control/query meeting minutes (heading 1) web @lantanagroup.com. eric. ... rob mcclure, seconded...

41
Vocabulary Working Group Meeting Minutes September 2014 1 Attendees Given Name Family Name Affiliation eMail Elaine Ayres NHI [email protected] Jos Baptist NICTIZ [email protected] Woody Beeler Beeler Consulting [email protected] Wendy Blumenthal CDC [email protected] Nathan Bunker AIRA [email protected] Jim Case NLM/ NIH [email protected] Mariam Cather [email protected] Hector Cintron ONC Contractor [email protected] om Chris Chute MAYO [email protected] Angelique Cortez [email protected] Ruben Daniels Nehta [email protected] Farrah Darborze ONC [email protected] Scott Deerwester [email protected] Gary Dolin IMO [email protected] Attila Farkas CHI [email protected] Evelyn Gallego ONC Heather Grain eHealth Education [email protected] Russ Hamm Lantana [email protected] Eric Haas APAl [email protected] Rob Hausam Hausam Consulting [email protected] Peter Hendler Kaiser Permanente [email protected] Erin Holt TDH [email protected] Julie James Blue Wave Informatics [email protected]. uk Cho Jan-Chang KWHIS [email protected] Anthony Julian Mayo [email protected] Andrew Kanter IMO [email protected] Sarita Keni Dept Vet Affairs (USA) [email protected] Ted Klein KCI [email protected] Jay Lyle VA [email protected] Rob McClure NLM/ONC [email protected] document.docx

Upload: phamdat

Post on 07-Feb-2018

215 views

Category:

Documents


0 download

TRANSCRIPT

Vocabulary Working GroupMeeting MinutesSeptember 2014

1 AttendeesGiven Name Family Name Affiliation eMail

Elaine Ayres NHI [email protected] Baptist NICTIZ [email protected] Beeler Beeler Consulting [email protected] Blumenthal CDC [email protected] Bunker AIRA [email protected] Case NLM/ NIH [email protected] Cather [email protected] Cintron ONC Contractor [email protected] Chute MAYO [email protected] Cortez [email protected] Daniels Nehta [email protected] Darborze ONC [email protected] Deerwester [email protected] Dolin IMO [email protected] Farkas CHI [email protected] Gallego ONCHeather Grain eHealth Education [email protected] Hamm Lantana [email protected] Haas APAl [email protected] Hausam Hausam Consulting [email protected] Hendler Kaiser Permanente [email protected] Holt TDH [email protected] James Blue Wave Informatics [email protected] Jan-Chang KWHIS [email protected] Julian Mayo [email protected] Kanter IMO [email protected] Keni Dept Vet Affairs (USA) [email protected] Klein KCI [email protected] Lyle VA [email protected] McClure NLM/ONC [email protected] McKinney Frank McKinney Group [email protected] McKnight Seimens [email protected] Macumber Apelon [email protected] Markwell IHTSDO [email protected] Martinez NIST [email protected]

document.docx

Riki Merrick APHL [email protected] Millar IHTSDO [email protected] Moldwin [email protected] Mykkanen HL7 finland [email protected] Narcisi American dental association [email protected] Oemig HL7 Germany [email protected] Nelson LOTS [email protected] Nelson [email protected] Pitkus [email protected] Rashed AJCC [email protected] Roche ONC [email protected] Roberts TN Dept of HealthStefan Sabutsch [email protected] Savage NGC/CDC [email protected] Seabury IHTSDO tse@ihtsdoorgAbdulmalik Shakir Shakir Consulting [email protected] Shah ONC [email protected] Skapik ONC [email protected]

Anne SmithNational Committee for Quality Assurance On phone

Robert Snelick NIST [email protected] SoldbrigSandra Stuart Kaiser [email protected] Taylor NIST [email protected] Unangst [email protected] Vreeman Reigenstrief PhoneMead Walker MWC [email protected] Wang [email protected] Williamson CDC [email protected] Jong Yu HL7 Korea Seungjong.yu@gmail

2 Sunday Q3 Chair: Jim CaseScribe: Heather Grain

2.1 Value Set ProjectThe mechanism for harmonising terminology and quality improvement was discussed. HL7 should be referring to vocabulary content, referencing not creating the vocabulary.

This issue needs to be reinforced within the HL7 community at every opportunity. The HTA could be helpful to assist. The message is that if you need to develop

document.docx

HL7 materials, for clinical stuff we want to stick to existing terminological resources.

FHIR has acknowledged that it is difficult to find and determine the external content required.

There is a need for the HTA to raise the visibility of the Vocabulary Tables project and its phases.

Phase 1 is the quality of content, Phase 2 is merging into a single repository with V3. To undertake this

merge there needs to be a parallel walk through of what should be external content. The HTA would be actively involved in this part of the process. Some of the earlier tables are expected to be difficult and time consuming as the content is of highly inconsistent quality.

2.2 Meeting Activities and AssignmentsChairs and scribes for each session were established and the spreadsheet of Vocabulary agenda was updated accordingly.

Project review: Validate milestones and dates in the time allowed. If incomplete review in the calls.

Action: All open projects will be reviewed on the next conference call agenda.

Tuesday Cross – Standard (all HL7 Product) value set activity was discussed to ensure that the major approaches to null flavours and exceptional values and issues were clear. These issues include: o Need to look at the use caseo Include other, unknown as part of the constrained list of permissible codeso Other cases where a null flavour property would be more appropriate.V2 needs to converge with the V3 CD data type use.

In 2.9 CWE should be retired from 2.10 so that conformance binding and usage are not be mingled together. Alternative representations also need to be addressed. Is a CD like thing going to be able to be used when there are multiple alternative representation? We need to represent the best format for CD in FHIR as a mechanism leveraging improved value set qualification. Related to flavours of null there is a need to convey information about why it is null. Empty reason is the mechanism in FHIR, a mechanism which is supported by the Vocab WG. The objective is to have the same value sets used everywhere. To make it common across all technology value types we need to have multiple value sets – one for the standard values and one for the supplementary values.

document.docx

It is not possible to make all the in convenience go away.

3 Sunday Q4 Chair: Ted KleinScribe: Heather Grain

The Vocabulary Tutorial Framework was reviewed and updated.

Modifications include:Removal of Core Principles, addition of Binding.Addition of a specific tutorial to introduce different terminological resources and assist people to find further information if needed.

Vocabulary 2Russ is giving Vocabulary 2 – he reports that it needs changes and review, still too much material. Hands on examples needed. This review should be undertaken in conjunction with a review of the tutorial specification. January meeting to schedule a quarter on this topic. Also plan to deliver in January.

Topics for inclusion: Use case for using coded datatype (was already covered in Vocab 1), but

refresh – you need these elements – here is where they are) Datatypes in v2 / V3 / CDA / FHIR – includes XML examples Best Practices

Tutorial Competencies for review include: Authoring value sets that are fit for their purpose (suggested but already

covered in authoring value sets) Adheres to good practices (covered in in authoriing value sets) Appropriate use of code systems How to manage IP protected terminologies How to manage the daunting maintenance of evaluating code system

updates and how that affects existing value sets Conformance testing Understand effective vocabulary use for interoperability (use the same

version of underlying code systems – how to recognize this, how to deal with it once known

explain these to decision makersHeather will update the overall plan and draft the first specification for the

Action: Heather to put on January Education timetable (done)Heather to prepare draft update of specification for review on teleconference calls and informed by Russ’ recent delivery.Action: Jim to schedule discussion on Vocabulary 2 tutorial for Sunday Q4 in January. This discussion incorporates consideration of the tutorial framework and specifications of the tutorials in that framework.

document.docx

Heather, 23/09/14,
To be done

A SNOMED CT tutorial was to be given by David Markwell at this meeting but there were insufficient rooms.

New Tutorial added to the Framework: External Terminology Resources. Introduction to commonly used code systems (ComCodes) Develop an Intro class including LOINC – SNOMED CT ICD-10, and 11, WHO classifications and one on ISO units of measure, country codes, Other common sources. The intention is to trial this in January and then to develop short webinars on the general intro and then the specific terminological resources.Skills from this new tutorial:

Process by which a subset of full terminology should be selected How to identify the subset mechanically and how to use this identifier in

connecting it to the terminology model

Audience: any implementer of simple or complex terminology in healthcare systems or those developing HL7 content products.Topics:

Value set and binding in v2 / V3 / CDA / FHIR Common features Known issues Use case

Possibly break up into 4 - 90 min webinars.

Action: Heather to prepare for delivery in January, Ted to assist. Heather has had this included in the Education timetable.

Binding Course – to be developed for MayAction: Heather and Ted to review and prepare specification

Action: Heather to update the Vocabulary Tutorial Framework document

Authoring a value setThis course is already established and includes content describing code system, value set, terminology and classification, and the appropriate design of content Where to go for support is also covered in this tutorial the outcomes of the value set definition project are being incorporated as this project develops.

4 Monday Q3 Value Set Definitions ProjectChair: Ted KleinScribe: Heather Grain

The principal activity to be undertaken at this meeting is comment disposition. There were 16 major negatives, and 54 minor negatives. There were an

document.docx

Heather, 23/09/14,
To be done
Heather, 23/09/14,
To be done

additional 62 minor negatives provided by IHTSDO which are also being resolved.

The intention is to address comments but there is no need to withdraw comments as this is a draft standard. The goal is to turn this into a normative ballot in two years recognising feedback and comments obtained.

Ballot required 50 affirmative votes, 49 in favour votes were received, the hope is to obtain withdrawal of those votes during this week.

Scheduling: Meeting on Tuesdays at 3:30 – 5pm eastern USA time.

Issues 197: Binding of expression language The purpose of the specification seems to want to separate the functions of expression language (not yet defined) and it is difficult to interpret the specification without that expression language.How can we implement if we don’ t have an expression language to use and build a parsing application to create the expansion.Issue (number unknown): Secondary expressionIf a comprehensive view of all things which could be done, then something more robust such as OWL might be needed. Want to definitively say how to do simple things simply that would be helpful while we establish more complex things that will need to be more complicated over time. An extension point could be defined and developed further.

Discussion:Requirement was generally agreed. It was also identified that there should be a single approach to syntax for all terminological resources in a manner which fits the requirements of each resource. There is a need to avoid mapping between syntactic mechanisms and representations. It would be more appropriate to declare the specific syntax to be used in defined instances.

Agreed that: There is a need to identify the syntax being used. MIF could be an interchange format. This means that anything you

do in IHTSDO to craft ref sets using an IHTSDO syntax, if you were to exchange in a VSD compliant manner you would have to map that to the MIF. We were saying that there would be such a syntax specified. We also need to describe a default, which could be based upon MIF, particularly for the simple components. If you have a different syntax you will need to say which syntax is being used. This is the classic restriction of being able to do what you need, and the need for flexibility. An implementation guide could indicate how this is done.

A value set of declared syntaxes may be required.

document.docx

Motion: Rob McClure, seconded Sandy StuartPersuasive with modification.Add a model element to capture the syntax being used. Allow for the use of externally defined syntax such as the IHTSDO query language or owl or others.There can be a MIF based code system default which means support of CLDs which are rendered in a specific syntax such as the upcoming IHTSDO or OWL syntax.Vote: In Favour: 23 Abstain: 1 AgainstCarried

Issue: Organisation / person who makes use of the value set vs. A software module for which it is intended.Need to define the difference between the technical intended use of value set – the technical utility. Another is the intended community activity – quality measure, query, Scope, User and Usage are a group which require clarification. User and Usage

Motion: Rob McClure.Persuasive with modRemove the two subcomponents of “user”, “usage” and make “use” 0 to many, which can be used as required.

When implementing an implementation may use more specific adaptations.Vote: In Favour: 22 Abstain: 0 Against: 1CarriedA question was raised about how a user community might implement this requirement.

5 Monday Q4 Tooling Requirements and Project Status ReviewChair: Jim CaseScribe: Ted Klein

Discussion from Russ on the rewording of the scope of the old IHTSDO Workbench PSS. The title and scope were reworded. Ted identified the tooling currently used for vocabulary authoring and maintenance in the different standards silos. We got hung up on the success criteria. Action: Carol and Russ will complete this updateTed to arrange a slot on an upcoming conference call will be allocated to discuss and approve the updates.

The requirements are being captured in an excel sheet. Jim pointed out that we will need to review those, prior to the January meeting in San Antonio.

document.docx

Heather Grain, 09/16/14,
HL7 considering in the value set project creating information about the data set specification in which a value set is used – Australia already maintains those details in MeteOR though HL7 materical isn’t updated with all HL7 content. This is related to a single repository.

We moved on to the projects in Project Insight. We went through them, as described below.

CTS2N: a few negatives were missed and not yet withdrawn. Russ has reached out to those with negatives to withdraw them and he is working with Lynn on the next step. It is likely that this will be a reaffirmation ballot. Jim requested that Russ update the Project Insight entry with the new status.

Binding syntax is on hold, but we need to proceed. We ran out of resources, it needs to be on hold, but not canceled as we will resume as soon as we can get resources freed up. Action to be done by Ted (ask Dave Hamill).

Project 839 CTS2 Implementation Guide: we have heard nothing. Action: Russ to contact Dave and get the project withdrawn.

Core Principles Project R2 864: contact MnM (Ted) Wednesday about this.Project 874 Tables: active and rolling

Project 945 NLM USCRS evaluation: Action: Ted to contact Dave/Lynn to withdraw this project.

CTS2 SVS incorporation Project 946: Action: Russ to request to withdraw this project, and also follow up with Kevin and Craig

Project 947: new PSS (item 1 above)

Project 948: education, active project under work

Project 1022 HL7 LOINC DOTF: Action for Ted: reach out to Nancy and Russ Ott on this project, the PSS and the fact that it is specified right now as a DSTU and why this still is around or if it as an HL7 project that should be withdrawn.

There are a collection of other projects on our 3 year plan with are just ideas; these can be discussed on future calls.VSD active:

Project 1056: active, but Ted does not yet have a Round Tuit.

SDC FHIR profile project: still active

Project 2.8.2 is still active

The negation project is also still active

document.docx

6 Tuesday Q1 TermInfo Ballot ReconciliationChair: Rob HausamScribe: Heather Grain

Call Schedule: 9am Eastern USA time for ongoing ballot reconciliation progress.

Comment Disposition. Some of the general discussion is captured here, but the details are available in the comment disposition documentation.

Issue: 148/149 We cannot prohibit the use of mood code. Where we have overlapping information in SNOMED CT would be highly restrictive and mean that there are many codes and expressions which could not be used. The guidance is that the content of mood is a ‘restatement’ rather than a combinational meaning. Context should only be added via the vocabulary where it cannot be defined using the information model.

Motion: not persuasive (voting details are in the ballot reconciliation document.

Discussion on the scope of TermInfo occurred. There was considerable discussion on the fact that the focus of the use of SNOMED CT focus is intended, as this version of the document explicitly attempts to resolve SNOMED CT issues across HL7 products knowing that we will move towards the issues associated with LOINC, ICD etc as a next phase. The other issue of changing the focus of content where existing code systems are in place

Clem McDonald has asked for in person discussion and resolution on several of his comments. This meet agenda included such discussion but he was not at the meeting in this quarter. We will leave his issues open until Q2 and Q3. He has also contributed and his position was clarified on conference calls. The feeling of the meeting was that his issues were very real, and well understood and appreciated, however the need to move towards improved practices is known and every effort is being made in this project, and others to support the inherent difficulties known to be associated with such changes. If he is unable to attend any of the other quarters available, the community considered the issue to be non-persuasive. Efforts will be made to contact him to seek his attendance and contribution. Organizational code and value will be influenced by the agreements between IHTSDO, LOINC and HL7.

document.docx

7 Tuesday Q2

7.1 Terminfo DSTU ballot reconciliation

pending

7.2 TermInfo DSTU publishing plan

pending

8 Tuesday Q3 IHTSDO Chair: Jim CaseScribe: Heather Grain

IHTSDO collaboration agreementThis agreement is in the final stages and includes consideration of a with shared work program associated with value sets and SNOMED CT implementation. The 4 areas of the agreement cover:

Licensing clarity associated with the use of SNOMED CT in HL7 artefacts Value Sets – working together to develop terminology content to meet

requirements, development of methods to work on the quality of content, binding and metadata requirements. HTA is the lead and single point of connection

Tooling – use of IHTSDO tooling to support HL7 content where appropriate.

Education – IHTSDO will provide support and oversight of SNOMED CT related educational materials and provide education activities for HL7.

HL7 publications will be required to include licensing statements for SNOMED CT content included in HL7 artefacts and guidance is being developed based upon a range of use cases to clarify use of SNOMED CT content and the relationship of that use to licensing requirements. The public good agreement for the use of SNOMED CT is associated with this process but is governed by IHTSDO licensing agreements and arrangements and is not directly in the purview of HL7.

Tooling:Combined content management system and request management system, and improved browser are the major new initiatives.Details are available at http://www.ihtsdo.org/develop/tooling/

Content change request. IHTSDO is moving to a request submission tool using Jera to support a maintainable tool process. Browser development will provide core and extension accessibility. This allows you to view all of the content developed by different countries, but also who has

document.docx

developed the content. The intention is to assist identification of national extension overlap and potential candidates for the international core.

A replacement for the IHTSDO workbench is also in development with better integration of request submission and editing management.The browser can be seen at: http://browser.ihtsdotools.org/

Licensing download: Salsa – the old one (SNOMED CT affiliate licensing service agreement) a new internet based approach is being developed.

Continued minimal support for the workbench will be provided. The old tooling will be retained. Any tool may be used to manage your extension. For countries which do not have the resources to support content development IHTSDO will provide those services. Providing this service is highly labour intensive for IHTSDO and the desire is to provide tools which are easier to use and leverage for specific functionalities.

Family of languages – computer processable languages.

Update is being undertaken to the compositional grammar to add features which allow constraints. A query language is being developed to allow you to specify the members of a construct

Aligned with expressions, syntax etc. You can say I want concepts that have an finding site attribute of foot and a causative agent of streptococcus – and generate a set based upon this. You may identify your starting point and build a set based upon the core, plus x.

Also to turn the hierarchy into classifications with alternative linearization rules to provide a similar syntax but alternative branches depending upon a specific use case.

The UML model for the overall structure is an ABNF representation of the syntax. You can represent it as being a model or use a different syntax for a specific purpose if needed. There is also a link back to the tooling you need to be able to demonstrate this functionality.

The querying work will need to go through the organisations processes which will take some time, but is likely to progress and be available next year, hopefully the first half of next year. Member countries could adopt this into their processes.

Compositional grammar update is likely to be available this year. A draft of this package of activities (generally referred to as the family of languages).

Query language specifications have been produced before – this work is no longer current and was not accepted by the organisation as it was not consistent

document.docx

with other things done previously. DO NOT USE THIS. The new work has internal consistency across the range of functional requirements for the ‘family of languages’.

Request that IHTSDO should list documents that have been withdrawn.

The starter guide is ‘published’ and is under normal publication review.

Benefits realization work is being undertaken to more effectively quantify the evidence and utility of SNOMED CT in different environments.

Starter Guide – status – differentiates between different status could be of value – it would be useful to know the circulation/availability and review cycles to support accurate use. Status includes; In Review, Search and data entry guide – is considered better than what is in the TIG is to formally accepted at this point in time. The process of document management is being considered at the moment.

Alergy and Reaction value sets submitted through Rob Hausam. Elaine is also an interested party. Patient Care have been developing models for allergy and intolerance. The DAM has been balloted and are now ready to consider the terminology to support the information model. There are specific value sets needed to support that modelling, some is not available yet.

Procedurally:Vocabulary should give guidance on the relevant terminological resource (code system) for use in a specific environment.

9 Tuesday Q4

10 Tuesday Q4 Cross HL7 Product value setsChair: Rob HausamScribe: Heather Grain

Implementation Guides for CDA and FHIR currently have conflicting value sets.Value set for both guides with some exceptional values (no information.... etc) we need some but not all of the V3 exceptional values.

Alternative 1: The V2 value sets have exceptional values in them but the V3 will not have these values in them.

Alternative 2: Have the same value sets in both and include the null values in the value sets. That includes including no information in the code value in v3 which breaks the rule of observation value as opposed to observation code.

document.docx

Heather Grain, 23/09/14,
Establish a specific task list for review of value set content

Alternative 3: Same set of values for both V2 and V3 but in the guide to V3 we put in a note that says that if you have one of these exceptional values put it in the null flavour

Whatever answer for V3 must also be used with CDA. FHIR is considering having both options. Use of ‘empty reason’ in FHIR covers this. A data empty reason is being considered for the other products.

Implementers don’t want to maintain two value sets. Heather identified that the use of a consistent approach to this across products and moving towards this consistency may benefit vendors in the longer term.

Other – it was agreed that there is a need to discuss other, but that this is not the focus of this specific problem. It was agreed that Other is not really a ‘null flavour’.

If choosing null flavours which are needed / not needed this is often not undertaken consistently or with clearly defined rationale or requirements.

Suggestion from Rob H. To have the basic value set containing non exceptional values and what ever set of additional exceptional values (unknowns....) in a separate value set – for those products you identify a value set which is a defined aggregate value set.

Functional value sets a)b)c)

Exceptional value sets123

For item B (a + 1)

In V3 this would mean that the exceptional values would need to be constrained. In V2 there is no mechanism to represent aggregated value set content. Today this would be handled by publishing into a single table, but in the future this would need to be able to be handled by value sets and a union statement.

Given the way the models work at the moment, correct but not too onerous

document.docx

Alternative 3: Same set of values for both V2 and V3 but in the guide to V3 we put in a note that says that if you have one of these exceptional values put it in the null flavour and that it use HL7 null flavour code system.

Question: which code system is ‘other’ coming from? If you have other in your code system and you add a new concept the value of other has changed – you have a new code system.

If value sets were versioned, would this manage this problem?

Straw Poll:

Alternative 1: two separate value sets – exceptional will be in the V2 but not the V3. 4 votesAlternative 2: same value set including exceptional values and put all the values in the same value set in V2 and V3.No votes

Alternative 3: same values used in v2 and v3 but the v3 implementation indicates that these are in the null flavour part.8 votes

Abstain: 5

Alternative 4: produce aggregate value sets made from the two separate components. Considered after the vote.

Another Alternative: Assume we are doing things according to some higher level and the rules allow us to do what is in 3 when you ask for the value set in V2 you get V3 constructed from that higher level specification. The use of option 3 would move toward this solution.

Option 3 is the lesser of most evils

Motion: AMS, Michelle.Request Vocabulary does not oppose the implementation strategy (Alternative 3).

Alternative 3: Same set of values for both V2 and V3 but in the guide to V3 we put in a note that says that if you have one of these exceptional values put it in the null flavour and that it use HL7 null flavour code system.

Discussion: How would we express only these null values. This will require Vocabulary to consider the impact and implications of this approach.

document.docx

In Favour: 13Against: 0Abstain: 4

Action – Conference Call- consider the issue of the values of flavours of null – Jim

Vocab to review null flavor value set members – especially otherV2 does not have capabilities to support hierarchy, and one cannot create a hierarchy when the underlying code system does not support a hierarchy, such as Trauma in CDA implementation guide which is waiting for LOINCs to complete.

Action: Provide details to HTA to support their followup

document.docx

11 Wednesday – Meetings held with other Work Groups – non hosted by Vocabulary

12 Thursday Q1

13 Value Sets Pilot test of the aspects which need to be done. We are trying to put in place

o When people want to put a constrained set together in a value set we are seeking that

Someone with the HTA can see if it is appropriate match

Include in the process for new items o Implement in SNOMED CT rather than

just making up new content Where estisting things have been used for

‘decades’o Jim has looked at specimen type – was

abandoned because it was an intractable process

Transition plan –o Focus to sort out the ones that hae been

around a long time to sort them out.o Request processo Development of a worklist process

Currently in test cases Family member Another case needed with different types Establish workplan and priorities

Then move into workplan Project scope statement approval process

o Vocabulary facilitator Requirement to identify the (review V2.9 requires a vocab facilitator How do we find the value sets being updated through

the PSS process Indicate individual value sets, or families of

vocabulary content. FHIR have approach directly – and IHTSDO – use of SNOMED CT

– that’s fine.o Don’t send the document to IHTSDO but send it first to the

HTA. Updates to project insight.

document.docx

Heather Grain, 17/09/14,
Include this

Update agenda for Friday.

Value Set issuesTo Heather, Jane, jDavid,

We could pass quality assurance activities to IHTSDO. Where value sets are broader than HL7 considerrations.

We need timeframes on required answers.

Are there key principles we can share to improve quality of decisions.

Use cases that lead to the process and guidance principles for authoring value sets.

14 Tuesday Q3 IHTSDO Chair: Jim CaseScribe: Heather Grain

IHTSDO collaboration agreementThis agreement is in the final stages of agreement, with shared work programs associated with value sets and SNOMED CT implementation. The 4 areas of the agreement are:

Licensing clarity associated with the use of SNOMED CT in HL7 artefacts Value Sets – working together to develop terminology content to meet

requirements, development of methods to work on the quality of content, binding and metadata requirements. HTA is the lead and single point of connection

Tooling – use of IHTSDO tooling to support HL7 content where appropriate.

Education – IHTSDO will provide support and oversight of SNOMED CT related educational materials and provide actual education activities for HL7.

HL7 publications will be required to include licensing statements for SNOMED CT content included in HL7 artefacts and guidance is being developed based upon a range of use cases to clarify use of SNOMED CT content and the relationship of that use to licensing requirements. The public good agreement for the use of SNOMED CT is associated with this process but is governed by IHTSDO licensing agreements and arrangements and is not directly in the purview of HL7.

Tooling:Combined content management system and request management system, and improved browser are the major new initiatives.

document.docx

Details are available at http://www.ihtsdo.org/develop/tooling/

Content change request. IHTSDO is moving to a request submission tool using Jera to support a maintainable tool process. Browser development will provide core and extension accessibility. This allows you to view all of the content developed by different countries, but also who has developed the content. The intention is to assist identification of national extension overlap and potential candidates for the international core.

A replacement for the IHTSDO workbench is also in development with better integration of request submission and editing management.The browser can be seen at: http://browser.ihtsdotools.org/

Licensing download: Salsa – the old one (SNOMED CT affiliate licensing service agreement) a new internet based approach is being developed.

Continued minimal support for the workbench will be provided. The old tooling will be retained. Any tool may be used to manage your extension. For countries which do not have the resources to support content development IHTSDO will provide those services. Providing this service is highly labour intensive for IHTSDO and the desire is to provide tools which are easier to use and leverage for specific functionalities.

Family of languages – computer processable languages.

Update is being undertaken to the compositional grammar to add features which allow constraints. A query language is being developed to allow you to specify the members of a construct

Aligned with expressions, syntax etc. You can say I want concepts that have an finding site attribute of foot and a causative agent of streptococcus – and generate a set based upon this. You may identify your starting point and build a set based upon the core, plus x.

Also to turn the hierarchy into classifications with alternative linearization rules to provide a similar syntax but alternative branches depending upon a specific use case.

The UML model for the overall structure is an ABNF representation of the syntax. You can represent it as being a model or use a different syntax for a specific purpose if needed. There is also a link back to the tooling you need to be able to demonstrate this functionality.

The querying work will need to go through the organisations processes which will take some time, but is likely to progress and be available next year, hopefully the first half of next year. Member countries could adopt this into their processes.

document.docx

Compositional grammar update is likely to be available this year. A draft of this package of activities (generally referred to as the family of languages).

Query language specifications have been produced before – this work is no longer current and was not accepted by the organisation as it was not consistent with other things done previously. DO NOT USE THIS. The new work has internal consistency across the range of functional requirements for the ‘family of languages’.

Request that IHTSDO should list documents that have been withdrawn.

The starter guide is ‘published’ and is under normal publication review.

Benefits realization work is being undertaken to more effectively quantify the evidence and utility of SNOMED CT in different environments.

Starter Guide – status – differentiate between different status could be of value – it would be useful to know the circulation/availability and review cycles to support accurate use. Status includes; In Review, Search and data entry guide – is considered better than what is in the TIG is to formally accepted at this point in time. The process of document management is being considered at the moment.

Alergy and Reaction value sets submitted through Rob Hausam. Elaine is also an interested party. Patient care have been developing models for allergy and intolerance. The DAM has been balloted and are now ready to consider the terminology to support the information model. There are specific value set sneeded to support that modelling, some not yet available.

Procedurally:Vocabulary should give guidance on the relevant terminological resource (code system) for use in a specific environment.

15 Tuesday Q3 Cross HL7 Product value setsChair: Rob HausamScribe: Heather Grain

Implementation Guides for CDA and FHIR currently have conflicting value sets.Value set for both guides with some exceptional values (no information.... etc) we need some but not all of the V3 exceptional values.

Alternative 1: The V2 value sets have exceptional values in them but the V3 will not have these values in them.

document.docx

Alternative 2: Have the same value sets in both and include the null values in the value sets. That includes including no information in the code value in v3 which breaks the rule of observation value as opposed to observation code.

Alternative 3: Same set of values for both V2 and V3 but in the guide to V3 we put in a note that says that if you have one of these exceptional values put it in the null flavour

Whatever answer for V3 must also be used with CDA. FHIR is considering having both options. Use of ‘empty reason’ in FHIR covers this. A data empty reason is being considered for the other products.

Implementers don’t want to maintain two value sets. Heather identified that the use of a consistent approach to this across products and moving towards this consistency may benefit vendors in the longer term.

Other – it was agreed that there is a need to discuss other, but that this is not the focus of this specific problem. It was agreed that Other is not really a ‘null flavour’.

If choosing null flavours which are needed / not needed this is often not undertaken consistently or with clearly defined rationale or requirements.

Suggestion from Rob H. To have the basic value set containing non exceptional values and what ever set of additional exceptional values (unknowns....) in a separate value set – for those products you identify a value set which is a defined aggregate value set.

Functional value sets a)b)c)

Exceptional value sets123

For item B (a + 1)

In V3 this would mean that the exceptional values would need to be constrained. In V2 there is no mechanism to represent aggregated value set content. Today this would be handled by publishing into a single table, but in the future this would need to be able to be handled by value sets and a union statement.

Given the way the models work at the moment, correct but not too onerous

document.docx

Alternative 3: Same set of values for both V2 and V3 but in the guide to V3 we put in a note that says that if you have one of these exceptional values put it in the null flavour and that it use HL7 null flavour code system.

Question: which code system is ‘other’ coming from? If you have other in your code system and you add a new concept the value of other has changed – you have a new code system.

If value sets were versioned, would this manage this problem?

Poll:

Alternative 1: two separate value sets – exceptional will be in the V2 but not the V3.4 votesAlternative 2: same value set including exceptional values and put all the values in the same value set in V2 and V3.No votes

Alternative 3: same values used in v2 and v3 but the v3 implementation indicate sthat these are in the null flavour part.8 votes

Alternative 4: produce aggregate value sets made from the two separate components.

Another Alternative: Assume we are doing things according to some higher level and the rules allow us to do what is in 3 when you ask for the value set in V2 you get V3 constructed from that higher level specification. The use of option 3 would move toward this solution.

Option 3 is the lesser of most evils

Motion: AMS, Michelle.Request Vocabulary does not oppose the implementation strategy (Alternative 3). Alternative 3: Same set of values for both V2 and V3 but in the guide to V3 we put in a note that says that if you have one of these exceptional values put it in the null flavour and that it use HL7 null flavour code system.

Discussion: How would we express only these null values. This will require Vocabulary to consider the impact and implications of this approach.

18 in the room – 1

document.docx

Heather Grain, 23/09/14,
Check riki’s numbers

Against: 0Abstain: 4Yes: 13

Action – Conference Call- consider the eissue of the values of flavours of null – JimOther: e.g. other diabetes, iwould be other if it is a list of diabetes, then other diabetes is

16 Tuesday Q4 Cross HL7 Product value setsChair: Rob HausamScribe: Heather Grain

Implementation Guides for CDA and FHIR currently have conflicting value sets.Value set for both guides with some exceptional values (no information.... etc) we need some but not all of the V3 exceptional values.

Alternative 1: The V2 value sets have exceptional values in them but the V3 will not have these values in them.

Alternative 2: Have the same value sets in both and include the null values in the value sets. That includes including no information in the code value in v3 which breaks the rule of observation value as opposed to observation code.

Alternative 3: Same set of values for both V2 and V3 but in the guide to V3 we put in a note that says that if you have one of these exceptional values put it in the null flavour

Whatever answer for V3 must also be used with CDA. FHIR is considering having both options. Use of ‘empty reason’ in FHIR covers this. A data empty reason is being considered for the other products.

Implementers don’t want to maintain two value sets. Heather identified that the use of a consistent approach to this across products and moving towards this consistency may benefit vendors in the longer term.

Other – it was agreed that there is a need to discuss other, but that this is not the focus of this specific problem. It was agreed that Other is not really a ‘null flavour’.

If choosing null flavours which are needed / not needed this is often not undertaken consistently or with clearly defined rationale or requirements.

document.docx

Suggestion from Rob H. To have the basic value set containing non exceptional values and what ever set of additional exceptional values (unknowns....) in a separate value set – for those products you identify a value set which is a defined aggregate value set.

Functional value sets a)b)c)

Exceptional value sets123

For item B (a + 1)

In V3 this would mean that the exceptional values would need to be constrained. In V2 there is no mechanism to represent aggregated value set content. Today this would be handled by publishing into a single table, but in the future this would need to be able to be handled by value sets and a union statement.

Given the way the models work at the moment, correct but not too onerous

Alternative 3: Same set of values for both V2 and V3 but in the guide to V3 we put in a note that says that if you have one of these exceptional values put it in the null flavour and that it use HL7 null flavour code system.

Question: which code system is ‘other’ coming from? If you have other in your code system and you add a new concept the value of other has changed – you have a new code system.

If value sets were versioned, would this manage this problem?

Straw Poll:

Alternative 1: two separate value sets – exceptional will be in the V2 but not the V3. 4 votesAlternative 2: same value set including exceptional values and put all the values in the same value set in V2 and V3.No votes

Alternative 3: same values used in v2 and v3 but the v3 implementation indicates that these are in the null flavour part.8 votes

document.docx

Abstain: 5

Alternative 4: produce aggregate value sets made from the two separate components. Considered after the vote.

Another Alternative: Assume we are doing things according to some higher level and the rules allow us to do what is in 3 when you ask for the value set in V2 you get V3 constructed from that higher level specification. The use of option 3 would move toward this solution.

Option 3 is the lesser of most evils

Motion: AMS, Michelle.Request Vocabulary does not oppose the implementation strategy (Alternative 3).

Alternative 3: Same set of values for both V2 and V3 but in the guide to V3 we put in a note that says that if you have one of these exceptional values put it in the null flavour and that it use HL7 null flavour code system.

Discussion: How would we express only these null values. This will require Vocabulary to consider the impact and implications of this approach.

In Favour: 13Against: 0Abstain: 4

Action – Conference Call- consider the issue of the values of flavours of null – Jim

Vocab to review null flavor value set members – especially otherV2 does not have capabilities to support hierarchy, and one cannot create a hierarchy when the underlying code system does not support a hierarchy, such as Trauma in CDA implementation guide which is waiting for LOINCs to complete.

Action: Provide details to HTA to support their followup

document.docx

17 Wednesday Q1 and Q2 with MnMSee notes from MnM

18 Wednesday Q3 and Q4 in meetings hosted by Publishing and Tooling

document.docx

19 Thursday Q1 conformance

19.1 CNE vs. CWE

Ted described the objective of bringing together the artefacts and value sets and their representation across HL7 products and the difficulties and confusion which arises between CNE and CWE. The suggestion is that conformance is defined in conformance guides and assertion of ‘exactly this set’ and only this set. This should be specified in the conformance rules in the implementation guide. By putting conformance in the international specification above the level of implementation guide leads to people breaking the rules because they must. Vocab are considering removing CNE as a data type and identify how to accomplish the requirements in implementation guide conformance statements.

Questions:Extensibility rules are associated with the value set. The extensibility rules also exist in the Data Set specification. There could be dissonance between the these.

Simple coded element and a complex data element is one potential to handle this problem. Data type flavours such as these would cover this problem. This would be applied by specifying the data type which define the requirements on the data structure itself. E.g.: first element is required, in another it could be conditional, CNE talks about whether the value set is extensible.

Motion: Rob Snelick, Rob HausamReplace data types IS and ID with a data type for simple coded element (which might be called), and replace CNE, CWE, and CE with Complex coded element (which might be called CEC) and to extend these with data type flavours.

Discussion:It will be incumbent upon the conformance group to document the constraining process for implementation of a V2 interface. Vocabulary, MnM and Conformance will need to work through the level of constraint rules needed, and where do they best fit. There is also a need to improve how we educate people. This will need to be added to the education framework. This change would target V2.10Publishing also needs to consider updates to general guidance for V2 to include these concepts to recognise that those using the older versions of V2 mean

Vote: For 11 Abstain 0 Against 0

This project will be lead by Conformance CGIT with co sponsors of Vocabulary and InM

document.docx

Action: Tony Julian will develop the PSS. Ted will get this on Vocab calls when ready.

19.2 Binding Discussion

Work occurring to clarify binding syntax in V2 and specifications in other products. In conjunction with the first motion Rob Snelick offered to work on how we connect the binding requirement to converge the different ways binding is done across the families. It was agreed that the first step would be to collect the current mechanism details for each product.

Action: Ted to schedule a call on this topic through doodle poll aiming for a meeting in November after the background documentation has been gathered. This call should go to the Vocab and CGIT lists and the InM Chairs. Once the time is established the Vocab and CGIT lists will be informed of the meeting.

Value set project – must finish identification of problems, receive fixes back from work groups (who are told of these issues and asked to address them). Consider fixes required for earlier versions than 2.9.

Phase 2 assessment is to ensure correct analysis and validate decisions from phase 1. This work should be completed within the next 3 months if regular calls continue. Stage 3 is the harmonisation across products and it is intended that the value set project with relevant OIDS all be created by October 2015.

Action: Ted to find the place in the V2 spec that sais to use a specific OID type This needs to be found and removed.

20 Thursday Q2

20.1 Glossary Loading to SKMTHeather Presented the PSS which was reviewed and agreed.

Motion: Jim Case, Heather GrainApproval of the updated PSS to be sent to Tooling. Vote: Yes 9, Abstention 0, Against 0

Note: the PSS has been sent to tooling.

document.docx

20.2 Value Set Content

A table may be a concept domain only, which has no suggested values.A value set with a concept domainOrA value set drawn from an external vocabulary (eg country codes – use the 3 digit values of the ISO country codes).

Where content is published in HL7 standards, with a few exceptions, these are treated s a code system in the HL7 vocabulary model and we populate these values in the code system and will progressively create the required OIDs.

The end of this project will publish the code system definitions, value sets and OIDs in version 2.9. Vocabulary will no longer maintain the values manually. They will be rendered only in chapter 2C and will be generated out of the database ONLY and that database will be the single source of truth.

The general approach to V2 tables which reference external code systemsRules:

If all codes from the code systemWe will define the value set definition – the table represents a value set only. This will apply to 141 Military rank and grade and ISO country codes.

External code system – in these cases – still create a value set identifier, describe the intended definition, and provide a URL that points to the external code system.

Is it possible for folks to go to HL7 and get any codes from that code system?

For intensional definitions Hl7 policy is that we do not publish value set expansions, but building the definition

HL7 will not provide reference to external value set content. It is not that it is wrong for us to do it, we need an infrastructure to support this decision which we don’t currently have. We should not provide access to external system content unless we have a fully vetted approach to support conformance with access to the content which is conformant with the source owner license.

For items which are blank – table stewards (table editors in work groups) are responsible for updates.In many cases V2 content was simply imported to V3. V3 harmonization requires this information is not left blank. The V3 usage and definitions may already exist

document.docx

in V3 code systems that could be used to provide ‘starting point’ data to assist development for V2 and reduce the workload.The tables process must discuss a process for the requirement where there is no alignment with what is incorporated in other HL7 products, but the intent is to harmonize. V3 ballot and CDA (hosted in Trifolia) are known to hold potentially conflicting value set content.

Part 1 aligns models and identifiers, content is the next phase through the work groups. Critical part of alignment is code system alignment.

Action: Request of HTA: Can we look at some of these fundamental content issues.

There is not only a need to define the code system, but also where to find the content. In FHIR, V3 each have different places, different model formats, but same technology. V2 and CDA have different places, model formats and different technologies. There is a desire to manage this tooling to make content available to the community.

Activities: Cleaning up databases Request for tooling from HTA to be able to obtain a list of truth from each

product.o V2 source of truth is still the MS Word drafts

There is a set of tools that exports MS Access content into the MIF in the V3 repository but the database model is different

As a user for VSAC 2.1 source, Frank O has been progressively correcting identified errors in the database so the best quality V2 vocabulary is in Franks database coming out of th tables project. The database is not in sync with the standards (it is consistent but not the same as the published standards).

Action:Request HTA to consider how to obtain funding to clean up our act across our products. – Rob McClure to introduce the discussion.

Suggest meeting with tooling on this topic to consider requirements and definition of management. Key people: Jim, Wendy, Frank, Ted plus others. Travel is a more significant issue – online is acceptable but would need commitment from employers that key individuals would be available/consultants receive some funding to allow rapid progress and the specification of the next requirements.

There is also a need for support to undertake the manual review of the process – this requires a group of 4 – 6 people working together to progress this activity.

document.docx

Support – venue for face to face time before the next HL7 meeting – on this topic – and request key people to be present. Include funding for contractors who are the key people, and subsequent request to support their attendance and then do the work defined during the face to face meeting. The meeting would need to specify the estimated time requirements.

It is currently estimated that this activity could be completed with 20 hours of work. To support meaningful use the completion of this 20 hours AND the specification of step 2 requirements (time and resources) needs to be defined, and funded and DONE by February 2015 to meet meaningful use requirements.

21 Thursday Q3

21.1 International Patient Summary PSSpending

21.2 ONC SDC/FHIR profile IG for commentpending

22 Thursday Q4

22.1 Workgroup planning: Agenda planning and action items, WG health items

pending

document.docx