Skip to end of metadata
Go to start of metadata

HL7 Clinical Genomics Weekly Call - 11 Dec 2018 11:00 AM (US Eastern)

 Agenda


Attendees Sign-in

(Presiding co-chair:  Bob Milius - NMDP/CIBMTR- bmilius@nmdp.org  )

  1. Caterina Lasome - iON Informatics - cat@ioninformatics.com
  2. Kevin Power - Cerner - kpower@cerner.com
  3. Joel Schneider - NMDP/CIBMTR - jschneid@nmdp.org
  4. Patrick Werner - MOLIT Institut/ Heilbronn University - patrick.werner@molit.eu
  5. Dora Finkeisen - MOLIT Institut - dora.finkeisen@molit.eu
  6. James Jones - BCH - james.jones.bch@gmail.com
  7. JD Nolen - Children’s Mercy Hospital - jlnolen@cmh.edu
  8. Liz Amos - NLM - liz.amos@nih.gov 
  9. Lloyd McKenzie - Gevity - lmckenzie@gevityinc.com
  10. Scott Robertson - Kaiser Permanente - scott.m.robertson@kp.org
  11. Ning Xie - BCH - ningxie2018@gmail.com
  12. Deepak Sharma - Mayo Clinic - sharma.deepak2@mayo.edu        
  13. Bob Freimuth - Mayo Clinic - freimuth.robert@mayo.edu
  14. Jamie Parker - Carradora Health - jamie.parker@carradora.com
  15. Mullai Murugan - BCM - murugan@bcm.edu
  16. Clem McDonald - NLM - clemmcdonald@mail.nih.gov        

Previous Minutes Approval

Agendas and Important Dates

 

Date

Co-Chair

Agenda

Important Dates

9-Oct-18

 

No Meeting

 

16-Oct-18

Kevin P

WGM Review

2-to-FHIR PSS - vote

Committers channel on Zulip

NIB for IG - vote

Code Systems / Value Sets

 

23-Oct-18

Kevin P

FHIR IG Block Vote - vote

Committers channel on Zulip

Code Systems / Value Sets

 

30-Oct-18

Kevin P

Tracker 19453 - Rename Sequence -> MolecularSequence

Tracker 19440 - Update to ‘stu note’ on FHIR Core profiles and artifacts

Tracker 1900 - Relax cardinality for reference.windowStart/windowEnd

Previous FHIR Core Trackers with no or inadequate resolution documented

Code Systems / Value Sets (Discussion: Patrick/Julian)

Oct 31: All R4 STU comments reconciled in gForge with votes and no “tracker issues”

Oct 28: NIBs due - IG must be “feature complete” and building in HL7’s continuous integration environment.  If not, the NIB will be refused

6-Nov-18

Bob M

 

 

13-Nov-18

Bob M

 

Nov 18: IG Content deadline – only QA changes after this point

Nov 18: IG Ballot reconciliation spreadsheet due - two ballot items left to reconcile

20-Nov-18

Kevin P

Topic 0: Ballot Sign Up

Topic 1: May 2018 Ballot Reconciliation Complete

Topic 2: Non-ballot tracker items still remaining

Topic 3: Code Systems / Value Sets (Discussion: Patrick/Julian - Clem/Liz)

  Nov 23: Co-Chair Nominations Close at 5:00 pm Eastern (email nominations to NOMINATIONS@HL7.ORG)

27-Nov-18

Bob M

 

Dec. 2: IG Final freeze deadline

4-Dec-18

Bob F

Topic 0: Ballot Sign Up

Topic 1: Jan 2019 WGM Agenda Planning

Topic 2: Non-ballot tracker items

 

Dec 6 - Consensus pool signup closed

Dec 7 - Ballots opened

 

Dec 7: Co-Chair election statements due to HQ

11-Dec-18

Bob M

 

 Dec 14: Deadline to post your WGM agenda on the WGM information page (WG Health metric)

18-Dec-18

Kevin P

 

 

25-Dec-18

Holiday

meeting cancelled

 

1-Jan-19

Holiday

meeting cancelled

 

8-Jan-19

Bob M

 

Jan 7 - Ballots closed

2019 January Working Group Meeting

Health Level Seven International January 2019 Working Group Meeting, San Antonio, TX

  • Location: San Antonio, TX United States
  • Start Date: January 12, 2019
  • End Date: January 18, 2019
  • Early bird discounts end December 21! 


External efforts

  • GA4GH Genomic Knowledge Standards (GKS) (leads: Bob Freimuth, Andy Yates)
  • DIGITiZe (aka National Academies)  (Grant Wood, JD Nolen)
    • Still attempting to schedule a follow-up meeting (Sandy, JD, Grant) for next steps.
    • will report to FHIR Foundation about DIGITIZe
  • ClinGen/ClinVar (Larry Babb, Bob Freimuth)
    • Internal follow-up on outcomes of the in-person modeling meeting, vetting ideas
    • Significant consideration given to GKS/VR approaches, may help drive that work
  • CDISC PGx (Dorina B.)
    • no report
  • ONC Sync for Genes (Bob Freimuth)
    • Collected example genetic test reports from each pilot site, mapping to FHIR profiles
    • Planning for Jan WGM Connectathon
      • S4G will conduct a number of tests under Scenario 10 of the CG genomics track
    • Discussion regarding VCF attachments on genetic test reports (with JD)
      • O&O discussed last week, will initiate a zulip chat thread to finalize
      • Patrick: Observation.attachment was removed in R4, will likely be back in R5
      • OO and II will have a joint call before the WGM to discuss

 Subgroup reports

Topic 0: Ballot Open!

Topic 1: Jan 2019 WGM Agenda Planning

Topic 2: Non-ballot gforge issues

Tracker

16412

N/A

David Poloway

Summary

IG should be US realm,

Links

http://build.fhir.org/ig/HL7/genomics-reporting/general.html

Resolution

Notes

Not persuasive as stated - out of scope for this IG (currently balloted as Universal) but US-specific may be created later, re: regulations, etc. if deemed necessary

Details

Have it be US realm (got feedback from other countries wanting flex on ontology to be used for constraining profiles)- as was done for the "HL7 Version 2 Implementation Guide: Clinical Genomics; fully LOINC-Qualified Cytogenetic Model, Release 1 - US Realm"

Follow-ups

-Mon, 07 May 2018 - by Lloyd McKenzie-The only terminology we're forcing in this IG is the use of LOINC codes to disambiguate types of Observations.  LOINC is freely available in all countries without payment.  If organizations want to use additional code systems (e.g. SNOMED), they're free to do so.  We already have the precedent of universal profiles requiring the use of LOINC codes with the vital signs profiles.  There's nothing in the intended scope that's U.S. specific.  Opening up the terminology would significantly reduce interoperability across countries and reduce the ability to leverage common data analysis, decisions support, detailed clinical models and other artifacts, so this should only be done if it's clear that implementers *cannot* use the terminology currently mandated rather than merely "would prefer not to send translations".

 

Patrick: motion to recommend as not persuasive to Tuesday group

Discussion/objections: None - recommendation: Not persuasive

  • Reviewed tracker (see link on “tracker” text in above table)
  • Motion: accept resolution
    • Move / 2nd:  Patrick/Bob F
    • Discussion:   none
    • Vote: (Abstain / Against / For):  0 /0  /  14
    • Result:  Motion passes

 

Tracker

16409

N/A

David Poloway

Summary

Link definitions to standard ontological terms,

Links

N/A

Resolution

Notes

Duplicate of 16513 - need glossary.

Details

Link definitions to standardized ontological terms (may be done via LOINC links to other ontologies)

Follow-ups

Discussion/objections: None - recommendation: Not persuasive None/mark as duplicate

  • Reviewed tracker (see link on “tracker” text in above table)
  • Motion: accept resolution
    • Move / 2nd:  Kevin P/Patrick
    • Discussion:   Bob F - question how is resolution recorded?; Kevin P mark as resolution as None, and mark as duplicate
    • motion to make ballot resolution field in gforge as "None" instead of "not persuasive"
    • Vote: (Abstain / Against / For):  0 /0  /  14
    • Result:  Motion passes

 

Tracker

16395

N/A

David Poloway

Summary

Add definition for certain types of data 1.2 Guiding principles,

Links

http://build.fhir.org/ig/HL7/genomics-reporting/index.html

Resolution

Notes

Not persuasive (non-substantial) - Referenced text already updated with previous tracker changes, AllelicState context mentioned on http://build.fhir.org/ig/HL7/genomics-reporting/obs-described-variant.html

Details

“It tries to ensure that data is captured in a manner that's consistent regardless of the type of testing that was done to ensure data can be consistently queried even if captured differently (e.g. variations identified in assay tests are reported in the same manner as those identified through direct sequencing)”-- In some cases, data/annotation information comes from statistical databases, not directly from testing. Therefore, should add a definition for them and a guide on how to use them correctly. For example, AllelicState info comes from an ethnic database.

Follow-ups

Further Discussion/objections: None - recommendation: Not persuasive

  • Reviewed tracker (see link on “tracker” text in above table)
  • Motion: accept resolution
    • Move / 2nd:  Patrick/Kevin
    • Discussion:  
      • Bob M: is this a duplicate?
        • Not clear where, so use "previous changes";
      • Scott: if previous change was done and is consistent with tracker #, then is should be 'persuasive' with already applied.
    • Vote: (Abstain / Against / For):  Clem / 0 /  15
    • Result:  Motion passes

 

Tracker

16388

N/A

David Poloway

Summary

Clarify definition for Panel,

Links

http://build.fhir.org/ig/HL7/genomics-reporting/general.html

Resolution

Notes

Persuasive with mod (non-substantial) - Add a couple examples of panel use in text to the section for a better picture of intended use.

Details

Make a clear definition for Panel, because it seems like Container but it is not functional terminology, unlike finding, impacting, interpretation, etc

Follow-ups

Clem: panels in v2 are similar to component collections in FHIR

Andrea: from lab reporting it is determined by order requisition. seems like addressed in key reporting elements. and not persuasive...

Kevin: to refresh, these panels are to group Observations

Lloyd: without this structure there would be no way to group Observations within a Genomics Report. (by drug/ordered vs supplemental/particular variants/etc)

 

Original proposal: provide link to Background section, where usage of “panel”, “finding”, “impact” and “interpretation” are further outlined.

Having already addressed the text, we should move on to an example or 2 of various panel configurations.

 

Andrea: examples are good

 

recommendation: persuasive with Mod - Add a couple examples of panel use in text to the section for a better picture of intended use.

Further Discussion/objections: None

  • Reviewed tracker (see link on “tracker” text in above table)
  • Motion: accept resolution
    • Move / 2nd:  Patrick/Clem
    • Discussion:   none
    • Vote: (Abstain / Against / For):   0/  0/  16
    • Result:  Motion passes

Tracker

16415

N/A

David Poloway

Summary

Add text to 1.3 sequenced variants Observation-genetics,

  Links

http://build.fhir.org/ig/HL7/genomics-reporting/sequencing.html

Resolution

Notes

Not persuasive (non-substantial) - This IG design prefers components over extensions and should not reference the core profiles for this guidance

Details

Add text for Observation-genetics profile perhaps at the end of the section (new section):

 

The Observation-genetics profile Observation-genetics profile is used to interpret variants from sequence resource. Clinical usage may need more specific representation of variant at locus or structural variant in whole genome. Some of the attributes of the profile follow: The observation-genetics Sequence extension will refer to the Sequence resource for sequence information related to this variant. The observation-genetics Interpretation extension will refer to an Observation instance which contains clinical interpretations for the variant described. The code, effective[x], issued, performer, method, specimen elements can be used to describe how the genetic observation (variant and sequence data) is obtained. Other extensions are used to describe attributes of this variant such as Genomics Source Class, Amino Acid Change Type, etc. These are mappings from v2 and lonic code reference with details can be found in this list .

Follow-ups

Further discussion/objections: None - recommendation: Not persuasive

 

  • Reviewed tracker (see link on “tracker” text in above table)
  • Motion: accept resolution
    • Move / 2nd:  Patrick/Kevin
    • Discussion:   none
    • Vote: (Abstain / Against / For):   0/0  / 16
    • Result:  motion passes

Tracker

16396

N/A

David Poloway

Summary

Add text to 1.3 sequenced variants,

Links

http://build.fhir.org/ig/HL7/genomics-reporting/sequencing.html

Resolution

Notes

Persuasive with mod (non-substantial) - add text to desc. Var. paragraph: “The underlying clinical concept behind each component is mapped to a specific LOINC code.”

Details

Add text: This includes three basic types of gene mutation. The most fundamental type is Described Variant which is the information carrier of complex variant structure and contains the differences of lab and clinical sequencing data. It describes clinical features of each variant (in Allele). Those features are in Component and are coded in the LOINC standard to which they are exclusively matched. Most of the features concerning gene variants are integrated into Described Variant. Therefore under many circumstances, it would provide essential explanation and makes it easy to be compatible with V2 standard.

Discussion

Discussion/objections: Patrick-- first comment isn’t that necessary, can just suggest the second one

(resolution updated above)

Further discussion/objections: None - recommendation: persuasive with mod

  • Reviewed tracker (see link on “tracker” text in above table)
  • Motion: accept resolution
    • Move / 2nd:  Clem M/James J
    • Discussion:   none
    • Vote: (Abstain / Against / For):  0/ 0 /  16
    • Result:  Motion passes

 

Tracker

16391

N/A

David Poloway

Summary

Add introduction and examples for Reporting,

Links

http://build.fhir.org/ig/HL7/genomics-reporting/general.html

Resolution

Notes

Persuasive with mod: Already have a placeholder page to build out querying but currently just has the DAM: http://build.fhir.org/ig/HL7/genomics-reporting/domain.html,

could update pending further review (particularly examples for querying within FHIR servers based off of DAM use cases, guidance on Region-Studied queries, and potentially emerging integration with GACS) but not ready to address fully yet.

Details

Lots of reminders for "searching" in all Reports. May be better to make an introduction and examples for searchable functions in all kinds of Report.

Discussion

Bob D: we may have to delay this until we determine a good way to query described variant. It has so many optional components so a search Operation may be required.

Clem: querying is tricky as they try to minimize indexed fields

Lloyd: operations may not help much here. The challenge is variability in what labs capture.

Bob D: example--lots of field in DV could use to convert to canonical representation. Rather than querying for any of these we could convert it to canonical and search for that, using an Operation. Build, refseq + position has been our best bet so far. I want to query a genomic server for all the information that addresses a particular gene.

Lloyd: essentially moving work around, 1: lab/server does this work, 2: user does a bunch of querying. Is one more reasonable than the other? May be better providing guidance on the different types of queries you would need to make to get to what you want. Labs housing the data would have to fully adopt the search operations in order for them to be useful.

Andrea: will likely take some time to get there from the lab side of things.

Patrick: we won’t be able to see how every lab will work but can focus on providing sample queries for specific use cases.

Bob D: would be nice to define a set of parameters a genomics server should need to support, but it will take some time to get there.

 

Summary: Not ready to address yet

  • Patrick - will provide examples, and include search examples, takes time to develop, keep in backlog,
  • Bob D - agree, will develop examples, not mature enough to develop a formal proposal
  • defer for further analysis
  • Bob F - discuss after connectathon? during WGM?
  • Clem - wants to see a real straw man proposal
  • Bob D - working on this, plan on testing during connectathon
  • Kevin P - updated tracker with follow-up discussion, removed "ready to vote" status

 


Harder Trackers about Ontologies and Code Systems (from Mon CG FHIR call)

 

Tracker

16407

N/A

David Poloway

Summary

Use standard ontology consistent with genetics community consensus,

Links

N/A

Resolution

Notes

Not persuasive - can already use additional codings in component.code

 

Details

Anchor the component "text" field and observation/other profile-constraining fields to more standardized/commonly-used ontologies in genetics, internationally used instead of LOINC that community is in consensus with (e.g. SO, SNOMED, etc). LOINC genetics terms have not been widely used in field- most recently introduced in v2- with limited adoption. Not an ontology (partial with panels)- so reasoning on knowledge would be harder if we don't use standard ontology with genetics community consensus definitions

Follow-ups

 

  • Reviewed tracker (see link on “tracker” text in above table)
  • Motion: accept resolution
    • Move / 2nd:  Kevin P/Patrick
    • Discussion:  
      • Patrick: slicing on "text" is bad,
      • Lloyd - "text" is for human, not computable
    • Vote: (Abstain / Against / For):  0 /  0/  16
    • Result:  motion passes

 

Tracker

16408

N/A

David Poloway

Summary

Anchor profile to more standard/common genetics ontology,

Links

N/A

Resolution

Notes

Option 1: Continue to require LOINC codes (while still supporting additional codes being sent in different systems/ontologies)

 

Option 2: Change required LOINC bindings in Observation.code and component.code to an Internal (HL7) Code system based off of SO/SNOMED for faster development and easier links with standardized ontologies

 

Option 3: Use Option 2 for development purposes but also request corresponding LOINC codes when near publication for integration with other clinical data

Details

Anchor the profile constraining field (currently a LOINC code) to more standardized/commonly-used ontologies in genetics, internationally used instead of LOINC that community is in consensus with (e.g. SO, SNOMED, etc).

Follow-ups

Pending Discussion

  • no time to review this 

Chat

  • Kevin Power Host 10:03AM
  • 0Anonymous10:29AM
    • Bob F needs to step away for a few minutes - will be back shortly
  • 0Anonymous 10:42AM
    • I’m back

Clinical Genomics Docs

  • No labels