MnM Minutes WGM 201305 Atlanta

Return to MnM Minutes for 2013

Sunday May 5 Q3

Agenda

• Hot topic triage
• Agenda planning

Attendees

Beeler Jr, George (Woody) (Beeler Consulting LLC)

Grieve, Grahame (Health Intersections Pty Ltd)

Hay, David (HL7 New Zealand)

Kramer, Ewout (Furore)

Loyd, Patrick (ICode Solutions)

McKenzie, Lloyd (HL7 Canada)

Shakir, AbdulMalik (City of Hope National Medical Center)

Stechishin, Andy (CANA Software & Services Ltd.)

Walker, D. Mead (Mead Walker Consulting)

Hot topics

1. RIM ballot reconciliation
   • Fixed values for collections of repeating attributes
   • What's the RIM default
   • How do we make it work in the schema generator

2. Lloyd will close the "social media" hot topic

3. Discussed "Model to Template Relationship", but no need to bring it up as a hot topic

4. Data types vocabulary schemas - what are we allowed to do within the framework of the ISO spec

5. FHIR RIM mappings
   • What will RIM mapping look like?
   • What do we do when the RIM can't do it?
   • What will it achieve?

Will tackle #1 and #4 in this meeting, #5 slotted for joint FHIR/MnM discussion

For Wed Q3, will host FHIR (will release room 219)

Release rooms for Tue Q1, Tue Q4

Move SAIF Artifact Definition Project from Tuesday Q4 to Thursday Q2

Motion: approve agenda for WGM - Woody/Patrick

RIM ballot reconciliation

Two Negatives from Dale Nelson

Findings documented in Reconciliation spread sheet. Vote: Not persuasive 7-0-1 (Woody/Grahame)

Beeler's Negative regarding blocked context conduction attributes

Steps to resolve:

• We should create value sets containing conductable ActRelationship type codes
• Designers can create a new value set that constrains the base value sets as per the rules in Core principles
• Designers can create a fixed value that is a "list of codes" that draws from the codes in the value set
• Tooling needs to verify this will propagate properly through the tools to the schemas
• MnM will document this in Core principles

Motion on these steps: Woody/Grahame 5-0-0

Woody takes action item to update Core Principles

Based on the above, Woody withdraws his negatives

Data type vocabulary schemas

Discussed issues that arise from mismatch of vocabulary terms listed in Data Types Specifications; those assembled into vocabulary schema; and those in Vocabulary content. Agreed on a strategy for isolating these and correcting them. Beeler will work with Grieve to accomplish this.

Monday May 6 Q1

Agenda

Chair - Beeler & Grieve

• FHIR Methodology

Attendees

Abrahao, Marivan (HL7 Brazil)

Beeler Jr, George (Woody) (Beeler Consulting LLC)

Cabrita, Rolim (Interfaceware)

Duteau, Jean (Duteau Design Inc)

Fine, Steven (Cerner Corporation)

Glinski, Steve (Surescripts)

Grieve, Grahame (Health Intersections Pty Ltd)

Henderson, Mike (Eastern Informatics, Inc.)

Jewell, Gabriele (Cerner Corporation)

Konishi, Yukinori (HL7 Japan)

Liu, Andrew (HL7 Canada)

Lynch, Cecil (Accenture)

Moehrke, John (GE Healthcare)

Parker, Ron (HL7 Canada)

Ryal, Tod (Cerner Corporation)

Shaver, Dave (Corepoint Health)

Smithies, Rik (HL7 UK)

Takasaka, Sadamu (HL7 Japan)

Walker, D. Mead (Mead Walker Consulting)

Wong, Nat (HL7 Australia)

Worden, Robert (Open Mapping Software)

Kavanagh, Richard (NHS England)

Walia, Randeep (Grafight Scratch, Inc.)

Murty, Kurella (Roche)

Topic: mustUnderstand

Discussion of mustUnderstand - Motion from G Grieve/Rik Smithies to "Rename mustUnderstand to isModifier, and add sentence saying that you must explain why it is that you think it is a modifier and/or what it modifies."

Approved 25-0-0

Topic: Various

Discussions around "messaging-like" paradigm with FHUR similar to V2, can be defined and has been tested.

Discussions on "Profile" resource and what is does. Is rather larger than desired. Was noted that although it is defined as a Resource, it is unlikely to be exchanged as instances.

Security and FHIR

There is a Wiki document in the "FHIR Publication" that is being built with the guidance of Security Work Group lists security-related issues when using FHIR, and how and where these might be addressed.

Monday May 6 Q2

Ewout Kramer and Woody Beeler for FHIR and MnM

Agenda

• Host FHIR
• FHIR Methodology

Attendees

Beeler Jr, George (Woody) (Beeler Consulting LLC)

Brancato, Chris

Cheung, James (Kaiser Permanente)

Fine, Steven (Cerner Corporation)

Greim MS RN-C, Patricia (U.S. Department of Veterans Affairs)

Jewell, Gabriele (Cerner Corporation)

Kavanagh, Richard (NHS England)

Kramer, Ewout (Furore)

Liu, Andrew (HL7 Canada)

Lynch, Cecil (Accenture)

Moehrke, John (GE Healthcare)

Ott, Russell

Parker, Craig (Intermountain Healthcare)

Pupo, Erik (Deloitte Consulting LLP)

Ryal, Tod (Cerner Corporation)

Shaver, Dave (Corepoint Health)

Smithies, Rik (HL7 UK)

Walia, Randeep (Grafight Scratch, Inc.)

Walker, D. Mead (Mead Walker Consulting)

Wong, Nat (HL7 Australia)

Methodology Issues that Came out of QA Assessment of FHIR

Distinction between "Type" and "Code" and Other field names

Rationale for discussion: "type" and "code" have been used as attribute names for what is likely to be an encoded result.

Discussion suggests that: "code" commonly adds NO understanding; should not repeat name of element in attribute; type works where it is the type of the element; code is not useful; if need further distinction and expand "type"

Similarly, "date" as an attribute name will beg the distinction of which kind of "date" (e.g. see V3 Act) is intended.

Usage of Terms like "Subject", "Author"

There were notes about inconsistencies of use of these terms. Discussion here suggests that:

1. Whatever definition for a term like "author" (for example) that FHIR selects should be consistent throughout FHIR
2. A Glossary would be very difficult (these tend to be more fine-grained than the usual glossary)
3. The RIM vocabulary (participation type codes, for example) provide fairly carefully thought out (and reviewed) set of definitions for things like "author", etc.

Suggest that FHIR consider adopting upper and lower camel case for names - lower for parameters and attributes, upper for resources and elements; etc.

Also the uniqueness criteria for names should be case agnostic. Thus "LastMan" and "lastMan" are not unique.

Discussed McKenzie Discussion of "Subject"

Appears focused on context conduction and the need to clearly state what if ANY assumptions can be made thereabout. What if a contained object has no subject??

Discussion recognized that FHIR's light-weight reference mechanism makes it reasonable to represent the subject in (for example) each member of a battery of tests.

Subsequent Summary by E. Kramer on this discussion and the topic above it

What the notes don't reflect is that there was no clear consensus about the use of subject/author etc. Most agreed that it was not necessarily a good idea to use them everywhere, because e.g. Order.subject is ambiguous.

There was concern that because of the v2/v3 background, users of the spec would assume they understood what "subject" and "author" means, though in FHIR the meaning of that word may change from resource to resource.

So, the consensus was that it should be used consistent across all resources if it is used on occasion, but that it is not necessarily a good idea to try to use these terms in all resources.

Discssion of mustUnderstand on optional Booleans

The line item 6 from McKenzie, is moot if all mustUnderstand have minimum cardinality of 1, which is currently on the "warning" track for becoming a rule.

Monday May 6 Q3 and Q4

Agenda

• Join Joint w/FHIR, ITS
• FHIR Infrastructure issues

Minutes by ITS

Tuesday May 7 Q2

Agenda

• "Ask MnM"

Attendees

Beeler Jr, George (Woody) (Beeler Consulting LLC)

Luensman, Genny

Shakir, AbdulMalik (City of Hope National Medical Center)

Occupational Data for Health

Discussed how to assure that appropriate data about patient's occupations and occupation history are captured and retrievable within electronic health records. Started with a data model (entity-relationship diagram) titled "Occupational Data for Health" from Genny Luensman of CDC/NIOSH/DRDS/SB.

While the model is readily representable in RIM semantics, the core question is how to assure these data are considered in profiles being on Consolidated CDA for consideration in Meaningful Use. The discussion was useful, the proponents are currently engaged with projects providing data requirements for meaningful use and therefore, at this point, do not need further help with regards to RIM and or methodology.

Tuesday May 7 Q3

Joint w/ Structure Documents (refer to their minutes)

Wednesday May 8 Q1

Agenda

• Host Vocab

Attendees

Beeler Jr, George (Woody) (Beeler Consulting LLC)

Case, James (National Library of Medicine)

Couderc, Carmela (Siemens Healthcare)

Hamm, Russell

Hausam, Robert (Hausam Consulting)

Huang, Wendy (Canada Health Infoway Inc.)

Klein, William Ted (Klein Consulting, Inc.)

Liu, Andrew (HL7 Canada)

McKenzie, Lloyd (HL7 Canada)

Representation of Human Language ConceptDomain

Problem is that the ConceptDomain HumanLanguage is bound in the RIM and data types. The Data Types binding causes a Universal binding to a ValueSet (from the ConceptDomain referenced in Abstract Data Types). There is a need presented to have the data type for RIM languageCode be CD (or CE) which would require a different binding.

Recognize that a different Concept Domain could be used for the RIM binding of languageCode in RIM, that then might be used for a different binding in a realm.

Challenge is that we are seeking to relax a ling-standing binding that is usable, except that the proposal seeks to use only 3-character codes that is counter to the IETF requirement that communication use the 2-character whenever a single concept is represented by both a 2-character and 3-character code.

Alternate concern is that we are taking a CD data type and forcing all realms to use a single binding. Note that this is the ONLY RIM attribute of type CD that is universally bound, as a result of the happenstance of the same concept domain being used in Abstract Data Types.

Resolution

Motion made by Huang, seconded by Klein that a harmonization proposal be prepared to:

Create a new Concept Domain be used as the RIM constraint on attribute LanguageCommunication.languageCode, including a definition of that concept domain; and to create a Representative Binding for that Concept Domain that matches the existing universal binding on Concept Domain HumanLanguage.

Approved - 7 Aye- 1 Against -0 Abstain

Core Principles Release 3 Content

Note that Core Principles R2 has not completed reconciliation. Nevertheless, given methodology and ISO changes, it is worthwhile developing a list of items for future inclusion in Core Principles, including:

• Binding Intensity - already agrred to by Vocabulary & MnM
• Need to better document how to express (in designs) and constrain the codes for the "blockedContextConduction..." attributes, including how they appear in schema
• Clarify what happens when a MIN ValueSet is not provided in a binding
• Management of text between Glossary and Body of document, which is currently confusing
• Use of URN as ValueSet identifiers

Wednesday May 8 Q2

Agenda

• Host Vocab, FHIR
• FHIR Vocabulary

Attendees

Beeler Jr, George (Woody) (Beeler Consulting LLC)

Canu, Nicolas (HL7 France, L'Atelier du Soft)

Couderc, Carmela (Siemens Healthcare)

Estelrich, Ana (Phast)

Hamm, Russell

Hausam, Robert (Hausam Consulting)

Hirai, Masaaki (HL7 Japan Voter #2 - Nihon Kohden Corp, Eng. Oper.)

Huang, Wendy (Canada Health Infoway Inc.)

Huff, Stanley (Intermountain Healthcare)

Jewell, Gabriele (Cerner Corporation)

Klein, William Ted (Klein Consulting, Inc.)

Konishi, Yukinori (HL7 Japan)

Liu, Andrew (HL7 Canada)

Lyle, Jay (Ockham Information Services LLC)

McClure, Robert

McKenzie, Lloyd (HL7 Canada)

Ott, Russell

Parker, Ron (HL7 Canada)

Snelick, Robert (National Institute of Standards and Technology)

Stuart, Sandra (Kaiser Permanente)

Takasaka, Sadamu (HL7 Japan)

FHIR Vocabulary-related "QA" Issues

FHIR has a "QA" review list that it is working through in preparing the DSTU. Several of these are vocabulary-related and were discussed here. Agreements/results from this meeting were documented in the FHIR QA pages.

FHIR Review of Value Sets

Agreed that there should be a formal review process established to review the value sets being authored as part of FHIR resource development process, in order to be prepared for dealing with Negative comments on the DTSU ballots of FHIR being prepared for the upcoming ballot cycle.

Agreed to following (loosely defined) process

1. Will establish a pool of volunteer reviewers. Seek volunteers to do the reviewing

2. Individuals wail review designated portions of the list of Value Sets (chunks) to make recommendations to the FHIR developers, with an independent review by a second volunteer.

3. FHIR Management Group (FMG) will assign each Value set to two of volunteers in chunks, with report on each chunk expected back in a week

4. Will be two classes of review
   1. Where not have code lists looking for volunteers who might identify external source

   2. Where do have a code list, review for consistency and possible substitution from external source

5. All recommendations will be managed and tracked by FMG

6. FMG will escalate tougher issues to Vocabulary call

Discussion of Terminology Concerns in the FHIR "Profile" Resource

McKenzie led a discussion of the elements in the FHIR "Profile" resource that impact on terminology and vocabulary interests.

Wednesday May 8 Q3

Agenda

• Host FHIR
• FHIR Core

Thursday May 9 Q1

Agenda

• RIM mappings for FHIR resources

Attendees

Beeler Jr, George (Woody) (Beeler Consulting LLC)

Fine, Steven (Cerner Corporation)

Glinski, Steve (Surescripts)

Grieve, Grahame (Health Intersections Pty Ltd)

Hay, Christian (GS1)

Henderson, Mike (Eastern Informatics, Inc.)

McKenzie, Lloyd (HL7 Canada)

Shafarman, Mark (Shafarman Consulting)

Shakir, AbdulMalik (City of Hope National Medical Center)

Worden, Robert (Open Mapping Software)

FHIR RIM Mappings

Mapping Purposes

1. Allow rough assertion of equivalence (hand waving)
   1. Where should an analyst start if they need to transform
   2. Discovery
   3. Ensure that knowledge of specification is reflected in FHIR – what have we missed?
   4. Just type stuff in a spreadsheet
2. Provide validation of rough assertion
   1. Ensure we don’t have typos
   2. Produce reports of “completeness”
   3. Add a limited syntax to allow extraction and validation of RIM terms, but not full association tracing or vocabulary
   4. Act.code, Act.classCode, ActClassCode=PROC [classCode must be Procedure, Code constrained to some type of ….]
3. Provide full mapping with ability to analyze (e.g. OWL)
   1. Overlap (sibling FHIR elements have overlapping semantics)
   2. Duplication (multiple elements, extensions that mean the same thing)
   3. Discovery
   4. Consistency of definition (same FHIR name, different FHIR definition)
   5. Full expression of semantics (negationInd, inversionInd, other structural attributes)
   6. Leverage languages like OWL (though not necessarily for authoring)
4. Full mapping with ability to transform
   1. Ideal = Round-trip between FHIR and RIM ITS
   2. Basis for round-trip of RMIMs

Discussion

For ballot, most we can likely accomplish is type #1. Can possibly accomplish #2 before DSTU publication. #3 should be investigated to see if we can achieve desired benefits and how much work it is. No benefit to #4 (at the RIM level).

For non-RIM mappings, most will exist only at level #1. Some projects may chose to develop level #4 mappings but this will typically be done at a project level, not an HL7 Int'l level.

Thursday May 9 Q2

Agenda

• SAIF Artifact Definition

Attendees

Beeler Jr, George (Woody) (Beeler Consulting LLC)

Kreisler, Austin (SAIC - Science Applications International Corp)

Shakir, Salimah (Shakir Consulting)

Steps to "Vitalize" SAIF Artifact Definition Process

Discussed steps we could take to re-vitalize this process. Shakir has a list of definitions that are in some stage of completion, and he has done a rough analysis of the percentage of completion. Further review of the candidate artifact list reveals several that could and perhaps should be started as well.

Consensus on management includes:

• Agreement that effort should focus on artifacts that are in use and for which developers can use the definition to guide their creation efforts.
• Agreement to advance these by slotting periodic MnM Meetings whose focus is Artifact Definition, and for which the participants will be told which Definition is to be advanced and will come to the meeting:
  • Prepared to help "knock-out" a new definition for those that are barely started;

  • Prepared to help complete items that are mostly done; and

  • Prepared to accept items that are considered finished.

• Will triage the existing into three states and build Wiki Categories to hold -
  • All intended definitions, with sub-groups:
    • Complete

    • Under Development and Review

    • Listed, but deferred

• Will once again solicit for people to join the effort on the conference calls to allow these to advance.

Discussion on the "primary purpose" on these reflected a split in interest between:

• belief that these would improve consistency between designs and requirements traceability;

• belief that they will facilitate better defined and more tightly documented models, for which traceability will be easier to define, but may not affect either consistency or traceability.

Thursday May 9 Q5-Q6

Agenda

• MnM Roundtable

Attendees

Abdul-Malik Shakir (chair), Andy Stechishin (RIMBAA), Austin Kreisler (SDocs), Ted Klein (Vocab), Wendy Huang (Patient Admin), Melva Peters (Education), Iryna Roy (Observer), Jean Duteau (Patient Care), Joginder Madra (PHER), Grahame Grieve (FHIR), Kevin Coonan (Emerg Care), Rob Hausam (Vocab, O&O), Lloyd McKenzie (MnM), Dale Nelson (ITS), Paul Knapp (Financial Management), Hugh Glover (Pharmacy), Julie James (Pharmacy), Woody Beeler (M&M, Publishing)

Harmonization

Proposal Deadline  (Initial) - June 23, 2013
Proposal Deadline (Final) - July 14, 2013
Harmonization - July 22-26, 2013

Motion: Accept the dates that have been proposed above. Woody/Austin, 16/0/0

• Groups must send a representative or send an email indicating that they have reviewed the content and indicating that they have no problems with the submitted proposals.
• For Affiliate proposals, the approved date has to be when the appropriate Affiliate body approved it

Roundtable

Structured Documents

• 3 harmonization proposals related to value sets for CDA R3
• 1 harmonization proposal joint with O&O for aligning V2 ObservationInterpretation with V3 ObservationInterpretation
• 3-4 harmonization proposals related to HQMF
• Methodology Questions
  • Ballot Comment on "Why Datatypes R1 not used?"
  • How to convey a datatypes constraint in the specification that is not covered by the current datatypes specification?

Vocabulary

• ActRelationshipType cleanup harmonization may be brought forward
• There may be a large proposal from Security as well

Patient Admin

• PA/PHER will be harmonizing 4 joint CMETs
• Harmonization Proposal to create a new Language Concept Domain

Patient Care

• Harmonization proposals from the Allergy Model work
• Harmonization proposals from the Drug and Food Preferences Model work
• Care Plan DAM work is continuing so no harmonization proposal

PHER

• No harmonization proposals anticipated this cycle

FHIR

• Had Vocabulary discussions about how to document V2 and V3 value sets in FHIR documentation
  • URL vs OID?
• 1 Harmonization Proposal from PA
  • Terminology Authority has been proposed which might have an effect on this proposal
• FHIR only defines code systems that are intimately tied to the resource design
  • There will be some methodology on when FHIR needs to bring vocabulary needs to harmonization

Clinical Statement

• Harmonization Proposal on the subject of ActMood

Orders&Observations

• see harmonization proposals from Structured Documents

ITS

• Undertaking a project for specifying HL7 over HTTP
• Datatypes R2 ITS passed and is submitted for publishing

Pharmacy

• How do we take a current Normative section and move it into a different section?

Publishing

• Cleanup of class code was undertaken

Vocabulary Workgroup Report

• Worked on resolving representations of value sets in Implementation Guides
• Worked on Tooling projects - Submission Request System for Terminology Updates, IHTSDO Workbench
• Discussion on value sets in national repositories
• CTS2 Ballot Reconcilliation
• FHIR Vocabulary Discussions
• Discussed V2 Table Cleanup requirements
• Discussion on Binding and Conformance
• Presentation on RIM and OWL representation

M&M Workgroup Report

see File:MnM Roundtable 201305.pptx given by Lloyd McKenzie

Other Business

• HQMF R2 Ballot had a negative ballot about why Datatypes R1 was not chosen?
  • TSC has ruled that Datatypes R2 must be used for new published material
• How to provide guidance on datatype equivalence that contradicts the existing datatype specification?
  • Record these in a project-specific implementation guide or the underlying RMIM
• Terminology Authority and its effect on HL7 use of existing value sets
  • Ted explained how Terminology Authority works
• Moving pieces of a Normative Topic into a new balloted topic
  • You can move them and they take on the release date of the new topic
• Should 'Ask M&M' continue?
  • If we can find the personnel to man it and we should publicize it
• Maintenance of multiple Vocabularies
  • We do have multiple vocabularies and some of the reasons are due to non-Vocabulary workgroups having requirements

Friday May 10 Q2

Agenda

• M&M Business
• FHIR Inter-version Compatibility
• FHIR REST Paths

Attendees

Gaby Jewell, Jean Duteau, Lloyd McKenzie, Ewout Kramer, Grahame Grieve, Jim Harrison, Jeff Ting, John Moerhke

M&M Business

MOTION: Approve the submission for publication for RIM Release 6. Woody/Jean, 8/0/0

ACTION: Jean will submit the WGM Schedule. FHIR Tutorial will be Sunday Q3/Q4. Ensure that tutorial pre-requisites have been met. MnM needs a room for Sunday Q3. Thu Q5 Facilitator's Roundtable needs to be listed as more general than MnM. Vocab sessions will be hosted by MnM. RIM Reconciliation will become Ask MnM.

FHIR Inter-version Compatibility

• Discussion of rules of what changes are backwards compatible
• Wiki will be updated with a list of issues and rules that need to be consulted by implementers
• FHIR Specification will point to the Wiki (http://wiki.hl7.org/index.php?title=FHIR_interversion_compatibility)
• We will provide guidance to implementers in DSTU suggesting that using tags may provide a mechanism to manage changes between DSTU versions - since those will not be guaranteed to be wire-format-compatible

FHIR REST Paths

• Grahame began up a discussion of alternate URLs to access patient-related resources

ACTION: Include in the specification a common use case that allows

To get a problem list:
/problem/search?patient=[@id]
/patient/[@id]/problem

To get a specific problem:
/problem/@[pid]
/patient/@[id]/problem/@[pid]