Race mapping between C-CDA and FHIR is more nuanced than raceCode -> category / sdtc:raceCode -> detail.

Each value in C-CDA sdtc extension needs to be compared to the hierarchical top-level category:

• If it is a descendent of the category value found in the base raceCode element, then it is a detail code
• If it is a different category value, then a 2nd category sub-extension should be used.
• If it is a descendent of a different category value, then its parent category should be added as a 2nd category sub-extension, and the sdtc:raceCode should be added as a detail.
• then this is a detail code and can be added to the detail sub-extension. If not, then a 2nd category sub-extension should be added. If the value is not a category value itself, then its category is added as a 2nd category and it is added as a detail.

Meanwhile .... a variety of approaches could be used to populate the "text" sub-extension. Redox concatenates all the displayNames, filtering out categories where both a category and a detail value exist. But there could be other valid approaches.

(And of course, this is all subject to complete overhaul once the new regulations for capturing race and ethnicity are implemented; but this is how C-CDA mapping works today)