Date: February 6, 2020
This session was hosted by Clinical Decision Support (invited CIMI and CQI). See CDS minutes for details. 2020-02-06 Q1 CDS WGM Minutes. Highlights posted here.
|Bryn chair; Ken scribe|
Problematic when authoring CDS or eCQMs when there are different versions of FHIR and each EHR implements differently. At the last WGM, Claude showed a logical model approach that translates to different versions of FHIR. The next step in discussion is about whether CMS or others would support infrastructure requirements / tooling. Motivation is because of multiple versions of FHIR and support is different depending on the implementation. Therefore, writing a knowledge artifact requires writing it in multiple versions of FHIR - either write to a specific FHIR version or use a logical model (QUICK) and the author and the implementer can use QUICK to then handle the mapping to the FHIR version(s) in use. Challenges are gaps in definitions in different FHIR versions. FHIR also provides multiple versions of freedom and QUICK may help reduce those concerns by providing constraints.
Claude Nanjo: Wrote knowledge artifact on QUICK for validation (Claude) - also supporting OMG - were able to generate recommendations. Model development was a subset of FHIR as a proof of concept. Can leverage mappings from prior versions but the mapping is not always one-to-one. Needs to be tested in a real production setting at University of Utah before a recommendation to HL7. The modeling led to many recommendations for FHIR R5.
Discussion about challenges:
FHIR Shorthand (Mark Kramer) may help with this challenge - SUSHI is the tooling. FHIR Shorthand provides a mechanism to express data from different FHIR forms in usable information.
The cost needs to be paid either in the mapping or in the tooling.
Logic written against logical model is cleaner.
Perhaps another approach is to develop a mechanism to cross FHIR resources for specific metadata (e.g., start/stop times, body site, etc.).
Needed for eCQM and CDS to share content:
Note - to use QUICK for implementers to extract data would add significant implementer burden unless the scope or eCQMs and CDS artifacts is limited to US Core content (i.e., without the extra content covered in QI-Core that does not exist in US Core.
Next steps -
Remaining QI-Core trackers (all non-ballot trackers - all September ballot trackers have been resolved and the reconciliation spreadsheet uploaded to the September ballot desktop):
Has a vote but still says triaged Persuasive: Mover/Seconder/vote: Floyd Eisenberg/Ken Kawamoto: 14-0-2
There was insufficient time to address these issues during Q1.
|Bryn||Bryn created a new page on the CDS Confluence site: Clinical Decision Support Standards. This site will provide information about all of the CDS knowledge artifacts supported and how they might related to each other and plans for the future. Discussed migration of the CDS Knowledge Artifact Specification to the FHIR based Clinical Reasoning specification.|
See CDS minutes which include a link to the presentation provided by Mohammed Afzal. 2020-02-06 Q3 CDS WGM Minutes
The work includes:
Adaptive CDS: A system which adapts, that changes and its content and elements are based on personalized and contextual information without changing the intent of recommendation.
For the rest of the discussion, refer to the presentation available in the CDS WGM minutes - 2020-02-06 Q1 CDS WGM Minutes.
This session was hosted by Clinical Decision Support (invited CQI). See CDS minutes for details. Highlights posted here. 2020-02-06 Q3 CDS WGM Minutes
The meeting was cancelled due to advanced indication of lack of attendees and content.