Population Participation and Other Factors that Impact the Compilation and the Utility of Resulting Databases

• Author: Henry T. Greely
• Position: J.D., Professor, Stanford Law School
• Pages: 79-90

Page79

J.D., Professor, Stanford Law School. The author thanks Daven Williams of LSU Law School for her work with the references. This work is based on a live presentation made on February 5, 2004, and does not necessarily reflect events and changes thereafter.

Let me begin by briefly addressing the social, ethical, and legal issues with respect to biobanks generally, and then the special considerations, the special problems, that arise from biobanking and particular human populations. It is important to remember that for almost every purpose, and particularly for purposes involving medical use, biobanks have to be more than just collections of physical specimens. DNA is not very useful for medical purposes unless you know something about the person's health history. You cannot find the gene for blue eyes from DNA unless you know whether people who gave you the DNA had blue eyes.

So, to be useful, biobanks really are more than just biological banks. They are combinations of physical samples of DNA-- which may be tumor tissues or something else--plus medical information about a person. I have tried calling them phenotypegenotype resources, an extremely clunky term that has not caught on, but it has the advantage of being accurate. They are resources that have both physical samples of genetic material and have information about each donor's phenotype; genotype is what your genes say about you, and phenotype is your body as it exists. So a biobank needs both phenotype information and genotype information to be really useful.

Now, one important thing to know about these large genotypephenotype resources is that their value is promising but speculative. No one really knows whether we are going to make great discoveries as a result of having 200,000 samples of DNA coupled with health information about people. It seems like a good way forward, in part because the traditional way of doing genetic research in health seems to have stalled to some extent. As we Page80 have gotten past the simple genetic diseases, the ones that are caused by a single gene, the so-called simple Mendelian disorders, we have discovered, first, that many of those simple Mendelian disorders were not as simple as we thought and, second, that there are known genetic connections to much more common and significant diseases, but those connections are turning out to be really hard to tease out.

We know that there are genetic connections to schizophrenia, which affects one percent of the population,¹ to diabetes, which affects six-to-eight percent of the population,² to various cancers, to high blood pressure,³ to Alzheimer's Disease,⁴ to lots of diseases that affect lots of people. But finding out how the genes connect to those diseases has turned out to be really hard, and it is hoped that collecting huge amounts of data from large numbers of people will give us a way to better make those connections. At least that is the hope, and that hope may prove true.

There are two things we know for sure about such genotypephenotype resources though. The first is that they are really expensive. The Iceland Project, if it ever actually happens in its entirety, is likely to cost several billion dollars to create a Page81 genotype-phenotype resource encompassing just 250,000 people.⁵

The U.K. is planning to sample 500,000 British citizens, subjects of Her Majesty the Queen, and they are budgeting several billion dollars for that project.⁶

Ironically, the genetic side is the cheap part; it is not that expensive to get samples from people with a cheek swab or a blood draw. The expense is attributable to getting good health information, inputting it, and making sure that it is good information. So, first, we know biobanking is going to be expensive.⁷ Second, we know that it is going to have lots of tricky ethical, social, and legal issues.⁸

About eight years ago, Bartha Knoppers, Ken Kidd, and I were at a meeting, I think in the suburban D.C. area when Bartha was Chair of the Ethics Committee of the Human Genome Organization, an international, non-profit, non-governmental group. The Committee put together a set of principles to govern research in human population genetics.⁹ I am afraid that I cannot remember these principles now except that they all started with "C" and that they were Bartha's Ten Commandments.¹⁰ I do remember at least five that are particularly important in this area: Page82 consent, control, confidentiality, communication, and commerce.¹¹

For each, biobanking adds new ethical, legal, and social complications to the existing complexities of research.

First, with respect to consent, if you are going to try to get health materials, health information, and physical genetic samples from people, from a lot of people, that takes a lot of consent, and it is very tempting to short-circuit that labor and time-intensive effort to get people to say "yes." Iceland did that short-circuiting. The Icelandic legislation does not require the consent of people to be in the health data side of the private company's (deCODE) database.¹² And it is that failure of informed consent that has, in part, kept that database tied up in public controversy and litigation in Iceland.¹³ More than five years after the passage of legislation that authorized deCODE to set up that database, it still does not exist.

Control, I think in some ways, is even more important. Control over what uses biobank material can be put to is an important concern for subjects, or at least it should be. I am happy to give my samples and my health information for research on a wide variety of medically related topics. And in traditional research, someone would be interested in premature grey hair and come to me and say, "Boy, do we need your sample." And I'd say, "Sure. No problem." With biobanking, in part because it is so expensive, nobody wants to create a set of 200,000 samples and information just to look at one disease, even one affirmative, enhancing condition like premature grey hair, no matter how important it is. You want to be able to use it for everything in order to amortize that several billion dollar cost.

So if I give a sample to be biobanked, I am giving it away not just for the study of premature grey hair, but for the study of all sorts of things, some of which I may not like. Personally, I would not want my samples to be used for studies of, say, race and intelligence, because I think that would likely be misused in bad Page83 ways. I might be reluctant to have my samples used for studies of Irish heritage and alcoholism, two things that run in my family, and, frankly, something that my year in New Orleans with Judge Wisdom threatened greatly to exacerbate. (I should make clear, it was a function of New Orleans, not of Judge Wisdom.) So, what makes biobanking different is that you have this large resource to be used for many purposes and, under most algorithms, when you make the decision to be a participant, you no longer hold control over what your samples are going to be used for.

Third is confidentiality, preventing unauthorized uses of information that a patient or a subject handed over in confidence. ("Confidentiality" is a narrower term than "privacy," which can include attempts by others to learn information that the individual has not voluntarily handed over to someone, as well as more fundamental issues of bodily autonomy, such as those involved in reproductive...