These are notes from a meeting with the Redox team on June 1st, 2021 to better understand how they implement their filters for PHI.
Originally, Redox implemented “filters” - essentially blacklisting values and nulling them out
Problem is, different clients send different fields and filters were implemented on a per-client connection basis.
Some SwipeSense customers filter fields on their end, but some just send everything unfitlered
Redox implemented custom logic for swipesense to whitelist values instead of blacklist them.
New config has been tied to our existing destinations, meaning that new integrations that utilize the current set of lambda endpoints will automatically adopt this whitelist filtering.
If a new lambda destination is created, this whitelist filter will not be automatically applied
The config is also tied to the current data models - if we add new data models, they will also not be filtered