George Church, PhD, is Professor of Genetics at Harvard Medical School, Director of the Center for Computational Genetics, and founder of the Personal Genome Project. The Personal Genome Project has collected genomic information and health records from over 1,000 informed participants, all shared openly with the research community and the general public.
GeneWatch: The Personal Genome Project requires that all participants are "informed." What does informed participation mean in this setting? What does a participant need to know before becoming involved in medical genetic research, particularly when their information may become public?
George Church: In principle, all medical research has the participant sign a consent form. In practice, many-if not most-principal investigators acknowledge that they don't feel the participants have read it or understood it, which we didn't feel was a good idea. So we have an entrance exam, 20 questions, and participants have to get 100%. The exam covers what would be in the consent form, maybe a little more elaborate than a consent form but a little easier to read. Consent forms tend to be written in legal language to protect the institution, so our exam precedes that with a very readable set of scenarios. For example, it asks: What if one of your relatives doesn't like that you are participating in this study? Have you checked with them? What if you learn that you have something incurable, or not actionable?
Does each of these questions have a "correct" answer?
There are answers that would indicate that this particular study isn't a good one for the individual to participate in. We're not providing health care, we're doing research. For example, if you're under 21 years old, you shouldn't participate in this study. If you have dementia and you can't finish the exam, you shouldn't participate in this study. If you really don't want any of your information to be public, there is a chance of being re-identified in any medical study that's public, and you need to consider that. So there's a long list of potential complications for someone participating in a medical genetic study. And the fact is that we only need a small fraction of the population to participate in medical research-there's no reason to be coercing people into signing a form that they don't understand.
Do you find that there's usefulness in participants having a preliminary understanding of how DNA works going into this kind of study?
I'm sure it's more helpful in some studies than others. I think in general, as a research participant, it's more important to have curiosity than to have all the answers. Knowing the difference between recessive and dominant or knowing how many base pairs are in your genome may not be crucial for deciding whether you're ready to participate in research on medical genetics.
It is crucial that you have thought through the likelihood of your information becoming public. Many studies either waffle or disingenuously promise that they will do their best to keep your information private, which gives you sort of a warm fuzzy feeling that it will stay private. We felt it was more forthright to say, "Look, if we say it's going to stay private, even on a secure computer system, there is still a very good chance that it could become public and even re-identified." Things like that, I think, tend to be more important in the decision making process for a potential participant than whether you know how many chromosomes there are.
And I think it's very important to do this education upfront. A lot of studies say "we'll educate them later, if they need it," but at that point it's too late. Then you have to say: "Hey, we found something really bad in your genome, do you want to hear about it?" At that point, they're hosed either way, right? Either they say they don't want to hear it and they potentially worry about it for the rest of their life, or they do hear it and they potentially worry about it for the rest of their life. What's better is to tell them up front, walk them through the scenarios, and ask specifically: "What if we found you have a high likelihood of getting Huntington's disease? Would you like to know about it?" They might say "no, that would worry me sick for the rest of my life." And you might have to say that this probably isn't the study for them.
Now, some studies address this by just saying they're not going to return any results to the individual. That's probably bad practice on two counts. Number one, it's naïve to think that you can keep the data private; but more importantly, if you do find something that could save someone's life, it doesn't seem morally defensible to withhold that information. But you can't wait until you've found something to give them the choice of whether or not to hear it. You have to find that out upfront.
You mentioned the idea that it's naïve to think that you can definitely keep the data private. I take it that's something that comes across in the education for PGP participants?
Oh, it comes across loud and clear. But it doesn't come across so clearly in most other studies, because they're concerned that they won't get participants if they don't fib a little bit and assure people that their information isn't going to get out, and that if it does get out there's no way it could be re-identified. They won't say "there's no way it can happen," they'll say "we will try our best" or "it's very unlikely," but the end result is that they give the would-be participant a very strong impression that their information can be kept private. And for some people the consequences of having it public when they've been promised that it would be kept private is worse than just having it known upfront that the information could become public. The problem is that the participant is being lied to, whether it's inadvertent or intentional.