Is it important that we should stick to the “concept_class” criteria while mapping lab terms?
Meaning for ex: If I have a lab measurement, I have chosen the “Measurement” domain but the name of the measurement is available under “lab test” and I also have similar terms under “Procedure” concept_class of the “Measurement” domain.
Am I good as long as I map it to the “Measurement” domain and it doesn’t really make any difference - the concept_class it is mapped to? I can map it to either “lab test” concept_class or “procedure” concept_class?
For measurement_concept_id, the only constraint is that the concept has to belong to “Measurement” domain. concept_class does not matter, meaning it can be “Lab test” or “Procedure” or any other value.
For value_as_concept_id field in either Observation or Measurement table, my understanding is that there is no constraint at all, meaning values that go into these two columns can belong to any domain.
Dmytry, I know that the value_as_concept_id field in Observation table has no domain constraint, meaning it accepts standard concepts of any domain. But is it also true for value_as_concept_id field in Measurement table?
Yes, this is exact rule for MEASUREMENT.value_as_concept_id.
Personally, I don’t like this restriction. We have a bunch of use cases when it’s better to concepts with different domains. For example, there is Measurement - bacterial culture, its results are bacterias, it’s better to use Organisms (Observation domain), because they have hierarchy, but Meas Value concepts don’t have, see this concept for example Athena