It sounds like what should happen is, the data should be supplied with the identifiers seperated by the original submitter.

Like a standard that everyone follows where when you write the reports in the first place, everyone is subject A who works at employer X and lives in county Z, and the legend is in a seperate part of the document which is trivial to withhold. Put both the job of segregating the data and the liability for doing it correctly on the submitter when they are creating the data in the first place. Then the FDA or other similar body only has to do a reasonable minimal due dilligence to realease the anonymised data.

If there are mistakes, make that the applicant's responsibility.Ĭherry picking two specific things I want to reply about.