Stream: implementers
Topic: Patient link negation
Peter Bernhardt (May 11 2016 at 16:02):
In reviewing an internal use I see a need for considering how a system may express the assertion that two patient records within its control are NOT the same person.
Peter Bernhardt (May 11 2016 at 16:02):
Currently the link type enumeration only allows for qualifying an affirmative link.
Peter Bernhardt (May 11 2016 at 16:03):
But what happens if the system later wants to say "These two patient records are NOT related."
Peter Bernhardt (May 11 2016 at 16:03):
Particularly useful if you implement a patient registry as a read-only event store.
Peter Bernhardt (May 11 2016 at 16:05):
And you need to handle the dreaded "un-merge" use case.
Grahame Grieve (May 11 2016 at 16:06):
there is no unmerge only unlink
Peter Bernhardt (May 11 2016 at 16:09):
i understand, Grahame.
Peter Bernhardt (May 11 2016 at 16:09):
but there is definitely an "un-merge" use case for implementers
Peter Bernhardt (May 11 2016 at 16:10):
i'm adding a change request to clarify and get this in front of the PA WG
John Moehrke (May 11 2016 at 16:12):
Isn't that just an exercise in some Application twisting apart the parts, and doing a RESTful -create and update? Thus REST already handles it.
Grahame Grieve (May 11 2016 at 16:15):
it goes far beyond that
Peter Bernhardt (May 11 2016 at 16:15):
That's right, John. In a homogenous FHIR system this would be easy to handle. What I'm looking at is a use case that involved handling ADT-A40 messages and then someone trying to back that out.
Peter Bernhardt (May 11 2016 at 16:18):
Tracker item 9993
Brian Postlethwaite (May 11 2016 at 19:27):
Yuk
Peter Bernhardt (May 11 2016 at 19:37):
;)
Last updated: Apr 12 2022 at 19:14 UTC
