Existing Wording: The other difference is that, for FHIR, the human-readable aspect is not limited to the document - it's driven by the resource itself. This may result in slightly less flexibility in rendering approach if a document is compiling existing resources.

Comment:

Is the red text true? Since the List resource can have its own narrative, which is not necessarily equal to the sum of its collection of resource narratives, it seems to have the same flexibility as a CDA section, which can be derived from its entries (analogous to resources) or have the narrative be independently authored by a human or machine.