CRM Taskforce Meeting Minutes - 1 March 2012

From CIMI
Jump to: navigation, search

Attendees

  • Linda Bird (Ministry of Health Holdings, Singapore)
  • Michael van der Zel (Results4Care)
  • Josh Mandel (SMArt)
  • Gerard Freriks (EN13606 Association)
  • Stan Huff (Intermountain Healthcare)
  • Thomas Beale (Ocean Informatics)

Apologies

  • Richard Kavanagh (NHS Connecting for Health)
  • Grahame Grieve (Health Intersections)

Agenda

Review draft report, including:

  • Recap on last week’s meeting
  • Reference Model Motivation
  • Architectural Framework
  • Requirements
  • Candidate Reference Model Options

Brief Summary

Membership – Stan reported that Galen has been recommended to represent the ONC on the taskforce

Motivation – Additional wording suggested for the reference model motivation, including:

  • “The reference model helps to define repeated semantics”,
  • “The information model of the core reference model defines the things that you need to be represented in tangible software, to guarantee that you can implement in your physical structure what has been constrained in the clinical model”,
  • “The reference model provides the common language”, and
  • “The reference model is the computer processable model used to place constraints on.”

Architectural Framework – Comments included:

  • Need two additional diagrams – an overview, and more detail on the Clinical Model specialisation levels (i.e. layering of models)
  • MOF and UML rectangles should not be layered.
  • Should mention ‘Templates’ on this diagram (or perhaps on another)

Requirements: Comments included

  • Some requirements are not reference model requirements, but instead may be requirements of the resulting clinical models. It was decided to leave these requirements in for the time being, so as not to lose them, but reassess their suitability at a future time.
  • RM-REQ-GN-03 (Multiple Outputs): Limitations include meaning can be lost, or some manual intervention may be required.
  • RM-REQ-GN-04 (Multiple Languages): This requirement was removed, as it is not considered a reference model requirement.
  • RM-REQ-GN-05 (Realm-Specific Specialisations and Extensions): Should instead be a requirement of the overall framework.
  • RM-REQ-GN-06 (No Clinically Relevant Attributes or Specialisations):
    • Should only contain attributes or classes whose meaning is invariable across the domain
    • We may not know how to apply this principle until we get into the reference model design
    • RM-REQ-GN-07 (Model Mapping): Should be a requirement of the RM formalism instead?
  • RM-REQ-GN-09 (Versioning) & RM-REQ-GN-10 (Approval Status): Unclear on purpose of these requirements
  • RM-REQ-GN-11 (Stability): This is close to the domain-invariant requirement of RM-REQ-GN-06
  • RM-REQ-GN-02 (Clinician Verification): More of a clinical model requirement
  • RM-REQ-ST-01 (Data Elements): Definition has been clarified by rewording.

Candidate Reference Models – UML Class diagrams for the proposals to be collected during the week. Proposals included:

  • ISO 13606-1 (or simplification)
  • openEHR reference model
  • openEHR/ISO13606-1 model
  • DCM reference model (ISO13972 based, Dutch)
  • Parts of Intermountain Clinical Element Model
  • EN13606 Association proposal [GF]

Actions

  • Linda Bird to:
    • Distribute minutes
    • Welcome Galen to the taskforce
  • All to:
    • Review draft document, and provide feedback and rewording suggestions
    • Assist in rewording the ‘Motivation’ section
    • Distribute UML class diagrams of candidate reference models listed above