CUIMC is attempting to provide distinctions between patient reported drugs and prescription drugs in our drug exposure table (currently no distinction is made). We see in the following Themis convention that these patient reported drugs should go in drug exposure domain with the corresponding appropriate type_concept_id concept to reflect patient reported.
However, some of the patient reported drugs in our source system do not have corresponding dates. We were thinking about populating the start date and the end date in drug exposure with the date of the encounter in which that patient reported use of that drug. The date obviously won’t reflect reality but the primary goal here is to be able to tell the difference between patient reported drugs and provider prescribed. Is there any convention for this use case?
If you know they were exposed to the drug recently, then it can go in the Drug Exposure table. If it’s a “past use of”, then it goes in the Observation table along with past medical history, past surgical history, etc. IF you don’t know if they are reporting drugs they are currently taking versus have taken, then err on the side of caution and put them in the Observation table. As might be the case with the following:
It they are reporting the drugs they are currently taking, then utilizing the visit / encounter date as the stop & start date is the most accurate reflection of what is actually happening to the person.
There is not a convention for imputing start and/or end dates for patient reported drugs, but it is something Themis could add. Want to create an issue here?
