I often wonder about the ethics of EHR de-identification.
Assuming we can make de-identification perfect, with no risk of undoing this de-identification, should we be doing it?
Well, as a starting point, and just for fun, let’s see what the Hippocratic Oath would say about it. (And yes, I am aware of the issues with the Hippocratic Oath in today’s world, but this is simply a fun experiment.)
Here is the relevant text:
And whatsoever I shall see or hear in the course of my profession, as well as outside my profession in my intercourse with men, if it be what should not be published abroad, I will never divulge, holding such things to be holy secrets.
To me, this suggests I don’t talk about my patients, regardless of whether or not I name them.
Put another way, if I go to a cocktail party (by the way, who actually goes to cocktail parties anymore?) I don’t talk about patient evaluations, which are often highly personal, no matter how interesting or how relevant to a topic of conversation. And it doesn’t matter how unlikely it is that this patient will ever be identified. It’s simply wrong.
So here the next question. If it is wrong for an individual physician to divulge de-identified patient information to a fellow cocktail guest, why is it OK to divulge 3 million patient EHRs to one hundred or so 30-somethings in Silicon Valley? (Keep in mind that it’s unlikely a single one of these patients gave informed consent to do this.)
Well, at this point the interested parties, which often have a vested financial interest in the outcome, will often discuss the subjects such as “the greater good” or the legalities of data ownership.
That’s ok. Just keep in mind that what you may be doing is divulging a person’s “holy secrets” to strangers, without the person’s consent, 3 million times in a row.