Skip to end of metadata
Go to start of metadata


Bryn Rhodes

Maria Michaels

Matt Burton

Dyann Matson-Koffman

Floyd Eisenberg

Brian Fengler

Jen Clark

Peter Muir

Jill Shuemaker

Joe Bormel

Keith Toussaint

Steph Hoelscher

Steve Bernstein

Josh Richardson

Jeff Danford

Dave Carlson

Jim Jellison

Barry Blumenfeld

Derek Ritz

Dwayne Hoelscher

Ira Lubin


  • Review progress on updates
  • Continue discussion on evidence grading representation


  • Profile organization (FHIR-27150)
    • Changes made to definitional profiles
      • Builds on shareable ActivityDefinition – extensions for CPGKnowledgeCapability and CPGKnowledgeRepresentation
      • Each profile can then indicate what level they are: shareable, computable, publishable, executable
    • Will do the same kind of changes (i.e., by category) in the alphabetical list
    • Will also do with activity profiles (which Bryn is still adding in)
  • Terminology (FHIR-27152)
    • Strengthening documentation – start with compose, then go to an expression, rules text as a last resort
    • Non-FHIR expression and terminology grammar is ok, but language must be sufficiently open and accessible so that the user can construct the value set following the semantics without needing to use a proprietary implementation
  • Grading (FHIR-27151)
    • Options included defining CPG grading system & provide maps all the way to providing in the FHIR base specification and include content/documentation for what grading system is being used and why (which should be consistent across the content but not a requirement)
    • Question from Dyann: Will the grading be consistent with the narrative guideline?
      • Bryn: Yes, the idea is that those producing the content state which grading system they are using, it’s not tied to one specific grading system.
    • Using a binding established in base FHIR specification (Clinical Reasoning Module) – includes a synthesis of common grading systems rather than binding to one specific grading system since not everyone uses the same grading system (therefore, it is an “example” binding in the Clinical Reasoning Module and is done in the same way in CPG-IG) – both strength of recommendation and quality of evidence – will continue to use until there is a specific terminology
    • GRADE has definitions but not aware of a terminology that we can point to
    • Derek Ritz will be on a call with folks from McMaster University, so he will ask if a terminology exists for GRADE (will confirm whether there is a code list for GRADE)
    • Code list – URL that uniquely identifies a code system
    • May need a discussion between HL7 and GRADE terminologists to establish a URI if there is or will be a code list for GRADE & will revisit with EBM-on-FHIR
    • Question from Derek: Can we host the URL with the GRADE codes? Per Bryn, that’s why we’re reaching out to the HL7 terminology authority (Carol or Rob) – will take time (similar issue to health worker classification with International Labor Organization specification)
    • Derek: GRADE is an international standard which has been built on consensus over a couple of decades (unlike AHRQ, for example), Laura M. agrees
  • Dispositions – should now be complete, just need to get everything applied in the IG
  • Ongoing work from the group or additional questions:
    • Don Casey – still working on document related to quality measurement
      • We should be clear about relationship between GRADE taxonomy & “level of decision support” (e.g., lower quality of evidence should be less prescriptive on what should be done in a computable artifact)
      • ACTION: Don can send some additional info in a book chapter that was published last year – Sent book chapter in chat: Personalized and Precision… (PDF)
    • Question from Dave Carlson – Do we need to ballot again or just respond and publish?
      • Bryn: Depends how substantially different ballot vs current content is + how broadly used trial use is adopted – based on status, pursuing publication and subsequent ballot (ultimately, decision will be up to CDS WG and TSC) – can ballot in Sept 2020 but would need to decide by July 5 deadline for Notice for Intent to Ballot (can do this if we don’t feel like we have enough unique content by then) – can still publish as long as all reconciliation content is applied
      • Derek – Publishing as a next step makes the most sense. There have been substantive changes/re-engineering (all advantageous and usefully reflecting lessons learned) therefore also makes sense to re-ballot, especially because of patient safety risks.
      • Bryn: Need to get to stability and be able to demonstrate and address patient safety issues.
      • Floyd Eisenberg: Having a published first version helps with testing because it provides a more stable version than the “build” version (which may be constantly changing) to help get the right kind of input from those testing to get to the next ballot.

Next meeting in 2 weeks: June 17, 2020