Comment:

It is unclear how one can represent a 'dependent' observation in FHIR. For instance, some observations may exist in a panel but may also exist independently of a panel - the systolic component of blood pressure. On the other hand, some observations are associated with a specific condition and can only exist in the presence of condition statement. For instance, the diameter of a wound on the left hand (and not that of the one on the right foot). Yet, it is unclear (1) how one may tie these to the condition (a diameter attribute? an attribute relating the condition to a set of observations, one of which is wound diameter?) and (2) whether the observation resource is appropriate for representing such dependent observations? I would recommend that FHIR and CIMI coordinate their activities on that front.

Summary:

How are dependent relationships represented?