The immunization refusal reason template is currently an observation with no value. Would it not make more sense either as an `<act>` with no value or as an observation with code = 71801-5 Reason for Intervention Refusal and the current valueset in value?

In the C-CDA<->FHIR mapping project, we discovered a difficulty in distinguishing this template from Indication, both of which are attached via entryRelationship[@typeCode="RSON"]. In fact, Indication itself is particularly problematic as it only has an example binding to the Problem Type value set. The only way to reliably differentiate these templates is via templateId.

Complicating this is the following recommendation from the companion guide, admittedly about Procedures, but if indication can be used to indicate why a procedure was not done, why not why an immunization was not done?

When representing that a procedure was not performed, the Indication (V2) (identifier: urn:hl7ii:2.16.840.1.113883.10.20.22.4.19:2014-06-09 template MAY be used to represent the rationale for not performing the procedure. More than one indication template may be contained within a Procedure template.

Suggest remodeling the Immunization Refusal Reason with code = 71801-5 and moving the current value set to value.

This doesn't completely solve the problem of ambiguous templates - in fact, in either case (today or whether this change is made), you could attach the Indication templateId to Immunization Refusal Reason, and the observation would conform to both. Perhaps there is a different typeCode that could be applied. Could also add a requirement in the Immunization Activity that the Indication template SHALL contain code = 59785-6 Indication for Immunization. That would make it completely unambiguous.