CG WG Call Notes leading to 2015 September WGM

June 16, 2015 -- Weekly call

Agenda

• FHIR stuff

Attendees

• Amnon Shabo (Shvo), Mollie Ullman-Cullere, Siew Lam, Larry Babb, Jonathan Holt, Perry Mar, Bob Freimuth, Dan, Lori, Travis

Draft Minutes

• Larry:
  • went through his FHIR ballot comment
  • we should not be bundling everything into a single flat Observation structure
  • Need a roadmap to show how this work is going to evolve
  • Need to support clinical decision support
• Amnon:
  • The interpretation of the base Observation resource should be explicitly disallowed
  • Clinical decision support is an important use case

June 9, 2015 -- Weekly call

Agenda

• Misc.

Attendees

• Amnon Shabo (Shvo), Mollie Ullman-Cullere, Siew Lam, Larry Babb, Jonathan Holt, Jeremy Warner, Dan Rutz, Bob Milius, Jeffery Karp, Scott Bolte, Joseph

Draft Minutes

• Comments on the FHIR genetic Family Health History FHIR Profile
• Decided to contact Lloyd and Grahame about issues in the current revision of the FHIR profile of genetic family history as explained by Jonathan Holt

June 2, 2015 -- Weekly call

Agenda

• Misc.

Attendees

• Amnon Shabo (Shvo), Mollie Ullman-Cullere, Grant Wood, Larry Babb, Jonathan Holt, Bob Freimuth, Jeremy Warner, Dan Rutz, Bob Milius, Jeffery Karp, Gil Alterovitz, Travis Pittman

Draft Minutes

• Dan Rutz introduced himself (representing Epic)
• Grant described the genomics policy meeting in DC
• The Global Alliance plenary meeting was mentioned
• Not clear if and which FHIR version is going normative
  • Later confirmed that there will be at least one more DSTU (DSTU 2.1) prior to normative balloting
• Bob Freimuth:
  • perhaps it's not possible to get to an overarching CG model covering all use cases
  • focus on messaging specifications
• Amnon:
  • HL7 main mandate is about exchange of information
  • however, the model of exchanged information impacts the internal representation of the data at the sending and receiving sides
  • if the internal representations are very different than the exchanged representation, then it might make semantic interoperability much harder
  • We have to work both bottom up (smaller-scale models for specific use cases) and top down (conceptual and broader modeling) so that we archive progress and learn lessons on the one hand, and still converge on the important principles to be kept in any CG model
  • the DIM effort is currently about two narrower models - Clinical Genomics Statement (just one single gen-phen association) and family history
  • the overarching models meet the concrete specifications by means of constraining, preferably using model-driven tools