> Last year, commenters’ main complaint was that it is impossible guarantee privacy. And this is true. In principle, any piece of info you publish about someone could be the last little clue someone else needs to uncover a great secret about them.
In principle? The research on de-anonymizing large data sets has gone a lot further than 'in principle'. I found a ton of papers and results when I spent an hour or two reading on the topic for http://www.gwern.net/Death%...
If you want to argue the results will be more useful than damaging, fine, but don't try to handwave away the damage as merely theoretical. That's as dishonest as advocates of regulation ignoring prevented gains and only counting prevented losses.
It sounds like your students are authoritarian. This isn't about not wanting change -- it's about the choice to use state power to make everything just right (often consistent with traditional values in some way).
I'd be curious to see the reactions to concrete definite proposals without this hidden mine style uncertainty about future revelation. Like if you instead had asked whether hospital aquired disease numbers or percentage of X procedures adjusted for socioeconomic class that result in successful patient treatment at a given hospital. These don't seem to hide unknown lurking dangers the citizen can't reliably predict but do seem to capture the more obvious benefits.
I think the word ALL is causing problems here. Given that you take data identifying patients to have the narrow meaning of data that directly identifies patients this suggests publishing the entire file consisting of all appointments by individual X except names is presumably to be published if I take you literally and we can link across providers with the notes listing other medications currently taken. Correlate births with facebook announcements, appointments with any public information on travel or moves. Link families by shared illnesses and linked appointment times. At the real extreme a psych medical record would include any major life changes like job loss, divorce, etc.. And we haven't even started to get clever and use traffic data with late arrivals, socioeconomic class, or vaccination times against the standard age schedule for their use.
---
Now that we've finished that logical exercisce it should be clear that releasing literally all data with names blanked out leaves no practical option to prevent everyone from learning your entire medical history. So it should come as no surprise that given such a broad unshaped proposal with easily imaginable worst cases clearly within the literal meaning of the statement people tend to react negatively.
Alright maybe you simply meant summary info for each practice. This is really no good either as people become patients at discrete times and will have to change doctors as they move. So suppose you one day glimpse an asthma inhaler or see them use an injector to stop a potentially fatal allergy to a bee sting. Dig up their life history of moves and at each move check for other maladies that a practice in the new local which increases it's count of asthma/bee allergy sufferers are also likely to add. So if you find out Joe has asthma and Joe has moved 6 times in his adult life and every time he comes to a new community a GP increases both his asthma patient and erectile disfunction patient counts in a correlated fashion. Light chit chat to reveal minor medical issues of wives or children will make analysis more exact.
---
I think the valid reason people are afraid of the greater release of 'anonymized' medical data is that they wish to keep some medical information private while by necessity (sometimes requires public use of medication), accident or simple desire not to be totally paranoid be able to share seemingly harmless info like 'Do you have a doc in town you like and would recommend' or 'My doc prescribed me that once didn't do me any good either so try asking for X." However, what they can and cannot share without revealing their secrets is a total cipher to them depending on subtle correlations and statistical likelihoods they haven't the faintest idea about not to mention more powerful data mining techniques not yet invented.
I’ve complained that regulation usually slows innovation. For example, huge driverless cars gains seem needlessly delayed by excess regulation (Tyler agrees). The problem, however, is not government per se, but the citizens to whom government defers. Politics is not about policy; voters are far more interested in showing off symbolic stances than in giving citizens more of what they want.
The problem is not the citizens. The citizens can't reasonably be blamed for being rational (and they are - since they are rationally ignorant and rationally status-seeking). Nor is it reasonable to expect them to change, since the necessary change would be against their own interests, since it would involve an irrationally large investment of time educating themselves.
It is pointless to blame a causal factor unless it is reasonable to expect better of the causal factor. For instance, small children are forgiven many errors and wrongs that older children are not, and the latter are forgiven for errors and wrongs that adults are not - because it is reasonable to expect better from one group but not another, reasonable to expect the practice of laying blame to be beneficial overall (e.g. by adding an incentive to do better).
It is unreasonable to expect the public to get better at this stuff, because it is not in their individual self-interest to make the effort. It is much more reasonable to expect there to be a better system than our current system, since there is such a large space of possible systems and so few systems have so far been tried.
But aside from all this, it is surely mistaken to suppose that the voting public has much real input into the content of regulation. Regulation is largely cooked up by people who never get elected. Your point that people are resistant to change applies not only to the average voter, but also to the average bureaucrat.
And since info supposedly only visible to government employees are often leaked via bribesReally? Do you have any evidence or statistics on this? My (obviously biased) knowledge based on news reports suggests that many leaks are due to negligence, such as accidentally selling old usb drives with sensitive information.
There's a bit of a "who/whom" question in that your students are much younger than the typical voter. The median 20 year-olds has basically no health problems and so the "con" side isn't personally salient.
I'm surprised that your students don't seem to be reading your blog.. then they should know better!
> Last year, commenters’ main complaint was that it is impossible guarantee privacy. And this is true. In principle, any piece of info you publish about someone could be the last little clue someone else needs to uncover a great secret about them.
In principle? The research on de-anonymizing large data sets has gone a lot further than 'in principle'. I found a ton of papers and results when I spent an hour or two reading on the topic for http://www.gwern.net/Death%...
If you want to argue the results will be more useful than damaging, fine, but don't try to handwave away the damage as merely theoretical. That's as dishonest as advocates of regulation ignoring prevented gains and only counting prevented losses.
Don't insurance companies have data on all the medical care that their insured have received?
It sounds like your students are authoritarian. This isn't about not wanting change -- it's about the choice to use state power to make everything just right (often consistent with traditional values in some way).
I'd be curious to see the reactions to concrete definite proposals without this hidden mine style uncertainty about future revelation. Like if you instead had asked whether hospital aquired disease numbers or percentage of X procedures adjusted for socioeconomic class that result in successful patient treatment at a given hospital. These don't seem to hide unknown lurking dangers the citizen can't reliably predict but do seem to capture the more obvious benefits.
I think the word ALL is causing problems here. Given that you take data identifying patients to have the narrow meaning of data that directly identifies patients this suggests publishing the entire file consisting of all appointments by individual X except names is presumably to be published if I take you literally and we can link across providers with the notes listing other medications currently taken. Correlate births with facebook announcements, appointments with any public information on travel or moves. Link families by shared illnesses and linked appointment times. At the real extreme a psych medical record would include any major life changes like job loss, divorce, etc.. And we haven't even started to get clever and use traffic data with late arrivals, socioeconomic class, or vaccination times against the standard age schedule for their use.
---
Now that we've finished that logical exercisce it should be clear that releasing literally all data with names blanked out leaves no practical option to prevent everyone from learning your entire medical history. So it should come as no surprise that given such a broad unshaped proposal with easily imaginable worst cases clearly within the literal meaning of the statement people tend to react negatively.
Alright maybe you simply meant summary info for each practice. This is really no good either as people become patients at discrete times and will have to change doctors as they move. So suppose you one day glimpse an asthma inhaler or see them use an injector to stop a potentially fatal allergy to a bee sting. Dig up their life history of moves and at each move check for other maladies that a practice in the new local which increases it's count of asthma/bee allergy sufferers are also likely to add. So if you find out Joe has asthma and Joe has moved 6 times in his adult life and every time he comes to a new community a GP increases both his asthma patient and erectile disfunction patient counts in a correlated fashion. Light chit chat to reveal minor medical issues of wives or children will make analysis more exact.
---
I think the valid reason people are afraid of the greater release of 'anonymized' medical data is that they wish to keep some medical information private while by necessity (sometimes requires public use of medication), accident or simple desire not to be totally paranoid be able to share seemingly harmless info like 'Do you have a doc in town you like and would recommend' or 'My doc prescribed me that once didn't do me any good either so try asking for X." However, what they can and cannot share without revealing their secrets is a total cipher to them depending on subtle correlations and statistical likelihoods they haven't the faintest idea about not to mention more powerful data mining techniques not yet invented.
I’ve complained that regulation usually slows innovation. For example, huge driverless cars gains seem needlessly delayed by excess regulation (Tyler agrees). The problem, however, is not government per se, but the citizens to whom government defers. Politics is not about policy; voters are far more interested in showing off symbolic stances than in giving citizens more of what they want.
The problem is not the citizens. The citizens can't reasonably be blamed for being rational (and they are - since they are rationally ignorant and rationally status-seeking). Nor is it reasonable to expect them to change, since the necessary change would be against their own interests, since it would involve an irrationally large investment of time educating themselves.
It is pointless to blame a causal factor unless it is reasonable to expect better of the causal factor. For instance, small children are forgiven many errors and wrongs that older children are not, and the latter are forgiven for errors and wrongs that adults are not - because it is reasonable to expect better from one group but not another, reasonable to expect the practice of laying blame to be beneficial overall (e.g. by adding an incentive to do better).
It is unreasonable to expect the public to get better at this stuff, because it is not in their individual self-interest to make the effort. It is much more reasonable to expect there to be a better system than our current system, since there is such a large space of possible systems and so few systems have so far been tried.
But aside from all this, it is surely mistaken to suppose that the voting public has much real input into the content of regulation. Regulation is largely cooked up by people who never get elected. Your point that people are resistant to change applies not only to the average voter, but also to the average bureaucrat.
voters are far more interested in showing off symbolic stances than in giving citizens more of what they want.
What do you expect? The vast majority of voters have only the vaguest idea which policies will give citizens more of what they want.
So they tend to vote for whatever looks like a good idea at the time. What else can they do?
And since info supposedly only visible to government employees are often leaked via bribesReally? Do you have any evidence or statistics on this? My (obviously biased) knowledge based on news reports suggests that many leaks are due to negligence, such as accidentally selling old usb drives with sensitive information.
There's a bit of a "who/whom" question in that your students are much younger than the typical voter. The median 20 year-olds has basically no health problems and so the "con" side isn't personally salient.