code system versionb. From a specified code system versionc. Where the single code is not from a code system

13. Preferred vs Requireda. What may or can change. This is about allowing people to do what they need to do

without breaking conformance. This is about being allowed to not use what you are given in the Standard. Preferred value set is considered to be not exhaustive but comprehensive enough to be implemented and satisfies the use case.

i. ID VSii. IS Value set version

b. Is Text allowed

Note: Need to define what is most recent, latest available, current version of a code system?Note: 11 a,- b and C code come from a code systemNote: 11 a - d and e would require a data type where the implementation of that data type allows optional conditional – data type where the text description is required if the code is not populated.

Note: 11 you cannot send a code which has the same meaning as a code in the ‘preferred’ value set. This requires further consideration.Note: that when ‘other’ is used this causes meaning to change.

IN a binding – what SNOMED CT calls a model binding – you have a particular meaning binding for that context – using concept domain. Separately we have the notion of once you have the model meaning you have implementation choices, rules and restrictions. Some of those choices are from a concrete set of codes all the way up to you do the best you can and represent the detailed meaning in text strings. Is it easier or harder to separate these notions?

Thursday Q3 – Vocab (host)/CIMIFacilitator Rob McClure Note taker(s) Heather Grain


P David Booth

PC Jim Case NLM

P Rachel Esmond PerRad

P Richard Ettema AEGIS


PC Rob Hausam

P Steve Hufnagel


P Jay Lyle


P Susan Matney

P Galen Mulrooney VA

P Jackie Mulrooney

P Claude Nonje CIMI

Stan Huff

P Ammon Shabo

P Michael van der Zel

P Dan Vreeman

Quorum Requirements Met: Yes 18 plus chair

Conformance Assertion and Testing of FHIR terminology resources and serviceFrom a testing definition perspective, CIMI and FHIR are trying to come up with (from a FHIR perspective) we have test scrips to define those tests along with resource instances (fixtures) which feed into the test. We are seeking test scrips which are defined to support conformance testing of terminology services. CIMI need assistance from Vocab to provide valid terminology content for conformance testing.

Connectathons have had terminology as a track for some time. Test scripts have been developed to test the patient resource in read, update, create etc. What is the next track they could consider? FHIR could take the suite of test scripts which exist for the terminology track and see what would need to be done for formal testing of terminology conformance. Richard has bought forward those conformance requirements originally developed by Grahame. This is not just about current use cases but also exercising what the resource says against version management and other use cases.

CIMI would like the models being produced in testing as a starting point to use in real applications. Stan asked what is needed? FHIR profiles with CIMI modelling. Options include:

CIMI template for problem list which is not easy. Basic admission observations

Would rather use what is already needed and focus on what has immediate practical utility.

Where we can find things which CIMI is looking at in multiple situations these could be logical candidates. As an initial set a representative sample would be helpful. Susan Matney indicated that there are some components which are live which might suit.

There is a need to include:- Simple cases - Service vs resource capability- Updates of versions

o Are the attributes in the resource sufficient to manage the update requirements?

Use cases are also need to say – what happens when this ‘’unexpected or uncommon event’ happens.

The concept map resource needs to identify version of terminologies and how this is tested and managed.

Helpers – Peter, Jay and Richard are willing to help.

HL7/CIMI namespaceThe CIMI namespace was created before CIMI was part of HL7 and is now owned by HL7. Concepts in the HSPC namespace can be used or HL7 can determine whether to maintain the existing HL7 namespace. The HTA have already established that anything needed by HL7 internationally will be asked to be promoted to SNOMED CT international. For things that are in the US edition that CIMI that HSPC needs they will contact Jim and ask for it to be promotedNew concepts;

- if needed right away – will create in the HSPC creation- If the HSPC concept is a child of the US concept then there is a request to promote made to

Jim.- If we can wait long enough we will ask for concepts to be directly added to the international

core.

It is agreed that the namespace is a scratch pad for quick use.- A solution to this needs to be resolved. The reason we want to move quickly may need a

more sophisticated HL7 policy. We can produce and make them available in ballots.- We need to meet our formal agreements

Keeping up with the load will be of a volume which is higher than a volunteer community may be able to cope with.

There is a need establish a clear process so that we do not duplicate activities. Heather requested that any discussion on content promotion or additions or management – the HTA be involved.

Request HSPC to have the discussion with SNOMED International there will be a need to clarify what can / cannot be published.

The idea of maintaining SNOMED CT, LOINC etc in a single environment is also an issue. It was agreed that to pre-suppose the problems before we get there has the potential to waste considerable time. We will deal with problems as they arise.

Rob McClure stated that if CIMI publish any SNOMED CT content it must follow the processes associated with HL7 and SNOMED International’s agreement. It may be US realm specific which might be the US version with HSPC extension then this is a US realm specific publication. E.g. the VA is building on the US extension

If the process is through a realm specific content it needs to be published and used in that realm only (within the licensing agreements of SNOMED International)

Calls2 – 3 pm Eastern Time for CIMI/Vocab Calls. These calls are to confirm exactly what is needed from Vocabulary. SNOMED CT, LOINC, Cancer links there is a URI with synonyms of LOINC content. There are some details to be resolved.

Thursday Q4 – Planning and GlossaryFacilitator Ted Klein Note taker(s) Heather Grain


PC Jim Case NLM


PC Rob Hausam



Quorum Requirements Met: Yes 6 with chair

Action: All co-chairs with responsibilities to check status on project insight and check TSC updates.

Joint Meeting requests Jim reported:

All rooms reserved and done.

Responsibilities – see Ted’s colourful document.

Conference CallsCall Day schedule Time USA

EasternLength Starting Date

Vocab General Thursday Fortnightly 3:30pm 1.5 hours 25th May TQA calls Ad hoc calls when requiredBinding Semantics Tuesday fortnightly 2pm 1 hour 23rd MayValue set expansion

Tuesday fortnightly 2pm 1 hour 16th May

Unified terminology governance

Tuesday fortnightly Noon 1 hour 20th June

TermInfo Tuesday Weekly 4pm 1 hour 16th May

Action: Vocabulary General Calls - Chair for 25th May - Jim. To arrange call and agenda. Rob and Russ will need to arrange the call on the 8th of June. Otherwise Ted will undertake this.

Meeting Adjourned 5pm.