Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Do no harm petition: Don't give big tech access to our medical records

act.wemove.eu

Post reply on HN
Print view View original

211–220 of 231 posts

Re: Do no harm petition: Don't give big tech access to our medical records

#211

Post by nmca »

photochemsyn wrote: post #135
nmca wrote: post #129

Earlier quoted context omitted.

This plainly isn't true, just to briefly articulate: * Hypothesis generation sped up * Candidate participant discovery sped up - including finding controls * large classes of analysis possible on just the data, including difference in difference and natural experiment

If the data is rigorously anonymized, then it loses all value for the health insurance / advertising sector, who want to exclude high-cost patients from coverage and target individuals with health conditions, respectively. Do you think Big Tech shareholders (with additional holdings in pharmaceuticals, petrochemical, agribusiness etc.) would be interested in a database of rigorously anonymized patient data intended t…

I basically don't believe non-degenerate (psued)anonymization is possible, although that complicated af homomorphic encryption stuff makes me a little uncertain.

Re: Do no harm petition: Don't give big tech access to our medical records

#212

Post by nradov »

TheCaptain4815 wrote: post #181
iandanforth wrote: post #13

I strongly oppose this kind of law. HIPAA is a massive burden on medical and research organizations of all sizes, I've personally spent hundreds of hours navigating both human and technical bureaucracy related to HIPAA and I wouldn't wish that on my worst enemy. Ultimately however the worst part of these laws is that they are so harmful to research in the long run. With easy and ready access to medical data we could…

I've always thought this about HIPPA as well, having worked in the healthcare field years ago. I imagine some "hero" leaking an enormous healthcare dataset after anonymizing some info and furthering research. Too bad I never had access to that.

There is nothing legally stopping healthcare organizations from asking patients for informed consent to share their anonymized data with researchers. The government even provides clear guidelines for de-identification.

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

Re: Do no harm petition: Don't give big tech access to our medical records

#213

Post by EatingWithForks »

cloverich wrote: post #207
rangerelf wrote: post #205

Earlier quoted context omitted.

> So maybe we should consider re-working HIPPA so the information can be used by the good guys instead. Who are these "good guys" you speak so highly of? CA gov? FL gov? Federal gov? FBI? Planned Parenthood? Some church? I'm curious about these mythical "good guys" that you, the NRA, and other people with an agenda repeatedly handwave about.

> Who are these "good guys" you speak so highly of Anyone who wants to use the records to improve healthcare. There are plenty of these folks in the medical field, based on my time in medical school and brief stint in a medical research group. I would also include anyone involved in your direct care, i.e. the front-desk personal at the various doctors offices we visit, who have no way of easily obtaining your medical…

> Anyone who wants to use the records to improve healthcare.

This is so nebulous. Anyone can claim they want to improve healthcare. The removal of abortion access has also been repeatedly been justified as improving healthcare. The Tuskeegee experiments (where they just let black men live with a curable, debilitating disease and continually lied to them about their suffering, some of them to their graves) was justified as improving healthcare. "Wanting to improve healthcare" does not preclude people from horrifically unethical behavior. Get better standards if you want to be accessing the private information of people without their consent.

Re: Do no harm petition: Don't give big tech access to our medical records

#214

Post by bpodgursky »

boring_twenties wrote: post #30
bpodgursky wrote: post #4

I don't think most people appreciate that medical research moves, at minimum, 20x slower than it would with open data access. Doing the simplest possible regression across two institutional datasets is a $100,000 project involving lawyers, de-identification, consents, and a host of bureaucracy. Since most projects end in null results, most don't happen. These laws, while well-intentioned, kill people. If you think th…

How much slower does medical research move than it would if experiments on live human beings were permitted?

We do a lot of experiments on live human beings, and I agree that medical research would move very slowly if we were only allowed to test on mice.

Re: Do no harm petition: Don't give big tech access to our medical records

#215

Post by tengbretson »

iandanforth wrote: post #37
freeone3000 wrote: post #33

Earlier quoted context omitted.

Doctor-patient confidentiality is important so that the doctor can act in your best interest (yes, yes, very funny) with full information. Very many things doctors need to be aware of to diagnose conditions have significant social stigma. Examples include alcoholism, mental disorders, substance abuse, genital status, sexual status, pregnancy status, the list goes on. The average person is reluctant to share this info…

[flagged]

> I'm not willing to let tens or hundreds of thousands of people die because of social stigma

How about a comprise then? What if we made it so that you could get unrestricted medical record access for any patient for whom you are willing to to guarantee (under severe financial penalty) will not die?

Re: Do no harm petition: Don't give big tech access to our medical records

#216

Post by peteradio »

iandanforth wrote: post #13

I strongly oppose this kind of law. HIPAA is a massive burden on medical and research organizations of all sizes, I've personally spent hundreds of hours navigating both human and technical bureaucracy related to HIPAA and I wouldn't wish that on my worst enemy. Ultimately however the worst part of these laws is that they are so harmful to research in the long run. With easy and ready access to medical data we could…

If medical records are open by default then there will be a dark research market that easily operates outside the purview of the law. Medical data, hell most data about people should be hidden by default. Your idea is wildly naive and you ought to read about chestertons fence.

Re: Do no harm petition: Don't give big tech access to our medical records

#217

Post by nradov »

cloverich wrote: post #207
rangerelf wrote: post #205

Earlier quoted context omitted.

> So maybe we should consider re-working HIPPA so the information can be used by the good guys instead. Who are these "good guys" you speak so highly of? CA gov? FL gov? Federal gov? FBI? Planned Parenthood? Some church? I'm curious about these mythical "good guys" that you, the NRA, and other people with an agenda repeatedly handwave about.

> Who are these "good guys" you speak so highly of Anyone who wants to use the records to improve healthcare. There are plenty of these folks in the medical field, based on my time in medical school and brief stint in a medical research group. I would also include anyone involved in your direct care, i.e. the front-desk personal at the various doctors offices we visit, who have no way of easily obtaining your medical…

I don't know why you're bringing up direct care. HIPAA specifically allows access to PHI for treatment, payment, and health care operations.

https://www.hhs.gov/hipaa/for-professionals/faq/authorizatio...

When providers fail to obtain medical histories for new patients that's not due to any legal restrictions. In most cases they simply haven't bothered to implement the open interoperability industry standards which make it easy to exchange such data, or haven't joined networks that facilitate such exchange. For example, the majority of physicians still don't have a publicly listed DirectTrust Direct Secure Messaging address even though they could have one at a very low cost.

Re: Do no harm petition: Don't give big tech access to our medical records

#218

Post by tmcb »

denton-scratch wrote: post #197
tmcb wrote: post #145

Earlier quoted context omitted.

The worst outcome is that individuals may have sensitive information in their medical history that could be poorly received by their social circles if it becomes public. In extreme cases, this could even lead to threats on their life. Consider this: the same argument could be applied to our endeavor to safeguard the secrecy of voting. What is the worst outcome if everyone discovers who you voted for? Such conclusions…

> The worst outcome is that individuals may have sensitive information in their medical history that could be poorly received by their social circles if it becomes public. That's far from the worst outcome. You could lose your job, be sent to prison, be denied insurance.

Yes. It was poor wording on my part trying to encompass all these situations under a generic description. But these are excellent examples of what I was trying to convey. Thanks.

Re: Do no harm petition: Don't give big tech access to our medical records

#220

Post by ClumsyPilot »

ianburrell wrote: post #148
chopin wrote: post #134

Earlier quoted context omitted.

If the liability is so great that an LLC is necessary maybe it shouldn't be done in the first place.

The potential liability for any company is big enough that limited liability is necessary. No one would start company if they could lose all their assets not just ones invested in company. Are people confused that limited liability limits the liability of the company? Because limited liability means that the liability of the owners is limited to their investment. The company can go bankrupt from losing lawsuit.

> No one would start company if they could lose all their assets not just ones invested in company.

Companies operated before the concept of Limited liability was even invented.

We have records of successfull companies from year 578, like Kongō Gumi, in Japan.

Post reply on HN