The "Guidance for the use of Negation Profiles" section indicates that authors should be aware that a single code from the not-Done value set can satisfy a negation profile query. Thus, when querying with a negation profile, results don't necessarily indicate that NONE were done, but rather that AT LEAST ONE was not done.

I'd suggest that the IG provide further guidance on how authors using negation profiles in eCQMs / CDS can confirm that NONE were done. I can think of two feasible approaches:

1. Add CQL logic that explictly checks the data for the presence of the qicore-NotDoneValueSet extension w/ the specified value set canonical as its value. This is likely the most explicit and correct approach.
2. Add a companion query that asserts that there is no existing result for the positive form of the query. While absence of data is not the same as negation of data, this may be a close enough proxy for some use cases when paired to assertion of the negation case as well.