For instance, suppose a patient is taking medication on a certain schedule (let’s say, once a day), then it makes sense to convey the first day on which medication has been taken as the ‘initial exposure’.
But suppose a patient takes medication for several days in a row, then pauses for a few days and then starts taking the (same) medication again and within that second cycle has an allergic reaction. Would the ‘initial exposure’ be equal to the first day of the first cycle of taking medication, or the first day of the second cycle? (I’d expect the latter.)
And the same question if the intermediate period is quite longer, e.g. if a patient takes the same medication twice for a certain antibiotic treatment, with several years apart.
Background of this question: I am modeling a data set that contains an ExposureDateTime element to FHIR and I was wondering to what degree I could make use of the existing openEHR-exposureDate extension (see Extension: exposureDate - FHIR v4.0.1). In this data set the ExposureDateTime is seen as the ‘relevant’ exposure date/time, i.e. the first time in a cycle the patient has been exposed to the substance.
Your question is a good one, and the comment to the ‘Initial exposure’ element also gives an indication that the “initial exposure” may not be as simple as a single datetime. The description for the element states
The date and/or time of the first exposure to the Substance for this Reaction Event.
In my thinking, the “for this Reaction Event” qualification is key. In your example where a patient had been exposed to the medication in a previous course without a reaction, but then had a reaction during a second course of the same medication, I’d expect like you that the initial exposure was the first day of the second course.
As an example, I’m using this archetype in the context of allergen immunotherapy. In this treatment pathway a patient is given injections of an allergen multiple times over three years, and on the initial visits several injections on the same day. In this case, say the patient has a reaction after the second injection of birch pollen allergen on the second day of injections. I then set the “initial exposure” to the time of the first birch pollen injection on the second day of injections.
I agree - the key to understanding is that this is inside the reaction event. In the original joint modelling effort with the FHIR community, I’m pretty sure this was the intent. It looks as if this concept has been removed from the FHIR allergyIntolerance resource and handled as an extensiion
Thanks Silje for your elaborate answer! It seems like this element in openEHR is indeed suitable for our use case as well, which means that we’ll adapt the openEHR-exposureDate extension in FHIR (probably together with the exposureDescription extension to give additional detail on which time instance has been chosen for the exposure).
Thanks Ian! The openEHR-exposureDate extension is implemented as an extension on AllergyIntolerance.reaction, so in FHIR it’s used inside the reaction event as well.