5 ms·

This is an excellent use of the data, a proof that it is worthwhile for the data to exist, and a terrifying demonstration that we need to grow the hell up with

by sfink 5y ago

This is an excellent use of the data, a proof that it is worthwhile for the data to exist, and a terrifying demonstration that we need to grow the hell up with respect to data privacy. It actually sounds like the anonymization and aggregation was done pretty well for this. There's no such thing as 100% anonymity, but there are useful levels of anonymity. The fact that we don't have a standardized way of blessing data with varying known levels of privacy, checked and enforced by law, is kind of ridiculous in 2022. If you want to maximize privacy here, then let's pass laws to prevent doctors from knowing the gender, age, and weight of any of their patients. I 100% agree that privacy is handled way, way too loosely in general, and this does look like the CDC was a lot more interested in a grab bag of use cases than it was with maintaining anonymity. But surely that's an indictment of our societal failure to regulate data brokers? It shouldn't be up to the CDC to act counter to their mission.

zdragnar 5y ago
> If you want to maximize privacy here, then let's pass laws to prevent doctors from knowing the gender, age, and weight of any of their patients. I am fine with my doctor knowing this about me. I am not fine with any doctor who does not see me as a patient knowing anything at all about me. > This is an excellent use of the data, a proof that it is worthwhile for the data to exist The use of the data is fine, but as for the data existing, I also find that to be questionable (in its current form). The particular issue that I have is that the data is used without true consent, and as such is tainted. OTOH, I suspect that "being OK with the CDC tracking your movements via GPS" would eliminate a very important cohort from the study. Which brings me back to your original statement: > a terrifying demonstration that we need to grow the hell up with respect to data privacy I don't genuinely don't know that we can have both (informed consent) and (quality data).
- sfink 5y ago
  >> If you want to maximize privacy here, then let's pass laws to prevent doctors from knowing the gender, age, and weight of any of their patients. > I am fine with my doctor knowing this about me. I am not fine with any doctor who does not see me as a patient knowing anything at all about me. I am ok with my doctor knowing there was a 12% chance that I broke curfew during a pandemic lockdown, given my zip code or whatever. I don't think it's a black and white thing. Especially since you can't have the extreme of no data—the data is already out there, in massive quantities, and mostly uncontrolled. >> a terrifying demonstration that we need to grow the hell up with respect to data privacy > I don't genuinely don't know that we can have both (informed consent) and (quality data). Oh, we can't, not anywhere near the optimal balance anyway. Because people. (And therefore governments, and laws, and everything else.) But I guess I still think that whatever crappy balance we can come up with is better than organizations that can make good use out of collections like this sticking their fingers in their ears and pretending there's no way to know certain things that all the bad actors already know because they're willing to buy this stuff on the black market. It leads to dogs and cats living together and people writing really bad run-on sentences like the previous one.
  - zdragnar 5y ago
    > I am ok with my doctor knowing there was a 12% chance that I broke curfew during a pandemic lockdown, given my zip code or whatever I am asking this out of genuine curiosity: Why? What actionable information does that tell your doctor about you, specifically?
    - sfink 5y ago
      Because, at least in my opinion, doctors should be considering base rates carefully. If I come in with an unusual set of symptoms, then I don't want to be tested for something that is vanishingly unlikely and not tested for something that has an unusually high frequency of occurring given my specific situation (demographic, race, vaccination rates in the community I live in for diseases that can produce my symptoms, etc.) Medicine should not be one size fits all when better relevant data are available. Different but related: I personally use similar information to determine the risk of various behaviors. Going to a meeting in someone's house where people won't be masked? Not a big deal if the current rate of community infection is very low in my local community. Playdate for my kids with friends who I know are not vaccinated for measles or pertussis or anything else? Depends on whether those are going around—and if they are, then I'd like to know how many kids are unvaccinated whether or not they're lying about it. If there's a way to get that aggregate information.
      - zdragnar 5y ago
        Running with the 12% figure, I'm still curious as to how that is useful. Have you traveled recently? Been to a large gathering? Never left your house? Visited with family from out of town? Do you commute for work as a front-line worker? Risk profiles for less common diseases make sense, but these are questions your doctor should ask regardless of the infection rate of your zip code in a pandemic scenario, however. > Medicine should not be one size fits all I don't think it's unreasonable to expect tests for most obvious fit of your symptoms (with followups for less likely causes if those are negative) rather than needing to do statistics puzzles for every patient that comes in through the door.
klabb3 5y ago
I get that CDC is extremely interested in measuring effects of policy and PR. What bothers me is how shady this is. Government shouldn't reach out to gray-area thugs for supply, and they should not try to keep their methods secret. Keep in mind public health authorities are absolutely dependent upon their reputation and track record. Having high trust is critical for ensuring voluntary compliance when shit hits the fan. They seem entirely unaware of this and operates more like an intelligence agency. Even if you're still trusting of CDC, can you at least see how these actions risk eroding that for others?
- oofbey 5y ago
  I totally hear you - this seems SUPER SHADY. But I think the truth is it's really not. These data brokers are totally legit - legally. And also pretty common. Cell companies cell this stuff all the time to just about anybody. They're not loud about it, and people frequently freak out when they realize it. But this stuff happens ALL THE TIME. I also have heard that this has been a common technique amongst academic epidemiologists during the pandemic. Because it's actually all above board, and super useful. Still sounds SUPER F*NG CREEPY. But it's commonplace.
panny 5y ago
>and a terrifying demonstration that we need to grow the hell up with respect to data privacy Once again, "conspiracy theorists" were right. I'm beginning to think the governments of the world are reading them for ideas. >The fact that we don't have a standardized way of blessing data with varying known levels of privacy, checked and enforced by law, is kind of ridiculous in 2022. Sure, and then as soon as that happens, someone who has patented/copyrighted the process coincidentally is handed a government mandated empire worth millions in licensing fees. Look into ICD-10 codes if you need an example. It's the classic college professor scam where they write the textbook and force everyone to buy it. >It shouldn't be up to the CDC to act counter to their mission. If CDC can't do the job without violating 4th amendment, it shouldn't exist. It's not a "private company" and the bill of rights exists for a good reason.
- maxerickson 5y ago
  This is commercial tracking data, not COVID app data. It's been well known for a long time that plenty of phone apps sell data to aggregators, it's not a conspiracy theory. That commercial data isn't something any court would call a violation of the 4th amendment.
  - panny 5y ago
    To me that sounds a bit like >It's not rape if you consent to sex but I slip off the condom when you aren't looking. If the government is explicitly not allowed to collect this data, then buying it is simply a different form of data collection.
    - maxerickson 5y ago
      Read closer. I didn't express a personal opinion on the topic, I expressed an opinion/assessment about how courts handle the topic. Like you can have a legal theory about something and it doesn't mean shit if courts don't accept it.
      - panny 5y ago
mc32 5y ago
Doesn't the way they went about this have the same issue as the NSA, etc., going to the UK or other jurisdictions to get data on Americans when they can't legally obtain that data? If that is shady, then why isn't this also called out? It looks like this is a method to circumvent the law.
chrismcb 5y ago
This is not an excellent he just of the data. And the ways that cdc want to use it is proof the data shouldn't exist. I don't have a problem with the CDC creating an app to collect such data. An app the clearly says what it is used for. But to collect the data through some other so and decide if someone broke curfew or visited a neighbor? You think that is ok? Pretty much the definition of Big Brother
- sfink 5y ago
  If they can identify individual people breaking curfew, then yes that's way out of line. If they can estimate the percentage of people breaking curfew within an entire large city, that seems like important policy input. And there's a huge gray area in between. In actual fact, I'm not entirely comfortable with the "large city" granularity either. I don't want a city that is majority black to be treated differently using the excuse of their aggregated health data, for instance. I think these types of data collections need to be treated very seriously, requiring justification for specific uses, tracking where it lives and enforcing penalties when it shows up elsewhere. Because these data are going to be out there and used. If they aren't managed responsibly, they'll be sold on the black market and used in completely uncontrolled ways.

by sfink 5y ago

zdragnar 5y ago

> If you want to maximize privacy here, then let's pass laws to prevent doctors from knowing the gender, age, and weight of any of their patients. I am fine with my doctor knowing this about me. I am not fine with any doctor who does not see me as a patient knowing anything at all about me. > This is an excellent use of the data, a proof that it is worthwhile for the data to exist The use of the data is fine, but as for the data existing, I also find that to be questionable (in its current form). The particular issue that I have is that the data is used without true consent, and as such is tainted. OTOH, I suspect that "being OK with the CDC tracking your movements via GPS" would eliminate a very important cohort from the study. Which brings me back to your original statement: > a terrifying demonstration that we need to grow the hell up with respect to data privacy I don't genuinely don't know that we can have both (informed consent) and (quality data).

sfink 5y ago
>> If you want to maximize privacy here, then let's pass laws to prevent doctors from knowing the gender, age, and weight of any of their patients. > I am fine with my doctor knowing this about me. I am not fine with any doctor who does not see me as a patient knowing anything at all about me. I am ok with my doctor knowing there was a 12% chance that I broke curfew during a pandemic lockdown, given my zip code or whatever. I don't think it's a black and white thing. Especially since you can't have the extreme of no data—the data is already out there, in massive quantities, and mostly uncontrolled. >> a terrifying demonstration that we need to grow the hell up with respect to data privacy > I don't genuinely don't know that we can have both (informed consent) and (quality data). Oh, we can't, not anywhere near the optimal balance anyway. Because people. (And therefore governments, and laws, and everything else.) But I guess I still think that whatever crappy balance we can come up with is better than organizations that can make good use out of collections like this sticking their fingers in their ears and pretending there's no way to know certain things that all the bad actors already know because they're willing to buy this stuff on the black market. It leads to dogs and cats living together and people writing really bad run-on sentences like the previous one.
- zdragnar 5y ago
  > I am ok with my doctor knowing there was a 12% chance that I broke curfew during a pandemic lockdown, given my zip code or whatever I am asking this out of genuine curiosity: Why? What actionable information does that tell your doctor about you, specifically?
  - sfink 5y ago
    Because, at least in my opinion, doctors should be considering base rates carefully. If I come in with an unusual set of symptoms, then I don't want to be tested for something that is vanishingly unlikely and not tested for something that has an unusually high frequency of occurring given my specific situation (demographic, race, vaccination rates in the community I live in for diseases that can produce my symptoms, etc.) Medicine should not be one size fits all when better relevant data are available. Different but related: I personally use similar information to determine the risk of various behaviors. Going to a meeting in someone's house where people won't be masked? Not a big deal if the current rate of community infection is very low in my local community. Playdate for my kids with friends who I know are not vaccinated for measles or pertussis or anything else? Depends on whether those are going around—and if they are, then I'd like to know how many kids are unvaccinated whether or not they're lying about it. If there's a way to get that aggregate information.
    - zdragnar 5y ago
      Running with the 12% figure, I'm still curious as to how that is useful. Have you traveled recently? Been to a large gathering? Never left your house? Visited with family from out of town? Do you commute for work as a front-line worker? Risk profiles for less common diseases make sense, but these are questions your doctor should ask regardless of the infection rate of your zip code in a pandemic scenario, however. > Medicine should not be one size fits all I don't think it's unreasonable to expect tests for most obvious fit of your symptoms (with followups for less likely causes if those are negative) rather than needing to do statistics puzzles for every patient that comes in through the door.

klabb3 5y ago

I get that CDC is extremely interested in measuring effects of policy and PR. What bothers me is how shady this is. Government shouldn't reach out to gray-area thugs for supply, and they should not try to keep their methods secret. Keep in mind public health authorities are absolutely dependent upon their reputation and track record. Having high trust is critical for ensuring voluntary compliance when shit hits the fan. They seem entirely unaware of this and operates more like an intelligence agency. Even if you're still trusting of CDC, can you at least see how these actions risk eroding that for others?

oofbey 5y ago
I totally hear you - this seems SUPER SHADY. But I think the truth is it's really not. These data brokers are totally legit - legally. And also pretty common. Cell companies cell this stuff all the time to just about anybody. They're not loud about it, and people frequently freak out when they realize it. But this stuff happens ALL THE TIME. I also have heard that this has been a common technique amongst academic epidemiologists during the pandemic. Because it's actually all above board, and super useful. Still sounds SUPER F*NG CREEPY. But it's commonplace.

panny 5y ago

>and a terrifying demonstration that we need to grow the hell up with respect to data privacy Once again, "conspiracy theorists" were right. I'm beginning to think the governments of the world are reading them for ideas. >The fact that we don't have a standardized way of blessing data with varying known levels of privacy, checked and enforced by law, is kind of ridiculous in 2022. Sure, and then as soon as that happens, someone who has patented/copyrighted the process coincidentally is handed a government mandated empire worth millions in licensing fees. Look into ICD-10 codes if you need an example. It's the classic college professor scam where they write the textbook and force everyone to buy it. >It shouldn't be up to the CDC to act counter to their mission. If CDC can't do the job without violating 4th amendment, it shouldn't exist. It's not a "private company" and the bill of rights exists for a good reason.

maxerickson 5y ago
This is commercial tracking data, not COVID app data. It's been well known for a long time that plenty of phone apps sell data to aggregators, it's not a conspiracy theory. That commercial data isn't something any court would call a violation of the 4th amendment.
- panny 5y ago
  To me that sounds a bit like >It's not rape if you consent to sex but I slip off the condom when you aren't looking. If the government is explicitly not allowed to collect this data, then buying it is simply a different form of data collection.
  - maxerickson 5y ago
    Read closer. I didn't express a personal opinion on the topic, I expressed an opinion/assessment about how courts handle the topic. Like you can have a legal theory about something and it doesn't mean shit if courts don't accept it.
    - panny 5y ago

mc32 5y ago

Doesn't the way they went about this have the same issue as the NSA, etc., going to the UK or other jurisdictions to get data on Americans when they can't legally obtain that data? If that is shady, then why isn't this also called out? It looks like this is a method to circumvent the law.

chrismcb 5y ago

This is not an excellent he just of the data. And the ways that cdc want to use it is proof the data shouldn't exist. I don't have a problem with the CDC creating an app to collect such data. An app the clearly says what it is used for. But to collect the data through some other so and decide if someone broke curfew or visited a neighbor? You think that is ok? Pretty much the definition of Big Brother

sfink 5y ago
If they can identify individual people breaking curfew, then yes that's way out of line. If they can estimate the percentage of people breaking curfew within an entire large city, that seems like important policy input. And there's a huge gray area in between. In actual fact, I'm not entirely comfortable with the "large city" granularity either. I don't want a city that is majority black to be treated differently using the excuse of their aggregated health data, for instance. I think these types of data collections need to be treated very seriously, requiring justification for specific uses, tracking where it lives and enforcing penalties when it shows up elsewhere. Because these data are going to be out there and used. If they aren't managed responsibly, they'll be sold on the black market and used in completely uncontrolled ways.