Stream: Profiling Academy
Topic: Modelling results from Skin Prick Tests
Pétur Valdimarsson (Mar 11 2021 at 10:25):
Hello!
Ran into an issue that I thought I'd share to see if you have input. (Hope this is the right stream)
As a part of a IG I'm working on, I am modelling an allergy investigation based on among other things skin prick test and provocation.
To communicate the results of the skin prick test, my thoughts were to use diagnostic report with observations, since the result of the test is a list of observations on sizes of possible rashes from the substances tested.
The problem is expressing; what is being observed (skin rash), size of skin rash, and the known causation substance.
By using Observation.partOf I can keep a reference to the procedure of the skin prick test, which in turn has a usedReference, but is this the right way to express this?
Feels like the Observation resource becomes a tad "thin", i.e. lacking in data.
I also looked at the option of using AllergyIntolerance but, seeing that we are communicating quantifiable measurements for clinical decision making, it did not seem quite correct.
As always, the problem is solvable with a extension here or mishandling Observation.compartment, but if there are less ambiguous ways of solving the problem, I'm all ears (eyes?).
Thoughts on this?
Lin Zhang (Mar 11 2021 at 11:47):
You might want to use a panel structure pattern to model the test (an Observation for the panel) and its member Observations (See the example at http://hl7.org/fhir/observation-example-vitals-panel.html ).
Pétur Valdimarsson (Mar 12 2021 at 10:37):
Lin Zhang said:
You might want to use a panel structure pattern to model the test (an Observation for the panel) and its member Observations (See the example at http://hl7.org/fhir/observation-example-vitals-panel.html ).
Yes, that would be another way of expressing the grouping, but I would still need to force one of my three data points in by extension on Observation.
The three data points being; skin rash which is the symptom being observed, the value as a quantitative measure or coding, and finally the known causation substance.
Richard Townley-O'Neill (Mar 14 2021 at 23:48):
If the rash size and causative agent are not independent observations you could use something like.
Observation.code=skin-rash-from-test
Observation.component[0].code=rash-size
Observation.component[0].valueCode=big
Observation.component[1].code=rash-cause
Observation.component[1].valueCode=egg
Last updated: Apr 12 2022 at 19:14 UTC
