The QDM (as of the December 2013 version) does not indicate cardinality for any attributes. It is obvious that not all attributes are applicable to every instance of a datatype, so one could reasonably assume that the cardinality of each attribute is 0..1 (zero or one). There are, however, measures such as CMS141v3 and CMS167v2 which require multiple results for a single event. See CQM-660, CYPRESS-90, and CYPRESS-169 for related discussion. If such cardinality is allowed, it should be explicitly stated in the QDM. That in turn could pave the way to the necessary changes to the QRDA specification to support corresponding cardinality for results so that we need not resort to such kludgy approaches as creating multiple instances of a datatype template, identical except for the result attribute. The QDM is the key to it all. If an attribute can have multiple instances, the QDM should say so and the proper support can flow to the MAT, HQMF, QRDA, Cypress, and so on.

Likewise, if there are attributes which should always be present for a given datatype, the QDM should state that. For example, if start datetime and stop datetime should always be present (and not null) for an Encounter Performed instance, the QDM should indicate that via the appropriate cardinality.