MnM Minutes WGM 201609 Baltimore

M&M Conference Call 4:00PM Eastern Time (Date above)

Return to MnM Minutes

Tue Q3

See FHIR_Infrastructure_Minutes_WGM_201609#Tue_Q3 FHIR-I minutes

Wed Q1

Attendees

• Rob Mcclure
• Ted Klein
• Jennifer Puyenbrock
• Mark Shafarman
• Ron Shapiro
• Nick Radov
• Alexander Henket
• Russ Hamm
• Sandy Stuart
• Michelle Miller
• Susan Barber
• Greg Gustafson
• Wendi Castillo
• Carmela Couderc
• Lloyd Mckenzie
• Oyvind Aassre
• Carol Macumber
• Rob Hausam

Agenda - FHIR Ballot reconciliation tasks:

• task 11363: Add designations to expand.contains; designations must be derived based on the value set definition. Grahame / Alexander vote 19-0-1. ***Action***: need to work on best practice
• task 11279: Add root code related with comment 'encourage people to be more specific'. Rob McClure / Grahame Grieve 20-0-0
• task 9535: move the valueset import to concept Set and document how the imported value sets are interpreted. Grahame Grieve / Rob Hausam 19-0-1
• task 11820: endorse addition of this break out session, and recognise the need for and deliver canonical definitions code systems, starting with SNOMED CT, LOINC and RxNorm Grahame Grieve / Rob McClure: 20-0-0

Wed Q2

MnM Hosting Vocabulary (Chaired by Lloyd M)

Attendees

[20160922-q2-attendees.jpg]

Unified HL7 Terminology Governance Process

• Review of work done to date.
• Problems have been identified:
  • HL7 curates multiple terminology objects
  • Multiple different models
  • Existing tools are very fragile and close to not-maintainable
  • Etc.
• Suggestions to address the problem(s):
  • Re-engineer the existing V3 HL7 harmonization process
    • Time consuming
    • Long turnaround time
    • Not open
  • Create a unified process for all terminology that is curated and published in HL7 products
    • Improve quality
    • Improve productivity
    • Reduce cost
    • Improve access
    • All products!
• Separate ‘structural’ (bound to model) (e.g. V2 message codes) vocabulary processes that are specific to a product family, from other vocabulary.
• Objectives
  • Unify terminology guidance
  • Single set of tooling and processes
  • Broaden community participation
  • Support HL7 processes for development & publication
  • V2, V3, CDA, FHIR
  • Ease process, increase scalability & reduce cost
• Challenges
  • Tooling
  • Cross product line
  • Change management
• Proposed approach:
  • Create and propose new process
  • Identify tooling requirements
  • Identify roles in the process, requirements and use of tooling
  • Define migration strategy
  • Secure funding
• Next steps
  • Evaluate the strawman developed over the summer as input to achieve the objectives
  • Create a PSS for a new project to develop a process
  • Retain the existing tooling project to evaluate open source available vocabulary management tools
• Strawman highlights/assumptions/suggestions
  • Harmonization must be a continuous process
  • Cross product, and run in parallel
  • Asynchronous – not dependent on face to face
  • Web-based, broad deployment
  • Acknowledge combination of automated & manual processes
  • Leverage social media
  • Leverage FHIR resources & services
  • QA and timely response to proposals is critical
  • Requires updates to HL7 policies
• Next Steps
  • Create PSS – define new process (Vocab sponsor, MnM co-sponsor)
  • Ted and Lloyd will be project leads
  • Identify core team to lead – intent is to create a small, very focused team
  • Start seeking funding
  • Start immediately
• Parallel efforts
  • Identify and make small adjustments to models to harmonize (V2 tables project success)
  • Evaluate tooling that may be candidates
  • Goal: select/develop tooling to use in the process
• Discussion:
  • Question: Are there existing vocabulary management tools available?
    • Answer: Yes, although not that satisfy all the HL7 requirements. People acknowledge that no matter what tool is selected or developed, there will always be some very specific HL7 requirements to satisfy – generating engineering tasks.
  • Governance tasks will be a key part of the process – this type of feature (among other requirments) might be difficult to find ‘off the shelf’.
  • Goal is to validate as much as possible via tooling, and leave the non-computable work to humans.
  • Question: Assumption that we can, and how to define the process, to get external funding?
    • HL7 has a tooling budget – and it was never spent. No tooling project has made it to the point where funding would be approved.
    • Attendees feel that the available $$ is not enough to come close to finish a project of this scope.
    • Reality – the $$ has been re-allocated, no $$ is available from HL7; we need to pursue outside funding. However, tooling has been acknowledged as a top priority. The board needs enough detail to allocate the $$.
    • Suggestion: target internal funding as the primary funding source and then expand
    • Obtaining outside funding should not be a barrier to starting the project.
    • External funding is beneficial to the HL7 bottom line
  • Question: Will part of the migration strategy continue to deliver what we have today, before things are fixed?
    • Answer: the implementation of the new process must be incremental – specifics unknown at this time
  • Topic: The fragility of the existing tools and the risk that introduces, makes adoption of this project urgent.
  • Question: Are there licensing issues:
    • Answer: Yes, HTA is working to address any issues related to developing standards
  • Question: How can people help?
    • Answer: Talk to (or email) Ted and/or Lloyd about getting on the core team. (2 volunteers spoke up at the meeing)
      • Carol M. – Apelon
      • Raj Mehra Cerner
    • Invite people who have been participating for the product families.
  • Topic: The core team should have members from organizations that use terminology management tools, develop terminology management tools. ***True. However, for the governance process, we need people experienced with harmonization.
  • Question: Difference between co-sponsors and interested parties (publishing, FHIR infrastructure, etc.). Is it really going to take months to get co-sponsors to agree to PSS? Not if the co-sponsor list is constrained.
    • One approach is to have Vocabulary, Tooling and MnM as co-sponsors – risk exists that this might delay the PSS from moving forward.

Motion: Ted K: Develop a PSS to define the new unified vocabulary governance process for HL7 as quickly as possible

Second: Jenny Puyenbroek.

Motion Passes: 13/0/0

Action Items

1. Ted to send a PSS draft to Lloyd
2. Review of PSS will be on 1st Vocab call after Baltimore meeting

Schedule next harmonization meeting

• Vocab, MnM, International must attend

Propose: November 9, 10 with potential of 11 as an overflow day.

• This overlaps with HL7 Education Summit
• Later than this will not leave enough time for ballot
• Prior to this will squeeze the people preparing proposals

Need to resolve: MnM and International coverage. Ted to follow up.

Action item: vocab agenda: San Antonio joint quarter, Wednesday Q2, Vocab to host

Tracker Items

[http://gforge.hl7.org/gf/project/fhir/tracker/?action=TrackerItemEdit&tracker_item_id=11029 11029

• A Boolean choice for ‘user-selected’ seems insufficient. If the code is user-selected, might we need to distinguish between selecting the system display and the local display. More nuance might be required for some use cases. E.g. – the display name might have defaulted in, or it might have ben translated.
• Is the issue really related to coding.Display? (a representation of the meaning of the code in the system, following the rules of the system). There is no statement in the definition as to whether this contains the ‘official’ representation as defined by the code system, or the ‘preferred’ name, etc.
• Action item: ask Jay L. for some specific examples of ‘not user selected’ and what would be done with that information? Would it be commonly used? (Lloyd to follow up)

Thur Q2

No quorum

Thur Q5-Roundtable

Attendees

• Mead Walker (RCRIM, BRIDG)
• Virginia Lorenzi (Education)
• Brian Peck (Publishing, V2 manage)
• Susan Barber (PHER)
• Josh Mandel (FHIR Infrastructure)
• Lloyd McKenzie (MnM, FHIR)
• Alexander Henket (PA, SD)
• Andy Stechishin (ITS, Publishing)
• Ewout Kramer (FHIR-I)
• Calvin Beebe (TSC, SD)
• Ted Klein (Vocab)
• Michel Rutten (observer)
• Brian Postlethwaite (PA)
• Riki Merrick (OO)
• Abdul Malik Shakir (MnM)
• Grahame Grieve (FHIR-I)
• Wayne Kubick (CTO)
• Rob Hausam (OO)
• Paul Knapp (FM)

Reports

Question from TSC

• Request to consider making RIM from continuous approval to periodic
  • Will need to renew at least every 5 years
• RIM structures have not been updated since Nov. 2014
• RIM includes structural vocab which changes more often
• Concern with ramifications of an IG passing that uses new vocab that hasn't been balloted
• We don't think changing to periodic would affect our schedule

BRIDG/RCRIM

• No harmonization proposals from either
• BRIDG uses v3 r2 data types, have created additional classes to capture v3 identifier pattern
• Want to modify their data type to include id type - so custom version of the data types
• They have a RIM mapping - should they bother updating the RIM mapping?

SD

• CDA 2.1 - backwards compatibility with some updates
• Need to use the v3 generator, and need to "warp" the output to be backward compatible
  • E.g. customizing core MIFs prior to generation
• Will not fit will with current RIM thinking
• Q: Can we propose these as changes to the RIM? A: No, this about rolling back changes for backwards compatibility

Grahame Grieve and Wayne Kubick join

Vocabulary

• v2 Tables project is done
  • Hundreds of small changes, new publication format
  • Don't use 2.8.2 as source for your change proposals, use content published on the harmonization page (early next week)
    • Will announce on harmonization list
• PSS will be coming on new vocab maintenance & governance process (from MnM, Tooling, EST and possibly Publishing)
  • Will change how harmonization is done (but not this cycle or likely anything in 2017)
  • Talk to Ted or Lloyd to be involved
  • continuous, independent of ballot schedule
• All v2 and v3 vocab changes must come through harmonization process

FHIR

• Less ballot comments on core, more on the domains
• Will be preparing for publication and for balloting of several IGs
• Nov. 27th for substantive freeze on core, Dec. 4th freeze on everything
• That freeze will be ballot and also input for QA
• Slipping will be very difficult due to ballot, so please hit the dates
• Challenges with Quality level and implementation not matching (implementation ahead of quality)
• Want those that are going to go normative to move up the scale
• Intention is that much of the infrastructure and some of the other resources will go Normative
• Conformance is working with us on defining "good practices" around IGs
• Expecting to see certification process released early next year
  • Renewal process not yet figured out
• Data Access Framework (DAF) will be changing name to "US Realm IG"
  • Not clear whether US Realm will be taking custodianship
• Need to decide if US Lab should be part of US Realm IG or be separate IG built on top
• Have an approval process for US extensions
  • Will need to figure out how to deal with existing extensions
• 2 changes coming that are effectively FMM 5 changes. Notification will go out seeking feedback
  • One is uncontroversial, non-breaking; One is non-breaking and controversial
• SMART on FHIR spec is being contributed to the FHIR spec. Have a PSS for that and it will be handled by FHIR-I initially
  • Appendix to the STU 3 publication
• Discussion tomorrow am (7:30 breakfast room) about testing
• FluentPath is being renamed to FHIRPath so it can be trademarked but will try to make clear it's intended for broader use
• Probably a harmonization proposal from Kathleen

Rob Hausam joins, Paul Knapp joins

OO

• 3 harmonization proposals
  • Issue with message structure codes related to event codes that were fixed last time
  • Table 78 (Observation interpretation)

Harmonization Dates

• Proposals due Oct. 14
• Corrections due by Nov. 4
• Harmonization Nov 9 & 10
• Cannot accept late proposals

Discussion topics

• Can BRIDG create a customized data types spec with modifications to identifiers
  • In a DAM you can use whatever you want
  • Could propose a change to the ISO data types, but low probability of it going forward
  • Can't claim to be conformant with the ISO data types if you make changes, but could find appropriate wording to say that it's "based on"

• Should BRIDG bother updating RIM mappings
  • General consensus - unless there's someone who's asking for it, low value
  • might be better to map to FHIR (which inherits RIM mappings too)

• New RIM data types
  • Can create re-useable complex data types - ask FHIR-I if you think this might be useful to you