4 ms·

I reject the framing. It's about responsibility vs selfishness and fear. Medical information is a byproduct of something you do anyway that can save lives. It i

I reject the framing. It's about responsibility vs selfishness and fear. Medical information is a byproduct of something you do anyway that can save lives. It is our responsibility to make it available. If you want to opt out, fine, but basic decency demands that as a society we get over our collective hangups and make this information available.

dmbche 3y ago
Are you a researcher? I find it alarming to hear this kind of talk, especially after so many have raised the very serious issues that can arise from this being public. It is not our responsibility to make this public. Your logic is similar to the gov trying to get access to all exchanges over the web to scrape them for terrorism and violent behavior. Maybe take a step back and think about the harms that would come out of this?
- delfinom 3y ago
  There's a group of individuals that have been lobbying the US government that if we got rid of HIPPA, then we could cure every disease on the planet with all that medical data they could now harvest.
arp242 3y ago
I don't mind aggregate data, or anonymized where I'm "Patient b15-2gty", which is what you seem to be talking about, but I do mind "here's all the data about patient Martin Tournoij, for everyone to read". I wouldn't want any future employers to see my medical data for example, as they may use this to discriminate (theoretically anyway; my medical history thus far is essentially non-existent). Also: in the late 90s my mother worked for the city to digitize a lot of social security records and such. She had a good friend who had trouble walking (crutches, wheelchair); the story she told was that she was hit by a car, but my mother read her records during her job and found out she had simply fallen and was never hit by a car. Much drama ensued. I have no idea why anyone would lie about that and I'm fuzzy on the details as I was about 12-13 at the time, but fundamentally I think people should have the right to lie about things like this, if they so choose, for whatever reason.
- hedora 3y ago
  Mapping “Patient b15-2gty” back to your real name has been trivial for decades. My first job involved a one hour lecture about how people had repeatedly accidentally deanonymized and accidentally leaked data at other institutions, leading to divorces and worse. We had somewhere between 10 to 100 bytes of entropy on each patient, and it would have been enough for any of their acquaintances to map back to real names and also severely violate privacy.
- JohnFen 3y ago
  > I don't mind aggregate anonymized data where I'm "Patient b15-2gty If you're ""Patient b15-2gty", then the data isn't aggregate. Aggregated data means that there are no individual data points at all, only aggregated ones, so there is no need for any sort of individual identifiers. If there's any sort of individual identifier, then the data cannot be effective anonymized. If the data is aggregated, and the original records that were included in the aggregate figures is deleted, then I think that's adequately anonymized and I would have no problem with it. Short of that, though, "anonymization" is a thing that doesn't actually exist.
  - arp242 3y ago
    Oops, an "or" got lost in editing; fixed it! I don't agree about anonymization; "true" anonymity is probably impossible in most areas of life, even for simple things like a walk in the forest as there's always something a significantly advanced sleuth can use.[1] There are no "true" one-way hashes either – you can always brute-force). It's about it being too infeasible to actually do that. [1]: https://arstechnica.com/science/2023/05/human-genomic-bycatch-our-dna-shows-up-in-environmental-samples/ https://arstechnica.com/science/2023/05/human-genomic-bycatc...
hedora 3y ago
Blaming individuals for not trusting medical and tech companies that have repeatedly abused private information is a strange take. Instead of taking the information by force (which health care conditioned on opt in certainly is), we should figure out how to build a trustworthy medical research and care industry. For instance: They could only make this information available to universities, and ensure the results of the research were public domain. Or, they could create separation of concerns for companies involved. If google wants to store medical data, then they have to spin off their ad business.
- JohnFen 3y ago
  > If google wants to store medical data, then they have to spin off their ad business. I don't think I'd find that sufficient.
JohnFen 3y ago
> demands that as a society we get over our collective hangups What you're calling "collective hangups" are, in fact, real and serious risks. I am distressed at how many people have forgotten (or perhaps were never aware) of the real abuse and harm that was happening with the sharing of medical data before legal protections were enacted. Heck, those harms still happen now, but to a reduced degree.

dmbche 3y ago

Are you a researcher? I find it alarming to hear this kind of talk, especially after so many have raised the very serious issues that can arise from this being public. It is not our responsibility to make this public. Your logic is similar to the gov trying to get access to all exchanges over the web to scrape them for terrorism and violent behavior. Maybe take a step back and think about the harms that would come out of this?

delfinom 3y ago
There's a group of individuals that have been lobbying the US government that if we got rid of HIPPA, then we could cure every disease on the planet with all that medical data they could now harvest.

arp242 3y ago

I don't mind aggregate data, or anonymized where I'm "Patient b15-2gty", which is what you seem to be talking about, but I do mind "here's all the data about patient Martin Tournoij, for everyone to read". I wouldn't want any future employers to see my medical data for example, as they may use this to discriminate (theoretically anyway; my medical history thus far is essentially non-existent). Also: in the late 90s my mother worked for the city to digitize a lot of social security records and such. She had a good friend who had trouble walking (crutches, wheelchair); the story she told was that she was hit by a car, but my mother read her records during her job and found out she had simply fallen and was never hit by a car. Much drama ensued. I have no idea why anyone would lie about that and I'm fuzzy on the details as I was about 12-13 at the time, but fundamentally I think people should have the right to lie about things like this, if they so choose, for whatever reason.

hedora 3y ago
Mapping “Patient b15-2gty” back to your real name has been trivial for decades. My first job involved a one hour lecture about how people had repeatedly accidentally deanonymized and accidentally leaked data at other institutions, leading to divorces and worse. We had somewhere between 10 to 100 bytes of entropy on each patient, and it would have been enough for any of their acquaintances to map back to real names and also severely violate privacy.
JohnFen 3y ago
> I don't mind aggregate anonymized data where I'm "Patient b15-2gty If you're ""Patient b15-2gty", then the data isn't aggregate. Aggregated data means that there are no individual data points at all, only aggregated ones, so there is no need for any sort of individual identifiers. If there's any sort of individual identifier, then the data cannot be effective anonymized. If the data is aggregated, and the original records that were included in the aggregate figures is deleted, then I think that's adequately anonymized and I would have no problem with it. Short of that, though, "anonymization" is a thing that doesn't actually exist.
- arp242 3y ago
  Oops, an "or" got lost in editing; fixed it! I don't agree about anonymization; "true" anonymity is probably impossible in most areas of life, even for simple things like a walk in the forest as there's always something a significantly advanced sleuth can use.[1] There are no "true" one-way hashes either – you can always brute-force). It's about it being too infeasible to actually do that. [1]: https://arstechnica.com/science/2023/05/human-genomic-bycatch-our-dna-shows-up-in-environmental-samples/ https://arstechnica.com/science/2023/05/human-genomic-bycatc...

hedora 3y ago

Blaming individuals for not trusting medical and tech companies that have repeatedly abused private information is a strange take. Instead of taking the information by force (which health care conditioned on opt in certainly is), we should figure out how to build a trustworthy medical research and care industry. For instance: They could only make this information available to universities, and ensure the results of the research were public domain. Or, they could create separation of concerns for companies involved. If google wants to store medical data, then they have to spin off their ad business.

JohnFen 3y ago
> If google wants to store medical data, then they have to spin off their ad business. I don't think I'd find that sufficient.

JohnFen 3y ago

> demands that as a society we get over our collective hangups What you're calling "collective hangups" are, in fact, real and serious risks. I am distressed at how many people have forgotten (or perhaps were never aware) of the real abuse and harm that was happening with the sharing of medical data before legal protections were enacted. Heck, those harms still happen now, but to a reduced degree.