The the most performant way to search for a resource will vary greatly based on what the server supports (for example, a questionnaire checking for history of a condition likely wants to search by code, but in US Core searching by code is optional). How do we want to create library content / authoring patterns / engine guidance to help deal with this problem?
A related challenge is that you often want to search on a value set of possible codes, but the way US Core is setup generally searches by a single code (ex: GET [base]/Condition?patient=1032702&code=[http://snomed.info/sct|442311008]). An engine could deal with this by expanding the value set and inlining it into the URL, but then you might run into limits and need to split the query. This also gets into a heuristic of whether it's even worth doing that, as compared to just searching on Patient and filtering down the results afterwards.
The the most performant way to search for a resource will vary greatly based on what the server supports (for example, a questionnaire checking for history of a condition likely wants to search by code, but in US Core searching by code is optional). How do we want to create library content / authoring patterns / engine guidance to help deal with this problem?
A related challenge is that you often want to search on a value set of possible codes, but the way US Core is setup generally searches by a single code (ex: GET [base]/Condition?patient=1032702&code=[http://snomed.info/sct|442311008]). An engine could deal with this by expanding the value set and inlining it into the URL, but then you might run into limits and need to split the query. This also gets into a heuristic of whether it's even worth doing that, as compared to just searching on Patient and filtering down the results afterwards.