Vocabulary Working Group
Meeting Minutes
May 2012
1 Attendees
Given Name / Family Name / Affiliation / eMailMeya / Achdiat / Standards Australia /
Elaine / Ayres /
Virinder / Batra /
Woody / Beeler / Beeler Consulting /
Linda / Bird / MOHH Singapore /
Adri / Burggrcaff / HL7 Netherlands /
Giorgio / Cangiou / HL7 Italia /
Nicolas / Canu / L’Atelier du Soft /
Jim / Case / NLM/ NIH /
Chengjian / Che / Lantana /
Adam / Chee /
Lee / Coller / Oracle /
Carmela / Couderc / Seimens / carmela.couderc@siemens
Jane / Daus /
Gaye / Dolin / Lantana /
Elaine / Ayres /
Ana / Estelrich /
Steve / Fine / Cerner /
Finnie / Flores / CIHI /
Sarah / Gaunt / Lantana /
Christoff / Gessner / HL7 Germany /
Adel / Ghlamallah / Infoway /
Heather / Grain / Llewelyn Grain Informatics /
Russ / Hamm / Apelon /
Monica / Harry / GPI /
Rob / Hausam / Hausam Consulting /
Peter / Hendler / Kaiser Permanente /
Masaaki / Hirai /
Piers / Hollott / Sierra Systems /
Wendy / Huang / CHI /
Julie / James / Blue Wave Informatics /
Gaby / Jewell / Cerner /
Dong Il / Jeong /
Hetty / Khan / CDC /
Graeme / King / Logica UK /
John / King / Kingdomain Corp /
Ted / Klein / KCI /
Andrzej / Knafel / Roche /
Beverly / Knight / Canada Health Infoway /
Marc / Koehn / GPI /
Jim / Krotz / Samhsa /
Austin / Kreisler / SAIC /
Jay / Lyle / VA /
Anil / Luthra /
David / Markwell / IHTSDO /
Brett / Marquard / Lantana
Bonny / McAllister / Iatric Systems /
Vince / McCauley / HL7 Australia /
Frank / McKinney /
Carol / Macuuber / Apelon /
Chris / Millet / NQF /
Senthil / Nachimmuthu / 3M /
Frank / Oemig / HL7 Germany /
Rajah / Rai /
Sarah / Ryan /
Stefan / Sabutsch / Hl7 Austria /
Sue / Schneider / eHealthOntario /
Rita / Scichilone / AHIMA /
Peter / Seiften /
Dave / Shaver / Corepoint Health /
Craig / Stancl / Mayo Clinic /
Andy / Stechishin /
Sandra / Stuart /
David / Sundaranstaled / Epic /
Asif / Syed /
Ryan / Van Dyne / CDC-CCID Information /
Hendry / Wijaya / MOHH Singapore /
Laura / Wood /
2 Sunday Q3 Facilitator documentation review working session
Chair: Ted Klein
Scribe: Heather Grain
Agenda modifications
Q2 and Q3 on Monday have been swapped to enable Ted to be present as tutorial schedules have changed.
No other changes to the agenda were identified. Chair and Scribe responsibilities were identified.
FHIR - Ted has agreed to develop a scope statement to ensure that vocabulary is correctly operating in FHIR. This will include appropriate constraints are in place. Ted to report on Thursday on the status of this work.
Core Principles R2 are version 3 models and are normative. A status report was received from Ted Klein. There was not a single comment on the vocabulary chapters. There are no ballot comments to be addressed.
A caveat in the notes for the reviewers included that the definitions for terms like term system etc have been put into skmt. Heather checked the SKMT and the content from the ballot has been added, but the metadata is not well structured and needs review.
The agenda schedule was updated to Version 7 and distributed to the Vocabulary list.
Facilitator materials
The problem:
We need to understand what facilitators need to know and do.
There are problems
which relate to harmonisation and harmonisation proposals, and
outside harmonisation which still relate to vocabulary, where the Vocabulary working group provides advisory input.
We need documentation about what a facilitator needs to do. We have documentation about the artefacts required for harmonisation. The total list of tasks which involve terminology is quite extensive and was originally listed in the 16 point document. This document needs to be reviewed. Information to support the first problem has been incorporated into the harmonisation policies and preparation documents.
Jim indicated that he has done an initial review of the existing 16 point document and found that it is not clear how it is intended to be used. The committee has confirmed that there is a need to update this document. Some of these points have detailed documentation. For further information see the updated wiki page: http://wiki.hl7.org/index.php?title=Vocabulary_Facilitator_Responsibilities
Changes suggested include:
Remove names associate with training documents
Remove how to make submissions for imported vocabularies.
It was Agreed that there is a need for a Vocabulary Facilitator Guide.
Action: Ted to raise the issue of vocabulary facilitator attendance at Vocabulary Facilitator meeting session would be a helpful health measure at the Co-chairs meeting.
A discussion on the role and relevance of a facilitator occurred. The essential issues will be included in the facilitator document. Include:
What is a vocabulary facilitator and relationship to modelling facilitator
What is the action within your working group
What is the action externally (harmonisation. OID registries for example)
The facilitator tasks roles and responsibilities include
Intrinsic: attending facilitator meetings,
participating in development of facilitator training materials documents and processes
participate in development of vocabulary facilitator resources - listed on the Vocabulary Facilitator Guide wiki.
Identification of training requirements to support facilitators
to development skills required to achieve the required tasks (keep up to date with changes)
Attend vocab facilitator meetings
Attend tutorials at least every X years.
Must subscribe to the vocabulary list.
Must be familiar with parts 4 and 5 of the Core Principles document.
Internal Domain workgroup focus: tasks to function as a terminology expert within the working group domain. (these will need to be enumerated)
Act as an expert resource to the domain working group on vocabulary and terminology principles and issues
Advice the working group on good vocabulary practice
Assist the work group with the construction / selection of value sets for coded attributes.
Liaison with the vocabulary working group on behalf of their working group.
Educate the domain work group of the role of the facilitator and be actively involved in the working group activities
Coordinate with the work group modelling facilitator on work group items
Inform the domain workgroup about licensing and IP issues on proposed terminology use.
Assist the domain work group with how to bind terminology to their models including:
understanding the implications of binding choices,
relationship to localisation.
Assist the domain work group in creation and modification of concept domains.
Externally focus (representing your working group to other communities.
focus on HL7 harmonisation process, working on behalf of the working group. E.g." SDO interactions, other HL7 organisational bodies, Steering division,
External roles attend harmonisation meeting for which you are making a submission
Prepare harmonisation proposals
Facilitate communication between work group and external experts and SDO's.
Register OID using HL7 OID registration process (after the domain working group has determined that registration is required.
Action: Ted to update the existing Facilitators page on the wiki to point to the Vocabulary facilitators guide page.
3 Monday Q1
Chair: Jim Case
Scribe: Heather Grain
HL7 use of SNOMED CT under the public good agreement.
Does a warning need to go out to users of HL7 standards that if they are using SNOMED CT for value set content for the international balloted international releases that they may only use extensional definitions.
IHTSDO written clarification on distribution of intentional reference sets is required.
E.g. SNOMED CT concept id of X and all of its subtypes. The definition syntax as defined in the MIF. This is a single xml attribute that says - include decedents. In order to resolve a reference set (to build the content required by using a reference set mechanism).
HL7 may publish the list of concepts and definitions in a transaction and in an artefact.
Clarification is also required from IHTSDO in writing on the use of post coordinated expressions in the public good agreement.
The action list spreadsheet was updated and confirmed and impacts upon the agenda identified. It will be reviewed again at the end of the week and circulated at that time.
Structured documents requirements and ideas about binding have been received from Bob Dolan. This will be discussed further this week with Structured Documents. The action item has been completed as we now have the questions and requirements.
4 Monday Q1
Chair: Ted Klein
Scribe: Heather Grain
4.1 Use of HL7 materials through the IHTSDO workbench
Russ Ham reported on the status of this technical alignment between the HL7 Version 3.0 Vocabulary Model Interchange Format (MIF) and the IHTSDO workbench.
The current state of HL7 vocabulary maintenance tooling is a concern. An option to resolve the numerous problems ion this area is to align vocabulary maintenance tooling with IHTSDO and to map HL7 content.
The approach is to consider data sources and the mapping approach. The mapping tables are quite large.
The meeting considered, but did not resolve whether a warning need to go out that if they are using SNOMED CT for value set content for the international balloted international releases that they may only use extensional definitions, they may not have value sets.
Action: Russ Hamm to provide the relevant URL.
Current state of terminology maintenance tooling:
Put together with a combination of xml (Vocabulary Maintenance Language) which uses a relational database and JAVA APIs.
Additionally there are manual merges of material which cannot be automatically maintained.
There is no current maintenance or upgrade to existing tooling, and this is increasing the number of manual updates to make things happen. This is a high risk strategy.
The current tool requires significant expertise to use accurately and there are only a handful of people on the planet who can do it and it is still prone to human errors.
This project has undertaken a conceptual mapping between the HL7 content and how this would be structured in the Workbench.
Version 2.1.6 of the MIF implementation of HL7 had instance data available and formed the basis of the review. The use of an eConcept was considered. This serves as an idea which has descriptions, annotations and relationships etc in a way that the workbench understands to be a concept - a building block that the workbench can understand or represent.
The project decision has been made to map to eConcepts and not to RF2 as eConcepts offer a better conceptual mapping for HL7 through the Workbench. This also means that it will not be necessary to add SNOMED CT specific information to our existing information. It will be possible to use the workbench to export HL7 content in a representation that is consistent with RF2.
The process that might be used to maintain HL7 content in the workbench was discussed.
Each HL7 primary terminology object is represented as a concept (concept domain, code system, concept, value set, binding and value set assertion)
Mapping focused on representing aHL7 MIF vocabulary elements and attributes as concepts within the IHTSDO workbench.
Auxiliary concepts are essentially a grouper to support terminology components, such as properties. This separates these from the primary terminology objects so that they don't get confused. The diagram below indicates the metadata created to represent HL7 content (as a prototype).
Note: inverse name is the property of a relationship (as each relationship is bidirectional). It was noted that this is not necessarily in the right place in this diagram (which is after all a prototype).
Some of the existing metadata used in HL7 vocabulary processes may not be required as it may have originated as a mechanism to support the current 'broken' elements of our tooling. One of the description types identified is the concept of a description (it was acknowledged that this concept might work better with a different title-e.g. print name).
The workbench offers the opportunity to add descriptive text, which was not able to be maintained in the existing tooling.
Value sets are handled similarly to metadata components. HL7 persists value sets with a definition through a value set definition which is a description of what the content of the value set is supposed to be. There are clauses used in the MIF which define the actual codes in the value set. The review found that the clauses in the MIF map very well to the clauses used in the Workbench to create a reference set. The Workbench actually offers greater functionality that what was available in the existing HL7 tooling.
An HL7 value set has no underlying structure it is just a 'sack' of concepts. The terminology provides the relationships.
There are discussions within IHTSDO as there is not yet a standard for display of relationships and input from the HL7 Vocabulary community on this issue was welcomed by David Markwell.
Migration process
There is a need to develop a MIF export function for the workbench. HL7 would put content in through the workbench, then export from the workbench and compare the two export files to validate the migration. The current requirement is to use both the workbench and the MIF / Rosetree as the existing tool supports publication. The change after migration is that data capture is through the workbench and MIF content is updated through export from the workbench.