Many measures are concerned with a certain number of events having occurred. While the “COUNT” operator provides this functionality, in several cases the temporal spacing of the events must also be considered as part of the calculation of the count. For example, consider the case of medication administration that must occur at least four times. If we want to ensure that those events were at least a day apart we would have to use specific occurrences as follows:

AND: "Occurrence D of Medication, Administered: X" >= 1 day(s)
starts after end of ("Occurrence C of Medication, Administered: X" >= 1 day(s) 
starts after end of ("Occurrence B of Medication, Administered: X" >= 1 day(s) 
starts after end of "Occurrence A of Medication, Administered: X"))
AND:
    AND: "Occurrence A of Medication, Administered: X"
    AND: "Occurrence B of Medication, Administered: X"
    AND: "Occurrence C of Medication, Administered: X"
    AND: "Occurrence D of Medication, Administered: X"

This does not make it clear that we care about the total number of administrations that were at least one day apart. Furthermore it obfuscates the actual rationale we are trying to reproduce, since the number ‘4’ does not appear within the measure logic. The “COUNT UNIQUE” operator can be used to simplify this logic to:

COUNT UNIQUE 1 day(s) >= 4: Medication, Administered: X

It is now readily apparent the threshold of events we are concerned with, as well as the relative temporal spacing of one day. The addition of the “COUNT UNIQUE” operator provides us a significantly simplified representation.