System may receive a patient's chart from two different sources such that it has two Patient resources that each have duplicate data (for example, the same "real world" encounters and associated notes are duplicated).

The $merge operation for patient should describe what behaviors are recommended for handling this duplicate data.

Right now, we say this:

The merge data processing SHALL update all references that refer to the source patient to reference the target patient.

However, we probably should allow servers to also perform de-duplication, including merging or deleting duplicate data, at the server's discretion, as this would prevent a search from returning duplicate clinical content post-merge.