Although the existing language is good, I feel that one additional piece of guidance is needed to indicate to implementers that some of the bound value sets are broad enough that a code can always be found to map to.

Here is a suggested point to add to this section to make that clear:

NOTE: For the US Core Procedure Codes value set and the US Core Condition Code value sets, the value set includes a number of high-level abstract codes.  For data not captured by the system transmitting the information, all effort should be made to automatically convert coded data to a fine-grained code from the specified value set. If automated conversion to a fine-grained code is not possible, the system can provide the existing code or the free text, but the system SHALL also provide a coding from a high-level abstract code, such as SNOMED CT 'Procedure', in order to remain conformant with the binding.