2 ms·
De-identification is a hard problem. All of the major enterprise healthcare organizations have dedicated analysts who work solely on de-identifying datasets so
by jollofricepeas 4y ago
De-identification is a hard problem.
All of the major enterprise healthcare organizations have dedicated analysts who work solely on de-identifying datasets so that the data can be used for commercial purposes.
- JohnFen 4y ago> De-identification is a hard problem. It's such a hard problem that it is not achievable except in very specific circumstances, such as if the data is just a single single datapoint or if the data is aggregated with a lot of other data, and the original individual datapoints are deleted. In practice, whenever you hear anyone claim they've "anonymized" or "de-identified" your data, if you assume that's not true, you're much more likely to be right than wrong.