Newly created language to satisfy related issue (FHIR-28719) implies that manual user action is required to identify which EHR data should not be sent to a payer. 

This commit: Added guidance around hook invocation https://jira.hl7.org/browse/FHI… · HL7/davinci-crd@597c853 (github.com) adds:

>Implementations *SHALL* have the ability to flag particular orders, appointments, encounters and other records as 'payer-sensitive' and ensure that such sensitive information is not queryable by payers and that such information does not trigger payer hooks or get included as part of payer hook invocation. The process to flag information as payer-sensitive should fit within standard workflow and minimize provider effort. The determination of the flag SHOULD be set as early in the process as possible in the workflow to ensure that decision support can be provided as early as possible.

This language has two problems:

1. It implies (perhaps requires) that manual action is required on behalf of provider staff to "flag" data that shouldn't be sent to payers. The CRD specification should not restrict the use of automated rules and processes for this purpose.
2. Needs validation, but I think that the intent of the linked issue, or at least the needs of the use-case in general isn't about identifying data/records that shouldn't be sent to any payer ever, but rather about only sending data to a specific payer. E.g. self-pay shouldn't send data to payer, generally payers should only receive data when appropriate and relevant.