Democratic governance of genomics: the case of UK Biobank

Abstract

Biobanks are collections of human biological tissue used for genomics research. This promises a better understanding of the gene-based contribution to common disease and development of a more personalized approach to healthcare with safer drugs and more effective treatment. However, biobanks are also controversial owing to the ethical, legal, social and political issues raised about their collection and use of biological samples. Therefore, their democratic governance is not only a normative challenge but also an empirical one. This paper is concerned with both challenges: it attempts to “evaluate” processes of democratic governance of genomics by focusing on the case of UK Biobank. The overall argument is that although the UK Biobank performs well in terms of general democratic governance structures, there are epistemological and practical limits to specific democratic governance processes.

Keywords:

• governance
• genomics
• UK Biobank

Introduction

Genomics is science and technology related to the study of the (human) genome. One of the most common perspectives of the genome is that it is “a body of information.” According to Parry and Dupré (2010, p. 4) “This might be as a series of nucleotides, the As, Ls, Gs, and Ts familiar in representations of genomics sequence, or as a series of markers, present or absent in a particular genome.” One innovative way of studying the genome in large populations is through the so-called biobanks. Biobanks are collections of human biological tissue that are used for genomics research. The latter promises a better understanding of the gene-based contribution to common disease and development of a more personalized approach to healthcare with safer drugs and more effective treatment (Hawkins and O'Doherty 2010).

However, as is well known, biobanks are also controversial owing to the ethical, legal, social and political issues raised about their collection and use of biological samples. Therefore, their democratic governance is not only a normative challenge but also an empirical one. This paper is concerned with both challenges: it attempts to “evaluate” processes of democratic governance of genomics by focusing on the case of UK Biobank. The latter is not just a resource for genomics research, aiming to “… improve the prevention, diagnosis and treatment of illness and the promotion of health throughout society” (UK Biobank 2003, p. 3), but also a voluntary association of people, aiming to promote the public good of population health.

This paper first of all draws on contemporary political theories of democracy in order to develop a set of broad indicators of assessment. Secondly, it applies these indicators to the specific case of UK Biobank, identifying “gaps” but also “good practices” in processes of democratic governance. The overall argument is that although the UK Biobank performs well in terms of formal democratic governance structures, there are practical limits to substantive democratic governance processes. The latter include representation and participation practices of decision-making.

The paper is structured as follows: section 2 explains the importance of democratic governance of large-scale population biobanks; section 3 presents the framework of conceptual understanding; section 4 describes the methodology of research, defining the broad qualitative indicators to be applied throughout the case of UK Biobank; section 5 summarizes this case; section 6 discusses democratic governance of the UK Biobank and its practical limitations. Section 7 concludes that improvement of democratic governance of the UK Biobank has to take on board its limitations.

Why democratic governance?

The successful decoding of the human genome at the beginning of the twenty-first century and subsequent advances in the new scientific-technological field of genomics revolutionized biomedical research and paved the way to a “post-genomic” era. In this context, many traditional small-refrigerator based biobanks have been transformed into sophisticated storage centers, containing DNA samples1 from large-scale populations. According to Tutton and Corrigan (2004, p. 1):

These databases are distinctive by the way that they combine genetic information derived from blood samples with various kinds of personal data about medical history, lifestyle or genealogy from thousands or hundred of thousands of ordinary people across national or regional territories.

Given the high sensitivity of DNA samples combined with health records, critiques of biobanks raise a number of legal, ethical and socio-political issues. Specifically, Rothstein (2005, p. 90) argues that “… the nature and degree of risk of biobank research depends on (1) the identifiability of the sample and any linked health information and (2) whether the samples are extant or to be collected prospectively.” Identifiable samples raise the risk of invasion of the donor's privacy while unidentifiable samples raise the risk of research that is not approved by the donor. Rothstein (ibid.) stresses that “Even where personal identifiers have been removed, labels on biological samples and linked medical records may contain demographic data such as gender, race and ethnicity.” This raises social risks because research findings can indicate which members of subpopulations are more likely to have some predisposition to a disease.

Fears of genetic discrimination in health and insurance are increased by the fact that several commercial interests are involved in biobanks. Salter and Jones (2005, p. 711) take one recent example, the Swedish biobank company UmanGenomics. According to them, the collapse of this biobank “… despite its much vaunted ethical foundations was due to a failure to work through both requirements of the scientists involved and the intellectual property requirements (IPRs) of a successful market venture in this field … .” Salter and Jones (ibid.) stress that “Where industry is content with a regulatory framework that facilitates its economic interests, civil society may feel that certain citizenship rights have been compromised in the interests of commercialisation.” There is no easy way to abolish this contradiction unless economic and social interests are reconciled through regulatory initiatives which go beyond benefit sharing to include protection of basic individual rights and liberties. These rights and liberties are primarily ethical demands for freedoms of “special importance” to all individuals.

The problem is even more serious in cases of biobankruptcy. For instance, Janger (2005, p. 80) invites us to consider a scenario in which a patient goes to his doctor for a cancer treatment.

… Doctor asks Patient if he will agree to provide a blood sample for a biobank (“Biobank”) that his hospital is building to help study genetic link factors in cancer … Patient agrees, authorising Biobank to use his blood sample and related information “for research devoted to identifying the role of genetics in cancer to develop new therapies.” For two years, the project runs along as expected … then Doctor's hospital finds itself in financial difficulty. A private pharmaceutical company, called Genscreen, offers the hospital … a significant amount of money for access to the information (excluding names) about those samples in Doctor's database. Instead of using the samples to identify cancer risks and develop treatment Genscreen intends to develop a cheap screening device which will identify genetic risk of cancer and allow insurance companies to exclude people with high risk from coverage. This is not the purpose for which Donor donated his blood nor is it a use for which he gave consent, nor would he give consent if asked.

Janger (ibid.) argues that in such a case, complex legal issues of property, liability and public enforcement have to be addressed. However, even if the legal procedure is successful in protecting the patient's rights, the question of public trust in biobanks in particular and the problem of legitimacy of genomics in general remain widely open. These technological innovations, as Salter and Jones (2005, p. 716) correctly point out, “… combine scientific and industrial promise with political danger.” That is the danger of failure to protect basic individual rights such as privacy in liberal societies of accelerated life sciences innovation.

In justification of large-scale population biobanks, advocates put forward the liberal argument of “public good” and “public interest.” As John Newton, the United Kingdom (UK) Biobank's chief executive, expressed it in 2003, “The [biobank] project is in the interest of public health and … will be relying on the public's good will to make it a success” (Wellcome Trust et al. 2003, p. 2). The conception of population health as a public good that includes life science innovations such as biobanks is based on the assumption that everyone, explicitly or implicitly, is interested in one rational good: combining genetic and health information for the whole of population that has the potential of “… developing scientific understanding of and therapeutic treatments for common multifactorial diseases, as well as, determining why groups of people have differing reactions to pharmaceutical drugs” (Tutton and Corrigan 2004, p. 1). This notion of public good points towards some positive scenarios. First, research results based on combined genetic and health information might play a crucial role in public health maintenance organizations, creating conditions for better planning and intervention in health of specific groups and people. Secondly, by using this type of information, state-sponsored systems of healthcare and insurance can anticipate costs, providing better and sustainable services to citizens.

However, even if, in theory, everyone might be interested in the public good of health, in practice, as has been already implied, individual members of civil society might still find biobanks much less attractive than other resources of life sciences research. The reason for this is that basic individual rights and freedoms are also in the public interest. Therefore, civil society formally has to assume the balanced protection of more than one rational good. This liberal assumption entails that no rational good is prioritized against another. Thus, government formally remains neutral towards different conceptions of the good. Neutrality though does not seem to be possible in substantive terms. The reason for this is that government often fails to balance the rational goods of public health and individual rights. The latter are also considered to be crucial in the case of large population biobanks because, as Salter and Jones (2005, p. 715) stress,

Databases of personal genetic information constitute a unique form of information about private individuals, which in turn poses unique concerns about its handling compared to typical medical records. Although conventional medical records may contain similar information about an individual's health to a genetic record, and both can be accessed without their presence, the information contained in genetic records is unique because of its certainty in revealing two things: the genetic health vulnerabilities of the individual, and their biological relationships to others.

Although this is by no means an undisputed proposition, it clearly illustrates the formal conflict between public health and individual rights. Thus, Capps et al. (2008, p. 26), referring to the UK Biobank, suggest that:

The possible conflict may … be seen in two senses: one despite UKB's commitment to benefiting the public good – and even in cases of extraordinary public danger – it has a prior role in protecting participants; and two, it may be undesirable to subjugate the public benefit to the extent that participants control the resource.

The challenge of achieving a formal balance between protecting each individual participant's right2 to privacy and pursuing the public good of health innovation raises the question of democratic governance of large-scale population biobanks. Specifically, the notion of democracy refers to “rule by the people” while the concept of governance refers to guiding or steering society and resources (Pierre and Peters 2000, Jordan et al. 2005). More recently “governance” has become conceptually autonomous and analytically distinct from “government,” marking a shift from the formal institutions of the political state to a new style of governing through informal participatory networks and non-governmental actors (Kooiman 1993, Rhodes 1996, Jessop 1998, Stoker 1998, Bache 2003, Jordan et al. 2005, Lyall and Tait 2005). Therefore, democratic governance of large-scale population biobanks implies democratic processes of politically sensitive associations that are not necessarily government-driven (Dahl 1989, p. 107). How can these processes be assessed? What are the criteria of assessment? What are the “gaps” and “good practices”?

In what follows we attempt to provide answers by doing two things: first, developing broad qualitative indicators of democratic governance of biobanks; secondly, applying these indicators to assess the specific case of UK Biobank. The overall exercise focuses on democratic governance. This is the area of decision-making and policy implementation for which the UK Biobank is held to account. The performance of the biobank in this area can be systematically assessed in terms of broad indicators of democratic governance.

A conceptual framework of democratic governance

In a recent article, Blaug (2002, p. 104) confesses that “Trying to understand democracy is like reaching into a black plastic bag. You can feel a large object but accurate description is difficult because the shape is complex.” Certainly, there are many theories of democracy. As Dryzek (2002, p. 28) points out “Democracy is not a static concept, whose essence could ever be decided once and for all. Rather, it is a dynamic and open-ended project.” Indeed, democracy has been interpreted in different ways in different times and places (Saward 2003, p. 161). However, in Western liberal societies, the nature of democracy is conceived in relation to the pursuing of public good and public interest (Mill 1951, Bentham 1960, Macpherson 1973). According to Capps et al. (2008, p. 14) the concept of public good refers to “generic features of everyone's well-being and freedom.” In this sense, public goods are necessary goods for the reproduction of the whole of society. By contrast the concept of public interest refers to procedures, including democratic procedures. In this sense, public interests are concerned with everyone's interests in the provision of the necessary frameworks and institutions within which certain public goods can be pursued (ibid.). Capps et al. (ibid., p. 14) insist that “it is … disingenuous to entirely separate interests from goods (although particular natures should not be conflated).” Although they are right in maintaining the separation between the public good and the public interest, they overlook that this separation is purely analytical. Moreover, they seem to fail to stress that the public good presupposes the public interest and vice versa. Indeed, one initial political definition of the public good is that members of an association have a collective interest in what is best for all and put it before their individual conceptions of the good. This definition is transcendental and derives from the ancient Greek perspective of democracy (Dahl 1989, p. 16). In modern liberal societies at least two more definitions of the public good can be found: first, an instrumental definition that accepts the existence of the public good but prioritizes the individual's humanity and the protection of her negative right to non-interference; secondly, an individualistic definition that accepts the public good as nothing more than the sum of different individual conceptions of the good or interests of an association's or polity's members (Elcock 2006, p. 104).

Throughout this paper, the transcendental definition of public good is adopted. Accordingly, democratic governance is viewed as a process of pursuing the public good of an association based on one central belief:

… no single member, and no minority of members, is so definitely better qualified to rule that the one or the few should be permitted to rule over the entire association … on the contrary … all the members of the association are adequately qualified to participate on an equal footing with others in the process of governing the association. (Dahl 1989, p. 31)

This is a belief in (political) equality and is shared across different models of democracy, including deliberative, direct, participatory, cosmopolitan and representative models. For instance, as Dryzek (2002, p. 1) argues “… increasingly, democratic legitimacy came to be seen in terms of the ability or opportunity to participate in effective deliberation on the part of those subject to collective decisions.” Participation is an essential criterion for a democratic process in general and (political) equality in particular. According to Dahl (1989, p. 109) “Throughout the process of making decisions citizens ought to have an adequate opportunity, and an equal opportunity, for expressing their preferences as to the final outcome.” Other criteria for a democratic process include: “voting equality” (i.e. members of the association must be ensured opportunities to express their equally weighted choices); “enlightened understanding” (i.e. members of the association must have equal opportunities for assessing information and validating the choice on the matter to be decided); “control of the agenda” (i.e. the association as a whole must have an exclusive opportunity to decide how matters are to be placed on the democratic agenda) (ibid., pp. 109–113).

Although Dahl's criteria for a democratic process are crucial for assessing an association, they are biased towards specific models of democratic governance such as republicanism (i.e. a constitutional model) and representation (i.e. a model of representative government). Therefore, they ignore other democratic models of governance like, for instance, those of deliberation (i.e. a discursive or communicative model), direct participation (i.e. a model of direct engagement with matters of decision-making) and cosmopolitanism (i.e. a model of cross-border or transcendental democratic governance). These particular models of governance appear to be more relevant in the assessment of large-scale population biobanks (Hawkins and O'Doherty 2010) than republicanism and representation. Specifically, deliberation requires citizens or members of an association to go beyond their private self-interests (Elster 1997), engaging in consensus-oriented discourse (Habermas 1996). The latter is the goal of deliberative decision-making. Consensus or indeed compromise can be achieved through discursive processes such as public discussion, involving free and equal citizens. According to Bohman and Rehg (1997, p. xviii) these are conditions necessary for democratic deliberation and they “… must be shown to apply … to current social conditions, including increasing cultural pluralism and complexity.” On the other hand, direct participation implies full democratic engagement of free and equal citizens in social institutions such as voluntary associations. This presupposes the abolishment of divisions based on social class, gender and race. Finally, cosmopolitanism goes beyond the conception of “sovereignty” of the modern liberal state, taking on board the impact of globalization and the growth of transnational networks on democratic participation in associations or polities. A decision made in one association within state borders might have advantageous or disadvantageous consequences for people throughout the world. For instance, a decision made in one biobank to investigate interactions between genes, environment and lifestyle of a particular race or ethnic group, say Blacks or African Americans (Kahn 2006), might have some advantageous but also disadvantageous consequences for Blacks or African Americans throughout the world.

To overcome the difficulty of orienting oneself among different models of democracy, Saward (2003) proposes a new approach to democratic theory, drawing on some provocative aspects of deliberation, direct participation and cosmopolitanism. Specifically, he offers a meta-theory of democracy that he calls “reflexive proceduralism.” According to him “reflexive proceduralism” is not another model of democracy but rather

… a framework or approach which recognises that multiple devices and principles can and will constitute democratic practice, and that we need systematically to stand back from existing models precisely in order to manipulate and combine their elements in democratically promising, tailored ways. (ibid., p. 167)

Indeed, Saward proposes four basic principles of democracy: political equality; inclusion; expressive freedom; and transparency. According to him, equality means that

… no system can be minimally democratic without institutionalising equality between citizens in some substantial way. Inclusion follows from political equality, but is entered here on its own because it is difficult to see how anything other than an inclusive, involving form of institutionalising political equality can be acceptable democratically. Freedom is included because the right to make one's choices in an uncoerced manner does not automatically follow from the previous two principles, and yet arguably is critical to a self-respecting democracy. And transparency because equality, inclusion and freedom, if valuable, must be seen to be done. (ibid., p. 162)

Saward argues that there are boundaries to the reasonable interpretation of all four democratic principles. He also stresses that “Democratic principles come alive (are ‘lived’) through the medium of decisional mechanisms or devices which are designed to activate them and which come to be justified in terms of them” (ibid., p. 166). Indeed, he provides a (rather incomplete) list of such devices, including “public hearings, debates and inquiries,” elections, majority rule, “systems of representation,” the referendum, “protected public spaces of civil freedom” and others (ibid., p. 167). In Saward's meta-theoretical approach to democracy, these devices and their sequences enact particular democratic principles. Also, there are variations and a number of sub-devices.

For the purposes of this research paper, “reflexive proceduralism” can be used as a general conceptual framework of democratic governance. This implies that basic principles of democracy can be methodologically transformed into broad qualitative indicators of democratic governance of large-scale population biobanks such as the UK Biobank. Other useful approaches might be also added so that devices of democracy such as “public hearings, debates and inquiries” can be methodologically developed as clusters of direct “measures” for these indicators, identifying “gaps” and “good practices.” The next section explains the main steps of this methodology.

Research methodology

Research for this paper was carried out in the UK between March 2009 and February 2010. The focus was on the UK Biobank. The latter is a voluntary association that operates within wider social and political institutions (e.g. government regulations). The core biobank processes under investigation were the processes of governance: structures and practices of decision-making and policy implementation for which the UK Biobank is held to account.

Broad indicators

Governance structures and practices have been studied and assessed in terms of four broad qualitative indicators of democracy: equality, inclusion, freedom and transparency. These broad indicators can be defined as follows:

•	Indicator 1: equality – institutionalized equality between citizens, directly or indirectly involved in the UK Biobank.
•	Indicator 2: inclusion – institutionalized inclusion of citizens in the Biobank's governance processes.
•	Indicator 3: freedom – every citizen, directly or indirectly involved in the UK Biobank, can make choices in an uncoerced manner.
•	Indicator 4: transparency – citizens, directly or indirectly involved in the Biobank, have access to information about governance.

Data collection

Data for each broad indicator have been provided through a case study of the UK Biobank. As has been explained elsewhere, “A case study consists of practices, attitudes and factors which take place in an organisation's history at a particular time” (Papaioannou 2002, p. 53). Two methods of data gathering have been used: first documentary analysis that includes academic journal articles, consultation documents, policy papers, reports, websites, brochures and press articles; secondly in-depth interviews (based on semi-structured questionnaires) with selected stakeholders involved in the UK Biobank governance structures and practices as well as members of the public (key informants), including one donor of a biological sample and one representative of a civil society group.

During empirical research, recent relevant documents were collected and seven face-to-face interviews were conducted with key informants. The interviews lasted 21 to 75 minutes, with a mean duration of 48 minutes. Textual representations of qualitative data have been analyzed and interpreted in such a way that provides an in-depth understanding of democratic governance of the UK Biobank. The most relevant text concerning public consultation and democratic deliberation was extracted and included in the STP case of the UK Biobank.

The general aim of our interviews was to map the structures and practices of democratic governance of the Biobank and identify the factors which influence decision-making for the sake of the public good of population health. The core set of questions for semi-structured questionnaire was as follows.

Indicator 1: equality

What does the UK Biobank do in order to achieve a certain level of equality between all citizens involved? For instance, which structures are involved in the decision-making process? Is there involvement of participants in that process? Are there consultations, public hearings and debates? Are there workshops, focus groups, etc.?

Indicator 2: inclusion

What does the UK Biobank do in order for its governance processes to be inclusive? For instance, does it support participants to be in the governance process from the onset of the project (design) and throughout its life? Does it encourage citizens' involvement in consultations, public hearings and debates? Does it include all different stakeholder interests and values (science, industry, civil society groups, etc.)?

Indicator 3: freedom

What does the UK Biobank do in order to ensure that the citizens involved make uncoerced choices? For instance, is there a process through which contemporary criteria of informed consent are applied? Is there a practice that promotes the rights of individuals (anonymity, confidentiality, etc.) and their autonomy? Can citizens withdraw from the process whenever they choose to do so?

Indicator 4: transparency

What does the UK Biobank do in order to promote transparency in governance? For instance, are there processes of publication of governance decisions and related information? Are there governance processes which ensure public scrutiny of management decisions? Are there processes of accountability?

The area of focus and the set of broad indicators are summarized in Table 1.

Table 1. Area of focus and broad indicators.

Areas of focus	Indicator 1: equality	Indicator 2: inclusion	Indicator 3: freedom	Indicator 4: transparency
Governance	Governance structures; participation in decision-making; representation of disciplines; consultations, public hearings, etc.	Support for participation; citizens' involvement; inclusion of different interests.	Informed consent; safeguard of individual rights; freedom to withdraw from the process.	Publication of management related information; public scrutiny processes; accountability.

Data validation

The issue of validity is crucial for our research findings. According to Seale and Filmer (1998, p. 134) “At its broadest, validity concerns the degree to which the findings of a research study are true.”In this research, the “truth” of our findings has been established by combining two methods: face validity and criterion validity.

Face validity is concerned with the process through which the researcher assesses the quality and relevance of his/her methods of data gathering. For the purpose of this study, the method of face validity has been applied to both documentary analysis3 and semi-structured interviews.4

Criterion validity refers to the process through which the researcher compares his/her gathered data with research findings of other empirical studies on the same topic. The method of criterion validity has been used in documentary analysis5 and semi-structured interviews.6

Data analysis

In order to process the qualitative data gathered during the interviews we focused on the core categories of democratic governance, i.e. equality, inclusion, freedom, and transparency. Various coding techniques have been employed to (1) label conceptual categories and properties and unveil logical connections; (2) identify, categorize and describe phenomena found in text; (3) distinguish between core and non-core categories and phenomena which help to synthesize ideas and provide answers to research questions.

The case of UK Biobank

The UK Biobank has been primarily established as a life sciences research resource, containing lifestyle, physical and genetic information as well as biological samples from people in the UK, aged between 40 and 69. Recruitment began in April 2007 and, at the time of writing, the goal of collecting data from 500,000 research participants (or donors) has been already reached. According to Laurie et al. (2009, p. 60):

Neither the objective nor the approach of UK Biobank are novel from an epidemiological perspective. The project design, in scientific terms, has been considered and approved by an international scientific review body; and in ethical terms, it has received approval from a multi-centre research ethics committee.

Indeed, what is novel about the UK Biobank is the scale of recruitment (almost 1% of the UK population) and the uncertainty surrounding the use of this resource for the public good of population health.

Although the UK Biobank is not a government initiative (but a resource commonly funded by the British Department of Health, the Medical Research Council and the Wellcome Trust) it is subject to government regulations, including the Data Protection Act 1998 that focuses on the processing of personal data. However, these government regulations constitute an external framework that does not answer the question of democratic governance in terms of internal practices of equality, inclusion, freedom and transparency.

Equality

The extent to which the UK Biobank provides citizens with equal opportunities to influence decisions depends on four internal governance structures:

1.	A Board of Directors (with representatives of the Funders).
2.	A Steering Committee (led by a Chief Executive Officer).
3.	An International Scientific Advisory Board.
4.	An Ethics and Governance Council (EGC).

These formal structures have been established throughout a series of public consultations that tried to equally involve a number of groups, including stakeholders, health professionals and industry. For example, the first consultation on Public perceptions of the collection of human biological samples (2000) involved 16 focus groups with members of the public, including minority ethnic groups, people of wide-ranging ages, socio-economic groups and regions (Wellcome Trust and Medical Research Council 2000). The second to eighth subsequent consultations equally involved more specialized groups, including healthcare professionals and potential participants aged 45–69 and industry. This demonstrates that the UK Biobank has been set up as a project that can improve science–public relations and restore trust (Haddow et al. 2008). Indeed, as Levitt (2005, p. 78) says: “… UK Biobank might be seen as a model for public involvement having commissioned eight consultations and provided open access to the findings. The public were consulted through surveys, focus groups and a peoples' panel before the plans were finalised.” Although these consultations cannot be seen as statistically representative of the views of the wider public, they provided broad legitimacy to the UK Biobank structures. Consultations offered evidence of support for formal governance, given concerns about misuse of samples and access by employers and/or insurers (Haddow et al. 2008).

However, as Hunter and Laurie (2009, p. 5) recognize “ … the consultations may be criticized for having adopted an expectation of a ‘passive public’ rather than one which would be more involved in the UK Biobank's governance and decision-making.” Indeed, there is no substantial involvement of the public in the UK Biobank governance through public hearings, debates and inquiries. For this reason Winickoff (2007) and Tutton et al. (2004) urge more public engagement with the Biobank's decision-making and governance processes. Our research reveals that there may be two reasons why the UK Biobank has so far been cautious about opening up its governance processes to the wider public. The first is the practical risk of competition between different interest groups. As one member of the Steering Committee confessed:

In principle [more engagement] is a very good idea [but] in practice the way … it works is you get interest groups involved. (Extract from interview 1)

Competition between interest groups often results in conflicts of values and interests, limiting the prospect of effective democratic governance. This is a particular concern for the UK Biobank: although in principle all groups have vested interests in shaping the democratic debate, in practice only some groups succeed in doing so through specific strategies of persuasion and/or vocal domination. This concern cannot be addressed unless legitimacy is established through equal reflection of values and interests. This does not presuppose involvement of interest groups but assumes a governance structure (i.e. EGC) that can equally reflect values and interests of different stakeholders. The role of the Ethics and Governance Framework (EGF) is to prevent the formation of “vocal minorities” which can “hijack” the debate about the UK Biobank.

The second reason against more open governance of the UK Biobank seems to be the conviction that the public good is better promoted through the application of the EGF. According to one member of the EGC:

The participants' interests are fairly well served if we do our job correctly and uphold trust that functions with the broad consent model. (Extract from interview 2)

The EGF is a living document (i.e. it can be revised over time as the project develops and new challenges arise) that formally “… outlines the commitments to participants, researchers and society at large, as well as, the practices to be adopted in the establishment and management of the UK Biobank resource” (Laurie et al. 2009, p. 63).

The EGF is guaranteed by the EGC, which is a structure that emerged out of the third consultation on the UK Biobank. In that consultation:

Participants generally recommended that some form of oversight body should be established and that the body should be able of acting independently of the users and sponsors … Participants said that this oversight body should agree the rules for access, use, royalty payments, guarantee destruction of sample of those who wish to withdraw, and set sanctions for those who abuse their right of access. (Medical Research Council and Wellcome Trust 2002, p. 4)

The EGC began its work in 2004. Since then its role has been to act as an impartial guardian of the EGF, advise on its revision, monitor the conformity of the UK Biobank with the EGF and advise on the interests of research participants and the general public (Laurie et al. 2009, p. 64). Both EGC and EGF constitute what Laurie et al. (2009, p. 51) call the “ethics+” approach to the UK Biobank. According to them, “ethics+” is the approach of “reflexive governance” that is

… democratic and systematic in gathering evidence about the range of values and interests at stake, in exploring where (and why) there are commonalities and differences and in developing policies or taking decisions which are ultimately acceptable to a body of “reasonable” stakeholders.

Despite the good practice of developing the “ethics+” approach for equal reflection of values and interests in the UK Biobank structures, this approach remains formal. That is to say, there is a need for more equal involvement with the UK Biobank at a substantive level of participatory practices. The equality gap can be only closed through encouragement of a more “active public.”

Inclusion

As has been already pointed out, this indicator follows from the indicator of equality. That is to say, unequal involvement with the UK Biobank at the substantive level of participatory practices leads to exclusion of certain groups of citizens and stakeholders. Our research shows that the most important stakeholder group excluded from governance is the research participants. As one member of the EGF confirmed:

… participants are not represented in the UK Biobank … (Extract from interview 3)

Exclusion of participants from democratic governance implies a bad practice that can lead to weak legitimacy. The importance of legitimacy cannot be underplayed since it is required in order to secure funding and participation of donors (Winickoff and Neuman 2005). The problem of weak legitimacy seems to generate conflicting views within the UK Biobank. For instance, one member of the Steering Committee said:

I think people should be represented, I really do. (Extract from interview 4)

However, one member of the Funders had a different view:

I think it would be quite difficult if you like a representative section of participants. (Extract from interview 5)

Finally, one member of the Board of Directors stressed:

… the EGC do hold regular public meetings and I guess some people who attend those are actually participants, donors but there is no formal way in which donors have routes into the advisory structure as it is at the moment. (Extract from interview 6)

The problem of exclusion of participants from the governance of the UK Biobank has been recognized by a number of researchers (Tutton et al. 2004, Levitt 2005, Wallace 2005, Winickoff 2007, Laurie et al. 2009). Among the solutions they have proposed, two are worthy of detailed discussion. The first is the so-called “shareholder model” of Winickoff (2007) and the second is the “stakeholder model” of Hunter and Laurie (2009). To begin with Winickoff's model, it is based on the notion of “partnership” that describes cooperative human relations with regard to shared conditions and aims (Winickoff 2007, p. 443). This form of relations is sharply distinguished from the notion of “ownership” that suggests full and undivided private property entitlements. According to Winickoff (ibid.), “In both the American and British legal systems, partnership involves joint control of assets and an equity interest in risks and benefits.” On this basis he proposes a corporate shareholder model of governance that requires a committee of direct representatives of research participants or donors. As Winickoff (ibid., p. 449) says, “Just as shareholders in a company are bound together by a common wish to make money, all donors will likely share the desire to maximise health goods for the public.” Therefore, his proposal includes the structure of a Donor Association with members who would elect leader(s) to represent them on the UK Biobank Board and in the EGC. Winickoff insists that his shareholder model can bridge the “agency gap” in the UK Biobank despite the practical limitations it faces because of large numbers and heterogeneous participants.

These limitations have been exposed by Hunter and Laurie (2009) who argue that Winickoff's shareholder model cannot be successfully applied in the realm of biobanks. In their paper, they provide three reasons against this model:

1.	The UK Biobank is a charity that has neither share capital nor shareholders.
2.	The Donor Association cannot be representative of the values and interests of 500,000 participants.
3.	There is a danger of the Donor Association being hijacked by vocal minorities.

Hunter and Laurie's reasons against Winickoff's shareholder model are realistic and therefore powerful ones. There are practical limitations to inclusive governance of the UK Biobank. Taking on board these limitations, Hunter and Laurie (2009) propose an alternative model of governance, the so-called “stakeholder model.” This model is designed to observe and respect the practical limits to democratic governance of the UK Biobank. Specifically, Hunter and Laurie argue that their stakeholder model is relevant because of its social orientation, its emphasis on inclusion and accountability. According to Hunter and Laurie, the basic premise of their model is that associations such as the UK Biobank should be governed for the benefit of and be accountable to stakeholders. The question is who are the stakeholders of the UK Biobank and what does it mean to involve them in governance? Hunter and Laurie (2009) identify the stakeholders of the UK Biobank as follows: research participants or donors; UK Biobank's Board of Directors; the EGC; funders and members of the charity; researchers; communities; the wider public or civil society; and future generations. The direct involvement of all these stakeholders in governance of the UK Biobank is rejected by Hunter and Laurie on the grounds that diversity and fluidity among stakeholders' communities render attempts at statistical representation disingenuous and impractical. Instead, they propose a strategy of engagement with stakeholders based on the idea of “partnership” as a real dynamic and deliberative process. This idea requires the UK Biobank to put in place mechanisms for ongoing dialogue on governance structures with multiple stakeholders. The latter's values and interests should be reflected in governance. In this sense Hunter and Laurie's notion of “partnership” is distinct from that Winickoff. The former is top down while the latter is bottom up.

Unfortunately, neither Hunter and Laurie's stakeholder model nor Winickoff's shareholder approach has been implemented so far in the substantive governance of the UK Biobank. As one member of the EGC said:

… we do not have a donor group structure … in the way that the stakeholder model would suggest, or the shareholder model would suggest. (Extract from interview 7)

This means that none of the above discussed models has been practically evaluated so far in terms of improving inclusion. As it stands, the UK Biobank needs to do more in order to close the inclusion gap, encouraging research participants to have better involvement in the project's decision-making process.

Freedom

This indicator does not follow from the indicator of inclusion but is crucial for the democratic governance of the UK Biobank. It might be argued that a good practice for ensuring freedom of all citizens involved in this project is the development of the EGF. As has been already said, the EGF is a living document. In the first section it clearly states that “Participation in the UK Biobank is voluntary.” From this it follows that all research participants have individual rights to “consent” and “withdraw.” These rights are neither absolute nor inviolable. Within the UK Biobank, consent is defined as a “general” or “open” consent (Fortun 2003). The reason for this is the inability of the Biobank to inform participants about future uses of their biological samples and lifestyle information. Even so, further consent is also proposed “… for research outside the assays covered by the original consent” (Tutton et al. 2004, p. 284). This has sparked some criticism by researchers such as Tutton et al. (ibid.) who argued that the UK Biobank sends out a “confusing mix of signals.” Tutton et al.

… advocate that the UK Biobank provide participants with information about all of the research as and when it becomes known. After receiving this information, participants should be given the opportunity to opt out of certain research or to withdraw from the project completely. (ibid.)

This particular view of consent has been challenged by Widdows (2009, pp. 180–181) who argues that the UK Biobank has adopted a broad (non-individualistic) consent because “… in population projects of the size of UK Biobank to insist on the ‘gold standard’ of fully informed individual consent would be impractical and unrealistic.” The current governance process of the UK Biobank provides participants with three options of withdrawal:

•	“No further contact”: UK Biobank would no longer contact the participant directly but would still have their permission to use information and samples provided previously and to obtain further information from their health-relevant records.
•	“No further access”: UK Biobank would no longer contact the participant directly, but would still have their permission to use the information and samples provided previously.
•	“No further use”: In addition to no longer contacting the participant or obtaining further information about them, any information or samples collected previously would no longer be available to researchers (Ethics and Governance Council 2007, p. 9).

In fact, the third of these options is impossible to implement in practice. As one member of the EGC stressed:

… it became clear that the nature of the IT systems within the UK Biobank meant that you could not destroy all the information the UK Biobank had on somebody. (Extract from interview 8)

This is particularly serious when one takes into account the issue of control of data use. As the UK Biobank's Chief Executive acknowledged, once data had been licensed out for one specific research project, it could be difficult to control the further uses of the data (Godfrey 2003). Given that complete anonymization of data is practically impossible, contradictions and problems of confidentiality and privacy have to be resolved not only through democratic governance of the UK Biobank but also through government. Thus, for instance, the UK Data Protection Act 1998 clearly provides that data must be fairly processed in accordance with the data subject's rights, used for limited purposes and not kept longer than necessary. Exemptions from these government regulations exist for safeguarding national security, for preventing crime and for prosecuting an offender (Elger 2010). A number of civil society groups, including Gene Watch and Liberty recommended steps for providing adequate protection against unauthorized disclosure and use of genetic information by either government or family (ibid.). The relationship between governance and government is safeguarding freedom of the UK Biobank participants. This implies that democracy at such micro-level cannot happen unless politics is providing an adequate regulatory framework at macro-level.

Transparency

This broad indicator is concerned with the availability of information about the activities of the UK Biobank so that the decision-making process becomes accountable to the public. It might be said that in order to promote transparency, the UK Biobank has developed formal governance structures. According to one member of the Steering Committee these structures include:

… a very formal structured management account system, budgeting, forecasting, and reporting of progress against targets … (Extract from interview 9)

Transparency of governance structures is crucial for the Biobank co-ordination and decision-making. However, this is just one type of transparency. There is another type that has to do with collecting biological data of participants. In order to meet the requirements of this latter type of transparency, the UK Biobank provides information through leaflets and a letter addressed to participants. Our research indicates that participants find this practice satisfactory. As one participant reflected:

[the information] was not in detail but I felt it gave a broad enough overview for people to make a choice. (Extract from interview 10)

The problem is more about elaborating the information and enabling participants to ask questions during the collection of their biological samples. According to the same participant:

They just bring you in [the local UK Biobank assessment center] and try to get you in and out of there like a factory … although in the beginning … they asked, do you have any questions … then they said there would be a chance to ask questions later. So I thought there would be a session … with a group of us … But basically what was referred to was the pamphlet and that was it, there was no any real engagement after that. (Extract from interview 11)

Failure to elaborate broad information about the collection of biological samples damages trust between participants and the UK Biobank. The same participant said:

So the process of actually making the donation I found much more troubling than the idea of going into it. I expected to be asked would you like to become involved in a group of other donors say or would you like to … [have] some kind of public engagement. (Extract from interview 12)

Certainly, we do not have data to claim that the majority of participants felt the same. What we can say is that there is an indication of the need for improvement of transparency and trust relations between the UK Biobank and its participants.

Probably, the most important type of transparency is the one concerned with the use of data for research. Given that the UK Biobank is a resource for the public good of health, no one in principle is excluded from its use. Thus, according to the EGF formal process, decisions about specific projects for the prevention, diagnosis and treatment of illness will be based on scientific and ethical evaluation of research proposals:

… all proposals will be reviewed by UK Biobank to ensure they are consistent with the participants' consent and this Framework, and that they have relevant ethics approval. All users, whether employed by universities, government, charities or commercial companies, will be held to the same scientific and ethical standards. (Ethics and Governance Council 2007, p. 12)

Although this formal transparency process is crucial for the UK Biobank's accountability to citizens, it cannot exclude two possibilities:

1.

A company or charity might obtain data for one purpose and use it for another. This is not necessarily because of hidden research agendas. Rather, as one member of the UK Biobank Steering Committee stressed: You cannot always tell what's going to spin out of a research program; there are many, many examples where something unanticipated came out of a research program, it was patented and was itself very useful. (Extract from interview 13)

2.

A company or charity may be better positioned than others to get access to data. As one member of the Board of Directors recognized: Some people who have been associated with the Biobank the whole way through will be able to put together proposals for the use of Biobank much more effectively than other people because they will know precisely what is in Biobank and they will know what its data quality is and all the rest. (Extract from interview 14) Both possibilities confirm that there are practical limits to transparent governance of the UK Biobank. These limits can even be epistemological, given the unanticipated outcome of scientific research.

Discussion: democratic governance and its limits

What does the case of the UK Biobank tell us about democratic governance of genomics? The answer is not straightforward. On the one hand, formal democratic governance structures such as the EGF and the EGC provide a general framework of equality, inclusion, freedom and transparency. On the other hand, substantive democratic governance processes such as representation and active involvement of research participants are limited by specific practical and epistemological issues, including the difficulty of accommodating all different interests and values, and the lack of knowledge about future research outcomes. Addressing these issues through the implementation of realistic models of democratic governance (e.g. the stakeholder model) might indeed improve the UK Biobank's performance against broad indicators of equality and inclusion. However, what realistic models cannot do is eliminate all barriers to democratic governance. As has been shown elsewhere (Lyall et al. 2009) governance of genomics is dynamic, context dependent and cannot stand alone. Our research findings confirm that the UK Biobank governance structures and processes respond to constant change of ethico-political values and interests in the context of British and international society. Thus, as Elger (2010) observes, these structures and processes have been influenced not only by domestic debates and critiques but also international experiences such as the Icelandic Biobank. Furthermore, as Petersen (2007, p. 33) also points out

… the rapid global sharing of information and experiences among countries, occasioned by the internet and the rise of supranational organisations, has facilitated learning across jurisdictions and some degree of international harmonisation of approach to biobank development, in relation to ethics, governance and public engagement.

Petersen provides examples such as the non-for-profit Public Population Project (P3G), an international consortium comprising members such as the UK Biobank.

Given the limits to governance, the UK Biobank is regulated by government (e.g. the Data Protection Act 1998). Government legislation provides another layer of protection to participants (Lyall et al. 2009). This facilitates and/or strengthens the development of trust throughout the project. Government is also crucial for preventing any kind of genetic discrimination by companies and employers. External problems such as genetic discrimination can only be dealt with through legislation. By contrast, internal challenges such as consent and confidentiality can only be dealt with through substantive democratic governance. The latter has to be developed in such a way that can allow research access to and use of the Biobank but at the same time respect participants' rights and freedoms. As Caulfield and Knoppers (2010, p. 3) note, “This is not a simple task.” For example, consent is directly linked to freedom of participants. Therefore, democratic governance practices under this indicator tend to promote consent. In the case of the UK Biobank, the EGF can clearly be assessed as a democratic governance structure that promotes general or open consent of participants.

The whole idea of democratic governance of genomics projects such as the UK Biobank is to give society (not only experts) the opportunity to democratically decide about the risks and benefits involved in the collection and use of genetic data and lifestyle information. According to Petersen (2007, p. 34) there are several arguments in support of the public's involvement in decision-making.

The instrumental argument is that participation is seen to lend legitimacy to decisions and lead to better results … The substantive argument is that lay judgements about risk are equally valid as those of experts … The normative argument is that there is an incompatibility existing between democratic ideals and the technocratic orientation.

However, as Petersen recognizes, there are limitations and unintended implications which cannot be ignored in democratic governance of genomics. Our research findings support his argument. Whether practical or epistemological, these limitations make the full realization of democratic governance an impossible ideal. Therefore, Petersen (ibid., p. 40) is right to identify the “… need for greater self-awareness of and on the relationship between science and its publics on the assumptions that currently inform ‘engagement’ practices.” After all, innovative genomics projects such as the UK Biobank have never been straightforward. A high degree of controversy has surrounded their scientific and commercial value. On the one hand, critics (Wallace 2005) have questioned the scientific validity of biobanks and their legitimacy, drawing on specific evidence. On the other hand, advocates have stressed the opportunities for commercialization. While the latter have argued that private companies should be granted exclusive rights (i.e. intellectual property rights) over the products of genomics research, the former have stressed the importance of open and free access for new knowledge creation and the public good (Haddow et al. 2007). Therefore, a realistic approach to democratic governance of biobanks might improve not only legitimacy but also judgments about risks and benefits, bridging the gulf between democratic and technocratic orientation.

Conclusion

Genomics research resources or projects such as the UK Biobank are envisaged as essential for understanding the complex interactions between genes and environment, paving the way towards a post-genomic future of personalized medicine and public health. Yet, ethical, legal, social and political risks and benefits involved in the collection and use of genetic data and lifestyle information raise the crucial question of their democratic governance for the public good of health. This paper has tried to address the question by formulating a set of broad indicators of democratic governance and applying it to assess the UK Biobank. The findings of our research suggest that although there is clearly a need for more equal and inclusive processes of representation and participation, these processes cannot overcome practical and epistemological limitations to governance. A realistic model of democratic governance of genomics has to balance the normative imperative for equality, inclusion, freedom and transparency with the empirical need for good practices that “work.”

Acknowledgements

The research for this paper was carried out within the ESRC Innogen Centre (the Open University and Edinburgh University), in the OU's Development and Practice (DPP) Group. I am grateful for financial support from this center and for intellectual support from colleagues, particularly David Wield, Gill Haddow and Ann Bruce.

Notes

Biobanking is a metaphor for the collection, storage, transfer and use of these samples (Palmer 2005).

Here we refer to research subjects and non-subjects alike.

Face validity of documents examines the validity or authenticity documents. Content of specific documents, word patterns, language, composition of paper, all provide checks on the validity or authenticity of relevant documents.

Face validity of interviews is concerned with whether the questions put forward indicate the intended concept (Seale and Filmer 1998). In this project, the assessment of the face validity has been helped by asking managers in the biobanks and civil society groups to assess how well the proposed questions indicate the concept.

Criterion validity of documents involves comparing documents such as policy reports, biobank websites, brochures, etc. with empirical data (observational studies or interviews) on the same topic. In this research, the criterion validity has been assessed by comparing relevant documents on democratic governance with semi-structured interview data.

Criterion validity of semi-structured interviews involves comparing the answers of interviewed subjects with established indicators on the same topic. In this research, the criterion validity of semi-structured interviews has been assessed by comparing the answers of interviewed subjects with other studies in the area of biobanks.