Principles organization and_operation_of_a_dna_bank


Published on

not mine

  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

Principles organization and_operation_of_a_dna_bank

  1. 1. Controlled Clinical Trials 23 (2002) 222–239 0197-2456/02/$—see front matter © 2002 Elsevier Science Inc. All rights reserved. PII: S0197-2456(02)00193-9 Principles, organization, and operation of a DNA bank for clinical trials: a Department of Veterans Affairs cooperative study Philip W. Lavori, Ph.D.a,b, *, Heidi Krause-Steinrauf, M.S.c , Mary Brophy, M.D.d , Joel Buxbaum, M.D.e , Jennifer Cockroft, M.P.H.a , David R. Cox, M.D., Ph.D.b , Louis Fiore, M.D.d , Henry T. Greely, J.D.f , Harry Greenberg, M.D.a,b , Edward W. Holmes, M.D.g , Lorene M. Nelson, Ph.D.b , Jeremy Sugarman, M.D., M.P.H., M.A.h a Palo Alto VA Health Care System, Palo Alto, California, USA b Stanford University, Stanford, California, USA c National Heart, Lung, and Blood Institute, Bethesda, Maryland, USA d Boston VA Health Care System, Boston, Massachusetts, USA e The Scripps Research Institute, La Jolla, California, USA f Stanford University Law School, Stanford, California, USA g University of California at San Diego School of Medicine, La Jolla, California, USA h Duke University School of Medicine, Durham, North Carolina, USA Manuscript received April 27, 2001; manuscript accepted December 10, 2001 Abstract The mapping and sequencing of the human genome promises rapid growth in understanding the ge- netically influenced mechanisms that underlie human disease. To realize this promise fully, it is neces- sary to relate genetic information to clinical phenotypes. Genetic tissue banking in clinical studies pro- vides opportunities to analyze the genetic contribution to variation in response to treatments. The challenges to progress are likely to come from the complex organizational, social, political, and ethical issues that must be resolved in order to put clinical and DNA bank information together. Concerns about subjects’ rights, informed consent, privacy, and ownership of genetic material require attention in the development of DNA banks. In this paper we describe one approach to the solution of these * Corresponding author: Philip W. Lavori, Ph.D., VA Cooperative Studies Program Coordinating Center, 795 Willow Road, Bldg. 205, Basement 151K, Menlo Park, CA 94025. Tel.: ϩ1-650-493-5000, ext. 22522; Fax: ϩ1-650-617-2605. E-mail address: Heidi Krause-Steinrauf was formerly with Palo Alto VA Health Care System; Edward W. Holmes was for- merly with Duke University School of Medicine.
  2. 2. P.W. Lavori et al./Controlled Clinical Trials 23 (2002) 222–239 223 problems that was adopted by one clinical trials group, the Department of Veterans Affairs Coopera- tive Studies Program. © 2002 Elsevier Science Inc. All rights reserved. Keywords: DNA banking; Ethics; Genes; Pharmacogenomics Introduction The mapping and sequencing of the human genome is a watershed event in medical science. We are in a period of rapid growth of the body of knowledge of human genetics and the cor- responding understanding of genetically determined mechanisms that underlie human disease [1–3]. Some genetic assays are already available for clinicians to use in diagnosis, treatment, and prognosis. To realize fully the potential of the new tools being developed by genome science, it is necessary to relate genetic information to accurate and detailed clinical information, including onset, course, and outcome of disease (i.e., the clinical phenotype) [4,5]. Randomized clinical trials and prospective observational studies that include storage of genetic tissue provide opportunities to gain insights into the genetic basis of variation in re- sponse to treatments [6]. Clinical trials provide the best evidence for treatment guidelines. However, future clinical practice may require that trial results be coupled with genetic infor- mation to determine treatment choices by genotype. The science of pharmacogenomics offers the prospect that information about the genetic determination of response to treatment can be used to individualize treatment selection and predict side effects [7]. Four essential elements or factors are needed for future progress: (1) high-resolution maps of the human genome with the ability to determine rapidly allelic variation in individuals, (2) high-quality, large-sample clinical datasets with well-characterized participants and longitudinal follow-up for effects of treatments, (3) stored DNA or other genetic tissue from the participants in the datasets and (4) advanced informatics for storage, retrieval, and correlation of complex clinical and genetic data. The genotypes that play a role in most diseases are unknown, even in many diseases with a rec- ognized genetic component. Stored DNA from persons with specific diseases can provide a rich resource for future genetic research in cardiovascular diseases, neurologic disorders, cancer, diabe- tes, respiratory disorders, and psychological disorders. Pharmaceutical companies and the biotech- nology industry have a keen interest in finding high-quality clinical datasets that are linked to ge- netic material to advance the development of new therapies. Most importantly, patients will benefit from research based on stored DNA as new knowledge affects the diagnosis of disease, the devel- opment of new therapies, and how these therapies are targeted. For these reasons, storage of ge- netic tissue is beginning to be adopted as a regular part of clinical trials and observational studies. If the use of genetic information were simply a matter of introducing a new diagnostic subtyping method, there would be little need for special consideration; genotyping would take its place beside blood chemistries, electrocardiography, and family history as a new tool for clinicians to use. However, the public perceives that genetic information is special. Many people believe that in some deep sense their genes carry their specific human identities, and the public is inclined to a greater belief in genetic determinism than even the most reductionist biological scientist [8]. Historical abuses of the science of inheritance, such as the eugenics move- ment in the early twentieth century, strengthen public concern.
  3. 3. 224 P.W. Lavori et al./Controlled Clinical Trials 23 (2002) 222–239 Progress in clinical genomics in the next decade is likely to come from resolution of the com- plex organizational, social, political, and ethical issues that arise when linking clinical and DNA information to create a resource for future scientific use (a “DNA bank”). Stored DNA without linked information about diagnosis, treatment, and follow-up is much less valuable, while a clini- cal dataset without stored genetic material is incomplete. Yet concerns about subjects’ rights (both as individuals and as groups), privacy, ownership of genetic material, and the stability of ar- rangements made to resolve these concerns require attention in the development of plans to store DNA [6,9,10]. In this paper we describe one approach to the solution of these problems that was adopted by one clinical trials group, the Department of Veterans Affairs (VA) Cooperative Stud- ies Program (CSP). Our decisions were determined in part by the special research context of our group. However, we hope this approach may be useful to others confronting similar problems. The Cooperative Studies Program The VA CSP is one of the oldest clinical trials organizations in the world. At any time, over 30 clinical trials and observational studies are ongoing in diseases that affect veterans. Many CSP studies enroll over 1000 patients, and follow-up may extend for several years, as- sessing outcomes such as mortality, major morbidity, and utilization of health resources. The multicenter clinical trials and prospective cohort studies in the CSP provide an opportunity for collecting and storing DNA for research into specific diseases. To respond to this opportunity, the CSP initiated study #478 “Genetic Tissue Banking in VA Clinical Research” aimed at providing a format for human subjects protection and a sci- entific, technical, and statistical infrastructure to support DNA banking in its studies. This ini- tiative aims to develop standard tools and protections that can be offered to investigators who want to integrate genetic information into CSP studies. The DNA Bank grew out of the CSP’s experience jointly managing a DNA bank with the National Heart, Lung, and Blood Institute. The joint DNA bank was part of the Beta-Blocker Evaluation of Survival Trial (BEST), a heart failure clinical trial, which collected over 1000 DNA specimens from BEST patients. The CSP DNA Bank was approved as a demonstration project in May 1999 and was approved as a regular part of CSP in May 2001. Currently two studies are collecting samples, and the Bank has provided planning support for another. Conceptual foundations for the CSP DNA Bank DNA bank versus genetics substudy Even in an individual clinical trial (single- or multisite), a DNA bank differs importantly from a genetics substudy. A substudy focuses on specific genetic hypotheses, which in turn define the extent of genotyping as well as the plans for analyses relating the genotypes to the clinical phenotypes and outcome of disease. In contrast, a DNA bank is oriented toward fu- ture hypotheses that may not be framed at the outset. It must justify itself on the basis of the latent scientific value of the clinical data in the context of substantial current ignorance about what will emerge from the genotyping. Thus, a DNA bank is not limited a priori to specific loci or genotyping techniques. Rather, the entire genome is in play.
  4. 4. P.W. Lavori et al./Controlled Clinical Trials 23 (2002) 222–239 225 DNA banking for a program of studies The CSP DNA Bank serves several different “parent” clinical studies, in different disease areas, over a longer period of time than would a DNA bank for a single study. Thus, the Bank must be able to call on a broad variety of expertise for planning and monitoring its ac- tivities. The scientific and ethical context of DNA banking is likely to evolve at a rapid pace. To respond, a bank must follow a dynamic and adaptive operational plan, rather than adher- ing to the fixed assumptions, methods, and design of a typical single scientific project. Scien- tific and ethics oversight committees play a central operating role in a multistudy bank and are the main routes for uptake of the changing rules and practices in the world outside the bank. In a bank serving a single study, these oversight functions might be served by some combination of existing committees (such as the study steering committee and the data and safety monitoring board) or even by the principal investigator in a single-site study. Guiding principles The CSP DNA Bank operations are guided by six key principles: Respect for autonomy The Bank requires a genetic tissue banking consent process that is separate from that of the clinical study through which the subject is recruited (the “parent study”). This gives the prospec- tive subject the opportunity to refuse to participate in the Bank while still taking part in the parent study. A separate process also provides an opportunity for complete disclosure of the scope of the linked clinical and genetic information. The Bank asks for consent to genetic tissue analysis by all present and future methods, including individual gene typing as well as genomewide scans. In contrast to the open-ended scope of genotyping, the Bank calls for limiting the clinical data to that collected in the parent study protocol. That is, without obtaining new informed consent the Bank would not collect or link additional phenotypic data on the subject beyond the parent study data- base. This limitation would also require new consent for post-parent-study follow-up and search of VA or other databases. Explicitly limiting the information base for future studies helps to en- sure that the participant’s specimen will not be used in research against his or her wishes. The decision to limit clinical data may reduce the value of the linked data; however, we are persuaded to do so by three arguments. First, the CSP trial datasets are particularly rich, since they are designed to support the needs of an intensive clinical investigation. Second, while the genetic analysis technology is still in flux, the clinical hypotheses are relatively familiar and likely to be well covered by the typical follow-up period in a CSP trial (often 3 or more years for each subject). Therefore, little is lost by requiring a unique and separate informed consent process in order to go beyond the originally contemplated data collection. Finally, asking permission to use the subject’s specimen for completely unspecified future study would appear to conflict with the need to provide truly informed consent and with the responsibility of in- stitutional review boards (IRBs) to weigh risks and benefits, since the possible risks associ- ated with such future research cannot be known at the time of consent or review (see page 65, National Bioethics Advisory Committee report Research Involving Human Biological Materials: Ethical Issues and Policy Guidance, August 1999).
  5. 5. 226 P.W. Lavori et al./Controlled Clinical Trials 23 (2002) 222–239 If a subject at any time chooses to withdraw consent, the Bank will destroy the subject’s specimen and delete linking information from its databases. The ability to guarantee that the subject’s wishes have been honored in this respect depends on the close control of linked data (see below). Protection of privacy In the terminology of the National Bioethics Advisory Committee (NBAC) report, the Bank is a repository of identified specimens, since it would be possible to identify the subject from whom the specimen was obtained. Genotyping will be performed by investigators who are provided coded samples by the Bank. The resulting genotypes are linked to the clinical data by the DNA Bank. Tissue samples or genotype data that are merged with clinical data may not properly be termed unlinked if a subject’s identity could be ascertained. These linked datasets are the source of the threat to privacy, and many of the Bank’s procedures are devoted to minimizing the risk of inappropriate disclosure of private information. It is important not to overstate the ability of the Bank to guarantee the research subject’s privacy [8,11]. But the investigator must be able to assure that the risks of inappropriate dis- closure are small. By controlling the availability of linked datasets, the Bank limits access to sources of information that might potentially be used to identify the individual subjects asso- ciated with the genetic tissue analysis. Taking steps to “de-identify” datasets before export- ing them to scientists outside the Bank may not be sufficient protection. The potential for “re-identifying” such data exists, by using unique combinations of clinical and demographic data to point to specific individuals. To avoid the risk of reidentification, the CSP DNA Bank proposes, as far as possible, to provide the service of genetic and clinical data analysis, in collaboration with scientists (“cli- ents”) who propose to access the specimens stored in the Bank. Such a service is a natural extension of the role of the CSP coordinating center in the conduct of its studies. It is possi- ble that some clients might make a strong case for obtaining and analyzing their own copies of a coded dataset, containing both clinical and genetic information. The Bank’s oversight committees will evaluate the rationale for such departures from normal operations and will verify that the released data do not pose a risk of breaching any subject’s privacy. If the ratio- nale is compelling, the Bank will release the information necessary to perform the analyses specified in the use proposal. The client must agree to maintain control over the dataset and limit use of the dataset to the stated purposes. This restricted access is one of the “working assumptions” that will need to be reevaluated over time, to test whether or not scientific productivity is compromised. At the outset we seek to minimize the number of people and institutions that the research subject is asked to trust with his or her genetic and clinical data. One cost of our position is that a sophisticated client would have to go through the steps of convincing the oversight committees that direct access to data is warranted. Another cost is that the CSP DNA Bank must maintain the necessary expertise to make the service available. We recognize that our decision is based in part on the special circumstance that the CSP coordinating centers are already charged with statistical and data management responsibilities for CSP studies. Other groups might believe that the reidentification possibility is sufficiently remote as to be an ignorable threat to privacy. The CSP DNA Bank oversight committees may come to this conclusion in the future, as well.
  6. 6. P.W. Lavori et al./Controlled Clinical Trials 23 (2002) 222–239 227 Appropriate disclosure The consent process and supporting documents must specify what will be done with po- tentially important emergent information about individual subjects [12–14]. There are three options: (1) never report information to anyone about an individual subject’s results, (2) of- fer the subject choices for each foreseeable contingency and let the subject decide, and (3) as a matter of policy, do not report back, but if unusual circumstances arise refer the specific in- dividual matter to the Bank oversight committees for a detailed plan. The CSP DNA Bank has rejected the rigid “no disclosure” option (1), even though it is a way to deal with a difficult problem. If something in the genetic profile of an individual dis- closes a substantial risk that could be averted by an intervention and the Bank has the ability to identify that person (through the patient’s enrollment site), it is obliged to act. Further- more, the richness of a clinical trials dataset makes it nearly impossible to truly anonymize linked data at the coordinating center. Even a randomly reordered data set (“scrambled iden- tifiers”) is relatively easy to decode. Therefore, although it is possible to guarantee that one will not seek to identify a patient, it is not possible to guarantee that one is unable to do so. Because the CSP is part of a nationwide medical care system and not a freestanding research enterprise, it has responsibilities that derive from its intramural role. Other groups might come to a different conclusion given different responsibilities and capabilities. Option (2) calls for the subject to determine the detailed rules for disclosure of test information during the consent process. Determining whether such a policy makes sense involves several con- siderations. First, the probability of discovering clinically relevant and reliable information in a particular individual is small given the present state of knowledge. The fact that an individual has a predisposing genetic profile may not be relevant if there is nothing that can be done about it. A genetic finding whose significance is speculative or purely statistical may not be helpful and may, in fact, be harmful in the sense that persons may have great difficulty interpreting the results [15]. If the information has been revealed by an experimental test, it would need to be repeated with standard procedures in a certified clinical laboratory. These options multiply confusingly, and each has a low probability of occurring. Similar (and more complex) issues arise if we consider disclosure of information that may be important to the relatives of a subject. This reasoning led the Bank to option (3), as described in the model informed consent document (see appendix). Since there is a very small chance of uncovering individual infor- mation that would be important for the subject to know, the subject should not expect to learn the results of his or her own genetic analyses. However, in the unlikely event that facts of individual import should arise, the model informed consent document states that the Bank will bring that case to the scientific and ethics oversight committees for their help in prepar- ing an individualized plan to use that information appropriately. A specific option to fore- close that possibility entirely will be made available to the subject. As the ability to use genetic information for the benefit of individual patients increases, the list of genotypes that might be disclosed will grow. The experiences of the oversight committees in this DNA Bank may contribute to ongoing consensus about proper actions in this area. At present, there are no foreseeable clinical benefits that can be promised to subjects who participate in DNA banking. However, as specific concerns regarding whether disclosure is appropriate arise, the potential for specific benefit will also increase, calling for continued assessment of ap- propriate disclosures in light of evolving knowledge.
  7. 7. 228 P.W. Lavori et al./Controlled Clinical Trials 23 (2002) 222–239 Protection from harm to individuals The risks to subjects from genetic research and genetic test results have been widely dis- cussed. The risks of discrimination in insurance and employment have been particularly stressed [8,14,16–19]. Also important are other possible harms that disclosure of information about their genomes might cause research subjects, such as familial discord or personal psy- chological problems [15,20]. The privacy protections outlined above should prevent un- planned or inadvertent disclosure. When, in exceptional circumstances, the information must be made available for medical reasons, the process would be planned specifically to avoid or minimize these kinds of harms. Avoiding harm to groups Of primary concern here are perceptions about the possible harm to racial or ethnic groups that might follow even scientifically well-grounded and well-intentioned proposals. Individ- ual members of those groups might not be willing to see their tissue used in studies contain- ing ethnic identifiers [14,21–23]. The Bank will seek out the views of individuals from po- tentially affected groups and include them in the review and oversight process. The oversight committees (see “Organization and operations”) will review all proposals for use with these sensitivities in mind and will consider whether additional consent requirements are appropri- ate for genetic analyses that focus on topics of particular sensitivity, such as genetic links to addictive behaviors. The Veterans Advisory Group (see “Organization and operations”) will be asked to advise the Bank on matters of policy and organization. Dealing fairly with commercial interests The DNA specimens and clinical information in the Bank could lead to the development of medically important and financially lucrative products. Whether any such value can or will follow from the Bank is highly uncertain; that uncertainty itself contributes to the con- cerns of some patients and others about possible exploitation [14,21,23,24]. Fairness to the research subjects may, in some cases, require that when tissues and information from an identifiable group (such as veterans) contribute to a commercially valuable product, some share of the value should be used to benefit the group. This principle suggests that if the Bank is able to recapture some of that value, the proceeds should be used for the direct bene- fit of VA patients. VA research and development staff may negotiate licensing agreements or other contracts to accomplish this goal. Other groups, such as a consortium of universities or managed care organizations, might see these opportunities in a different way. Organization and operations In this section, we describe the organization of the Bank and step through the standard op- erating procedures, beginning with informed consent and proceeding through handling, ship- ping, and storage of specimens, linking of clinical and genetic data, and review of proposals for use. The Bank has five components, providing it with adequate resources to store DNA, to conduct analyses that relate the genetic and clinical information, to manage and maximize the scientific use of the Bank, and to address the ethical, legal, and social implications of this
  8. 8. P.W. Lavori et al./Controlled Clinical Trials 23 (2002) 222–239 229 sensitive material. These components are the Genetic Tissue Core Laboratory, the DNA Co- ordinating Center, the Scientific Advisory Committee, the Ethics Oversight Committee, and the Veterans Advisory Group. Members of oversight committees are appointed by the VA Research and Development Office and serve overlapping terms. The Genetic Tissue Core Laboratory (GTCL), housed at the Massachusetts VA Epidemi- ology Research and Information Center (MAVERIC) in Boston, serves as a central reposi- tory for specimens. The GTCL trains and equips study personnel for the collection of genetic tissue, extracts and stores DNA, and performs some genotyping. The GTCL ships DNA specimens to investigators with approved requests for use of the DNA Bank. Such technical support for collection, storage, shipping, and processing of material is designed to minimize the extra burden on study personnel and to ensure that genetic material is handled and stored in a manner that preserves its value. The DNA Coordinating Center (DNACC), a subunit of the Palo Alto CSP Coordinating Center, administers the Bank and makes its resources available to CSP investigators who plan to collect DNA specimens. The director of the Palo Alto CSPCC is the administrative head of the DNA Bank. The DNACC is responsible for maintaining the protection of human subjects as standards evolve and for obtaining administrative assurances of confidentiality. It is responsible for updating regulatory documents and consent forms and handling IRB que- ries related to collection of DNA specimens. DNACC staff coordinate the committees that deal with the ethical, legal, and social implications of banking and using genetic tissue. Protecting the confidentiality of genetic information is a central responsibility of the DNACC, which maintains appropriate linkages to clinical datasets and controls all access to genetic information. The DNACC takes overall responsibility for assuring that the genetic tissue and clinical data repositories have physical, legal, and administrative security. The DNACC pro- vides statistical and data management support to investigators planning a DNA bank and to cli- ents who wish to access specimens and conduct analyses. The Scientific Advisory Committee (SAC), a group of individuals with expertise in genet- ics, epidemiology, molecular biology, and specific disease areas, helps set policy for the use of the Bank and provides technical and scientific advice to the DNACC. The SAC helps identify new studies that should be considered for genetic tissue collection. It advises on is- sues such as the type and quantity of specimens for banking, storage requirements, and addi- tional clinical information that should be included in the parent-study data collection to en- sure the utility of the DNA specimens for later analysis. The SAC meets periodically to review and recommend approval of proposals for use of the Bank. The Ethics Oversight Committee (EOC) is composed of experts in the legal and ethical implications of genetics research with humans, as well as experts in the relevant scientific disciplines. It meets regularly to review provisions for the protection of human research vol- unteers and to provide a disinterested review of the activities of the Bank, including propos- als for use of the stored tissue. The Veterans Advisory Group (VAG) consists of a group of veterans who will be asked to provide their perspectives on the ethical and social implications of DNA banking in veteran subjects of CSP studies. The advice of this group will help shape future operations of the Bank, which is designed to adapt to meet the changing social consensus on the use of such material and information. Some examples of the initial agenda items for such a group include
  9. 9. 230 P.W. Lavori et al./Controlled Clinical Trials 23 (2002) 222–239 the “disclosure” provisions (whether, when, and how to inform subjects about genetic infor- mation that may be important to them or their families), subgroup sensitivities to certain re- search topics (for example, minority group differences in genetic vulnerabilities and the ge- netics of stigmatized disorders), and commercial use of Bank information. Interaction of the Bank with the parent clinical study The involvement of the Bank begins during planning of a new clinical study to help to assess the added costs, risks, and benefits of a gene banking component and advise on the need for any special procedures or data collection in the parent study. The DNACC works with the parent- study investigators to adapt the generic protocol and template consent forms for the DNA compo- nent of the parent study. The parent-study investigators are responsible for framing any genetic hypotheses that are known at present, so that the initial plans for genotyping can be reviewed. It also identifies study-specific issues relating to disclosure, privacy, or other matters that might af- fect the design of the study or the study informed consent procedures. The DNACC also helps the local site investigators to obtain local approvals, lending its experience with such IRB reviews to investigators who might not have participated in a DNA bank before. Collection, handling, shipping, and storage of specimens The GTCL works with the parent-study investigators to prepare materials and supplies and handles the logistics of specimen preparation and shipping. The study subjects are re- cruited to participate in the DNA component after giving informed consent to enroll in the parent study. Only those who choose to participate and who provide a separate consent form are enrolled in the Bank. The blood specimen for DNA analysis is collected and sent to the GTCL using labels from the kit supplied to the site. No identifying information other than the “specimen code number” is placed on the biological specimen sent to the GTCL (see Fig. 1). A duplicate specimen code number is attached by the site to a data form that also has the “pa- tient code number” (the identifier of clinical information collected in the parent study). These data are sent to the DNACC. Once received at the GTCL the specimens are usually pro- cessed and frozen prior to DNA extraction in order to allow for high-volume batch extraction once all the specimens have been collected. The GTCL stores the specimen under a “bank code number,” which it assigns. After sending the log that matches specimen and bank code numbers to the DNACC, the GTCL destroys the log. Thus, the only links between the patient identification number and the bank code number are kept at the DNACC. Access agreement An investigator (the “client”) making a proposal for access to the tissue specimens or to the linked clinical and genetic database, first provides the Bank with a statement of research intent (SORI), outlining the hypotheses, proposed genotyping, and statistical analyses. The DNACC may provide statistical assistance to the client preparing the SORI. The SORI is reviewed by the SAC, and if approved, the investigator submits a full research proposal to the SAC and EOC for review. Bank staff work with the client to satisfy the requirements of the committees. Upon final approval, the applicant institution enters into a materials transfer agreement, which summarizes
  10. 10. P.W. Lavori et al./Controlled Clinical Trials 23 (2002) 222–239 231 the terms and conditions of the agreement to access the Bank. This agreement spells out the re- sponsibilities of all parties in connection with the receipt, handling, storage, and use of DNA specimens and the protection of the privacy and confidentiality of patient data. The GTCL sends samples to the analytic laboratory identified with the bank code number. The analytic lab returns the results of the genetic analyses (indexed by the bank code num- ber) in secure electronic form to the DNACC. Management of information at the GTCL and DNACC The coding data described above are used by the DNACC to construct a protected cross- walk database that links clinical and DNA information. After the DNACC receives genotype data from the analytic lab, it uses the key matching bank code number and the specimen code number, together with the key matching the DNA specimen code number and the clinical (parent) study identifier, to link the genotype data and the clinical data. The analyses speci- fied in the research proposal are then carried out by the statisticians at the DNACC. If the DNACC cannot perform the analyses, it may create coded datasets for export to certain cli- ents. The results of all genetic assays become part of the DNA Bank and are available for use by other researchers. The DNACC serves as the entry point for all interested parties and ac- tively seeks out such collaborators to maximize the utility of the Bank. Informatics and statistical analysis The informatics system of the CSP DNA Bank is designed to protect the patient’s privacy by securing the linked clinical and genetic databases against intrusion, unauthorized copying, loss, and other threats. At the same time, it is useful for potential clients of the Bank to be able to browse the protocol, forms, and procedure manuals of the parent studies to assess the suitability of the Bank data for testing their hypotheses. The Bank also will provide conve- nient access to aggregate summaries of the diagnoses, demographics, outcomes, and other pertinent characteristics of the participants in the parent studies of the Bank. The core of the Bank informatics system is a cluster of computers, a printer, backup de- vices, and other hardware that is physically isolated from other networks and computers in the coordinating center and located in a secured room. This private system hosts the linked databases and the linking information. The public face of the Bank is a web server providing access for the purposes described above. In addition to document browsing, the Bank will provide a system that would allow limited exploration of the clinical datasets without allow- ing access to potentially unique identifiers. Enabling a broad range of potential collaborators to perform these analyses by employing a user-friendly web-based interface will allow greater use of these datasets and foster increased collaboration among researchers. As discussed above, the staff of the DNACC is prepared to collaborate with the client in- vestigators to perform the analyses of linked genetic and clinical data. The clients would be able to test their hypotheses and publish the results without the need for a transfer of sensi- tive data outside the Bank. Each of the parent studies is coordinated by one of the statistical or epidemiologic centers of the CSP. Therefore the staff of the DNACC has ready access to the experience of the biostatistician coordinating the study. Familiarity with the clinical dataset can
  11. 11. 232 P.W. Lavori et al./Controlled Clinical Trials 23 (2002) 222–239 Fig. 1. VA cooperative study program DNA banking and data collection process.
  12. 12. P.W. Lavori et al./Controlled Clinical Trials 23 (2002) 222–239 233 be an advantage. However, exploration sometimes produces important results and increasing the number of researchers with access to data may be a better way to explore a large and well-constructed dataset. This trade-off is similar to the one that arises in the analysis of data from trials: whether to plan and test prespecified hypotheses to preserve the unique confir- matory strength of the data collected in a clinical trial or to explore and exploit serendipity. Client investigators who do not wish to avail themselves of the statistical services of the DNACC must provide a specific rationale for obtaining custom-coded datasets, and the pro- posal must be approved by the SAC and EOC, as well as the VA’s Chief Research and De- velopment Officer. The Bank takes the view that exported datasets are outside its control, and therefore it cannot guarantee that they will not be disseminated beyond the client. Re- moval of ostensible identifiers may not be proof against malicious attempts to link the data to specific individuals. If the risk of loss of privacy is even slightly increased by propagating copies of the linked data, the risk must be justified by specific scientific benefits. These concerns motivate the Bank’s current restrictive policy on export of linked data, but these decisions are not immutable. If the future of responsible, safe DNA banking leads to a more re- laxed, permissive environment, then coding may be seen as a sufficient guarantee of protection. Or, it may be possible to fashion a compromise, allowing limited direct access to parts of the database. Discussion The undeniable scientific potential of gene science is accompanied by persistent public wariness about the ends and means of genetics research. For public trust to flourish, rules, in- stitutions, and methods for the ethical and efficient conduct of genetics research must de- velop alongside the technical advances in genotyping. The CSP confronted the issues raised by DNA banking in clinical research as part of a more general VA research and development review of the implications of gene science for VA research and clinical care. This review led to a programwide planning effort resulting in the current CSP DNA Bank. The most difficult issues in planning the CSP DNA Bank emerged from the ethical, social, and legal implications of the design choices that at first appeared necessary to maximize the scientific value of the Bank. It is useful to make such apparent conflicts explicit, so they can be addressed. More importantly, public trust is best earned by openness about ends and means. The implications will also evolve over the life of the Bank, as some public fears abate and others increase. The Bank purposely avoids some of the more difficult issues by limiting its purview to new studies. The Bank’s procedures do not deal with the use of specimens collected in previous stud- ies, without informed consent for genetics research (the “legacy” problem) [13]. The prospective stance of the Bank is intentional. The primary aim of pharmacogenetics is to tailor treatments to genotype, making it desirable to study the interaction of genotype with current treatments. The Bank’s procedures have not been extended to the collection of genotype and phenotype informa- tion for family studies. All of the issues raised in the foregoing take on added complexity in the context of recruitment of family members for DNA banking. The Bank addresses scientific ques- tions best answered in specimens of patients with well-defined diseases treated in a controlled, ex- perimental context. On the other hand, we see no barrier to extending the Bank’s procedures to the collection and storage of data on the expression of genes in various tissues or to the study of
  13. 13. 234 P.W. Lavori et al./Controlled Clinical Trials 23 (2002) 222–239 the “proteomics” of disease (the expression of genes and translation into proteins). These will no doubt take on great importance in the years ahead as costs come down and techniques evolve. Acknowledgments The opinions expressed herein are the opinions of the authors and do not represent official position of the Department of Veterans Affairs. This work was funded by the Department of Veterans Affairs Cooperative Studies Program as CSP#478. References [1] Ellsworth D, Hallman DM, Boerwinkle E. Impact of the human genome project on epidemiologic research. Epidemiol Rev 1997;19:3–13. [2] Khoury MJ. Genetic epidemiology and the future of disease prevention and public health. Epidemiol Rev 1997;19:175–180. [3] Leiden JM. The genetics of dilated cardiomyopathy — emerging clues to the puzzle. N Engl J Med 1997;337:1080–1081. [4] Steinberg KK, Sanderlin KC, Ou C, et al. DNA banking in epidemiologic studies. Epidemiol Rev 1997;19:156–162. [5] Merikangas KR, Sendsen JD. Genetic epidemiology of psychiatric disorders. Epidemiol Rev 1997;1:144–155. [6] Report of the Special Emphasis Panel (SEP) on Opportunities and Obstacles to Genetic Research in NHLBI Clinical Stud- ies. Bethesda, Maryland: National Heart, Lung, and Blood Institute, 1997. [7] Evans WE, Relling MV. Pharmacogenomics: translating functional genomics into rational therapeutics. Science 1999; 286:487–491. [8] Bondy M, Mastromarino C. Ethical issues of genetic testing and their implications in epidemiologic studies. Ann Epide- miol 1997;7:363–366. [9] Greely HT. Genomics research and human subjects. Science 1998;282:625. [10] Greely HT. Breaking the stalemate: a prospective regulatory framework for unforeseen research uses of human tissue sam- ples and health information. Wake Forest L Rev 1999;34:737–766. [11] Siegler M. Confidentiality in medicine: a decrepit concept. N Engl J Med 1982;307:1518–1521. [12] Last JM. Obligations and responsibilities of epidemiologists to research subjects. J Clin Epidemiol 1991;44(Suppl 1):95S–101S. [13] Clayton EW, Steinberg KK, Khoury MJ, et al. Informed consent for genetic research on stored tissue samples. JAMA 1995;274:1786–1792. [14] Holtzman NA, Andrews LB. Ethical and legal issues in genetic epidemiology. Epidemiol Rev 1997;19:163–174. [15] Burgess MM, Laberge CM, Knoppers BM. Bioethics for clinicians: ethics and genetics in medicine. CMAJ 1998;158:1309–1311. [16] Billings PR, Kohn MA, de Cuevas M, et al. Discrimination as a consequence of genetic testing. Am J Hum Genet 1992;50:476–482. [17] Rothstein MA. Genetic testing: employability, insurability, and health reform. J Natl Cancer Inst Monogr 1995;17:87–90. [18] Lapham EV, Kozma C, Weiss JO. Genetic discrimination: perspectives of consumers. Science 1996;274:621–624. [19] Geller LN, Alper JS, Billings PR, et al. Individual, family, and societal dimensions of genetic discrimination: a case study analysis. Science and Engineering Ethics 1996;2:71–88. [20] Lerman C, Narod S, Schulman K, et al. BRCA1 testing in families with hereditary breast-ovarian cancer: a prospective study of patient decision making and outcomes. JAMA 1996;275:1885–1892. [21] Greely HT. Genes, patents, and indigenous peoples: biomedical research and indigenous people’s rights. Cultural Survival Quarterly 1996;Summer:54–57. [22] Greely HT. The control of genetic research: involving the “groups between.” Houston Law Review 1997;33:1397–1430. [23] North American Regional Committee, Human Genome Diversity Project. Proposed model ethical protocol for collecting DNA samples. Houston Law Review 1997;33:1431–1473. [24] Carey J. Genetically tailored treatments could transform medicine. Business Week. January 18, 1999. p. 98–100. Appendix Sample consent form
  14. 14. P.W. Lavori et al./Controlled Clinical Trials 23 (2002) 222–239 235
  15. 15. 236 P.W. Lavori et al./Controlled Clinical Trials 23 (2002) 222–239
  16. 16. P.W. Lavori et al./Controlled Clinical Trials 23 (2002) 222–239 237
  17. 17. 238 P.W. Lavori et al./Controlled Clinical Trials 23 (2002) 222–239
  18. 18. P.W. Lavori et al./Controlled Clinical Trials 23 (2002) 222–239 239