2020-09 Patient Care Virtual WGM Attendance
Attendance: 2020-09 Patient Care Virtual WGM Attendance
Monday (4pm Eastern) BRR/PC AdverseEvent
Recap last BRR/PC joint discussion
- 2020-02-27 Patient Care FHIR Conference Call
- Patient Care had found base resource changes
- Hugh Glover started a profile, but not finished or ready to share yet
- BRR moved off into ResearchStudy/Subject
- Rebecca Baker and Mike Hamidi are not present in this session, but were involved in CDISC mappings.
- Stephen asked if the mappings are available for public review? Subscribe to BRR listserv. Reviews through HL7 are coming throughout 2020 (and then the actual ballot process)
- Christine Denney - gaps identified and captured in gap document (with Lloyd McKenzie) TBD which gaps would result in JIRAs for base spec vs extensions in IG. Need to balance guidance in R4 vs what's coming in R5
- action taken / follow-up
- mitigating actions
- questionnaires (did headache occur?) - expected adverse events with meds
- understanding what part of body (body site/structure)
- anything that happens after they sign consent is considered an AdverseEvent (then investigate causality) - no distinction between event vs reaction
- Jean Duteau - blood products and vaccination work
- IBM to HL7 project
- Would like to use R5 version of AdverseEvent (based on clean-up that has since been done)
- Past work on BRR/PC work has really helped
- Mark Kramer - cancer drug trial
- looking for light weight way to capture unexpected adverse events
- Oncologists not interested in expected adverse events (nausea)
- BRR is very interested in going to R5 for AdverseEvent and a bunch of other resources
- Hugh inquired about Connectathon - Vulcan - not on the schedule for January (too soon?). Hugh asked if there was interest in a Connectathon and when?
- Robinette Renner is interested, but January is too soon. She will know more in the coming weeks on timing.
- Amit Popat - public health was looking at it in context of vaccination
- Will continue to poll for interest in May 2021
Patient Care FHIR Backlog - specific to AdverseEvent
- FHIR-27784 - Make AdverseEvent.outcome repeating RESOLVED - CHANGE REQUIRED
- FHIR-26436 - Explain the interaction between AdverseEvent.code and AdverseEvent.resultingCondition TRIAGED
- FHIR-26435 - The definition of AdverseEvent.resultingCondition restricts it to results to exposures. RESOLVED - CHANGE REQUIRED
- FHIR-26434 - Request to add attribute to capture latent medical harm WAITING FOR INPUT
Will reschedule monthly BRR/PC conference calls on Thurs at 5pm Eastern
Adjourned at 5:40pm Eastern
Thursday (4pm Eastern) PC + SDWG
David Riddle (presenter)
- Mapping Author and Informant Information to FHIR US Core Profiles for FHIR Resources Derived from C-CDA Documents
- CCDA docs can be used to exchange patient information
- Can automating fransformationof C-CDA data to FHIR resources
- Must have consistent mapping between CCDA data and FHIR resources
- Obtain comprehensive and consistent directions on how to map
- provision of level of detail required
- Focus is on CCDA participations
- Reality Encountered
- Mappings not intuitive
- different groups are proposing different mappings - e.g. FiveWs, SOA
- There may not be a single one size fit pattern
- Intuitive mapping at the header level - CCDA Healder to the CCDA-on-FHIR-US-Real-Header Composition Profile
- Mapping at the entry level for PAMPI Resources less intuitive
Reasons why PC did not feel condition recorder
- cardinality mis-match
- recorded date set proxy for the onset date. Had an open question around if recorded date is when the data was first entered into the system.
- PCWG thought that author should not map
- SOA cross-paradigm project has different mappings
- Ken Lord - built the SOA author mappings. Progress made on the cross-paradigm mappings was decided within SOA and Higher-ups and OO at HL7 that SOA may not have been the right place for this.
- Recommendation that OO and PH wanted to come up with normative mapping between V2.x and FHIR
- Hans has developed a bit of a methodology with very well founded FHIR concepts that can be used across CCDA to FHIR. Recommend looking at that to see if there is a methodology that can be derived.
- Lisa Nelson -
- did SOA ballot comments get completed? Yes. didn't receive any comments on the mappings.
- Can CDA Management group pull this out? The context and concepts of the mappings doesn't belong to SOA. It belongs somewhere else.
- V2 mappings is further along. Is there the concept of Author in V2 mappings? Yes.
- If the CDA Management group take ownership of this work and would like to take a look at what the V2 people did.
- Ken Lord - is important to also understand the semantic
- Stephen Chu - differences is in the different paradigm. For example, CCDA effectiveTime can be overloaded.
- Lloyd McKenzie -
- FHIR mappings are maintained in a single place. Mapping to specs in multiple places are not helpful to implementers. Will take this back to FMG as a governance question.
- Normative mapping is scary because we can't make a point-in-time assertion if people are populating the data as intended. Does not mean all the implementations are populating the data in the same way.
- There is no expectation in FHIR to author FHIR model to make mapping in FHIR as easy as possible.
- The objective is to keep the resources as tightly aligned as possible with the information in the implementer space and not to what is mapped.
- Mapping exercise is appropriate but when if gets hard does not mean make changes to the FHIR model to fix the mapping.
- There can be variations in how resources represent the same type of elements. But the variations should be diven by differences in implementer requirements and not by who is creating mapping and what their goal was at that time
- Gay Dolin -
- Use case - when multiple authors results from aggregated documents that comes from an HIE.
- Lloyd - what story is being told in the CCDA documents? If keeping recorder in Condition resources should be cleaned up. Means different people are authoring at the different times and may not best represent what provenance conveys.
- David - this is where issues with differences in provenance is run into
- a single author can be associated
- two authors with the identical timestamp is not supported in FHIR. Are there systems that does this? CCDA allows it but FHIR does not know what it means. Options would be to pick the first, or do extensions. The real question is what is the expectations of the behavior.
- David - have established there are CCDA docs with multiple authors.
- Lloyd - expectation that CCDA on FHIR team would provide guidance. If multiple could map the earliest one or do other things but need to consider the ramifications of doing these different things. There is always a scope bandwidth timeframe.
- Gay - The context is that CCDA on FHIR is still a document. Is this still the context when mapping from document to data?
- David - not mapping document to document. Mapping from a CCDA document
- Lloyd - mapping from CDA document to FHIR document should allow the resources to stand alone.
- Ken - to understand the general principal that the source is CCDA and the target is FHIR.
- Lloyd - This part of the implementation is 'iffy' because the standard is not super stable.
- Ken - agree but having at least some baseline adds value
- Lloyd - CCDA on FHIR and V2 on FHIR IGs provide this baseline.
- David - CCDA on FHIR IG does not go to that level of detail.
- Ken - Agree with Lloyd and there is a lot of work that has been done that is reusable
- Lloyd - the initial CCDA on FHIR shows some depth and that's okay. You can take this work and make it go deeper rather than put that detail of work somewhere else. Reality is what we have right now is not sufficient. We may never get to a point of just push a button and transform perfectly. But this does not mean we can't go further. Making a conscious choice to consolidate all our guidance is the best way to go.
- David - what Lloyd is saying make a lot of sense. Concern is less for a push button transformation but the more it's left up to implementers the more hindrance to interoperability
- Gay - maybe the next step would be for SDWG to do an update to CCDA on FHIR. Agreement from Vassil and Lisa.
- Vassil - look at the entries in CCDA and see how they map to Resources.
- Stephen - the few months of discussion, the concept of participation can be decomposed into a number of FHIR categories. Neither CCDA nor FHIR have decomposed these concepts adequately. Is the objective of your exercise is to do the mapping consistently and the endpoint is for a transform push button or is this to look a little deeper and see what changes in FHIR need to be done so there is not semantic loss when transformation occurs.
- David - can't make a separation between mapping and semantics. Would it make sense to narrow the focus and establish authoritative mapping for PAMPI.
- Stephen - agreement to scope. Patient Care owns PAP. Patient care would like to work with David on doing the mapping.
- David - it's logical for SDWG to allow the owners for the FHIR resource to provide feedback.
- Gay - the home will be SDWG and would be a collaborative effort or cross work group
- Stephen - in the past there have been disconnect between SD and the clinical domain.
- Gay - agreement that we have to figure this out. The main thing is there would definitely have to be collaboration. The work would inform a ballot that would be associated with CCDA on FHIR.
- Lisa - we could be building consensus on how this is done while it's being worked on. May need to do a project and propose some stuff. Could start with a confluence page under CCDA on FHIR.
- Gay - will need to bring this project to SDWG and see what they say.
- Stephen - Lisa has been the bridge between SD and PC and has done good work in keeping the 2 groups aligned.
- Lisa- this could be included in the PSS. Would also be the liaison for PCWG with other work groups. Suggest adding results to PAMPI (MAPPIR)
- Stephen - Emma is also involved with representing PC at SDWG. What type of assistance does David need from PCWG
- David - interested in asking implementers what they are doing. Have received some guidance from PC about recorder. May be forced from a timeline perspective to do my own thing for now.
- Stephen - what do you need from PC for now?
- David - may need to do more review with PC after re-grouping. Will align with Michelle. Have to thing thru other aspects of mapping that would need patient care input.
- Stephen - a good approach to is to put forth some JIRA tickets
- Gay - will need discussion with SDWG before putting together a PSS
- Meeting adjourned at 1733 EST.
Friday (10am Eastern) PC + Vocab
- CIMI: Pain model project & ballot plans
- Are CDA status issues closed?
- Extension to support pertainsToGoal
- Allergy code
- Model bindings? Goals, status, expectations.
- Status codes: seeing a similar pattern in Meds. Need codeableConcept, or code + extension.
- Publication of Representing Negatives (Representing Negatives)
- Review select re-writes; especially table
- 25, 87, 101: category alignment
- 75: outline
- 76, 101, 126, 128: pattern, method
- 90: next steps
- V2 coverage: anyone have examples?
- ANF: no actual examples to use
- Any more examples to add?
- Include ANF in appendix
- Remove dependency on Logica Covid-10 IG ballot
Nathan presented the Pain deck.
Pain is the most commonly documented symptom. The Nursing Knowledge Big Data conference has been a key driver to standardizing this data element. The project (sponsored by CIC, co-sponsored by CIMI & PC) has produced a mature LOINC panel to represent the state of the art information model for pain.
Lisa: Gravity has followed a similar trajectory: 1. assemble the experts, 2. review existing documentation and practice; 3. engage Regenstrief to help harmonize and iron out issues with putting the model into a Questionnaire context.
Next step: putting the Questionnaire into a FHIR IG, with guidance on inferring other resources (Observation, etc.) from the QuestionnaireResponse.
Looking for connectathon partners, since they must conduct a connectathon prior to ballot.
a. we need to clarify connectathon requirement. It clearly should not apply to informative specifications. And can it be fulfilled outside of HL7-sponsored connectathons? Question for Clinical Steering Division.
b. Once this is clear, the project team will need to decide whether to ballot an informative specification sooner or a FHIR STU later, and how much later.
c. MaxMD offered to assist with testing & possible "connectathon."
CIC as the sponsoring WG will lead these efforts.Russ suggested Anesthesia should be involved. Others: PACIO, Patient Empowerment.
Absence of Allergy
Russ points out that an allergy is a clinical entity in the world that is expected to persist. "No known allergies" is not. AllergyIntolerance should not be where we record "NKA."
Is this a patient safety issue? If an application were to assess a patient record and found "no allergy to X," it might misread "X" as a risk and cause the provider to choose another, less appropriate medication.
Michelle will check to see if this issue has been voted on in the past. If so, there may be a process to resurrect it.
Michelle further points out that NKA is in the US Core, so there is some momentum. Change is unlikely for purely semantic reasons.
Lisa has an idea for how to profile this.
CDA & FHIR condition/allergy status
Condition clinical status seems to align.
Action item: post list somewhere. Jay to put on confluence; if there's a better place, we'll discuss.
Action item: We need a concept map. Jay or Rob H
CDA has changed their binding to dynamic, so if the selected SNOMED codes are deemed incorrect, they can be updated. Rob thinks this is unlikley.
This is not in CCDA, but Giorgio Cangioli has proposed a problem template that consolidates CCDA and IPS which does incorporate a verification observation using the FHIR values.
Action item: STU comments in support of this
Action item: IAT topic on the element for implementers (no takers yet)
Action item: We need a concept map. Jay or Rob H
Verification may be condition-specific. Entered-in-error may result from an administrative rather than clinical process. It's not clear that either issue affects the element.
Lisa suggests a new template for Problems that elides the Concern act might be taken up enthusiastically by implementers. If so, the mapping problem goes away, for future data anyway.
We will re-start meetings to finalize edits for publication. Watch for announcements. (Jay)
Status of Term Info
Expired DSTU. Team debating whether to publish or retire. CDA quality criteria require conformance, but it's not clear that it's followed.
Action item: request removal of quality criterion (Lisa)
FHIR publication includes "maps," which include SNOMED maps. These seem immature. We'd like to understand expectations around these maps.