Viewpointviewpoint Consent and Anonymization in Research Involving Biobanks

viewpointviewpoint Consent and anonymization in research involving biobanks

Differing terms and norms present serious barriers to an international framework

Bernice S. Elger & Arthur L. Caplan

iological samples—such as tissues, expanding the definition of non-identifiable blood and cells—are an increasingly samples to include those that have been Bimportant tool for research into human coded. The growing lack of international diseases and their genetic and physiological consensus interferes with the efficiency of causes. To ease their storage and access, biomedical research that makes use of many of these samples are now stored in biobanks in several countries. We therefore biobanks. The number of human biological argue in favour of using general consent, samples in such collections amounted to together with the right to opt out and several hundred million in 1999 in the USA approval by Institutional Review Boards alone—about one sample per US citizen (IRB), and against an enlargement of the (Eiseman, 2000; Azarow et al, 2003)—and is term ‘non-identifiable’. increasing rapidly. Three-quarters of the clinical trials that drug companies submit to the he term biobank is relatively new. It US Food and Drug Administration appeared in PubMed for the first time (Rockville, MD, USA) for approval now Tin 1996 (Loft & Poulsen, 1996) but include a provision for sampling and storing was not used with any frequency until 2000. human tissue for future genetic analysis Although the term is used to describe vari- (Abbott, 2003). At the same time, there is a ous biological repositories, it originally boom of population biobanks, as more and referred to large population banks of human more countries establish new sample collec- tissue and related data. In this article, tions (Kaiser, 2002). Among the best known biobank refers to any collection of human are: the Icelandic Health Sector Database; biological material—organs, tissue, blood, the Estonian Genome Project; the UK cells and other body fluids—that contains at Biobank; the CARTaGENE Project in least traces of DNA or RNA that would Quebec, Canada; the Banco Nacional de obtained, vary widely. Many European allow genetic analysis. ADN in Spain; the International HapMap guidelines take the view that general con- Project; and several US biobanks, such as sent is acceptable to use samples for future, The growing lack of international the National Children’s Study, the Marshfield as yet unspecified, research projects; US consensus interferes with the Clinic’s Personalized Medicine Research and Canadian policy follows a more rigor- efficiency of biomedical research Project and the National Health and ous standard of consent. Until 2004, both that makes use of biobanks in Nutrition Examinations Surveys. Europe and the USA considered coded and This boom of biobanks has spawned a linked anonymized samples—in which a several countries ‘boomlet’ of regulations and guidelines, code links the sample to its donor—as iden- which has created controversies, particular- tifiable and therefore requiring participants’ The storage of tissue samples and data ly about the importance and definition of consent to future use. However, in 2004, the either linked to the samples or derived from informed consent. The consent of partici- US Office for Human Research Protections them needs to be clearly distinguished. pants is usually required before biobank (OHRP; Rockville, MD, USA) reduced the These data comprise information about the samples can be used in research, but the domain requiring consent to use samples donor of the material, such as demographic nature of this consent, and how it is and/or data that are identifiable, thus characteristics, the type of disease associated

TERMS USED FOR EITHER DATA (2004) and the Australian National Health Research Project in Wisconsin, USA, both OR SAMPLES and Medical Research Council (1999). of which are used to study gene–environ- Various scientific associations put together ment interactions. Typical research pro- Non-exhaustive list of terms used in the their own guidelines about DNA banking, jects that make use of these biobanks will literature to describe different degrees of and the Council for International use DNA from blood or other tissues, data anonymization of samples and data relevant to biobank research. Organizations of Medical Sciences, recog- from the participants’ present and future nizing the importance of biobanks for epi- medical records, and data from screening Completely anonymized demiological research, revised its guidelines questionnaires or physical and laboratory Unlinked anonymized to integrate relevant issues from the biobank examinations. This is combined with Irréversiblement anonymisé debate (CIOMS, 2005). The United Nations information about lifestyle and environ- Irretrievably unlinked to an identifiable person Educational, Scientific and Cultural mental factors that can be regularly updat- Anonymously coded Organization (Paris, France) adopted the ed by sending participants new question- Anonymous International Declaration on Human naires. When establishing such Anonymized Genetic Data in October 2003 (UNESCO, collections of samples and related data, it Unidentified 2003), and France, Germany, Canada and is often impossible to anticipate what De-identified De-linked Switzerland have all issued their own guide- studies might emerge, which leaves the Permanently de-linked lines for biobanks or genetic databases matter of participants’ consent to such Réversiblement anonymisé (CCNE, 2003; Nationaler Ethikrat, 2004; future studies very much in the air. Indeed, Not traceable Commission de l’Éthique de la Science et de a major ethical problem for prospective Traceable la Technologie, 2003; Schweizer Akademie biobanks is how to assure participants’ Coded der Medizinischen Wissenschaften, 2006). consent when it is not known what they Identifiably linked These guidelines contain clearly diver- are consenting to in terms of future Pseudoanonymisiert gent recommendations in important areas, research. The question of the importance Unlinked which interfere with international collabo- and meaning of informed consent is one Unlinked to an identifiable person ration. Not only do different systems exist main reason why international guidelines Encoded Encrypted for the collection of data and the processing on biobanks lack any consensus. Identified of samples, but also the guidelines reflect Nominative fundamentally different ethical frameworks he doctrine of informed consent has Directly identified (Knoppers, 2005). Consequently, scientists been a central component in research Fully identifiable have criticized the little that has been done Tethics since human-rights abuses— Confidential to maximize the potential of biobanks by such as experiments on concentration camp Linked to an identifiable person ensuring that samples and information can inmates in Nazi Germany, and the Tuskegee Identifiable be shared between them (Pearson, 2004). Syphilis Experiment in which US physicians Personal data What are the reasons for this profusion of left victims untreated to study the course of guidelines, and why is it apparently so diffi- the disease—resulted in worldwide abhor- with the sample, the outcome of the disease, cult to devise a single universal framework? rence and regulation. The idea of informed treatment and so on. In addition, DNA and As such a framework exists for clinical consent—the requirement to inform partici- RNA represent information, which has led to research ethics, why is the regulation of pants in a research study of all planned exper- the terms ‘genetic database’ and ‘population biobanks so varied? iments—has accordingly become the gold database’ that are regularly used as syn- standard of research ethics (Kegley, 2004). onyms for the term ‘biobank’. Not only do different systems However, when research participants Not only have the ethical issues related exist for the collection of data provide tissue and information to prospec- to these biobanks raised extraordinary pas- and the processing of samples, tive biobanks, they cannot give informed sions (Barbour, 2003) and stimulated an but also the guidelines reflect consent to future research projects that increasing number of publications, but also have yet to be specified. Consequently, the creation of new biobanks and the fundamentally different ethical according to classical research ethics expansion of existing repositories have frameworks (Annas et al, 1995), participants should be spawned new guidelines. One direct reac- contacted to give consent for each new tion to the heated controversies surrounding It should be noted that the most impor- research project after having been informed the Icelandic biobank is the Declaration on tant ethical questions are different for about the details. This approach is not only Ethical Considerations Regarding Health prospective biobanks as compared with costly (Korn, 1999), but also endangers the Databases, a draft produced by the World existing biobanks, which contain samples scientific value of the entire biobank pro- Medical Association (WMA; Ferney-le- stored before the discussions on ethical ject, as it is highly probable that a consider- Voltaire, France; WMA, 2002). Other exam- issues started. A characteristic of most able percentage of participants will be lost ples are guidelines from the UK Medical prospective biobanks is that samples and for future studies. Either participants cannot Research Council (MRC, 2001), the US data are collected for long-term future be located or they do not respond for vari- National Bioethics Advisory Commission use, not just for a single project. Typical ous reasons, including the simple nuisance (NBAC, 1999), the Council of Europe examples are the UK Biobank and the of reading all the details and reacting to Committee of Ministers (COE, 2006), OHRP Marshfield Clinic’s Personalized Medicine repeated letters asking for new consent.

Many European guidelines take the view that general or broad consent, although dis- ANONYMIZED CODED tinguished from blanket consent, is acceptable for “unspecified future research use” of samples (CDBI, 2006). For example, the Unlinked (irreversibly) Linked (reversibly) Coded Council of Europe’s Steering Committee on Anonymized (USA) Anonymized: a link exists, Researcher has access but the researcher does not to the code (Europe) Bioethics states in an explanatory memo- have access randum that “When biological materials of human origin and personal data are collect- Fig 1 | Communication barriers: the same terms are used with different meanings in ed it is best practice to ask the sources for various guidelines and journal articles. their consent to future use, even in cases where the specifics of the future research abandoned in favour of general consent—a However, a waiver alone is not projects are unknown” (CDBI, 2006). less strict standard. Some Asian countries sufficient to forego obtaining participants’ German guidelines similarly endorse gener- have a similar approach: Japanese guide- consent. If a waiver is granted, “it is still al consent (Nationaler Ethikrat, 2004), as do lines, for example, contain the idea of “com- appropriate to seek consent in order to the recommendations of the UK Human prehensive consent” (Council for Science show respect for the subject, unless it is Genetics Commission (2002) and laws in and Technology, 2000). Not so in the USA. impracticable to locate him or her in order Sweden, Iceland and Estonia, in which a to obtain it” (NBAC, 1999). As an “addition- “broad description of the purpose is lthough there is some support for gen- al measure of protection”, the NBAC men- allowed” (Kaye et al, 2004). eral consent (Grizzle et al, 1999), the tions the possibility for donors to withdraw This form of general consent is consid- Aprevailing opinion in the USA main- from a study that has been granted a waiver ered acceptable if two conditions are ful- tains the classical standard of informed con- of informed consent. filled: the approval of all future projects by a sent, which can be called the ‘American solu- In 2004, the US Office for Human research ethics committee or “competent tion’. In the USA and Canada, the model that Research Protection proposed a different body” (COE, 2006) and the participants’ is most often recommended is so-called solution by broadening the definition of right to withdraw samples at any time. The multi-layered consent, which asks research ‘non-identifiable’ (OHRP, 2004). US federal European Society of Human Genetics writes participants to make different choices on a regulations contains similar provisions to “… individuals may be asked to consent for detailed form. There is a tendency to obtain the Declaration of Helsinki, which states a broader use. In that case, there is no need limited consent, related to one disease or to that “Medical research involving human to recontact individuals although the sub- any specific description of future research subjects includes research on identifiable jects should be able to communicate should projects. An example is the Framingham human material or identifiable data” (WMA, they wish to withdraw” (European Society of Heart Study, directed by the US National 2004). It follows that any research using Human Genetics, 2003). The document Heart, Lung, and Blood Institute (Bethesda, non-identifiable samples does not create an describes withdrawal, sometimes also called MD, USA), which obtains consent for DNA obligation to obtain informed consent and ‘opt out’, as “[i]ndividuals should be given testing. Following legal advice, additional approval of the protocol from an IRB or a the right to withdraw at any time from the consent forms were designed to obtain new, research ethics commission. research, including destruction of their sam- written informed consent from all partici- ple”. Similarly, the Recommendation of the pants to allow RNA testing. hat does ‘identifiable’ mean in the COE Committee of Ministers states that indi- As one would expect, the strict require- context of biomedical research? viduals have the right to withdraw consent at ment of informed consent is a burden for WAgain, European and North any time (COE, 2006). The Ethics and research. It is thus not surprising that in the American standards differ. An analysis of vari- Governance Framework of the UK Biobank American context, there has been a search ous guidelines for a definition of ‘identifiable’ (2003a) has built in these three elements: for alternative solutions. In 1999, the reveals a multitude of different terms (see general consent, approval by an ethics com- guidelines of the US National Bioethics sidebar; Knoppers & Saginur, 2005). Almost mittee and the right to withdrawal. Advisory Commission (NBAC) proposed, every guideline uses separate terminology, among other things, a strategy of waivers, although there are some traditions: for exam- …a major ethical problem for the criteria for which are already defined ple, the Council of Europe’s recommendation prospective biobanks is how to by federal regulations. According to the adopts terminology from previous guidelines assure participants’ consent proposal, the requirement of informed of the UK Medical Research Council (MRC), when it is not known what they consent can be waived if: “[t]he research which is not surprising given that the primary are consenting to in terms of involves no more than minimal risk to the author of the first CDBI draft works with the subjects; [t]he waiver or alteration will not MRC. In the American tradition, the OHRP future research affect adversely the rights and welfare of uses terminology proposed in the NBAC the subjects; [t]he research could not be guidelines. Clearly, there are communication Clearly, this approach, which can be practicably carried out without the waiver barriers where the same term is used with a dubbed the ‘European solution’, changes or alteration; and, [Au: OK?] [w]henever different meaning in different guidelines (Fig classical health research ethics. When it appropriate, the subjects will be provided 1), and readers should examine with caution comes to biomedical research using with additional pertinent information fol- how the terms ‘anonymized’ and ‘coded’ are biobanks, classical informed consent is lowing their participation” (NBAC, 1999). defined in different texts.

The terminology used by the European not have access to the code (Fig 1; OHRP, s we have shown, the challenge pro- documents (CDBI, 2006; COE, 2006) is on 2004). These discrepancies are not limited duced by biobanks is immense: after the basis of five levels of anonymization for to different definitions of the same term, but Amore than 50 years of classical human samples: anonymous, unlinked even more seriously, involve the whole reg- health research ethics, regulatory agencies anonymized (French translation: ulatory framework. have begun to question fundamental ethical irréversiblement anonymisé), linked milestones. Europeans have abandoned anonymized (réversiblement anonymisé), ntil the OHRP revised its guidelines informed consent in favour of general con- coded and identified. If samples contain in 2004, all important regulations in sent and Americans have enlarged the defin- any trace of DNA, they are not truly anony- Uthe USA and Europe agreed on one ition of what constitutes non-identifiable mous, because it is always possible to iden- point: coded and linked anonymized sam- samples and data. As the two continents tify the donor through DNA fingerprint- ples were considered to be identifiable have chosen different ways to change ing—comparing DNA sequences at only both by the NBAC (1999) and in Europe research ethics, a global framework is 30–80 statistically independent single (CDBI, 2002; COE, 2006), because in both becoming impossible. nucleotide polymorphisms will uniquely cases a link exists. Only if this link is irre- What are the possible ways to solve this define a single person (Lin et al, 2004). versibly destroyed are samples and data dilemma? We propose that an analysis of Anonymous is therefore an appropriate considered unidentifiable and, thus, the arguments in favour or against each term only for archaeological samples. The research using such samples was not con- solution will help to find the most accept- term ‘anonymized’ means that biological sidered human subject research in accor- able one. Enlarging the definition of ‘non- material is stored alongside associated dance with the Declaration of Helsinki. identifiable’ clearly facilitates research: information, such as the type of tumour, However, in their new guidance, the OHRP there will be no costs to obtaining informed medical treatment, donor’s age and so enlarged the definition of non-identifiable consent, no delays and, particularly in the forth, but all information that would allow in the following way: “OHRP considers pri- USA, no need to obtain IRB or research identification of the research participant or vate information or specimens not to be ethics commission approval. Finally, and patient is stripped, either irreversibly individually identifiable when they cannot perhaps most importantly, it is possible to (unlinked anonymized) or reversibly be linked to specific individuals by the maintain the high standard of informed con- (linked anonymized). In the case of linked investigator(s) either directly or indirectly sent at least for identifiable data and tissue. anonymized samples, identification is pos- through coding systems” (OHRP, 2004). sible by a code, to which researchers or This is the case if “the investigators and the … after more than 50 years of other users of the material—as part of the holder of the key enter into an agreement classical health research ethics, definition of the term ‘reversibly/linked prohibiting the release of the key to the regulatory agencies have begun anonymized’—do not have access. Coded investigators under any circumstances, samples have the same characteristics as until the individuals are deceased (note that to question fundamental ethical linked (reversibly) anonymized samples, the [Department of Health and Human milestones the only difference being that researchers Services] regulations do not require the IRB and users have access to the code. Finally, to review and approve this agreement).” It is Conversely, there are a considerable samples are considered to be identified if also the case if “there are IRB-approved number of arguments against an expanded the information that allows identification— written policies and operating procedures definition of ‘non-identifiable’ and in name, address and so on—is associated for a repository or data management center favour of a less elevated standard of con- directly with the tissue, such as when the that prohibit the release of the key to the sent. First, biomedical research involving patient’s nametag is attached to the sample. investigators under any circumstances, biobanks implies risks for identifiable This is, for example, how pathology depart- until the individuals are deceased” or if groups and communities, because the ments usually store clinical samples. “there are other legal requirements pro- anonymity of the individual does not imply In European documents, the term hibiting the release of the key to the investi- the anonymity of groups. Second, if anonymized could mean either unlinked or gators, until the individuals are deceased.” researchers use coded samples without linked anonymized. In most US and English OHRP also specifies that “[t]his guidance having access to the code—to use the Canadian texts, anonymized refers only to applies to existing private information and European terminology, if they use linked unlinked anonymized samples. specimens, as well as to private information (reversibly) anonymized samples—as sug- Interestingly, however, the guidelines from and specimens to be collected in the future gested by the OHRP, this means that a link Quebec have adopted the terminology from for purposes other than the currently pro- exists. Using this link, it is possible to con- the French translation of former versions of posed research.” The advantage of enlarging tact the donors at any time. Those who have the CDBI guidelines (CDBI, 2002) and dis- the definition of non-identifiable is obvious: access to the code might find it difficult not tinguish between reversibly and irreversibly researchers can maintain high standards of to contact the donor if it could prevent anonymized samples. According to the informed consent, but are provided with a future harm—for example, by alerting them European terminology, ‘coded’ always simple means to escape strict regulations by to a hidden medical condition. means that researchers or other users have entering agreements that prohibit them from Third, there is the theoretical possibility access to the code, whereas the OHRP uses access to the code, without having to destroy that the code could be broken for less justifi- the term to refer to what Europeans and the link. Through these simple arrangements, able reasons. As a result, the ethical ques- French Canadians call ‘linked anonymized’ any type of future research is authorized tions are not the same as in the case of samples—a link exists but researchers do without the need for consent or IRB approval. unlinked (irreversibly) anonymized samples,

and simply circumventing present regula- In summary, the arguments seem to be in CDBI (2002) Proposal for an Instrument on the Use tions is not an adequate response to this favour of using general consent together with of Archived Human Biological Materials in Biomedical Research. Strasbourg, France: Council problem. Fourth, approval by an IRB or the right to opt out and IRB approval, and of Europe Steering Committee on Bioethics ethics committee for future research projects against an enlargement of the term ‘non- CDBI (2006) Draft explanatory memorandum to is desirable to ensure efficient use of identifiable’. It is evident that the lack of inter- the draft recommendation on research on resources such as tissue. Fifth, one might national consensus about the regulatory biological materials of human origin. Strasbourg, France: Council of Europe Steering question the sense of a solution whose main framework of biobanks interferes with the Committee on Bioethics goal is to escape existing regulations so that efficiency of research and delays projects CIOMS (2005) Revised version of the 1991 most biobank research can take place with- (Barbour, 2003; Frank, 1999) or even halts CIOMS International Guidelines for Ethical out further surveillance. them (Abbott, 2004; Normile, 2003). Thus, Review of Epidemiological Studies. Geneva, Switzerland: Council for International international collaboration becomes increas- Organizations of Medical Sciences ast, but not least, the preferences of the ingly difficult. To maximize the benefit of COE (2006) Recommendation Rec(2006)4 of the tissue providers must be considered. biobanks and genetic databases for both Committee of Ministers to member states on Several empirical studies have shown research and public health, a single ethical research on biological materials of human L origin. Strasbourg, France: Council of Europe that 90% of patients and research partici- framework is essential, which requires a har- Commission de l’Éthique de la Science et de la pants find general consent adequate monization of the terminology about Technologie (2003) Les Enjeux Éthiques des (Wendler, 2006; Stegmayr & Asplund, anonymity. The arguments above speak in Banques d’Information Génétique: Pour un 2003). For instance, many participants that favour of allowing general consent point Encadrement Démocratique et Responsable. Sainte-Foy, Quebec, Canada: Commission de are eligible for the UK Biobank have towards an enlargement of the term ‘non- l’´Ethique de la Science et de la Technologie endorsed general consent, because they do identifiable’. There is a clear need to discuss Council for Science and Technology (2000) not want to be re-contacted repeatedly (UK and formulate such a framework on an inter- Fundamental Principles of Research on the Biobank, 2003b). Empirical studies also national level. The only way to achieve future Human Genome. Tokyo, Japan: Council for Science and Technology show that a substantial percentage of progress in biobanking is through harmo- Eiseman E (2000) Stored tissue samples: an research participants and patients want to nization of the key terms and key norms. inventory of sources in the United States. In be able to approve the use of their samples, National Bioethics Advisory Commission, even if it is anonymized (Wendler & ACKNOWLEDGEMENTS Research involving Human Biological Materials: Funding was provided by the Swiss National Ethical Issues and Policy Guidance, Vol II pp Emanuel, 2002). Existing and ongoing leg- Science Foundation. This article has benefited D1–D52. Rockville, MD, USA: National islation reflects these preferences. In several from the first author’s work as a member of a Bioethics Advisory Commission European countries, the prevailing opinion research collaboration between the Institute of European Society of Human Genetics (2003) Data is that a form of consent should be required Bioethics at the University of Geneva, the Institute storage and DNA banking for biomedical research: technical, social and ethical issues. for Medical Ethics of the Charité, Berlin, and the to use anonymized samples, at least for Eur J Hum Genet 11 (Suppl 2): S8–S10 genetic testing. Legislation in the Department of Ethics, Trade, Human Rights, and Frank L (1999) Storm brews over gene bank of Law, funded by the Geneva International Netherlands, Switzerland and France, and Estonian population. Science 286: 1262–1263 Academic Network, entitled ‘Human Genetic Grizzle W, Grody WW, Noll WW, Sobel ME, legislation projects in Belgium, contain Databases: towards a global ethical framework’, Stass SA, Trainer T, Travers H, Weedn V, provisions that require informing tissue and in particular from conversations with the Woodruff K (1999) Recommended policies for donors and giving them the possibility to members of this project: Nikola Biller-Andorno, uses of human tissue in research, education, opt out as a prerequisite for any re-use of Andrea Boggio, Alex Capron, Agomoni Ganguli, and quality control. Ad Hoc Committee on and Alexandre Mauron. Stored Tissue, College of American Pathologists. their anonymized samples (Trouet, 2004). Arch Pathol Lab Med 123: 296–300 An example of such a law is the Swiss REFERENCES Human Genetics Commission (2002) Inside Federal Law on the Genetic Testing of Abbott A (2003) With your genes? Take one of Information: Balancing Interests in the Use of Humans. Article 20 of this law, which regu- these, three times a day. Nature 425: 760–762 Personal Genetic Data. London, UK: UK Abbott A (2004) Icelandic database shelved as Department of Health lates the re-use of biological material, states Kaiser J (2002) Population databases boom, from that a sample shall be re-used only for the court judges privacy in peril. Nature 429: 118 Annas GJ, Glantz LH, Roche PA (1995) Drafting Iceland to the U.S. Science 298: 1158–1161 Kaye J, Helgason HH, Nõmper A, Sild T, Wendel L purposes already approved by the donor. the Genetic Privacy Act: science, policy, and (2004) Population genetic databases: a Genetic tests shall only be carried out using practical considerations. J Law Med Ethics 23: comparative analysis of the law in Iceland, 360–366 biological material obtained for a different Sweden, Estonia and the UK. TRAMES 8: 15–33 Australian National Health and Medical Research purpose if the material is anonymized and [Au: Please provide the full name of this journal] Council (1999) Guidelines for Genetic Registers Kegley JA (2004) Challenges to informed consent. the donor (or his/her legal surrogate) has and Associated Genetic Material. Canberra, been informed about his/her rights and has EMBO Rep 5: 832–836 ACT, Australia: Commonwealth of Australia Knoppers BM (2005) Biobanking: international not expressed any opposition to its re-use Azarow KS, Olmstead FL, Hume RF, Myers J, norms. J Law Med Ethics 33: 7–14 (Swiss Confederation, 2004). Calhoun BC, Martin LS (2003) Ethical use of Knoppers BM, Saginur M (2005) The Babel of tissue samples in genetic research. Mil Med genetic data terminology. Nat Biotechnol 23: 168: 437–441 925–927 … one might question the sense Barbour V (2003) UK Biobank: a project in search Korn D (1999) Contribution of the Human Tissue of a solution whose main goal is of a protocol? Lancet 361: 1734–1738 Archive to the advancement of medical CCNE (2003) Avis No 77. Problèmes Éthiques knowledge and the public health. In National to escape existing regulations, so Posés par les Collections de Matériel Biologique Bioethics Advisory Commission, Research that most biobank research can et les Données d’Information Associées: involving Human Biological Materials: Ethical take place without further “Biobanques”, “Biothèques”. Paris, France: Issues and Policy Guidance, Vol II pp E1–E30. Comité Consultatif National d’Ethique Pour les Rockville, MD, USA: National Bioethics surveillance Sciences de la Vie et de la Santé Advisory Commission

Lin Z, Owen AB, Altman RB (2004) Genomic Most people are willing to provide informed research and human subject privacy. Science consent [In Swedish]. Lakartidningen 305: 183 100: 618–620 Loft S, Poulsen HE (1996) Cancer risk and Swiss Confederation (2004) Bundesgesetz über oxidative DNA damage in man. J Mol Med 74: genetische Untersuchungen beim Menschen 297–312 [Federal Law on the Genetic Testing of MRC (2001) Human Tissue and Biological Humans]. Berne, Switzerland: Federal Samples for Use in Research. Operational and Department of Justice and Police Ethical Guidelines. London, UK: Medical Trouet C (2004) New European guidelines for the Research Council use of stored human biological materials in Nationaler Ethikrat (2004) Biobanken für die biomedical research. J Med Ethics 30: 99–103 Forschung. Stellungnahme. Berlin, Germany: UK Biobank (2003a) Setting Standards. The UK Nationaler Ethikrat Biobank Ethics and Governance Framework. Bernice S. Elger is in the Département de NBAC (1999) Research Involving Human Cheshire, UK: UK Biobank Médecine Communautaire, Médecine Légale Biological Materials: Ethical Issues and Policy UK Biobank (2003b) UK Biobank Consultation on (Institut de Bioéthique), Centre Médical Guidance, Vol I. Rockville, MD, USA: National the Ethical and Governance Framework. Universitaire in Geneva, Switzerland. Bioethics Advisory Commission London, UK: People Science & Policy Ltd E-mail: [email protected] Normile D (2003) Japan guidelines under fire after UNESCO (2003) International Declaration on protest halts study. Science 301: 1039 Human Genetic Data. Paris, France: UNESCO OHRP (2004) Guidance on Research Involving Wendler D (2006) One-time general consent for Coded Private Information or Biological research on biological samples. BMJ 332: Specimens. Rockville, MD, USA: Office for 544–547 Human Research Protections Wendler D, Emanuel E (2002) The debate over Pearson H (2004) Summit calls for clear view of research on stored biological samples: what do deposits in all biobanks. Nature 432: 426 sources think? Arch Intern Med 162: Schweizer Akademie der Medizinischen 1457–1462 Wissenschaften (2006) Biobanken: Gewinnung, WMA (2002) The World Medical Association Aufbewahrung und Nutzung von menschlichem Declaration on Ethical Considerations biologischem Material für Ausbildung und Regarding Health Databases. Ferney-le- Forschung. Medizinisch-ethische Richtlinien Voltaire, France: World Medical Association und Empfehlungen. Basel, Switzerland: WMA (2004) The World Medical Association Arthur L. Caplan is Director of the Center for Schweizer Akademie der Medizinischen Declaration of Helsinki: Ethical Principles for Bioethics, University of Pennsylvania Medical Wissenschaften Medical Research Involving Human Subjects. School, Philadelphia, PA, USA Stegmayr B, Asplund K (2003) Genetic research on Ferney-le-Voltaire, France: World Medical E-mail: [email protected] blood samples stored for years in biobanks. Association doi:10.1038/sj.embor.7400740