Stream: implementers
Topic: Workflow pattern
Grahame Grieve (Feb 13 2017 at 19:16):
@Lloyd McKenzie I am unhappy about the manifestation of some of the workflow patterns. See task GF#12819 for example
Grahame Grieve (Feb 13 2017 at 19:26):
in fact, as of today, I'm officially in agreement with the CIMI people that we've created a mess in FHIR, around .actor.role
Grahame Grieve (Feb 13 2017 at 19:27):
and I think it's a self inflicted wound
Jose Costa Teixeira (Feb 13 2017 at 19:31):
just for immunization: I think that immunization is done in a way that hurts itself.
It is both the request and event in one resource. This is common for immunizations since v2, and it is the same as if we have a "medication" resource and use it for prescription and administration.
Grahame Grieve (Feb 13 2017 at 19:32):
that's a long running discussion in the PHER committee. You should take it up with them and explain why you think it's a problem with use cases
Jose Costa Teixeira (Feb 13 2017 at 19:32):
i would have defended harmonizing immunization to the workflow pattern, e.g. immunizationRequest and immunization"Event"
Jose Costa Teixeira (Feb 13 2017 at 19:33):
i will find them and bring it up there. thx
Grahame Grieve (Feb 13 2017 at 21:22):
Procedure.performer.role Example Procedure Performer Role Codes (FHIR) CareTeam.participant.role Example ParticipantType (v3+FHIR) MedicationAdministration.performer.role Example Procedure Performer Role Codes (FHIR) MedicationDispense.performer.role Example Procedure Performer Role Codes (FHIR) Immunization.practitioner.role Example Immunization Role Codes (v2) Encounter.participant.type Extensible ParticipantType (v3+FHIR) ChargeItem.participant.role Required Procedure Performer Role Codes (FHIR) Provenance.agent.role Extensible ProvenanceParticipantRole (FHIR) AuditEvent.agent.role Extensible Audit agent Role ID Code (DICOM) Consent.actor.role Extensible Consent Actor Roles (v3) Appointment.participant.type Extensible ParticipantType (v3+FHIR) Claim.payee.type Example Payee Type Codes (FHIR) ExplanationOfBenefit.payee.type Example Payee Type Codes (FHIR) Contract.agent.role Example Contract Actor Role Codes (FHIR)
Eric Haas (Feb 13 2017 at 22:52):
And all of OO's too.
Grahame Grieve (Feb 13 2017 at 22:55):
did I miss some? where?
Bryn Rhodes (Feb 13 2017 at 22:58):
There's a participantType element in PlanDefinition currently a code | Required | ActionParticipantType (FHIR).
Grahame Grieve (Feb 13 2017 at 22:59):
that's not the same pattern, so I didn't include it. But I suspect it will be shown that it needs to be finer grained than just resource type
Bryn Rhodes (Feb 13 2017 at 23:00):
Yes, we're in the process of expanding it.
Grahame Grieve (Feb 13 2017 at 23:00):
then you'll probably fall into the same pattern
Eric Haas (Feb 13 2017 at 23:09):
DR, Obs (extension).
Grahame Grieve (Feb 13 2017 at 23:13):
event-performerRole?
Eric Haas (Feb 13 2017 at 23:15):
yes
Lloyd McKenzie (Feb 13 2017 at 23:20):
I don't understand the problem with the pattern for most of these locations. Whether it falls into the 80% for all of them, I don't know, but having an unconstrained vocabulary for most of these is unavoidable. What makes sense is going to be context specific for different types of encounters, procedures, etc. Obviously if the work group can agree on mandating an extensible value set that draws on free terminology, great. But I don't think variablity here is any more problematic than variability in allergy codes or drug codes
Lloyd McKenzie (Feb 13 2017 at 23:21):
Note that the pattern was driven by what was happening in existing resources (e.g. Procedure and Encounter) and identifying whether it was reasonable/appropriate in a large portion of other resources. It's not something that's mandated - work groups can, should (and in several cases have) simplified the model when it made sense to do so.
Grahame Grieve (Feb 13 2017 at 23:26):
I'm doing more analysis. Provenance.agent.role... I'm not sure how to read the RIM mappings - is it mapped to Participation.typeCode?
Lloyd McKenzie (Feb 13 2017 at 23:38):
Yes, though that's a "creative" way of expressing the mapping. It's talking about Participation.typeCode and outbound ActRelationship.typeCode.
John Moehrke (Feb 14 2017 at 17:05):
you do realize that RIM mappings is a black-art... the vast majority of us have no idea how to spell RIM... So, I am sure everyone is very happy to have someone with RIM knowledge to correct or at-least question constructively
Grahame Grieve (Feb 14 2017 at 19:31):
yes we know that
Thomas Lukasik (Feb 14 2017 at 19:54):
..but lest we forget ;-)
Grahame Grieve (Feb 14 2017 at 19:55):
well, I'm willing to hold a wake for v3 at HIMSS ;-)
Dave Carlson (Feb 14 2017 at 20:01):
I would be happy to propose a toast for the removal of requiring RIM mapping for a FHIR resource to reach FMM 2 ;-)
Lloyd McKenzie (Feb 14 2017 at 20:55):
You only need it to reach FMM 3. And there's serious consideration to dropping that as a requirement, but it won't happen before STU 3 is published
Last updated: Apr 12 2022 at 19:14 UTC
