The current trend in federated query systems that use SHRINE is to dictate that the ontology used for SHRINE is identical to the one used in the local instance. Examples of this are found on the PCORI and ACT projects. The reasoning behind this is that it makes the process of making the SHRINE adapter mapping file much easier. In most case the mapping file becomes a direct 1 for 1 mapping from the SHRINE term to the term found at the i2b2 instance.
Highlighted in the tutorial below are approaches that each site can use to modify the ontology for use in their i2b2 instance. Multiple strategies are covered such as expanding the use of patient_dimension based queries for Demographics; updating concept_dimension based queries to reflect the concept and modifier basecodes that are in use locally; and finally, merging locally derived terms into standard ontologies.
Ontology Approaches for Federated Query Systems A power point presentation on organizing your ontology for use in federated query systems such as SHRINE