MnM Minutes WGM 200901

Sunday Q3

Agenda

• Agenda Topics Review
• Hot Topics Triage

Minutes

See updated agenda.

Sunday Q4

Agenda

• Tooling Discussion

Minutes

• SVN and Publishing
  • Directories
  • Branches and tags for ballots and normative edition
    • Tag - a "badge" applied to each member of a set of files saying those files belong to a set identified by the tag. Implemented in SVN as a "tag folder" with a set of pointers to the content. Seen as a "folder" under "tags" (sibling to "trunk") one folder/per assigned tag
      • Branch is defined "at a particular version" and is a new set of files and changes that may differ from the trunk.
    • Proposal
      • Use tags to identify each ballot.
        • Applied at the end of the ballot cycle (1 week before close).
        • Include a domain "manifest" for each domain.
      • Use branch to hold each Normative Edition, with NO INTENT to merge later.
    • Observation -- be CONSERVATIVE and only version that which
      • (a) is needed for publishing, and
      • (b) needs version management

Monday Q1

Agenda

• Business Meeting (SWOT, DMPs, etc.)

Minutes

• Reviewed SWOT

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

• Reviewed and updated MnM Strategic Plan, in particular to define the committee's role in SAEAF.

http://newgforge.hl7.nscee.edu/frs/?group_id=27

• Discussed the future role of MnM
  • Should MnM own the Data Types?
  • Is MnM active enough in SAEAF? The consensus was that we probably have not been.
• Woody will create a draft for a scope statement for the SAEAF related activities.
• Woody urged the other co-chairs to attend the Steering Division calls.

Monday Q2

Agenda

• RIM and Core Principles ballot reconciliation
• Proposed re-org of document from technical editors

Minutes

• We reviewed Ravi's comments on the RIM ballot (the ballot was for comment only).
  • COMMENT -- The codingStrength identifiers are moved into the vocabulary models. So at present user has to traverse back and forth between rim and vocabulary's to get this information which is very tedious.
    • RESPONSE -- Suggest that the representation for CNE of structural attributes be picked up from Vocabulary and displayed in RIM representation.
  • COMMENT -- The list of structural attributes needs to present which is missing. At the moment present the partial list is present in the XML_ITS specification document. CfH harmonization proposal dated FEB 2006 related to this (discussed and accepted) is never implemented so far.
    • RESPONSE -- Add immutable status after "Mandatory" in the attribute detail. Also include a list of immutable attributes in an appendix.
  • COMMENT -- The cardinality for these attributes (Attachment.id and Attachment.text) are 1..1 which similar to the structural attributes. Are these attributes Mandatory similar to Act.classCode or Do they have to be marked 0..1 as against to 1..1.
    • RESPONSE -- Prepare a harmonization proposal to either assert that these are required, or make them 0..1. Other option is to display "not required" for all attributes with a blank conformance indicator.
• Discussion of reorganizing the Core Principles document.
  • see:

http://wiki.hl7.org/images/b/be/CorePrinciples_notes_090111.doc

Monday Q3

Agenda

• RIM and Core Principles Ballot

Minutes

• Review of comments on the RIM ballot (comment only ballot).
  • Comment by Tom de Jong
    • We need to review Exposure and Procedure to make sure their definitions are appropriate since SubstanceAdministration. If needed, a harmonization proposal should be submitted to handle this.
  • Comments by Frieda Hall
    • See spreadsheet
  • Comments by Agha Kahn
    • Spreadsheet was empty.
  • Comments by Austin Kreisler
    • See spreadsheet
  • Comments by Rene Sponk
    • See spreadsheet

• We will propose to the Facilitators' Roundtable that we intend to do a harmonization call on February 27th.

Monday Q4

Agenda

• Transition to Datatypes (& Wrappers?) R2

Minutes

• Transition to Dataypes R2
  • SET → DSET
  • CE → CD
  • GTS → QSET<TS>
  • BAG<T> → COLL<T>
  • IVL<INT> → INT
  • IVL<PQ> → PQ
• Lloyd and Grahame will document mappings.
• The mappings will be documented as tagged annotations of the R1 attributes in the database.
• Motion: (Grahame/Andy) Accept the actions proposed in Interaction and Versioning (Hot Topic). Motion passes (8:0:0).

Tuesday Q3

Agenda

• Design Patterns:
  • Guidance on "Organizational Role is part of Organizational Role"
  • Request Responses (one receiver responsibility vs. two)
• OIDs Registry

Minutes

• OIDs Registry
  • The is discussion about managing OIDs internationally.
  • Sylvia Thun suggested a model where country specific registries manage country specific OIDs and coordinate with an international registry.
  • Woody asked whether it really matters if there is more than one OID for the same thing in a registry.
  • Multiple OIDs cause extra effort in determining sameness, but don't interfere with uniqueness.
  • There was general agreement that it would be good if an international body could manage healthcare OIDs globally. The WHO was suggested.
  • Cecil suggested that having an RDF based OID registry would be a Good Thing™.
  • Three options suggested by Lloyd:
    • Clarify that we are only interested in common public namespaces.
    • State that we are comfortable with affiliates overriding HL7 OIDs, but we would encourage them to communicate the override.
    • Would could change SHALL to SHOULD or MAY in the statement that there SHALL only be one OID for a given entity in the OID registry.
  • Decisions
    • We decided to leave the wording at SHALL in the Core Principles document with a footnote that this is something that is under discussion.
    • We will work with Germany and other countries that have OID registries to see if we can import their OIDs into the HL7 registry.

Wednesday Q1

Agenda

• Joint with Vocabulary
• Value Set Conformance and Constraint
• Coordination of Core Principles document
• Harmonization Process
• Presence of Structural Vocabulary in RIM Ballot

Minutes

• Presence of Structural Vocabulary in RIM Ballot (Beeler)
  • We will be balloting the RIM on an annual basis.
  • This is the first year since 2003 that we have balloted the RIM.
  • All changes resulting from the ballot process will be subsequently taken through the harmonization process.
  • Vocabulary constraints to the RIM are open to comment (e.g. ActClass, RoleClass, etc.).
  • Jane Curry asked about comments by David Markwell from 2005 or earlier. She will seek to find them and introduce them for discussion.
• Value Set Conformance and Constraint
  • This is a continuation of an discussion from Monday Q2 in the Vocabulary WG.
  • A maximum value set (MAX) contains everything that may be supported. For a specific instance you can define subsets that are used and that are ignored.
  • There was some question about whether "ignored/not ignored" is the right level of granularity. Greater granularity may be more realistic in the future, but the current distinction is okay for now.
  • There is a general consensus that it would be desirable (but not required) for the conformance statements about attributes and about vocabulary to be consistent and that changes be advanced together.
  • Begin of content considered by the motion below.
    • The UV binding statement consists of an assertion of one principal and two optional value sets. The principal value set (MAX) indicates the complete set of codes that are available for use (excluding CWE extensions note that CWE/CNE talks about whether or not the MAX value set is extensible). The remaining two value sets designate subsets of the principal value set that are:
      • required (MIN) meaning they must be supported (used in some useful manner by the system sending, receiving, or otherwise processing the model) or
      • ignored (IGNORED) meaning they are not used in any useful manner by the system.
    • The set of codes that exist within the maximum value set (MAX) but do not exist in MIN or IGNORED are considered to be "conformance undeclared". Undeclared conformance represents design-time optionality; the final implementation profile should have no remaining undeclared conformance. As a specification is tightened and approaches an implementation profile, the MIN and IGNORED value sets will tend to increase in size and the MAX value set may decrease in size such that, for a complete implementation profile, the union of the MIN and IGNORED value sets will contain the same set of codes as the MAX value set.
    • The rules:
      • When creating constrained artifacts, including realm context bindings (where a UV or other higher level binding exists) and constrained static models, the value set bindings must represent a narrowing of the content of the higher level value set.
        • When you constrain you can leave the MAX the same or make it narrower,
          • you can leave the MIN the same or make it bigger.
          • you can leave the IGNORED the same or make it bigger and
      • at all times MIN and IGNORED are exclusive and must be proper subsets of MAX.
  • End of content considered by the motion below.
  • Motion -- Woody/Jane -- MnM and Vocab endorse these decisions, and ask that they be included for ballot in Core Principles. Motion passes (16:0:0).

Wednesday Q2

Agenda

• Joint with Vocabulary
• Continuation of agenda from Q1

Minutes

• Coordination of Core Principles document
  • Woody gave an overview of the current state of the Core Principles document.
• Harmonization Process
  • When changes are made to harmonization proposals, work groups do not always realize these changes were made and sometimes generate their ballots based on incorrect content.
  • Ted suggested that design models that use the items addressed in a harmonization proposal be listed on the harmonization proposal. He also suggested that we provide more reference information for people creating proposals.
  • Rob stated that more attention needs to be given to vocabulary before models go normative; specifically all domains should have a binding in the Representative Realm.
  • Ted suggested that we need to give more attention to non-structural vocabulary.
  • Ted suggested that we need an explicit sign-off by submitters for harmonization proposals after the harmonization meeting. Woody pointed out that this could hold up subsequent steps if workgroups were slow to sign-off.
  • Woody suggested that we should use the bug tracking functionality of GForge to track problems (not necessary every proposal).
  • We need to ask harmonization proposal submitters to:
    • Check the notes from the harmonization meeting for their proposals to make sure they are accurate.
    • After the changes have been applied, to make sure it was implemented correctly.
• Russ stated a concern that there are interdependencies between code systems used for EntityName (qualifiers and types) and that there is therefore a risk for inconsistency.
  • Woody stated that MnM needs to exert control over concept relationship types to ensure consistency.
  • There was discussion about where tools could support the interdependencies in a consistent manner.
  • This discussion needs to be further discussed at a future date.
  • The alternatives seem to be that we either allow multi-axial code systems for structural vocabulary or we have to deal with interdependencies between code systems.

Wednesday Q3

Agenda

• Joint with TSC re: Enterprise Architecture (SAEAF)

Minutes

• TSC hosting, see TSC minutes.

Wednesday Q4

Agenda

• Joint with TSC re: Enterprise Architecture (SAEAF)

Minutes

• TSC hosting, see TSC minutes.

Thursday Q1

Agenda

• Joint with HSSP, IMN, and ITS
• Dynamic Model

Minutes

• Hosted by INM, see INM minutes.

Thursday Evening -- Facilitators' Roundtable

Agenda

• RIM Ballot/Harmonization
• Datatypes R2 and Wrappers R2 Strategies
• Harmonization Meeting, April 14-17
• Harmonization Call, February 27
• RMIM Iconography
• Context Conduction

Minutes

• RIM Ballot/Harmonization Strategy
  • We want all changes to the RIM to go through harmonization.
  • We will be having a harmonization conference call February 27th to assist in this.
• Datatypes R2 and Wrappers R2 Strategies
  • Patrick Loyd will be moving Wrappers R2 forward.
  • It is expected that Datatypes R2 will be ready by the end of September.
  • We will do a mapping between Datatypes R1 and Datatypes R2.
• INM and MnM are in agreement to move the responsibility of datatypes from INM to MnM.
  • MnM will create another co-chair position for accommodate a steward for the datatypes (Grahame).
• Normative Editions 2010 will be published with datatypes R2.
• Regular Harmonization meeting in Las Vegas, April 14-17.
• RMIM Iconography
  • Moving to an Eclipse-based static model designer.
  • Lloyd asked people to consider the pros and cons of changing the iconography to be more UML-like. This is because the Eclipse-based tool's graphics are based on a UML rendering engine and may have difficulty creating some of the graphics currently used by HL7. Lloyd will create a wiki page where discussion about this will occur.

Vocabulary -- Ted Klein

• Vocab WG has made refinement to vocabulary binding. These changes will need to propagate into the MIF.
• Ted plans to present the details of these changes at the next harmonization meeting.
• A couple of changes have been made to the harmonization submission form. The new form should be available shortly.
• Ted asked people to make sure they review the results of harmonization proposals to make sure that any modifications to the proposal that happen during the harmonization process are propagated back into their artifacts.

Patient Administration -- Norman Daoust

• PA will have a number of vocabulary items and one RIM change for harmonization.
• PA was under the impression that they should take context conduction out of their models. Upon further discussion, it appears that they were advised to take context conduction out of some specific models because it was not needed there, but not instructed to remove it universally. Patrick stated that removing context conduction from some of these models has caused unintended downstream consequences; in particular, the change made the Clinical Statement model for Encounter inconsistent with the PA model for Encounter. Hugh pointed out that Pharmacy has used context conduction inconsistently and, since it is recognized that context conduction is broken, have decided to remove it completely until such a time that it is fixed. Others echoed similar sentiments.

ITS -- Dale Nelson

• Has been looking at the idea of a new ITS around serialization.
• What are the rules for serialization? The committee could not find the algorithm that goes from an HMD to a serialized MIF.
• Who owns the expression of the abstract serialization? Answer - MnM owns the serialization algorithm. The algorithm is based on sort order. Rene stated that serialization is documented on the wiki with an action item to include it in Core Principles.

CTS II -- Russ Hamm

• The intent is for CST II to be balloted as DSTU in the next cycle.

PHER -- Austin Kreisler

• Some vocabulary harmonization items will be coming this round.

OO -- Patrick Loyd

• OO rewritten the project scope statements for its four major areas.
• The big push in OO and Clinical Statement is to achieve a consistent set of broadly used models.
• Jean asked why Clinical Statement does not use some models from other committees (e.g. CMETs). The response was that the scope of Clinical Statement is only clinical content and therefore many of the models cannot be used directly. Several people expressed concern about this scope being too narrow.

ARB -- Mead Walker

• Reported on the dynamic model discussion from Thursday Q1.
• There is a desire to have a dynamic model discussion at the harmonization meeting in April.

Patient Care -- Kevin Coonan

• Looking at updating and reballoting the care provision domain.
• Looking at modeling guidelines.
• Working on detailed clinical models (templates for observations).
• This issue was raised that we need a forum getting answers to modeling issues. Should we schedule a quarter each meeting for clinical design with participation from experts from a number of committees?

Publishing -- Woody Beeler

• Woody asked for greater participation in the publishing calls, especially from the committees with larger amounts of ballot content.

FM -- Kathleen Connor

• Met with SD on P2P.
• MITA will have an out of cycle meeting in Minnesota, sponsored by FM.

MnM -- Lloyd McKenzie

• Lloyd summarized MnM's activities for the week.
• When Lloyd recapped the OID discussion, Woody suggested that the topic of OIDs and OID registries needs to be addressed by the joint ISO/CEN/HL7 group.

• Motion (Lloyd/Gunther) MnM to endorse the creation of an additional co-chair position with a major responsibility of maintaining datatypes and nominate Grahame Grieve to fill that position in the interim until a formal vote can be held at the next WGM.
  • Motion passes (16:0:0).

Friday Q1

Agenda

• MnM Wrap-up and Planning for next Harmonization and WG Meetings

Friday Q2

Agenda

• Conformance in Choices
• Shared Messages

Minutes

Attendees: Lloyd McKenzie (chair), Grahame Grieve, Rene Spronk, Patrick Loyd, Wendy Huang

Datatypes Ballot

Motion: MnM endorses Datatypes R2 to return to 2nd membership ballot with the original ballot pool with scope limited to review of substantive changes since the previous ballot. GG to write up the necessary paperwork Moved: Grahame/Patrick: 4-0-0

Kyoto

Tuesday Q3 joint meeting on datatypes with Vocab now needs to be with MnM. Lloyd has sent a note to Craig allowing him to include this in our room bookings.

Conformance

Motion: MnM endorses the idea of migrating from "Not Permitted" to "Ignored" as a value for conformance, based on the discussion held with Vocab. The code of NP would be deprecated with direction that HL7-published models referencing it should either change the maximum cardinality to 0 (if they want to show it in the model for some reason, e.g. for documentation) or remove it from the model entirely. This will be taken to Implementation/Conformance and discussed on a conference call or at Kyoto. Moved: Grahame/Patrick 4-0-0

Motion: MnM endorses including "conformance" as property of "choice" associations (i.e. the relationship between a choice and the classes within the choice), allowing choices to be marked as "Required" (or "ignored"). This will be taken to Implementation/Conformance and discussed on a conference call or at Kyoto. Moved: Rene/Patrick 4-0-0

Request Interaction pattern

This will be listed as a hot topic by Rene Initial leaning is that best practice is to have 2 distinct receiver responsibility interactions because each has a distinct trigger event and the "Accept" trigger event may actually be the trigger for other interactions. It also makes it easier to see what's happened without drilling into the message and conveys this in a consistent way. Also, it allows for better re-use if the payloads need to be different for accept vs. refuse (rather than a single message type that's a choice of the two pieces). (Note that a Web Service wrapper can be used to allow a choice of any of the possible receiver responsibilities.)

Organizational Role relationship

Reviewed [Design_pattern:_organizations_part/partOf]. Will be discussed on a hot topics call

Context conduction

PA changes don't directly cause issues for Clinical Statement, but may indirectly.