I’m in 2 minds about whether to require Reference.reference.
There is a pattern where we only provide the display value and use this to create a natural boundary in the navigable resource graph that the consumer is obliged to respect:
prescriber: {
display: 'Dr AN Other'
}
However, for these references in GP Connect there is a presumption that the resources would (?should) be available, in which case cardinality 1…1 would be appropriate.
Agree that quantityText feels like extension on quantity. As quantity fields have all been constrained with cardinality 1…1, I wonder if this is intended as an alternative when the dispense instruction doesn’t have all the fields? @RKavanagh?
I’m also not sure if the constraints will create a problem for some cases. For example, do we expect the medication order to capture devices as well? On quick perusal, I can’t a UCUM code for ‘colostomy bag’ …
Even with aligned UoM, a number of dispense instructions are only held in plain text on source GP systems, and in any case may not be easily represented in structured form.
There’s a UoM included in dm+d that all (NHS Emgland) EPS prescriptions have to use as the UoM when describing the quantity. All GP systems have to be able to output that UoM (coded and text) as part of the EPS prescription. What I’m less sure of is whether they hold it in the system as such.
UCUM only specifically includes “measurable” units not “countable” things such as bags, tablets, capsules etc. But it allows all those things by using its curly braces syntax.
No that is right. They are considered as comments, with no semantic, and are somewhat discouraged.
The idea is that these things are not measurable, so are categorically different from what UCUM is for. Ideally countables and measurables would be separated but it’s common to treat them interchangeably.