Chair and Scribe: Floyd Eisenberg 

Minutes from last call: OMOP Quality Measurement 23 February 2022

Recording of this call (access requires passcode):

Next FHIR-OMOP All Hands Meeting - March 28, 2022 (Monday) 4-5 PM ET - Subgroup leads will provide updates on their progress and 2022 roadmap (primary goals, objectives and deliverables).  You can find login details below.

The FHIR-OMOP WG- subgroups are as follows:

    • Data Model Harmonization 
    • Data Quality Measurements 
    • Terminologies 
    • Oncology 

Zoom Meeting: https://us02web.zoom.us/j/83645856405?pwd=OHJRdVVtb3BBQTdVaWp5UlBXUEdvUT09

Agenda:

Discuss potential project proposal and diagrams: 

Reminder from last call regarding COVID Criteria for cohort: https://github.com/National-COVID-Cohort-Collaborative/Phenotype_Data_Acquisition/wiki/Latest-Phenotype

Identify:

  • Content for priority use case for dQM:
    • COVID patient cohort and Non-COVID matched Control Cohort (COVID Criteria for cohort: https://github.com/National-COVID-Cohort-Collaborative/Phenotype_Data_Acquisition/wiki/Latest-Phenotype)
      • Align cohort definitions - terminology - OMOP code list(s) to VSAC value set(s)
      • Develop Atlas cohort definitions
      • Translate Atlas cohort definition/expression to FHIR-CQL expressions
      • Access data store and retrieve data (synthetic patient data):
        • OMOP data store for Atlas expression
        • FHIR data store for FHIR-CQL expression
      • Consider: Mapping terminology (VSAC, external) to OMOP code sets; Creation of Atlas queries and convert to FHIR-CQL 
        • Johns Hopkins University (JHU) – create OMOP derivative (N3C process) for COVID – can run phenotype definition for N3C – ready for May Connectathon

          • What needs to be done to allow creation of a de-identified cohort for use?
          • Consider synthetic data set from Andrew Williams (Tufts) - exists in OMOP; availability in FHIR may meet April timeframe - not certain
        • Use VSAC as reference but value sets exist only in OMOP. 

        • Investigate Georgia Tech activities for FHIR-OMOP implementation of Hapi FHIR server – have some OMOP transforms that expose the data through a FHIR server – use a CQL execution service to run against the FHIR-OMOP transformation layer. 

          • Discuss performance issues – provides a way to demonstration that one can run a query against FHIR-OMOP data.  Uses transformations built into that implementation. 

        • Consider availability of Paul Nagy (JHU - available after mid-March), and NIH, Ken Gersling and Raju Hemadri (Raju was on the call)
        • Compare concordance of OMOP/Atlas query to OMOP data store with FHIR-CQL query to FHIR data store
  • Timing of capabilities for upcoming Connectathon activities (what is feasible when):
    • HL7 FHIR Connectathon May 2022
    • NCQA/HL7 Digital Quality Summit (July 2022)
    • HL7 FHIR Connectathon September 2022
    • HL7 Ballot January 2023: Ballot informative IG?
      • Reconcile informative document
      • May require additional HL7 Connectathons to further evaluate new requirements suggested by ballot comment Further iteration dependent on issues identified

Discussion:

Terminology:

  • Qi.Yang is working on the OMOP to VSAC conversion - currently manual - can be automated using programming to pull into Atlas at one time.  Comment from Qi: Good news! Jack Brewster from IQVIA team has successfully loaded the VSAC into Atlas programmatically using R package!  He can do a demo during our next meeting if anyone is interested.
  • Davera Gabriel - VSAC has FHIR API hosted by NLM. Is that the method (answer - no; the information was provided directly - the load into Atlas was manual, not the mapping)
    • N3C work did the phenotype and the value set
  • Determine how the architecture works for use with additional use cases
    • Qi demonstrated the information received from Thomas White (MedStar) - all value sets by concept code - vocabulary, code - 70,000 rows, 565 value sets for the eCQMs used by CMS programs. He is willing to share scripts with anyone interested. The project looked at all of the vocabulary from VSAC - evaluate which are supported in OMOP; only 2 not supported (CDCREC - manually mapped, and HSLOC - healthcare service location) - all of the others (563) had mapping to OMOP. Some names are different (e.g., RxNORM versus RXNORM) - provides corresponding OMOP concept_id as standard concepts. That allows use of the concept_id in Atlas by giving the set a name and adding the concept_ids - currently manual but working on doing this programmatically.
    • Davera Gabriel suggested a method to use the FHIR API will allow incorporating updates that VSAC performs periodically. ICD-10-CM is not a standard vocabulary in OMOP. 

Next steps:

  1. Describe requirements for the various components of the architecture - Seeking input from WG members:

    • Terminology:
      • Infrastructure:
        • VSAC value sets and FHIR API
        • Programmatic methodology for extracting VSAC value sets and mapping terms to OMOP terminology (current manual upload)
        • Recommendations for handling concepts missing from OMOP (such as examples Qi experienced for HSLOC and CDC Race and Ethnicity Codes)
        • Subscription or periodic requery when VSAC updates the terminology
      • Participants:
        • TBD
    • Expression:
      • Infrastructure:
        • Atlas - SQL capability exists - need Atlas SQL instance for pilot testing
        • FHIR - CQL capability exists - need FHIR CQL instance for pilot testing
        • Expression conversion (what infrastructure is required here):
          • Atlas phenotype to FHIR cohort conversion
          • FHIR cohort expression with CQL to Atlas phenotype conversion
      • Participants:
        • TBD
    • Data Query and Response:
      • Infrastructure:
        • Atlas data store for synthetic patients
        • FHIR data store for synthetic patients
        • Query capability for Atlas patients and retrieve of OMOP cohort(s)
        • Query capability for FHIR patients and retrieve of FHIR synthetic cohort(s)
        • Structure to compare OMOP cohort with FHIR cohort, identify concordance and challenges
      • Participants:
        • TBD
  2. Continue to build on Project Scope Statement PSS-1981 such that it can be progressed.
  3. Further discussions with Vulcan to understand potential Accelerator support.
  4. At this point, if folks want to move forward with the project and/or Connectathon piloting, we definitely need people to step up to do the work to define the elements, to create the prototypes, and to pilot them.

Adjournment - Adjourned at 11:01 AM ET

Attendees


AttendeeAttendee's OrganizationAttendee's emailPresent
1Abdullah RafiqiESAC

abdullah.rafiqi@esacinc.com

yes
2NCQA

3

Ben Hamlin

NCQA

hamlin@ncqa.org

yes
4Bryan Laraway 


5


6


7

yes
8Dipti Gandhi


9Dorothy LeeNCQA

10Edward Smith


11Evan Patrick Minty


12

Floyd Eisenberg

iParsimony, LLC

FEisenberg@iParsimony.com

yes
13ESAC

14James Brash 


15Jamie Smith


16


17John David Evans


18JP

yes
19Julia Dawson

yes
20


21Justin McKelvy


22Latasha Estrada


23Luis Alaniz 


24Mia Nievera 

yes
25NCQA

26Ming Dunajick


27


28

yes
29


30

yes
31


32

yes
33Raju Hemadri

yes
34


35@Sandra Siami


36Sebastian van SandijkOdysseus

37

Telligen

SRankins@telligen.com


38MedStar Health

39Vickie Reyes (GC) 

yes
40

yes
41Zhen Lin


42JD LiddilM2GENjames.liddil@m2gen.com
  • No labels