MnM Minutes WGM 201205 Vancouver
Return to MnM Minutes for 2012
- 1 Sunday Jan. 15 Q3
- 2 Monday May 14 Q1
- 3 Monday May 14 Q2
- 4 Monday May 14 Q3
- 5 Tuesday May 15 Q1
- 6 Tuesday May 15 Q2
- 7 Tuesday May 15 Q4
- 8 Wednesday May 16 Q1
- 9 Wednesday May 16 Q3
- 10 Thursday May 17 Evening
- 11 Friday May 18 Q1
Sunday Jan. 15 Q3
Woody Beeler, Jean Duteau, Austin Kreisler, Lloyd McKenzie, Rik Smithies, Andy Stechishin, Mead Walker, Grahame Grieve, Dale Nelson, Adel Ghemallah, Amnon Shabo, Lee Coller, Ann Wrightson, Adam Chee
Establish Agenda for WGM from Hot Topics and FHIR
- Reviewed the existing Hot Topics
- Reviewed when FHIR was going to be discussed in other committees
- Updated the Agenda with the proper quarters noted.
Monday May 14 Q1
Woody Beeler, Abdul-Malik Shakir, Lloyd McKenzie, Ann Wrightson, Ewout Kramer
SAIF Artifact Definition & Implementation Guide
Started discussing the list of SAIF Artifacts to be defined from the SAIF Artifact List page. Re-categorized this based on current views of requirements, including the anticipated inclusion of FHIR in the forthcoming cycles. Lloyd McKenzie collected and edited a Word Document of the result of this discussion. It was agreed to by those present and will be attached to these minutes and used to update the SAIF Artifact List page by gwb.
Abdul-Malik Shakir agreed to take on Project Lead for this project and Woody Beeler agreed to be the facilitator for the project. We will schedule a call to re-invigorate this definition activity.
Monday May 14 Q2
McKenzie, Beeler - Failed to reach quorum
Ballot Reconciliation for Core Principles R2
Discussed the issues raised on documentation of "isDocumentCharacteristic". Agreed to a strategy to correct, and that would issue second ballot on this item only.
Ballot Reconciliation for RIM R5
Considered how to respond to votes. Most related to "isDocumentCharacteristic" (above). Will treat as non-significant change, and not re-ballot.
All these decisions will be revisited on a Reconciliation Conference Call.
Monday May 14 Q3
Woody Beeler, Lee Coller, David Hay, Linda Bird, Henry Wijaya, Ewout Kramer, Rik Smithies, Tanya Achilles, Dale Nelson, Paul Knapp, Mark Tucker, Piers Hollott, Ann Wrightson, Bo Dagnell, Lloyd McKenzie, Brian Pech, Grahame Grieve, Rajan Rai
FHIR Data Types and Extant Data Types R3 Requests
- Do we do an R3?
- If so when? and
- What is the content difference?
ISSUES from FHIR:
- Will need to consider, with respect to FHIR, how to handle choices (selection of data type at instance time.)
- Grieve: "Lack of a typed identifier is a serious omission in V3" (most noticeable in CDA)
FHIR Changes in re R2:
- Structure of PQ with originalText
- Relaxing constraints on: TEL, CD, II
- Text content exists everywhere
- Additional nullFlavors - "process error", and "as text"
- Simpler version of GTS that might be of use in other settings
From R3 Page
The Data Types R3 Wiki page will be re-written to reflect these decisions. The following list is intended to summarize what is in scope for R3:
- Update IETF RFC number under Language data type (BASIC) from RFC 3066 to RFC 5646 (or the latest RFC number) (Wendy Huang)
- Address the issues raised by turn "II.scope into a mandatory attribute"
- Add note to integrity Check about it's utility with digital signatures
- Add ability to flag lists as unsorted.
- PIVL is unable to express that an interval starts at a time relative to another event.
- Make it explicitly obvious that relative URLs are allowed in CDA documents for URL types.
- Add a property to the CD datatype as "valueSetDefinition" with a data type of ED intended to contain the CTS 2 (or possibly other) definition of the value set.
Whence DT R3?
Should follow the first cycle of FHIR deployment to gain experience before trying to tie down the R3 changes upon which FHIR is dependent.
Ergo: Initiate Data Type 3 project proposal discussion in January WGM 2013.
Moved: Grieve/Beeler Unanimous
Tuesday May 15 Q1
Woody Beeler, David Hay, Tanya Achilles, Dale Nelson, Mark Tucker, Ann Wrightson, Lloyd McKenzie, Brian Pech, Grahame Grieve, Andy Stechishin, Charlie McCay, Steve Fine, Jay Lyle, Rik Smithies, Grahame Greive, Ewout Kramer, Abdul-Malik Shakir, Charlie Mead, Peter Hendler, Paul Knapp, Linda Bird,Even Osfer, Ndri nguyen, Peirce Hollot, Bo Dagnall
FHIR Methodology - Profiles and Extensions
Q: Are profiles necessary for the definition/use of extensions? A: Yes extensions are defined by profiles. Some Profiles will only be 'grab bags' of extensions
Q: Are profiles part of the core specification? A: Profiles are expected to be balloted separately from the core specification (much like implementation guides), the specification of profiles and extension is part of the core specification.
Q: Are profiles cascading? Do profiles apply to profiles? A: Yes
Q: What are drvinign design principles behind FHIR's extension mechanism A: Static wire format. 'Z' segment problem - I have received blob of data or collisions of content/naming. Leveraged experience of OBX segment.
Full principle: in order to deliver a specification for all user you need to deliver a specification that meets all users needs - must have an extension mechanism
Ann: this is really an open system but may need to explore implementation with different trade-offs Grahame: not essentially an open system but rather a system for governance.
Grahame; If you design an extension and do not vet it with HL7 there is a risk the extension will be orphaned by the community. HL7 may decide it needs to be part of the resource and update the resources. However, If HL7 blesses an extension it will never become part of the resource (this is proposed governance policy, not formalized)
Grahame describing the 'unrolling' or 'unpicking' mechanism for the definition of profiles. Note Grahame illustrated the profile in a spreadsheet but this is simply a way to define the profile (user data collection tool) the normative form is the resource/profile definition (XML)
Sidenote: FHIR is not tied to tools, there is no prescribed tooling stack. The specification should be easily produced with readily available, software that is expected to be installed on the majority of resource definer's systems.
There was a side discussion on the display format of the resource definition.
AMS asked about the mapping columns on the definition spreadsheets. Lloyd described the mapping for the Lab resources. Lloyd stated that the there is a desire to express the mapping in RDF.Ann raised the question of whether GELLO could be used for the mapping language. Lloyd detailed some of the challenges that occurred in the RIM mapping. AMS requested that we use something existing rather than inventing a new language. Grahame has a preference for using RDF. There was agreement in the room that we have need for two columns, one for the committee at large (textual) one for the Modelling facilitator.
Grahame presented a sample extension. There was discussion on the possible 'noise' injected into the instance from the inclusion of extension definition reference, but it was ultimately agreed that the mechanism was a reasonable trade-off between additional data and explicit-ness