You can't fully automate this because we require almost 100% accuracy. You need a manual review to ensure no data is leaked.
This is how medical providers share notes without manually reviewing them all.
[0] https://www.hhs.gov/hipaa/for-professionals/privacy/special-...
The FOIA request was for all of the documentation used to approve the vaccine.
> I’m sure there are techniques that can be purchased for entity resolution that can automate the scrubbing for PHI and PII.
Sounds like a good way to have "anonymized" data that isn't actually anonymized like most datasets ...
I think this is a big problem when it comes to stuff like data brokering and selling. Less so with issues like this because vaccine data researchers are less likely to cause harm to someone than others who buy anonymized data.
That's a feature, not a bug. Bureacracy is a game, and they are grandmasters.
If they are literally manually redacting patient data from pages and notes then that is a flaw in the study, I think.
I would expect patient data to be pretty limited in the trial data since they are purposefully double blind. Notes are a different story, but I’m sure there are techniques that can be purchased for entity resolution that can automate the scrubbing for PHI and PII.
Seems more like the FDA and Pfizer being silly than the judge.