Hacker News new | ask | show | jobs
by resource_waste 710 days ago
I have been quite anti-HIPPA since realizing how 'privacy' was the excuse to stunt science.

My conspiracy: With massive medical data, ML/AI would have been 'discovered'/built sooner. Limiting the data makes it so only a few people can be specialists under the supervision of medical cartels.
2 comments
nradov 710 days ago
You have misunderstood. The HIPAA (not "HIPPA") Privacy Rule doesn't stunt science. It's easy to request patient consent for using PHI, and properly de-identified data isn't even considered to be PHI.

https://privacyruleandresearch.nih.gov/pr_08.asp

https://www.hhs.gov/hipaa/for-professionals/privacy/index.ht...

link
KingOfCoders 710 days ago
Great, where can I find your medical data on the web? Care to give an URL? Would be perfect to include your salary.
link
zo1 710 days ago
Not OP. If my doctors and hospitals would give it to me in a good and easily-collatable format, then I and who knows how many other people would gladly donate it to science or for research purposes. Heck some people have this tendency to donate actual body parts and their entire bodies to science, so it's not a big stretch to say some would donate this sort of personal information for a purpose of their own choosing.

This isn't a fully settled debate (including your salary example) so you can't just assume one side is right and argue it as if it's some unquestionable human right.

link
nradov 710 days ago
Many healthcare providers now have APIs for patients to download their data in standard HL7 FHIR and/or CDA format. You should ask about this, and if they don't have it then consider switching providers. All modern EHRs have that functionality built in so for providers it's simply a matter of switching it on.

However, most of that data is useless for research purposes. Even if the format complies with industry standards the quality is often bad with many data elements lacking consistent coding. You can't just feed clinical data from a bunch of different random sources into a research project and expect to get accurate results: it's a garbage in / garbage out issue. That's why most clinical research studies involve just a few provider organizations so that the researchers can properly configure the systems and train the clinicians on consistent data entry.

link
KingOfCoders 710 days ago
Deanonymized? For sale?
link