Genes and Confidentiality: Tricky!*

A couple of weeks ago, the D–ly M–l** asked me to comment on the Personal Genome Project‘s call for 100 000 volunteers who’d be willing to have their DNA sequenced so that it could be correlated with their health records and used as a tool for research.  As it happens, my peals of wisdom never made it into print, but here’s an expanded version of the things I said.

First up, this project is superficially similar to that undertaken by the UK Biobank.  The idea behind both is that, since many illnesses have a genetic component to them, understanding those illnesses fully will require doing genetic research.  Sometimes that will be on cells in a lab; sometimes it’ll be population surveys.  Often, the idea will be to learn as much as we can about individuals’ genomes, and then to keep track of their health over a prolonged period.  If, across the population, we notice a correlation between a given gene and a given illness, the hope would be that we could work out more effective treatments.  Insofar as participating in this project might help with research into things like cancer, it’s tempting to think that it’s admirable – some even argue that participation in medical research is a moral duty (though others disagree: to and fro and to and fro***; cf this and this and this).

However, there are also moral problems to consider.  For one thing, the PGP does not promise anonymity and confidentiality – in fact, it’s explicit that it won’t be seeking to maintain it – and these are generally a default requirement for any medical or research procedure.  It’s important, then, that participants fully understand what is being asked of them, and that what is being asked is unconventional.  To its credit, the PGP is not backwards in coming forwards in respect of this:

A surprisingly small amount of data is necessary to uniquely identify a person by name. For example, a zip code, birthdate and gender may be enough to infer someone’s identity with a high degree of accuracy. One of the most identifiable pieces of information is human DNA sequence: it can be used to predict a variety of medical conditions and traits, like hair and eye color, facial features and even surname.

Because we cannot guarantee privacy and we are committed to sharing data for the advancement of science, we feel the most ethical and practical solution is to collaborate with individuals who are comfortable sharing their data without any promises of privacy, confidentiality or anonymity. We call this “open consent”.

I actually think that this is a brave move: it’s laudably candid.  And as long as it is understood from the start, loss of anonymity need not be a huge problem.  The website is clear on the risks, too:

Because personal genome sequencing is a new human endeavor, as a society we lack a detailed understanding of all risks posed by an individual’s genome sequence being known and widely shared. At a minimum, individuals should be comfortable with a variety of worst-case scenarios and known risks. In principle, anyone with sufficient knowledge could take a participant’s genome and/or other personal information and use them to:

infer paternity or other features of the participant’s genealogy

claim statistical evidence that could affect employment or insurance or the ability to obtain financial services for the participant

claim relatedness to criminals or incriminate relatives

make synthetic DNA corresponding to the participant and plant it at a crime scene

reveal propensity for a disease currently lacking effective treatment options

The improvement of our understanding and management of these risks is as important a contribution to the long-term utility of personal genomics as any other this project and its participants may achieve.

There’s something a little strange about the benefits listed (the participation in a cause that aims to advance medicine and global health; to contribute directly to a scientific endeavor; self-curiosity) inasmuch as that I’m not sure how one would weigh them against the risks – but that’s a minor thing.  Maybe I’d replace the “risks and benefits” language with “reasons to participate and reasons not to”.

Still: the focus is on the would-be participant, and ensuring informed consent, and all that stuff.  Yet I think that there are problems with that.  Important as informed consent from the participant might be, the problem with information about genes is that it’s not like other information: it’s not just about you or me.  If I have a cold and enrol into research on that, it tells us nothing about my relatives; if I have a particular gene, though, it does, because genes are shared between family members.  In a nutshell, a person allowing genetic information to be used in studies like this is in effect speaking on behalf of her identical twin sister, and other siblings, parents, children, and so on.  This raises a number of questions.  Is it possible for anyone to give adequate consent in this circumstance?  That is: should we be looking to gain the consent of the participant’s blood relatives?  That would be prohibitively impractical, and we’d have to weigh up their interests with the common interest in understanding illness; but given that genetic information can be stigmatising, it’s an important consideration.  Suppose, too, that a relative refuses consent: would that mean that the would-be participant should be barred from participating?  (Oh, nuts to it: I’ve pimped myself so much in this post already that one more instance can’t do any harm…)  Does it really make sense to talk about genetic information “belonging” to a person anyway, given that it is by its nature shared?  The answer that we give to that will make a big difference to how we answer the other questions.  For sure, we tend to think in terms of “my information”; but that doesn’t imply ownership, any more than the phrase “he took my photograph” implies that anyone has done anything to anything I own.

One thing I didn’t say to the M–l, and that’s only just occurred to me, has to do with why this project is necessary given that there’s already Biobank.  Is it possible that there’s an element of me-too-ism?  UKB is five times bigger; but it won’t, as far as I can see, make use of or release as much data – it doesn’t involve things like MRI scans, which the PGP does.  While the PGP is avowedly non-profitmaking, that’s compatible with there being big questions about data ownership.  Answers on a postcard, please.


* This title was brought to you by the letters G, A, C and T.  Oh, yes.

** Yes, yes.  I know.  I’m trying to make it better one tiny bit at a time.  Also, it might count as impact.

*** I’ve re-worked and expanded the fro arguments elsewhere: *cough*chapters8and9*cough*

  • Keith Tayler

    Not sure I would call the PGP’s warning that it is extremely easy to identify people from a few bits of information as ‘brave’. It’s truthful and puts to shame the UK Biobank that has been economical with the truth. I congratulate the PGP for their candor and hope people will think very carefully before volunteering.
    I think sometimes that those that take the ‘participation in medical research is a duty’ line appear to living in a parallel world where the biotech industry, Big Pharma and the state are doing research solely for the betterment of science and the good of humanity. Academic researchers and medical professionals are all honest altruistic people, and the history of medical research is presented as being thoroughly decent enterprise from which all have benefited. (This is clearly incorrect, it being at best a benefit to the majority in rich countries.) There is no attempt to address the problem of commercial research beyond saying that benefits should be shared. Privacy issues are swept aside because people should forget privacy and think about all the benefits to future people. Just in case we are not convinced by the thin concepts that underpin this duty, we are threatened that participation should perhaps be made compulsory if we don’t do our duty.
    It has taken hundreds of years to construct our concept of privacy, much of it having been constructed by technological changes. The last three or four decades have been marked by relatively slow technological change in the western world; but there are signs we are emerging from this, especially in technologies that greatly alter our privacy, so we better be sure what we are doing with technologies that redefine public/private divide and can be owned. Although I agree with you that genes should be treated as commons, I can’t see how they can be without relinquishing some essential aspects of our privacy. Part of the problem, as was the case with land commons in the 18th and 19th centuries in the UK, is that governments are now actively promoting enclosure. We should be greatly reducing the influence of intellectual property rights on research and development not strengthen them. That of course is not going to happen.
    Anyway, who needs a duty to participate in research when the UK government publishes all NHS patient records (see