Its nine million members make Kaiser Permanente a giant of American health care. They also put the health maintenance organization in a position to create one of the nation’s most expansive biological resources for medical research. Kaiser’s Research Program on Genes, Environment and Health houses one of the largest repositories of DNA samples in the United States, with samples from 200,000 patients. In the program’s 5,000-square-foot facility in Berkeley, Calif., there are freezers to hold blood and storage for extracted DNA and saliva kits. Sarah Rowell, associate director of the program, admits she only truly comprehended the enormity of this “biobank” when she stood in a warehouse jammed with letters inviting people to participate—the mailing filled a large room.
Because Kaiser keeps electronic medical records for all of its patients, researchers have access to detailed clinical information about the people whose blood or tissue they analyze. “The average length of membership of the biobank participants is more than 20 years,” Rowell says, which gives scientists the opportunity to link genetic information about someone to a long, rich clinical history. Another benefit is that Kaiser’s patient population is ethnically diverse, and researchers may be able to measure the impact of ethnicity on someone’s risk of developing specific diseases. Still another advantage, according to Rowell, is California’s extensive tracking of pollution and other environmental factors. “It’s kind of unbelievable how much information is available,” she says.
Kaiser is hardly alone in attempting to build potent resources for medical research. Institutions around the world are stockpiling tissue and data, and linking samples they collect with medical records. And Japan, Sweden and the United Kingdom are using their centralized health care systems to invest in biobanks on a national scale, amassing very broad collections.
Why invest in such resources? As one example, one of the findings to emerge from the Atherosclerosis Risk in Communities (ARIC) study, which began in 1987, was that 1 in 40 black subjects had a mutation in the gene PCSK9, which was associated with lower levels of LDL cholesterol (known to promote cardiovascular disease) and fewer heart problems over a 15-year period. That finding provided key evidence that PCSK9 is a good target for cholesterol-lowering drugs—and now PCSK9 inhibitors are making their way through clinical trials.
Researchers believe there are many more associations that could be discovered if they were able to sift through the DNA, tissue and medical information of vast numbers of people—populations large enough to give statistical power to observations such as that rare mutation in PCSK9. But unlike focused studies like ARIC, general DNA and tissue repositories can be used again and again for many purposes.
Big biobanks involve society in research on an unprecedented scale. Though you may never participate in a clinical trial or a research study, it’s likely you’ll be asked to contribute to a biobank in the coming years—if your blood or saliva isn’t already in one. Your DNA, for example, might be preserved indefinitely and used by researchers in ways that have yet to be invented.
It’s that prospect, among others, that is posing difficult questions about informed consent and privacy issues. Will people who agree to contribute to a biobank object if their biological material is used in ways they don’t sanction? The University of Arizona, for instance, recently had to pay damages to members of the Havasupai Indian tribe for using blood samples, originally donated for diabetes research, for studies on the genetics of mental illness. The later use violated the values of the tribe.
“You have on the one hand tremendous scientific interest, enthusiasm and investment in big biobanks,” says Timothy Caulfield, head of the Health Law and Science Policy Group at the University of Alberta in Edmonton. “But on the other hand, there are legal, social and ethical issues that haven’t been resolved.” Addressing those challenges may help redefine how patients get involved in research.
Although today’s increasingly large biobanks may represent a new phenomenon, they follow in the tradition of epidemiological studies that have tracked groups of people for years or decades. Much like the process of creating a productive biobank, prospective cohort studies (so called because they track events that happen during the study period) involve recruiting and tracking large numbers of patients for long periods of time, and such efforts have yielded invaluable information. The Nurses’ Health Studies, for example, which have been following more than 238,000 female nurses since 1976, have showcased the effects of smoking, alcohol,diet and birth control pills on women’s health. The ongoing ARIC study has probed differences in disease risk by race, gender, location and date.
Those prospective efforts have the advantage of letting researchers in a single study consider genetic, environmental and lifestyle factors involved in a disease, says Rory Collins, chief executive of the UK Biobank, a publicly funded national biobank with more than half a million participants. Of those factors, genes are the easy part; because a person’s genes are constant, researchers can perform analyses of people who already have a disease to determine whether certain genetic mutations may have put them at risk. But understanding nongenetic factors such as diet, socioeconomic status or activity levels is trickier, because any or all of those things may change. Piecing together what happened in the past can be difficult, so the most reliable method is to follow subjects to discover their fates.
Biobanks may be a boon to medical research as they grow in size and complexity, but they raise several challenges, says Pearl O’Rourke, director of human research affairs at Partners HealthCare in Boston (a system founded by Brigham and Women’s Hospital and Massachusetts General Hospital), which is developing its own biobank. O’Rourke explains that leftover blood or tissue can be used in research without patient consent if it is disconnected from identifiable information and there’s no interaction between researcher and patient. But most biobanks want identifiable specimens that can be linked to donors’ existing information and updated over time.
Traditionally, people who participate in research undergo informed consent, in which risks are disclosed to them. But informed consent was designed for subjects enrolled in single studies with specific aims. Biobanks, on the other hand, are enduring repositories of biological samples and data about their donors, information that may be “checked out” by different researchers with different agendas.
The dilemma of biobanking, says O’Rourke, is that “it is asking a lot of people, yet it is asking very little of people at the same time.” On the one hand, biobanking requires minimal involvement with its subjects, while on the other hand it expands the scope of traditional studies in terms of the variety of research and duration of someone’s participation.
Becoming a research subject in a biobank could be as easy as sending off a blood sample during a doctor’s visit. But rather than participating in a specific study, you are donating the blood for a range of current and future purposes. DNA could be extracted and sequenced. Blood metabolites could be identified and linked to information about your latest doctor’s visit. Your sample might be used to study cardiovascular disease or bipolar disorder. Your cells might be cloned and passed along to other laboratories. An organization could use your information to develop a profitable new drug, or the biobank itself might be transferred to a new location.
Biobanks have adopted new approaches to consent. Vanderbilt University’s BioVU project, for example, automatically adds leftover blood from routine clinical tests to its DNA repository unless patients opt out by checking a box on a form they must sign to consent to treatment. The majority of biobanks, in contrast, use an “opt in” model so that patients must actively choose to participate. But whatever happens when samples are collected, patients aren’t likely to be given specifics about how their contributions may ultimately be utilized. “We can’t really predict what future research will bring,” says Christopher Thomas Scott, a senior research scholar at Stanford University’s Center for Biomedical Ethics. Biobanks have been moving toward a type of broad consent in which subjects acknowledge that their samples could be used in a variety of ways—in essence donating their data “to science.”
But Alberta’s Caulfield believes that while there is growing consensus among biobank researchers and administrators about the concept of broad consent, surveys suggest the public holds varying attitudes on the issue. “I always call it a ship of consensus sailing on a sea of dissent,” he says. The Havasupai case highlighted the fact that some people may want more of a say in how their samples are used.
Privacy has also been a concern. Biobanks can keep samples anonymous or strip them of personal identifying information, and have even invested in encryption software. Still, a recent study in the journal Science demonstrated that it’s possible to identify some anonymous donors. Genetics researchers at the Whitehead Institute for Biomedical Research searched the database of the 1000 Genomes Project, which sequences genomes of anonymous participants and makes the information freely available online. By analyzing the Y chromosomes of male subjects, combined with some sleuthing on genealogy Websites, they determined the identities of about 50 people. The paper, published in collaboration with a bioethicist, cautioned that biobanks cannot guarantee anonymity. (Participants’ ages have since been removed from the database.)
But it’s unclear what the consequences of privacy breaches would be, and to what extent the information from a genome could be damaging to someone if seen by employers or insurers. Consent forms often mention the small risk of being identified, and in a world in which people post the details of their lives on social media sites, genetic privacy may become just another casualty of technological progress. Still, O’Rourke says, “medical information is held to a higher standard.”
Finally, one key question is whether patients should be told about research findings that are relevant to their health, such as a genome analysis that uncovers a disease-causing mutation. Last year an NIH-sponsored working group recommended that biobanks take responsibility for returning results to patients if the information is deemed scientifically sound and reveals a condition that could be medically treated. But such a requirement could be difficult and costly. Because most studies look for general associations about health, discoveries with clear implications for an individual’s health are rare. Rowell says that Kaiser, for instance, gives no information to patients except that of “medical significance.” So far, no such reports have been made, and the biobank has no mechanism for handling one. But as more research is conducted on more patients, such decisions are expected to be more common, and it may prove hard for biobanks to create a consistent, logical and ethical process for deciding when to notify patients.
Yet even as biobank operators debate issues of privacy and informed consent, some patients are realizing that they possess valuable biological material and information. In 1999, Joe Maas found out that his seven-year-old son, Kevin, had a rare genetic condition called pseudoxanthoma elasticum (PXE), which causes elastic tissue in the body to become mineralized and can lead to problems with the skin, eyes and blood vessels. A registered pharmacist, Maas began educating himself about the disease. That’s when he came in contact with Sharon Terry, who with her husband had founded a research and advocacy organization called PXE International after their children were diagnosed with the disease. As part of its work, the group established the first biobank to be created by patients.
Maas, his wife and their three children all provided samples for the biobank. “I’m zero hung up on the privacy matter,” says Maas, who now serves on the board of PXE International.
Though not a scientist, Terry has contributed to the scientific literature on PXE and also heads a second organization, Genetic Alliance, that builds resources to support research on all diseases. PXE International established a disease registry and biobank that was the basis for a larger Genetic Alliance Registry and Biobank, which Terry says includes 30,000 samples from people with inflammatory breast cancer, psoriasis, chronic fatigue syndrome or one of several other diseases. Like Maas, those participants are motivated to donate samples and information to foster research about a disease that affects them or a family member. But Genetic Alliance is betting that many other people will also want to take an active role in contributing to research. The group launched a project called Registries for All Diseases to encourage volunteers with or without particular diseases to submit clinical and health survey information as well as biological samples. The project’s goal is to make it easier for clinical trials to find patients, but the information will also be useful for more basic studies on disease.
While many biobanks have been moving toward broader consent and limited patient involvement, Genetic Alliance has taken the opposite approach. “The participant is at the center,” Terry says. Just as a Facebook user has the option of fine-tuning privacy settings, biobank participants get to choose how their samples and information are used—from allowing any researcher access to requiring permission for each study.
It remains to be seen whether such a generalized, patient- focused database can generate the level of participation and research activity that institution-led biobanks have produced. Krishanu Saha, a stem cell researcher at the University of Wisconsin, Madison, has argued that biobanks tend to view the public as a resource to be mined for data and tissues, and that 43 treating patients as partners might encourage more to participate. One solution, developed by Sage Bionetworks, a Seattle nonprofit, is a tool called Portable Legal Consent, designed to give patients a way to donate their genetic information to science in a broader way than before. After going through an informed consent process, patients attach that consent to any data they donate to a common database; their de-identified data becomes available to researchers who agree to a broad set of conditions. The data is not owned or held by specific institutions or investigators. It’s not clear yet whether or how this type of model could be expanded to banking of biological samples.
The goal of gathering millions of samples and patient records requires an unprecedented involvement of volunteers in research. But the wide variation in strategies of biobanks points to a shifting landscape in research and culture more generally. As science scales up, are stricter controls needed to protect patients, or are they simply a remnant of paternalism in research? Whatever the answer, biobanks need to face the growing emphasis on connectedness, transparency and data-sharing that has permeated science while negotiating an increasingly involved public. We each hold information in our cells that could lead to new discoveries about disease and health. The fundamental challenge for biobanks is to collect and deliver that information to science in a way that benefits all of us.