This wiki has undergone a migration to Confluence found Here
<meta name="googlebot" content="noindex">

MnM Minutes WGM 201505 Paris

From HL7Wiki
Jump to navigation Jump to search

Return to MnM Minutes for 2015

Sunday May 10 Q3 - Finalize Agenda/Review Hot Topics

Attendees

Jean Duteau, Chair/Scribe
Lloyd McKenzie
Austin Kreisler
Grahame Grieve
Josh Mandel

Minutes

  • Agenda Finalization

Walked through the agenda and kept the quarters as assigned. Determined chairs/scribes and suggested topics.

  • M&M Business

Discussed "forking of the RIM" for Structured Documents

Monday May 11 Q3 - FHIR Datatypes Reconciliation

Attendees

Lloyd McKenzie (chair), Gevity, lloyd@lmckenzie.com
Jean Duteau (scribe), DDI, jean@duteaudesign.com
Bill deBeaubier, Systems Made Simple, Bill.deBeaubier@systemsmadesimple.com
Richard Kavanagh, richard.kavanagh@hsck.gov.uk
Tim McNeil, SureScripts, tim.mcneil@surescripts.com
Michelle Miller, Cerner, mmoseman@cerner.com
Peter Jordan, pkjordan@xtra.co.nz

Minutes

Reviewed the following tracker items:

Created the following tracker items:

Wednesday May 13 Q1 - M&M FHIR Vocabulary Joint Session

Attendees

Minutes

Wednesday May 13 Q2 - M&M FHIR Vocabulary Joint Session

Attendees

Bill de Beaubien
Heather Grain
Jos Baptist
Sandra Stuart
Sylvia Thun
Tim McNeil
Ted Klein
Lloyd McKenzie
Rob Hausam
Russ Hamm
Rob McClure

Minutes

Single Source Terminology

There is a need for consistent and improved vocabulary governance as we move towards a single vocabulary resource across HL7 families and products.

Discussion explored whether we use V3 artefacts and processes as a quality mechanism.

It has become apparent that we have to stop allowing poor terminologies across product lines.

Harmonisation process is documented, supported (to a degree), has established operational practices and is well known in some parts of the community. Objective: · Supports resolution of issues between terminology content representation

· Identification of quality issues in these data components


Shared content · Process required for consistent review of content

· Address the V2 disparity and quality problem


Represent the V2 content in the V3 model will merge these two vocabulary – this will reduce us from 4 to 3 representations (V3, FHIR, and CDA). The process will require some modifications as the ballot processes differ. These are the first early ones to bring together.

Also want to bring FHIR into this and CDA without breaking the systems or causing undue overheads.

Processes to support cloud sourcing which support · Access

· Improve governance and review

· Support responsiveness to needs for change


Moved: Ted Klein that MnM cosponsor the Terminology Quality Assurance PSS. Vote: Yes: 10, Abstain 0, No, 0.

The TQA PSS must be approved by the FHIR management group for approval. – Action Rob McClure. · We will need to identify the current use of terminology across the families and we will seek that from the relevant cosponsors and working groups.

· Will capture issues that people know they have, and consider how this can be aligned

· Identify quality requirements and guidance (using V2 as a basis)

· Identify tooling requirements and issues, including timely input to gForge replacement decision making

· Identify ballot implications

· Define scope of what is included in quality assessment processes

· Define who will be responsible for corrections of different component elements

· Define priorities for development

These processes may occur concurrently. Knowing the broad vision, information can be assembled on processes and ideas and ways to pilot different options.

Timelines for calls on this project will be fortnightly on the week after VSD calls on Tuesday. TQA call will be fortnightly at 5pm (New York) starting Tuesday 26th May.

RDF

Need to be able to take the relevant conformance field which is structured and value sets and possibly maps and express instances of these as RDFs / OWL to define syntax for FHIR instances.

Seeking a single solution for a wide range of options and problems. The vision is that you should be able to say find me all of the patients who have blood pressures in this range who are not on a hypertensive in an RDF format. Want to make sure people will be able to use it if they have the horsepower, and use case and tooling.

We also have the ability to turn information into schemas and do the same thing and take profiles and value sets and turn them into OWL or RDF2.

In RDF you typically want to point to a concept with one URL but in FHIR RDF there will need to be able to assert concatenations to handle differences in what is inside a code.

Ontology which says a FHIR coding with a code system, url, version etc then this is a subtype of the codesystem – these could be generated if the concept syntax is known.

Some systems which have URLs some which use special characters etc – the ontology will have dependence between the properties and the concepts as defined.

The approach proposed would have a consistent RDF presentation of codes as every code system has the three elements required to assert meaning. These would be different for each code system but the mechanisms would be consistent in FHIR representation. The RDF will expose the mechanisms but not necessarily the actual code content. The RDF equivalent will always include all components.

CDA2.1

You cannot expect a great deal of assistance from MnM and Vocabulary for structured document implementation of CDA2.1 as we don’t have the background information or processes established at the moment. Take the schema which has the enumeration and run against the MIF to identify the discrepancies. Creation of a 4th RIM – will not include a sandbox with open ability to modify established structures.

Thursday May 14 Q1 - FHIR Methodology

Attendees

Lloyd McKenzie (chair)
Michelle Miller
Tim McNeil
Richard Kavenagh
Prashant Trivedi
Simone Heckmann
Ronn Archamboult
Jeffrey Ting
Stephen Royce
Peter Jordan
Bill de Beaubien

Minutes

Tracker Item 7622

Extensive discussion of issues around annotations. Considered defining this as an extension and whether it should be modifier or not. Also, if it was an extension, whether systems should communicate information using both the existing "note" elements as well as within the annotation extension. Eventually decided that the "least worse" solution was to add a data type to capture annotations and use it all places where notes/comments appear unless MnM agrees that a single string would be sufficient

Tracker Item 8049

Discussed adding XHTML as allowable for extensions. Several people in the room felt they would use it. See tracker item for disposition & vote

Thursday May 14 Q5 - Facilitators Roundtable - Joint with Vocab, FHIR

Attendees

Richard, Ted, Russ, Julie, Rob, Hugh, Eric, Riki, Lloyd, Paul, Alexander, Andy, Melva, Ron, Martijn, Kevin, Michel, Kevin, Grahame, Austin, Jean

Minutes

0) HARMONIZATION DISCUSSION July 21-22, 2015 Initial Proposals - June 21, 2015 Final Proposal - July 12, 2015

-We need to provide some means for all groups to know which proposals are accepted and which are now.

1) AROUND THE HORN O&O - 2-3 small-medium SD - some proposals dependent on RIM version FHIR - 2 small

2) VOCABULARY UPDATE -Terminology Quality Assurance Process that is pan-HL7 = projects will be created to design and then implement the process -the requirement for a display name for all codes in FHIR and v2 has met resistance and the Tables project is reconsidering the requirement

3) FHIR UPDATE -finish your triage of items (block vote, necessary, substantive changes) = assessment by each WG by end of May if mid-July publication is achievable and, if not, what will slip -additional DSTU2.1 balloting in January = covers order workflow processing and financial management resources. Contact FMG if you have new resources/profiles that you want to include DSTU2.1 -FHIR Infrastructure WG was created -Some of the ballot comments were actually against v3 or v2 Vocabulary and not FHIR-specific = If you have those, let Grahame know of the items. -Motion: Tell the committees that they have until mid-June to update their FHIR resources codes to have displayNames otherwise the FHIR Infrastructure group will set the displayNames and they will have to fix them after if needed. (Grahame will generate a list of suspect codes) Paul/Grahame, 18-0-1

4) M&M UPDATE -FHIR Resources will have notes/comment elements changed to note and be of type Annotation. Contact M&M if you want a simple string.

5) RIM FORKING StructDocs would take RIM 2.07 and an updated version of vocabulary and create a fork of the RIM "2.07a" that would combine those into a RIM release. SD has created a project to create the fork but there are no volunteers at this time.

6) FHIR RIM MAPPING -Resources were intended to be mapped back to the RIM. =Not many of the mappings have been done. Their value was very limited. One ballot comment about RIM mappings but a lot of comments about the v2 mappings. -Different levels of RIM mapping. Skeletal is the minimum and Executional is the largest. -Deprecate the existing RIM mappings and Grahame/Lloyd will create a new RIM mapping framework based on the RDF representation. We will then enforce Skeletal mappings.

7) RIM RDF REPRESENTATION -A new RDF representation will be produced about the RIM. This will make the FHIR RIM mapping doable. It will be based on the existing RDF representation. Discussions are on the ITS RDF sub-group.