As noted in FHIR-36383, CDS has already agreed to remove cqf-questionnaire because SDC's population mechanisms are better defined and more robust.  

The description here doesn't take into account the fact that questionnaires are hierarchical and can have repeating elements.  As such, there can be numerous items in a QuestionnaireResponse that correspond to a single Questionnaire item.  Furthermore, if any questions or groups are optional, then they'll be omitted in the QuestionnaireResponse if no answers are provided.  There's also no discussion of the possibility of multiple answers, what happens if the user overrides answers, dependency of answers on previous questions, difference between determining a 'default' value vs. calculating a score.  This guidance just isn't sufficient - and SDC has tons of guidance and examples, so we should just be pointing at it.