Stream: implementers
Topic: linking condition and observation
Sebastiaan Raap (May 24 2019 at 13:57):
Out of curiosity:
https://www.hl7.org/fhir/condition.html#9.2.3.5 states:
It is common as part of checklists prior to admission, surgery, enrollment in trials, etc. to ask questions such as "are you pregnant", "do you have a history of hypertension", etc. This information should NOT be captured using the Condition resource but should instead be captured using QuestionnaireResponse or Observation.
For example, the question is about bladderfunction. In the Observation you provide the answer: Yes my bladderfunction is abnormal. Suppose you wish to create a Condition instance for the compromised bladderfunction. You would link the BladderFunction-Observation Instance to the Condition through condition.evidence.detail (right?).
But let's say the patient has an UroStoma because of this Condition. Would it be logical to link this to the Condition? We've captured Stoma in an observation profile.
https://www.hl7.org/fhir/extension-workflow-reasonreference.html states
Indicates another resource whose existence justifies this event.
https://www.hl7.org/fhir/observation-definitions.html#Observation.focus states
The focus of an observation could also be an existing condition, an intervention, the subject's diet, another observation of the subject, or a body structure such as tumor or implanted device.
In the UroStoma instance, should I use Observation.focus for the reference to the condition? Or the Observation.reasonReference extension? Or should I link these resources differently?
Thanks in advance!
Lloyd McKenzie (May 24 2019 at 14:58):
Focus is what's observed. reasonReference is why you decided to make the Observation. Observation could be used to assert causation, but typically "A was caused by B" is handled using Condition (there's an extension that allows you to make causation links)
Sebastiaan Raap (May 27 2019 at 08:19):
I was thinking about using Condition at first, but now I'm in serious doubt: The bladderfunction (care, not cure) information model has a valueset attached with three options:
1. incontinent
2. Sometimes a little accident (max once a day)
3. continent
Now if 1 or 2 is the case, i'd be more than happy to set this information in a Condition, but to me it doesn't make sense to profile this in Condition if there's no actual condition (option 3).. This is why my first instinct is to profile this in observation.. But if a client in the Care setting would do a GET /Condition request, and the compromised bladderfunction wouldn't be returned I'd think that wouldn't be correct. Other way around, if the client would return "continent", that wouldn't be correct either. I don't think Condition shouldn't be used for stating unaffected body functions, should it?
Any ideas on how to tackle this?
Lloyd McKenzie (May 27 2019 at 13:18):
This very much sounds like an Observation. You could potentially infer a Condition record from the Observation, but it'd be pretty limited in terms of the amount of information available (when it started, who diagnosed, how severe, etc.)
Sebastiaan Raap (May 27 2019 at 13:40):
thank you!
Last updated: Apr 12 2022 at 19:14 UTC
