Proposed Wording: Disallow the use of the interpretation attribute of the base Observation Resource.

Comment:

For new types of observations like in genomics, interpretation is typically done by a different performer, using a different method and at a different time than those of the source observation; in addition, multiple interpretations could be created for the same observation, because the knowledge and evidence behind the interpretation process is rapidly changing or controversial and different approaches are available. Therefore, while this bundling is indeed convenient in obvious cases like glucose interpretation, in more complex cases, it promotes pre-coordination constructs, which is bad for analytics (e.g., clinical decision support, clinical quality measurement, etc.).