Consider how to approach CodeableConcept.text

[mikix]

When we denormalize CodeableConcepts, how should we approach text?

Scenarios

1. Coding.display is present but no Concept.text
2. Coding.display and Concept.text both present
3. No Coding.display but Concept.text is present
4. No Codings at all, but Concept.text is present

1 & 2 are probably easy - just use the Coding.display value as the denormalized display.

3 probably doesn't happen that often. But could.

4 happens in Cerner for inline Medications, Device.type, and probably elsewhere. It's not clear yet if EHRs generally offer a dropdown for that kind of situation and/or allow freeform text entry. In which case, the number of unique text strings is huge (a consideration for using this string in a CUBE).

Use cases

1. What to show in the dashboard as part of the CUBE?
2. What to offer to studies for inspection? (i.e. do we want to allow display and/or text as columns for studies to try to parse?)

[mikix]

Current status is that we ignore text right now and always use display