back to article Research bods told: Try to ID anonymised data subjects? No more CASH for you

Medical research funding bodies in the UK may withdraw support for projects where researchers attempt to work out the identity of individuals behind anonymised data without the subjects' permission. Cancer Research UK, the Economic and Social Research Council, the Medical Research Council and the Wellcome Trust said funding …

COMMENTS

This topic is closed for new posts.
  1. Destroy All Monsters Silver badge
    Paris Hilton

    But why would anyone do that except as a late-night fun-and-giggles project and/or to attempt to identify a particularly hot mate?

    1. lglethal Silver badge
      Go

      If researchers were to find a particular primsing strain of DNA (say a particular mutation gives extra protection against a disease), I can see there being a desire by the researcher to try and get a few more samples from that person or there immediate family.

      This legislation is designed to stop them trying to find these people, and instead go back to the governing body, who can decide if they want to contact the people or not.

      Someone donating blood/DNA/whatever to a research body, might be willing to do so again, but that doesnt mean they want to be approached by random scientists demanding their liver...

      1. Grikath

        Why would they need more samples from the original donor? If they find anything promising, the DNA sequence is known, so it can be synthesised and multiplied at much less cost than isolating it from a cell culture. Expression in vivo can be done using cloning techniques in cell cultures. ( This aside from any roohah over Moral Aspects of human cell cloning.) Once you have a pure sample of DNA, there's , technically speaking, not much need to ever know who the original owner is/was.

        If you need to do population/distribution research you need to cast your net a lot wider than "just family" , in which case going through the proper authorities is pretty much the only way to get your samples/data anyway.

      2. frank ly

        @Iglethal

        "This legislation is designed ..."

        This is not legislation. It's recommendations for possible future action by a recognised set of professional bodies. You may have been influenced by the use of the words, "... sanctions that are proportionate to the nature of the offence ..."; which seems to be about an offence to professional standards, not an offence in law.

        Sadly, the law lags behind the development of techniques in many areas of scientific development, especially information technology and data processing. Also, the companies that stand to make lots of money from this have access to politicians and senior civil servants; hence they can influence the law, or lack of it.

    2. JohnMurray

      http://www.telegraph.co.uk/finance/personalfinance/insurance/10667245/NHS-database-will-it-push-up-your-insurance-premiums.html

    3. John Smith 19 Gold badge
      Unhappy

      @Destroy All Monsters

      "But why would anyone do that except as a late-night fun-and-giggles project and/or to attempt to identify a particularly hot mate?"

      Ho ho.

      You really need to use that joke icon

  2. Will Godfrey Silver badge
    FAIL

    La La La La

    So if we discourage the (reasonably) good guys from poking around for weaknesses in our 'anonymity" process it makes certain the bad guys won't have a chance of doing so.

    1. JDC

      Re: La La La La

      Not really: this possible sanction is to cover people maliciously setting out to recover identities from supposedly anonymous data: I imagine that as long as you publicly state the intent is test the anonymisation process you could be granted an exception. An ethical researcher would also inform the data owner if they discover flaws.

      1. Anonymous Coward
        Anonymous Coward

        Re: La La La La

        As usual in such cases (there are many parallels) the authorities seem to think that if they can prevent legitimate, well-intentioned researchers from doing something that will somehow stop criminals from doing it.

        CLUE: Criminals are defined as people who don't give a flying fuck about the law, regulations, guidelines, or anything except PROFIT. (Yes, I know that also describes a large class of business executives: all I can say is that, if it walks like a duck...). So stopping law-abiding people from showing that such things can be done merely gives criminals (and crooked business executives) a free run. This is exactly the same mentality as the belief that forbidding white hat researchers to look for ways of breaking security will do anyone any good (other than criminals).

        "La La La La" sums it up perfectly.

        1. Tom 13

          Re: La La La La

          I don't read it that way at all. I read it as "if you've been authorized to have access to this information only for the purposes of doing medical research and you instead try to identify the people to whom we promised anonymity, we're gonna drop so much hurt on you you'll wish it was sharks with frickin' laser beams instead."

          And really, that's a weaker point than many people think it is right now.

          Should they also invest in the sort of white/grey hat analysis you are advocating? Sure. But that doesn't mean it's the only or route to go. Defense in multiple layers.

        2. Dave Bell

          Re: La La La La

          The people here control the purse strings. When you take out these four sources, you don't have much left in the UK. It's not going to stop corporate profit seekers, and these funding sources are not always going to agree, but pissing-off the Wellcome Trust is a very bad career move for a medical researcher. Who would employ somebody whose actions have killed the funding for a research project?

  3. John Smith 19 Gold badge
    Unhappy

    It's a *start*

    But I think you can bet that as long as the NHS will dish out most of their database for peanuts because insurance companies want this data they (or rather the middlemen they will "outsource" this to to keep their hands clean) won't be deterred.

    1. This post has been deleted by its author

      1. David Pollard

        How about prison sentences for the directors ...

        This seems to be a completely effective deterrent in the cases of misdeeds committed by journalists, MPs, bent coppers ...

      2. Tom 13

        Re: How about prison sentences for the directors

        While I share the sentiment that route is much harder to execute under our current legal system. Better direct financial penalties on the offending organizations. If there's a legal way to do it, attach them to the directors as well. As in: if they work for another medical research team after leaving a sanctioned company, that new company becomes sanctioned as well.

        That will focus their minds sufficiently for most purposes.

  4. SMabille

    Research vs commercial interest

    Most research project won't give a damn about the identity of the samples they use.

    The risk is de-anonymising data comes from commercial companies that will be able to buy them from the government, and those by definition are self funded....

    And to be perfectly honest I really care far less if Dr. XYZ in a lab trying to find a cure for what ever genetic orphan disease is able to find my name to contact me about is research than insurance ABC knowing I have that gene and more likely to develop a particular disease exclude it from my medical cover or increase the price to reflect the risk.

    On top of that to be able to ID anonymised data you need to cross reference them with other databases which are unlikely to be available to researcher but already in hand of commercial structure.

    So the real problem is that, in the current form, there is no option to opt-in for research only purpose, so it will be a complete opt-out for me (if I trust the government to respect my choice and not force me down the line in a couple of year....)

    1. Don Jefe

      Re: Research vs commercial interest

      It's kind of sweet you think medical research projects are so goal oriented they have no interest in potential commercial, non-medicinal, aspects of their work. That's grossly inaccurate.

      It doesn't matter if you are researching particle behavior in soft drinks exposed to gamma emissions or cancer from consuming those soft drinks, the primary goal of any research is never the end of the story. In fact, there's a strong argument to be made that the majority of new discoveries in any field are secondary aspects of only marginally related projects.

      Penicillin is one such example everyone is familiar with, but there's a lot of other things as well. In medical research psychological assessments of people in the study are enormously valuable. That information is tied to the physiological traits of the participants as well as hereditary information and their exposure to different environmental factors. That data is used in scads of other ways, but one of the biggest is in pharma marketing. It isn't an accident that ads for herpes medications always show a young white woman with an animal in a medium income home that obviously doesn't have a man living in it. With the exception of that horribly misguided ad for Viagra, old people meds always show a child in the ad. That isn't an illustration of continuity of family or anything, the child is exploiting the fact old people want to feel young again and often have the money to try new ways of achieving that (drugs).

      That list goes on, as does color preferences in targeted populations, the number and grouping of consonants and vowels in the drugs trade name for a given market, the use of deliberately misused phonetic triggers to 'help people remember' what drugs they take, and why. It's not uncommon for the sale of research participant data to actually make the study hugely profitable even when actual medicinal knowledge gained was less than zero.

      You know what makes medical research really, really valuable for marketing purposes? If you can identify who the data actually refers to: Subject 87 becomes Eric Bogle (or whoever). If you know that you can really fine tune your ads. Guess who figured out how to ID individuals based on anonymized data. Anybody, no? It sure as fuck wasn't corporate marketing people.

      It was medical researchers that not only identified the opportunity, they figured out how to do it. In what is actually a neat bit of modern history, the massive quantities of medical records created as a function of a globe spanning conflict and great strides in what we now call 'computer science', afforded technically inclined medical practitioners to create new ways to sort through anonymized data and identify previously unknown connections between things.

      The cool IT angle in this isn't just the computing, it's the fact the data was anonymized only for processing, due to limited computing resources, but was tied back to the actual, complete, records so validating the output was fairly straightforward. Then the researchers decided they could do this with other anonymized data from a variety of sources but they needed to validate the findings.

      Understandably, most all governments were wary of letting another government create lists of people for research. The recent, rather unpleasant, mess in Germany had captured a lot of attention you know. They would provide data on their soldiers and the soldiers families, but anonymized 'as a safety measure'.

      So off to work the researchers go. Summarized to say, they figured out how to ID individual data sources from anonymized data sets and validate the info. It was at this point that medical research went off the rails and we're all still stuck on the moving train. Extremely targeted marketing of today got its start in pharma. Just for fun go take a look at how the leadership of big pharma companies made an overnight shift from little bald guys with glasses and a stethoscope to the equivalent of Don Draper. They realized the best money was in playing on people's need for any kind of assurance about complex, scary things that can kill them or their loved ones. Pharma management was selected, in part, to appeal to scared mothers/housewives who had nothing to do but sit home and drink: 'Sexy medicine man to the rescue'.

      Look at where those first mega wealthy ad execs came from and who funded the agencies. Bayer was a major force in that. The were the first to figure out that you can sell stuff and make people forget unpleasant things from the not so distant past just by using the right words and pictures and putting them in a place where they would be seen by the intended audience.

      Take a look around at how many ad execs today came out of pharma, and how many ad execs are on the Boards of pharma companies. It's all a nasty little inbred bunch of assholes. The lot of them. Before you take off down the street proclaiming how medical researchers don't care about commercial opportunities you should take a few minutes and dig in a bit deeper. Everything most despicable about pharma, marketing and behavioral management came out of pharma in 1944 and that's still where the best research on general marketing and behavioral management is being done.

      Successful medical research long ago stopped using new medical discoveries as their sole performance metric. New methodologies and algorithms for 'better, more effective and less intrusive ads' are crazy huge money. That shit isn't developed solely in the bowels of the Google campus. Having you dancing around in the streets, preaching that medical researchers are just looking for medical science doesn't help anyone. Having you campaigning to arm the bane of modern existence with more advanced weapons is just fucking dumb. They invented it and they sure as shit haven't been resting on their laurels all this time. They've even branched out and started online ad agencies that get bought up for huge sums by companies like Goog.... Oh. Forgot to tell you, about that. Well, you can just look into it yourself.

    2. JohnMurray

      Re: Research vs commercial interest

      You just use the credit ref agencies. They already have all the info, gleaner from using the electors register, loans, bank details, etc. After all, even the DWP has a contract with them!

    3. Destroy All Monsters Silver badge
      Headmaster

      Re: Research vs commercial interest

      And to be perfectly honest I really care ... about insurance ABC knowing I have that gene and more likely to develop a particular disease exclude it from my medical cover or increase the price to reflect the risk.

      Doesn't sound very ethical to me. Basically, you want deluxe low coverage for a known shite car.

      The argument may well be "it's not my fault", but it's not the fault of the guy who will pay for it either.

      1. Graham Cobb Silver badge

        Re: Research vs commercial interest

        Doesn't sound very ethical to me. Basically, you want deluxe low coverage for a known shite car.

        I strongly recommend reading Tim Harford's "Undercover Economist". Whether it is ethical or not depends on whether you know you are at increased risk (and whether you are taking out the policy because of that knowledge). Unfortunately, that is not easy to determine!

        Harford explains the complexity of this (serious) problem very well. Unfortunately there is no good solution. The best solution for now seems to be for the health insurers to agree to deliberately forgo knowing much about you -- that has a chance of evening out the risk (and the premiums), at least until people become generally much better informed than they are today about what they are at risk of. At that point, the health insurance business will collapse altogether.

  5. Graham Marsden
    Facepalm

    So tell me...

    ... how much funding will be withdrawn and how does that compare to the commercial value of the data which could be extracted...?

    1. Don Jefe

      Re: So tell me...

      That, Mr. Marsden, is the single most messed up part of medical research. Sure, there are rules in place, but if results from breaking those rules seem positive then it doesn't take much to position the withdrawal of funding as a component of back room deals with other pharma companies.

      It doesn't even have to be true. People are always ready to believe that drug companies have no moral or ethical standards (I don't really know, so I make no judgement on the standards they uphold, or don't) but plenty of people certainly do believe it. Look at the whole childhood vaccination mess.

      Once, moderately, famous slags from the '90's wailing about governments colluding with vaccine providers to increase margins/research funding/remove the will of the populace/whatever is all it takes to cause enormous fear and suffering. Playing on beliefs founded in fear is not something I would put past pharma companies.

  6. ecofeco Silver badge

    Try?

    From what I've seen, they are already able to this.

    Barn door, horse, etc.

  7. Anonymous Coward
    Anonymous Coward

    Ah, back to the old and trusty standby, security through obscurity.

    It worked quite well in the past.

    Feel free to add in a long list of failures.

This topic is closed for new posts.

Other stories you might like