The reporter has a point. The intent is for “medication, administered not done” to cover all occurrences of administration. Since the logic does not have an sequence, if this piece of logic is checked before the medication administration, then the case may flow to the numerator for the wrong reason. The reason why we used “medication, administered not done” was that we wanted to capture patients for whom the order may have been placed, but the medication was never administered (in addition to patients for whom the medication was never ordered). I guess the question is if we are looking for “medication, administered not done”, is it implied that an order was placed? I’m not sure the QDM goes into this detail describing the relationships among data types and from HL7 RIM/CDA/ perspective I’m not sure what the assumptions are (or aren’t). So I don’t disagree that “medication, order not done” may have a place in the logic, but it doesn’t replace “medication, administered not done”. I’m not sure how we can represent the concept of ALL instances versus ANY instance in this case. Generally speaking, when we use negation rationale we are considering the scenario in which the recommended care was not performed; we typically look for a single instance of the recommended therapy (notable exception: VTE-3), so the inverse of that is NO instance meets the criteria. The current logic certainly doesn’t address this conditional scenario. In my opinion, the cleanest way to address this would be to move all the negation rationale criteria to the denominator exceptions section. That way “medication, administered not done” (and potentially “medication, order not done”) would only be checked if no instance of the medication was given to the patient. The reason why we didn’t do this in the first place was that these cases would flow to the numerator in the original measure and we were trying to maintain a base of comparability by at least maintaining criteria in the same “bucket” (denominator, denominator exclusions, numerator). Lately I’ve been thinking this may not have been the greatest idea, due to issues just like this one. For example, in the same measure, a patient flows to the numerator if they’re administered IV heparin. This is supposed to be an exceptional situation and IV heparin is not a recommended form of VTE prophylaxis. However, because the criterion is lumped together with appropriate prophylaxis options in the numerator section, I’m not sure everyone will understand this nuance. Again, the solution would be to move the criteria to the denominator exceptions section, since that would create a clear separation between the recommended care (numerator criteria) and exceptional situations In current VTE-1 specification, the numerator criteria do not provide VTE anticoagulant administration evaluation detail such as how an anticoagulant medication should be administered in terms of dosage, schedule, and duration. It is upon measure implementer and physician to determine the actual evaluation criteria of how to determine an anticoagulant is properly administered or not administered with reason. For example, let's presume that a patient with VTE risk should take 5 doses of heparin for best VTE prophylaxis outcome, however, in real world, following scenarios could occur:  1. the patient only took 3 doses (missed 2 doses without reason)  2. the patient only took 4 doses (missed 1 doses without reason)  3. the patient only took 4 doses, and had a reason of why the dose 5 was not administered.  4. the patient took all 5 doses. An EHR system may only record the administered heparin, or the reason of not administering heparin, it is likely that it has no information about the missed doses (scenario #1 and #2). Evaluating patients with these four different scenarios relies upon how the anticoagulant checking logic is designed and implemented as such logic is not clearly expressed in the VTE-1 CQM. Moving negation rationale logic to denominator exception helps but won't resolve the issue. We need further discussion.