Sharing with Strangers: Governance Models for Borderless Genomic Research in a Territorial World
One of the practices that has defined the ethos of genomic research to date is a commitment to open and rapid sharing of genomic data and resources. As genomic research evolves into an international enterprise, this commitment is being challenged by the need to respect the interests of those it involves and implicates, from individual scientists and subjects to institutions and nations. In this essay, we first describe the types of claims that different stakeholders are making about the disposition of genomic data and samples. Next, we illustrate the complexities of these multiple claims by applying them to the case of one ongoing international genomics initiative, the H3Africa Consortium. Finally, in the light of these complexities, we conclude by comparing and contrasting four governance models for future international data-sharing policy and practices in genomics.
Expectations are high around the world that more research on human genomic variation will improve the utility of “precision medicine” and help address population health disparities through “precision public health” (Juengst et al. 2016). In large measure, these expectations rest on the premise that researchers will be able to share human DNA samples and genomic data freely and widely across the international scientific community. The human genomics community pioneered polices of early deposit of genomic research data into open databases to facilitate the exchange and use of data during the mapping and sequencing of the human genome in the 1990s (Hilgartner 2013), and the success of those policies has made a commitment to scientific sharing fundamental to genomic medicine’s vision for the future of health research. As the Global Alliance for Genomics and Health now puts it, “The sharing of genomic [End Page 67] and health-related data for biomedical research is of key importance in ensuring continued progress in our understanding of human health and well-being” (cf. Global Alliance for Genomics and Health).
For the human genomics research community, a commitment to open data- and sample-sharing is not merely a practical strategy for undertaking their work efficiently. Their “borderless” ethos has also become a moral point of pride, grounded on principles of transparency, reciprocity, and solidarity, and bolstered by the intuition that “the collective human genome is a common heritage of humankind” (cf. Majumdar, Cook-Deegan, and McGuire 2016, 1). But this ideology still functions in a world structured by many types of borders. As we describe below, regulatory conditions placed on genomic data-sharing have been erected and defended to respect individual and familial privacy, to support the translation of ideas, data, and knowledge into commercial products, to protect communities from the risks of discrimination and exploitation, to promote group identity and solidarity, and to redress inequities between populations and nations. Sharing with strangers across those borders does not come naturally in a competitive world, and the complex ethical, legal, and social challenges that it creates for genomics research governance are only beginning to be mapped and reconciled. Realizing the aspirations of precision medicine and public health genomics will require understanding all of these moral claims, and adjudicating them without violating the legitimate interests they are meant to protect (cf. Kaye, Terry, Juengst, et al. 2018).
Those who make claims over the disposition of genomic information often do so with the language of “rights.” Appeals are made to a collection of moral and legal rights—privacy rights, property rights, rights to benefit, rights to research, fiduciary rights, sovereignty rights, and human solidarity rights, among others—to stake claims to decision-making authority over the disposition and use of particular samples or data. Taken as mutually exclusive claims they can each be asserted as policy trump cards, played by an individual or group with the expectation that they will be honored and respected. This tends to create policy stalemates that allow other non-moral considerations such as source of funding, relevant expertise, and material resources to dominate by default, usually to the advantage of the stakeholders who control other forms of power. But if any of the rights claims that challenge the scope of the borderless data-sharing ethos in international genomic research are as legitimate as their proponents insist, we will need to develop ways to take those claims more seriously in the research governance process. [End Page 68]
In this paper, we take three steps down that road. In the next section, we attempt to catalogue the variety of claims to control over the disposition of genomic data and samples, and unpack each of these positions looking for common ground. After that, we consider one prominent international genomic research consortium, the “H3Africa” initiative, to see how these different moral claims are being harmonized—or not—in actual practice. The H3Africa Consortium is particularly interesting in this regard, because it is designed as an African-led partnership between scientists in economically developed countries and those from lower and middle-income countries, and involves DNA sample sources from multiple African nations and communities. Then in the final section we consider the recent literature on “fiduciary,” “adaptive,” and “shared” models of research governance to explore possible ways forward through the complexities that our first two steps uncover.
STAKEHOLDER CLAIMS IN GENOMIC DATA-SHARING POLICY DEVELOPMENT
Those jostling for position around the international genomic research planning table can be sorted into at least six groups: 1) the individuals whose samples are studied, claiming rights of bodily integrity, informational privacy, personal autonomy, but also the rights to contribute to science, to benefit from science (cf. Knoppers et al. 2014), or simply to learn; 2) families (immediate and distant) asserting inheritance or other familial rights to control the disposition of new genetic information about their members; 3) scientists exercising their rights to free inquiry, and intellectual property rights to their discoveries; 4) institutions and organizations defending their fiduciary rights as designated stewards of DNA collections on behalf of both the DNA donors and their investors to benefit from their investments; 5) communities (ranging from nations to local identity groups) declaring sovereignty rights over their genomic patrimonies; and 6) international organizations like UNESCO who promote the solidarity rights of the whole human species to govern its genomic information collectively. In this section, we describe the claims that are made on behalf of each of these stakeholders, and the challenges they each face.
When asked about the disposition of the material they contribute to genomic research as participants, individuals often describe their specimens as extensions of themselves, to which their rights of bodily integrity, [End Page 69] informational privacy, and personal autonomy still apply (Terry and Cook-Deegan 2012). For example, in interviews with individual genomic research participants, Conley et al. report that participants describe their specimen donation as “surrendering a piece of your essence,” because DNA is
such a personal thing. It’s your genetics. It’s who you are. And . . . having that on file is just . . . more invasive than having your fingerprint on file or having your photo on file or something. I don’t know how to explain it better than that. It’s like an unsettling feeling I guess.(2012, 619)
In part, this perception reflects the fact that the risks of donating material for genetic research are not physical risks of bodily harm, as in other forms of medical experimentation, but rather risks to one’s social identity: risks of stigmatization, exclusion, discrimination, and the loss of opportunity. Some have criticized the ontological assumptions behind such perspectives for exacerbating these risks, treating biospecimens too essentialistically as incarnations or avatars of their human sources (Knoppers and Laberge 1995). In practice, many do not perceive their donations as extensions of themselves, as evidenced by those consumers of direct-to-consumer genetic tests who see no risks in sharing data about themselves widely (McGowan, Fishman, and Lambrix 2010), and scientists who advocate for cultivating a culture of “informational altruists” will to accept any risks of sharing for the greater good (Kohane and Aultman 2005). Providers of genomic material who are themselves patients with a genetic disease or their relatives are especially likely to see their genomes instrumentally as useful tools rather than essential to their identities as people (Lim et al. 2017). Neverthless, to the extent that important identity interests are implicated in genomic research, the individual’s concerns to control these sources of social risk are understandable, and well within the penumbra of individuals’ accepted privacy rights.
As a result, most international recommendations for the governance of samples and data in genomic research still express the conviction that individuals should control the disposition of identifiable information about themselves. For example, the UNESCO Universal Declaration on the Human Genome and Human Rights asserts that “prior free and informed consent of the person concerned shall be obtained” for any “research, treatment or diagnosis affecting an individual’s genome,” including secondary research on shared genetic samples, and that “genetic data associated with an identifiable person and stored or processed for the purposes of research or any other purpose must be held confidential,” and shared only with the individual’s permission (UNESCO 1997). [End Page 70]
Of course, even with the prospective collection of voluntary donations from informed participants, the scope of an individual’s rights to control the disposition of their samples often differs dramatically from their rights in other human research settings. Fundamental rights protected in the most international regulatory regimes for the protection of human subjects, such as the right to consent and the right to withdraw from research at any time, become complicated in the stored-tissue and data-based research setting. Genomic research initiatives may give individuals the right to withdraw from further involvement in studies, but rarely to have their donated samples withdrawn from collections or returned, especially after they have been shared with secondary researchers (Melham et al. 2014). Similarly, the ability to sustain the individual informed consent process over multiple data-sharing transactions quickly becomes impracticable.
Moreover, even if the logistical hurdles of individualized control over genomic data-sharing could be overcome, the idea also faces some moral challenges. After all, the genetic information contained in one individual’s biological samples is never about just that individual. As hereditary information, genomic information is already shared by its source’s extended biological family before it is ever collected for research. That means that claims to control over any specific bit of that information could be made by any individual whom the information might implicate, whether the cells that yielded it came from that individual’s body or not.
In many cultures and communities, family involvement in genomic research decision-making will come naturally—perhaps necessarily—just because of the ways in which individual’s identities and obligations are understood (Mahowald et al. 2001). But even in individualistic societies that prize independent self-determination, family members may claim a right to be at the table, since they stand to suffer or gain from the research just as much as the individual specimen sources might.
For example, when the whole genome of the “HeLa” cell line was sequenced in 2013, family members of Henrietta Lacks, the original source of that cell line, were affronted. “That is private family information,” asserted Jeri Lacks-Whye, Henrietta Lacks’s granddaughter, “It shouldn’t have been published without our consent” (Yandell 2013). The family argued that, as Mrs. Lacks’s heirs and offspring, they should have a say in controlling the use of their familial genetic information. In essence, they framed the genomic information generated by the sequencing of the [End Page 71] HeLa cell specimens as a familial legacy that should be governed like other parts of a private bequest, giving relatives and descendants inheritance rights to share in decision-making about the use of the HeLa cell sequence data. In agreeing to add Lacks family representation to the Data Access Committee, established to oversee the research uses of the HeLa genome, the NIH seemed to concur with the view that:
Genetic information is, spontaneous mutations aside, essentially and unavoidably familial in nature. It is this feature of genetics that allows individuals to benefit from genetic testing and diagnosis. When a patient attends a genetic clinic, or discusses genetics with his or her general practitioner, a family history will be constructed, drawing on familial information about diseases and illness supplied by the patient about other family members, often without their consent. In many cases an extensive family history is needed to assess the usefulness of genetic testing. Given this, there is no obvious reason why one family member should be able to benefit and yet, at the same time, be allowed to exclude others from access to such benefits.(Parker and Lucassen 2004, 165)
To make fair decisions about the disposition of any identifiable genomic information, then, the Lacks family’s story suggests that there is a prima facie expectation that the sources’ families will have to be involved in the deliberations as well (Lucassen 2007). Respecting and facilitating similar levels of familial control for the relatives of every participant in large-scale international genomic studies, of course, would be even more impractical than attempting to let individual sample donors make all access decisions. In making its arrangement with the Lacks family, the NIH was insistent that Henrietta Lacks’s celebrity as an icon of research injustice made this a unique case, and that their inclusion on the data access committee should not be taken to set a precedent (Kahn 2013). Moreover, it is not entirely clear whether the family’s claim is legally coherent. Their lawyers’ argument seems to conflate their biological inheritance of Henrietta’s genes with an intentional bequest, when it is not obvious how their lineage (and therefore estate law) establishes any familial rights to control the information in question. Most legal precedents in this area, in fact, suggest that the scientists who sequenced Mrs. Lacks’s genome have as much of a claim over the disposition of the data it contains as the family, since it is only through their decoding efforts that it acquires its value as information. As a result, courts have repeatedly sided with research institutions in disputes over biospecimen ownership rights, not with the tissue donors or their families (Charo 2006). [End Page 72]
One proposed approach to formally recognizing familial interests in genomic information management is to consider all DNA deposits in a biobank as “joint accounts” rather than “personal accounts,” “much like information in a joint bank account,” with decision-making conducted at the familial level (Parker and Lucassen 2004). Of course, members of the same family may not agree on the requirements for an adequate management of such joint accounts (Doukas and Berg 2001). Moreover, as the case of the Lacks family vividly shows, sometimes collective familial decisions about the disposition of their joint genomic data may conflict with the free inquiry and intellectual property interests of the genomic researchers themselves.
For biomedical investigators, genomic research is no different from any other type of biomedical science, and involves the same suite of intellectual property and scientific freedom protections they would enjoy doing other kinds of research. These interests are often oversimplified as a right to free inquiry—defended in liberal democracies and universities—even in the context of the many ways in which science is normally regulated: conditions placed on scientists by funders that require compliance with the terms of receiving a grant or contract, peer review requiremens by scholarly journals, or federal regulations that will place conditions of prior ethical review and approval by an IRB or other committee on the ethical conduct of research. But free inquiry to seek fundamental knowledge is but one of the rights that researchers claim in the genomic world. Researchers also see liberal access to genomic data as essential to applied research with tangible health applications. The attraction to treatment-driven research is seductive. Even Francis Collins, the director of the international effort to map and sequence the human genome, when testifying before the Senate Subcommittee on Labor-HHS-Education Appropriations, referred to the prospect of benefit to human health by the country’s investment in the HGP:
The rationale for the Human Genome Project, and the strong and sustained Congressional support for it, has been the promise of improving human health. We are already beginning to see the fruits of that investment.
Moreover, especially with the passage of legislation in the US (Bayh–Dole)1 we have seen a dramatic increase in the value of intellectual property [End Page 73] by researchers and the institutions in which they work (Markel 2013). These claims to intellectual property rights are intended to protect their ability to determine who can use and control genomic information and their interests in any professional or commercial benefits that may flow from it.
For as long as body parts (from the smallest building blocks of life to entire bodies) have been used for experimentation, education, diagnosis, and treatment, there have been institutions built to collect, store, and distribute them. With the advent of genomics all these collections become potential sources of human DNA samples for genomic research. But while many institutions—academic, not-for-profit, and private—all contain repositories of useful human tissue, their goals can be very different, and affect both their data-sharing policies and the relationships they have with the human sources of their samples.
With the completion of the human genome sequence and developments in high-throughput sequencing technology, the price of sequencing a full genome has been reduced to less than $1000 (cf. National Human Genome Research Institute). This has increased the incentive for more institutions to acquire more specimens, share them, sell them, and coordinate their optimal use with other institutions. But more than acquiring the physical specimens, these biobanks are proliferating across the globe, and with them, the questions about how institutions understand and protect their interests in these materials.
Biobanks are often described as acting as “stewards” (Willams, Sheps, Mcgrath, and Mitchell 2010; Henderson et al. 2013) of those resources, including the exertion of control over their distribution, use, and disposition on behalf of the tissue sources. These claims are sometimes framed as fiduciary rights, granted within the context of a trust relationship dedicated to the best interests of some “true owners” of the samples or data, regardless of whether those “owners” are understood to be individuals, families, or communities. The concept of “best interests” can be deployed here with some success, in much the same way it is used in medical care decisions for those who cannot express their own preferences.
Of course, a more traditional form of this kind of stewardship role is that of the political communities to which the individuals and families donate DNA samples, from nation states to local identity groups. Above and beyond any fiduciary promises that institutions may make, communities [End Page 74] themselves are increasingly active in seeking to protect the interests of their members in genomic research, and this trend has stimulated a growing body of literature discussing the merits of “respect for community” as a principle for genomic research ethics (Jones, Bush, and Macauley 2014).
Within representative democracies, governments can be empowered by their citizens to protect their personal, civil, and financial interests by regulating the activities with which they become involved, including international genomic research involving sample sharing (Muula and Mfutso-Bengo 2007). For example, in 2005, in anticipation of interest by foreign scientists in the patterns of genomic variation within its population, Mexico’s parliament passed an amendment to its health care law, declaring that “Mexican-derived human genome data are property of Mexico’s government and prohibits its collection, use and export without prior government approval” (Benjamin 2009, 344). Essentially, this law reflected the argument that, like other forms of indigenous knowledge or local natural resources, genomic variation information is the rightful communal “patrimony” of the population groups it describes. On this view, any sample- and data-sharing decisions should reflect the sovereignty rights of the communities from which these resources come, including rights to manage (and benefit from) how their resources are exploited (James et al. 2014).
A similar philosophy is embodied in a approach of one of Canada’s federal granting councils to soliciting genomic research samples from Indigenous Peoples of Canada, which they call the “DNA on Loan” model, in which ownership of the genomic information is explicitly granted to the specific Indigenous communities from which it comes, rather than to their individual members or the institutions that house their samples (cf. Canadian Institutes of Health Research). As the advocates for this model say,
In this way the individual and the community (or their designates) retain the ability to determine the future handling and use of biological samples. . . . Because the researcher is ‘the steward’ and not the ‘owner’ of the samples, if, for example, in 10 years after the initial research has been carried out new genes for the condition initially under research are revealed, the researcher is then bound to report back to the community.(Arbour and Cook 2006, 153) [End Page 75]
Of course, sovereign nations are not the only kinds of communities that seek to represent and protect their members’ interest; almost everyone belongs to an entire system of communities that might claim a legitimate role in policy-making over the sharing of genomic research samples, from kinship groups like tribes, place-based groups like neighborhoods, or socially defined groups like research cohorts, patient groups, religious affiliations, ethnicities, or races. Most of these populations do not have a clear system for representing the preferences of their individual members, but they can all, as collectivities, suffer the harms or enjoy the benefits of genomic research framed against their identities. Thus, along with the rising interest in engaging individual research participants and families in genomic research decision-making, there have been increasing calls to give communities voices in the control of genomic research, not just as vehicles for the preferences of their individual members, but in order to defend their own collective interests as well (Upshur, Lavery, and Tindina 2007).
“Community engagement” has become the portmanteau term for a wide variety of strategies to facilitate this collective voice in the design and conduct of genomic research, and almost all new genomic sample and data-sharing initiatives embrace it as an important value (Haldeman et al. 2014). Unfortunately, as the Mexican example shows, appeals to community interests out of a commitment to group solidarity rarely help support policies in favor of wider genomic sample and data-sharing. By definition, communities have to define themselves against the rest of the human population, and structure their allegiances accordingly.
On the other hand, just like families, the boundaries of communities are overlapping and fungible; it is not always easy to discern which communities to engage in a genomic research policy-making process. This is a challenge for biomedical research that aspires to be community based, but it is especially difficult for genomic research. This is because, while the potential harms and benefits of any particular genomic study will apply to the people with genotypes that research implicates, very few of the ways in which we humans organize ourselves into groups actually correspond to the distribution of specific genotypes across our species (Keita et al. 2004).
Given the diversity of communities with which researchers usually engage, under what circumstances could different communities be expected to serve as legitimate spokespeople for members of the genetic populations with interests at stake? As distrust of centralized national governments in populations and philosophical skepticism about the [End Page 76] cogency of individual autonomy both grow, the intuition that the groups in between sovereign nations and autonomous individuals, like families, tribes, or local communities, should be as involved in genomic research decision-making as the scientific community is becoming widespread. But it is also increasingly unclear how to operationalize that intuition in practical genomic research policies.
6. The Species
Finally, some argue that ultimately the human genome is one of the earth’s great global resources, and that no proprietary claims should be made about the information that it contains at all. On this view, the the human genome should be seen as part of the “common heritage of humanity,” like the atmosphere or seabeds, which should be explored, cultivated, and exploited only by universal agreement, on behalf of future generations. Unlike the other stakeholders’ interests in restricting the open sharing of genomic information, this approach encourages the widest forms of data and sample-sharing as possible. Since human genomic research seeks to harvest and exploit a natural resource that belongs to all of us, the argument goes, it should be governed by policies that encourage ensuring that its result are freely available to everyone. This idea is captured in the UNESCO 1997 Declaration on the Human Genome and Human Rights, which begins by asserting that:
The human genome underlies the fundamental unity of all members of the human family, as well as the recognition of their inherent dignity and diversity. In a symbolic sense, it is the heritage of humanity.(emphasis added)
The point of this symbolic common heritage, of course, is to dismiss any claims to proprietary control of human genomic information and relegates genomic research governance to global agreements like those that (attempt to) preserve other shared resources vital to our species’ survival, as a matter of human solidarity rights.
Critics of this approach push back in two ways. First, they question the conceptualization of “the human genome” as a static natural resource that could (even in principle) ever be preserved or cultivated at the species level; the human genome is either a scientific abstraction like “the human skeleton,” or a shorthand way of referring to the complete pool of genes and their multiple variants that our species displays at a given time (Juengst 2006). Attempting to regulate the idea of the human genome is a category mistake, and our collective gene pool changes with every human birth and [End Page 77] death. Moreover, unlike our attempts to protect other global resources, any serious attempts to police the ebb and flow of the human gene pool at the global level would have to involve draconian intrusions into people’s reproductive lives (Juengst 2006).
The assertions on behalf of these six different constituencies reflect different understandings of property, inheritance, gifting, community, and kinship, which will resonate differently across cultural contexts and genomic research projects of different scope. These assertions also have very different—perhaps even incommensurable—implications for the governance of genomic research domestically and globally. Nevertheless, they are deployed, implicitly or explicitly, every time a new international genomic research project, database, or biobank is established—indeed, whenever the challenges of developing resource-sharing policies are played out. If they all reflect parts of this overall moral terrain correctly, is there any way to integrate them into a meaningful map for those who seek to use it?
CASE STUDY: THE H3AFRICA PROJECT
To see how these six sets of claims might be best integrated in the governance of international genomics, it is instructive to consider how they are being adjudicated in practice, and the challenges they are raising for researcher planners. In this section, we examine the sample- and data-sharing policies of one such ongoing initiative, the Human Health and Heredity in Africa (H3Africa) Consortium. H3Africa is an international collaboration launched jointly by the Wellcome Trust in the United Kingdom and the NIH in the US, to sponsor a coordinated set of genomic variation studies across the continent of Africa. The project involves the collection of DNA samples from individuals from multiple sites across the African continent, their storage in several centralized biobanks, and their distribution to researchers from around the African continent—and eventually the world—to use in translational genomic research (cf. Nnamuchi 2017).
Like other genomic initiatives, H3Africa is committed to the wide research sharing of the genetic samples and data it collects. Yet, unlike earlier initiatives, the governance of this initiative is explicitly designed to address the concerns of stakeholders in developing its sample-sharing policies. In describing the H3Africa Policy Framework entitled “Negotiating Fairness in Genomics,” the H3Africa Consortium Steering Committee argues that, [End Page 78]
Of ethical importance is that H3Africa builds equitable partnerships between researchers and other key stakeholders. Equitable, or fair, partnerships can help build strong research systems. They are also a means to counter exploitation and promote mutual respect and trust, and offer an opportunity to ensure that research is responsive to local needs and that data interpretation is contextualized.
However, H3Africa’s understanding of the “stakeholder” with standing in these negotiations is selective in interesting ways. As we explain below, examing the Consortium’s “High Level Principles on Ethics, Governance, and Resource Sharing” (cf. H3Africa n.d.) suggests that they are committed to protecting the interests of several, but not all, of the six classes of the potential stakeholders identified above, and they give the stakeholders they do recognize very different roles in the Consortium’s governance and policy-making.
First, the guidelines give individual participants their conventional role, by insisting that, “It is essential that all research participants give full informed consent for the research uses for which access to their samples and data is sought.” But any more robust individual role in the control of donated samples or data is immediately undercut by the requirement that “the consent needs to be broad enough to allow for future and secondary uses of data, in line with the opportunities to use such data in advancing knowledge to improve health.” As the guidelines further explain,
H3Africa research studies will share phenotypic and genomic data for secondary analysis. This means that although samples may be collected for a study on a particular disease, the genomic data may be analyzed for many different diseases. Also, the data may be analyzed to look at population differences or to test new tools for analyzing genomic information. The consent information and processes need to make this very clear to potential participants.(H3Africa 2013, 7)
As H3Africa has begun to involve a wider array of stakeholders from different African settings, however, this initial stance has begun to shift. Based on more extensive stakeholder engagements, the project now endorses a model framework that builds considerably more flexibility into its notion of broad consent, explaining that,
Broad consent as understood in the framework allows for use of samples and/or data for unspecified future studies, but with conditions. These conditions can involve, for instance, a restriction on the types of studies or diseases that [End Page 79] samples/data can be used for; a specified oversight and approval process for future use; ongoing consultation with sample donors about future use, if possible; and a process allowing participants to withdraw samples or data from the storage facility that holds them.(Yakubu et al. 2018, 4)
This evolution relegates much of this decision-making to individual projects, empowering local investigators and participants to exert their own control over their data and sample-sharing plans. Whether this flexibility counts as a weakness or a strength of the H3Africa’s approach will depend on which theoretical model of governance one uses, as we shall see below.
Second, despite the importance of kinship ties and family identities in the literature on African interpretations of genomic information, the governance framework for H3Africa is silent on the status of familial claims to control over genomic information. In fact, beyond an individual’s voluntary donation of biospecimens to the H3Africa project, neither participants or their families are given a substantive role in the governance of the ongoing research. The Consortium’s rules are made jointly by the funding agencies and its expert committees of scientists, and there seem to be no consortium-wide mechanisms for ongoing contact with individual participants or their communities once a DNA donation is made.
The governance policies of H3Africa do seem path-breaking in protecting and privileging the voices of the African scientists involved in its research in the face of power differentials with their Anglo–American funders and collaborators. For example, the H3Africa “Negotiating Fairness in Genomics” framework focuses on rebalancing these background resource inequities and power differentials through its governing committees and collaborator engagement exercises, by ensuring “preferential access to funding, samples and data for African researchers” (deVries et al. 2015). But while the researchers and institutions involved in H3Africa are all given voices at the governance table, the closest that the governance framework comes to involving other stakeholders directly is the aspirational hope that “ideally,” the Consortium’s working groups should involve “patient advocates.”
On the other hand, H3Africa does aspire to address the interests of the communities from which they recruit. In declaring the “High Level Principles” behind its governance policies, H3Africa emphasizes that “research networks and programs should incorporate engagement with the communities participating in the research as in integral element,” and that “this should be an ongoing activity throughout the program which extends beyond the original consent.” By extending community [End Page 80] interactions beyond the one-time “broad consent to sharing” imposed on individual participants (and families?), this principle gives communities particular power as stakeholders, and the Consortium has underlined that commitment by developing specific guidelines that reflect its rationale for this prioritization:
In terms of intrinsic value, engaging the community is important as a sign of respect for the values, culture and traditions of the community involved in research. This may be particularly important when research involves participants and communities that are relatively untouched by science and modern medicine—as may be the case for communities that are unfamiliar with genomic studies. Community engagement can be seen as a means of protecting the rights and interests of relevant communities and avoiding any form of exploitation. Community engagement is also an important step in determining the (health) needs and expectations of the community, ensuring that there is an opportunity to consider whether and how these could reasonably be met within the scope of the project.(H3Africa 2013, 21)
Finally, while it does not go so far as to afford communities with the sovereignty to deny researchers access to their members, H3Africa does also seem to recognize the claims of genomic sovereignty at the national level, when it acknowledges that: “The transfer of samples must be in accordance with all relevant local laws, and appropriate research ethics approval for all research projects must be in place in line with relevant local regulations and best practice.” We take this to mean that while the Consortium may endorse the ideals of open sample and data-sharing, some of the nations in which the participating scientists live do not hold the same views. For example, the South African National Biotechnology Advisory Committee issued a position statement on “Genomic Sovereignty in South Africa” that endorsed the view that human genomic data about its citizens was a natural resource that could be protected under a doctrine of genomic sovereignty modeled on the Mexican law we described above. (deVries and Pepper 2012).
In making regulatory space for such national claims to biospecimen control, of course, H3Africa implicitly rejects the claims that the human genome is the “common heritage of humankind,” and any role that the human species as a whole might play in determining the Consortium’s sample-sharing policies. In fact, in recommending its approach as a “model framework” for other genomic research initiatives in Africa, the H3Africa Ethics Committee is frank in asserting that, while “recognizing that it likely also yields benefits to the global population,” “the primary [End Page 81] objective of African biobanking and genomic research should be to benefit African people” on the grounds that in such studies African participants are “shouldering the research burdens and risks” (H3Africa 2018, 3, 7).
The governance of international genomic research consortia is a phenomenon in flux, and no one should expect any particular example to satisfy the claims of all the kinds of stakeholders who have been nominated for standing within it. Indeed, the example of H3Africa is instructive in part for what it neglects: it may be that the moral claims of specific families over individuals or of the whole human species over particular human communities can never be made strongly enough to merit dedicated representation in the governance of genomic research. Moreover, H3Africa’s provocative principle of privileging the less powerful in negotiating fairness between scientists seems like an idea that could be taken even further, depending on the governance approach adopted by the Consortium. Several recent proposals for approaches to research governance bear particular attention in this regard.
ALTERNATIVE GOVERNANCE APPROACHES?
In 2008, the ethical issues posed by the governance of international genomic research were described by one commentator as a “nightmare for public policymakers” because of the “tower of Babel” created by the conflicting rights claims reviewed above (Mauron 2008). Over the subsequent decade, the literature has only grown (cf. Bull, Roberts, and Parker 2015), suggesting an array of different frameworks, principles, guidelines, and recommendations to help the scientific community move forward (cf. Mauron 2008; Kaye and Stranger 2009; DeVries, Tindana, Little, et al. 2012; Global Alliance 2014).
These different governance schemes can be segregated into four overlapping approaches, each of which is worth assessing in light of our review of the various stakeholder rights claims at play in international genomic research. For convenience, we refer to these four alternative approaches by the adjectives that seem to sum up their core virtues in their proponents’ minds: fiduciary, adaptive, shared, and dynamic. Proponents of all four approaches recognize that international genomic research collaborators rarely come to a research project with equal standing to debate and shape consortial data and sample-sharing policies. Some scientists will be advantaged by being the founders and initial architects of the initiative, while others will disadvantaged by coming on board after key decisions have already been made. Well-resourced scientists [End Page 82] and institutions will likely be more influential in decision-making than the scientists and institutions that must depend on them for support. Active members of the international scientific community will find more influential audiences for their views than those isolated from the larger world by language, politics, economics, or geography. Moreover, the same kinds of contextual inequities are only magnified when the stakeholders involved are the consortium’s lay research participants, their families, and their communities. For these stakeholders, the disenfranchising effects of nationality and wealth are only compounded by gaps in education and social status between them and the scientists that seek to recruit them.
While each alternative approach has merits in different genomic research contexts, what distinguishes them for our purposes is the extent to which they use the social determinants of negotiating power among different stakeholders to help reconcile their rights claims, by finding ways to amplify the most disenfranchised voices at the genomic governance table.
Fiduciary Governance approaches represent the most direct extension of traditional single biobank governance philosophy to international consortia. On this view, the locus of decision-making about the sharing of data or samples within and beyond the circle of sample-contributing researchers is the institution and its managers. In accepting that role, they take on fiduciary obligations to protect the best interests of all the other stakeholders who might have sample disposition claims: the sample sources and their families and communities, the contributing scientists and their institutions and funders, the nations and humanity as a whole. In practice, as a result, the role of individuals as the source of samples is usually limited to one form or another of giving “consent to be governed” by others (Koenig 2014), either through group consent or “informed cohort” policies (Holm and Taylor 2012) or arrangements in which researchers act as the sources’ “stewards” or “trustees” in research decision-making (Kaye and Stranger 2009).
One putative virtue of the Fiduciary model in domestic contexts is that it avoids having to address the legitimacy of different stakeholder claims beyond the scientists involved. Whether the DNA under study is owned by its sources, their families, or communities or is merely “on loan” from some set of these stakeholders is irrelevant to subsequent data-sharing policy decisions, as long as the participants have consented to permit the biobank to shepherd the disposition of their specimens in the interests [End Page 83] of the larger society. As a result, the Fiduciary Governance model is the form of governance that is easiest to relapse into when more participatory aspirations falter. In international contexts, however, the inconvenient fact that collections of data and tissue have to be maintained and managed somewhere quickly raises questions about which larger society will benefit from the research and how those benefits will further the interests of participants who live elsewhere. Because of this, new international genomic research consortia usually seek governance schemes that can provide a wider role for a range of international stakeholders in policy development.
The second broad family of international governance approaches, which O’Doherty and colleagues helpfully describe as “Adaptive,” are also drawn from efforts to manage large-scale domestic biobanks, but in situations that lack the background trust required of participants and collaborators that Fiduciary approaches requires (O’Doherty et al. 2011). The principal virtue of these approaches is to recognize the fluid nature of governance requirements as new stakeholders join a project, and the need to incorporate their voices when revising project policies. On these approaches, international genomic research initiatives would be initially designed and established by the scientific community, as they are under Fiduciary approaches. Then, as new collaborators and participants are recruited into the initiative, its governance structures would open up to include them or their representatives as members of the initiative’s decision-making committees, as equal “stakeholders” (Hunter and Laurie 2009) in the enterprise. For example, the Canadian biobank in British Columbia, BC Generations, maintains a “Participant Association” of interested individual tissue donors, who elect a “Participant Board” of representatives who are then engaged to speak for the whole participant cohort on the projects management committees. Even more importantly, the expectation is that project policies on matters such as sample-sharing might even change under the influence of these participant voices. As the architects of the BC Generations project explain, rather than simply being asked to entrust all decision making to the project’s founding management team (as in the Fiduciary model),
Participants would be asked to consent to a governance process as part of the consent to participate in the biobank. The consent process would include explicit invitations to participate actively in governance and follow biobank [End Page 84] activities in newsletters/web forums. The adaptive mechanisms would be made clear, including an explanation that the Board of Directors, which includes participant-members, has the authority to change existing policies and procedures.(O’Doherty et al. 2011, 367)
Some have elaborated on the Adaptive Governance model to imagine even more direct interaction between the participant cohort and the project management through electronically mediated “e-governance” platforms (Kaye 2012). For projects that seek to grow internationally, similar processes and technologies would allow new scientific partners to become involved in iteratively shaping the consortium’s policies in the same manner. We note as well, that Adaptive does not imply “expansive,” i.e., only by adding voices. Rather, adaptive models are intriguing, we think, precisely because they are flexible, which would suggest that as project goals evolve, changes in governance structure could be built in.
Adaptive governance models go some distance towards giving both research participants and scientific partners ongoing roles in the development and direction of genomic research, but they do not address all of the diverse interests of those that make rights claims against international genomic research consortia. There is no direct way for families, communities, or nations to be represented, and these models cannot address the interests of potential users or beneficiaries of the consortium’s resources and research. As the advocates of adaptive governance acknowledge,
There are issues such as intellectual property and community specific needs (for instance, involving identifiable communities or those with particular views of biospecimens), that may require additional components of governance, but are beyond the scope of this paper.(O’Doherty et al. 2011, 368)
Shared governance, on the other hand, includes models designed to be as inclusive as possible across the range of possible stakeholders. Instead of being ramped-up versions of domestic biobank governance schemes, they emerge from the literature of international public health research ethics and community-based participatory research, to focus from the start on addressing the power disparities between collaborators and potential beneficiaries. One recent articulation of this approach is Bridget Pratt and Adnan Hyder’s work on shared governance for international research consortia. Pratt and Hyder have brought Jennifer Ruger’s previous work on [End Page 85] fairness in the governance of health care delivery to bear on the research context (Pratt and Hyder 2016). First, they argue that a key feature of shared governance in research settings is a commitment to “shared sovereignty” that calls for an inclusive policy-making process where stakeholders from all levels “jointly develop health governance structures, set standards and rules, determine priorities, allocate resources, and establish different actors’ roles and responsibilities.” Moreover, they argue,
Inclusion further means that it is highly desirable for consortia policy-setting processes to be informed by the views of not only consortium members but also health research users and beneficiaries, especially those from disadvantaged groups that experience poor health.(Pratt and Hyder 2016, 38)
Second, Pratt and Hyder argue that a just approach to shared governance would go beyond giving all stakeholders equal standing, to take into consideration the background disparities that stratify the negotiating power of different stakeholders, with an eye to insuring that those differentials do not disadvantage the less powerful negotiators. Discussing this in the context of genomic research Pratt separately argues that,
an inclusive process would be attentive to power imbalances 1) between researchers, research users and research beneficiaries, 2) between researchers (eg. HIC versus LMIC, junior versus senior), and 3) between different research users and research beneficiaries (eg, DNA donors, their familial and community leaders, their national governments). This would involve first identifying between which relevant stakeholders there are large gaps in power, and then developing strategies to minimize the impact of those disparities on the deliberative process.(Pratt 2017, 1)
Pratt and Hyder’s “shared sovereignty” principle would widen the variety of voices involved in genomic research governance beyond the range of both the Fiduciary or Adaptive approaches, by going beyond even research participants to the communities and constituencies that might eventually benefit from the research. Like Community-based Participatory Research models in public health research (cf. Skinner et al. 2015) it aims to bring all of these voices into the process from the very beginning of the project planning, rather than integrating them serially as the project unfolds. While both of these features of shared sovereignty seem well designed to respect the whole range of stakeholder rights claims that might be made about a project’s data- and sample-sharing policies, operationalizing them does pose special difficulties for genomic research. [End Page 86]
First, for open-ended international genomic research consortia it will be impossible to identify the stakeholders who should be involved in designing the research. Unlike community-based public health and health services research, basic genomic research has no discrete geographical frame of reference, and thus no way to ascertain in advance the key constituencies to engage. The knowledge it yields describes and serves the needs of genetic populations threaded across the world in patterns that become apparent only as the research proceeds. Moreover, even if the funders of the H3Africa Consortium could have engaged the peoples of the global African diaspora—or even the inhabitants of the African continent—in a grassroots research needs assessment, basic genomic variation research would not likely take priority over research addressing more pressing African needs, like clean water, adequate nutrition, or the prevention of infectious disease (cf. WHO/AFRO 2014).
Pratt and Hyder’s second principle of structuring the deliberations so that the least empowered have an equal voice also would have interesting implications in the context of sharing governance for data-sharing policy. The Shared Governance model tends to be focused on equalizing the voices of the less and more powerful researchers, with an eye towards the ways in which socio-economic and professional disadvantages muffle the voices of less privileged scientists. But if the research participants are also understood to be equal stakeholders, the process will also have to be capable of redressing the ways in which their voices can be structurally muffled by multiple layers of (often legitimate) interpreters—their families, communities, nations, etc.—with stakes of their own in the outcome. Moreover, it would have to be able to engage the relevant participants whenever a new decision had to be made about granting another scientist access to their specimens or data.
Finally, what would be the result of “balancing” the power differentials in this deliberation? Arguably, the most powerful party will be the Consortium’s founders—both its funders and its original leadership. In the H3Africa example, for example, that would be the initiative’s big Western H3Africa funders, NIH and the Wellcome Trust, and the leadership of scientists from leading African universities. Using Shared Governance, they should have the weakest voice in the deliberations, no matter what policies they would have previously decided upon. After that might come governments with legal jurisdictional claims over the decision—geographical communities or nations. Their claims might be enough to overrule or compromise the “open sharing” aspirations of the [End Page 87] founders but not those of stakeholders with less power, like the external researchers seeking access to the consortium’s resources or the communities and families of its sample donors. And of course, given the constraints on their power that come with their roles in their communities and families, the individual sample sources are the least powerful of all—and thus, by the logic of the Shared Governance model, the stakeholders whose voices should be the most fiercely protected.
The upshot of this compensatory cascade seems to be that the best way of equalizing the influence of the least powerful in this situation is to let individual family members decide for themselves whether or not data from their specific donated samples may be shared more widely, regardless of what their spokespeople, their clinicians, their community, or the project’s funders and founders think. But of course, turning consortium governance questions into individual participant choices would require empowering potential sample donors to make those choices freely in the face of the whole power hierarchy, which will require individually tailored education about risks and alternatives and strong confidentiality protections. But now we have come full circle: this seems to be exactly the individualistic position that all the other stakeholders in genomic research complain is so impractical, unfair to them, and obstructive to science as to require new forms of research governance in the first place.
It may seem ironic that our exploration of alternative governance models ultimately returns us to the liberal tradition of individual research participants voluntarily negotiating—and renegotiating—what happens to the data and tissue they donate through an ongoing exchange of information like that envisioned by the architects of “dynamic consent” (Kaye et al. 2015). Of course, a proliferation of ‘personalized preference’ contracts regarding sample-sharing would create a record-keeping nightmare for sample managers and might blunt efforts to speed scientific discovery through wide data-sharing. To some extent, this trade-off may be the price of fairness in genomic research governance, assuming fairness is the first among the substantive values supporting biobank governance. But the purpose of governance is to enable the smooth and responsible functioning of an activity, and the defining activity of a genomic repository is research. If the costs to research progress of a “shared governance” approach that privileges individuals as the least powerful negotiators were greater than what it could achieve in terms of integrating and respecting stakeholders’ voices, it would be self-defeating. [End Page 88]
But perhaps “shared governance” need not be the end of the story. If the research governance process could be as dynamic at the collective level as “dynamic consent” aspires to be at the individual level, perhaps elements from Fiduciary, Adaptive, and Shared governance models could be combined to produce a form of “dynamic governance” that could achieve both goals of protecting fairness and facilitating research in a highly contextualized way. Consider the following scenario:
One could imagine a governance strategy that allows new international genomic research initiatives to be initially proposed and designed by their scientific architects, in line with the Fiduciary Governance model, including draft data- and sample-sharing plans reflecting the “borderless” ethos of the genomics community. Then, as the Adaptive Governance model prescribes, as new investigators, laboratories, or biobanks in different parts of the world become involved, they would be integrated into the governance of the initiative, and might bring different intellectual property and genomic sovereignty claims to bear to create exceptions and amendments to those plans. Even more importantly, as the project begins to recruit its DNA donors, the participant cohort would also gain seats at the project’s planning table through its own representative “participant board,” and together the research collaborators and the research participants would periodically review, extend, or amend the project’s sharing policies. The other non-participating stakeholder groups that become implicated in the project—participants’ families, communities, and beneficiaries—could be engaged as sources of information and perspective during the policy-making process as well, but the policy itself would be made by the project’s integrated governance structure.
One advantage of this approach is that genomic researchers could still embrace H3Africa’s “shared governance” principle of privileging the less powerful, by making the research participants “first among equals” as arbiters of the other stakeholders input. Such approaches would not be “participant driven” so much as “participant powered,” in the sense that, once involved, the participant cohort would be given the power to prioritize the other stakeholder’s rights claims within the collective policy-making process—including policy-making that privileges the others’ claims over individual participants—if that is what particular circumstances require. Ultimately, this means that the cohort of people volunteering their genetic information for research would be free to bring the interests of their families, communities, and countries to the decision-making table if [End Page 89] they see themselves as representing these other stakeholders, and would have the power to see them respected. Since those interests will vary as the participating cohort changes, this mechanism can help refocus the research on questions of critical importance to those involved, which will help ensure that research-enabling data-sharing policies are endorsed. On the other hand, the shifting needs of the participating cohort across different social contexts do also mean that scientific initiatives might be expected to evolve as they involve a wider range of stakeholders, and their data-sharing policies might legitimately change along with that evolution.
As as the era of large-scale, international genomic research accelerates, the scientific value of open sharing is making it more important to decide which stakeholders’ claims to having a say in setting the parameters of that practice are well-founded, and to develop governance approaches that can take legitimate claims more seriously. This paper’s review of such claims and the different approaches to addressing them points to some of the questions and considerations relevant to that task. If the H3Africa model can be extended to include the dynamic participation of researchers and research participants into the governance process, it may be that the most significant paradigm shift that genomic research catalyzes is not simply “precision medicine,” but in the way that the collaboration to achieve its goals is performed.
Eric T. Juengst, Ph.D., is Director of the Center for Bioethics and Professor in the Departments of Social Medicine and Genetics at the UNC-Chapel Hill School of Medicine. He did his masters and doctoral work at Georgetown between 1978–1984 and since 1997 he has been the principal investigator of a series of NIH-funded research projects examining the ethical, conceptual, and social policy issues that will be raised by the availability of genetic and genomic technologies.
Eric M. Meslin Ph.D., FCAHS, is President and CEO of the Council of Canadian Academies (CCA) a not-for-profit organization that conducts independent, evidence-based assessments of leading policy topics for the Government of Canada. Trained in bioethics at the Kennedy Institute of Ethics, he has held academic appointments at Indiana University, University of Toronto, University of Western Australia, and Université de Toulouse, as well positions at the National Human Genome Research Institute and the US National Bioethics Advisory Commission.
This essay has its origins in discussions of the ELSI 2.0 International Collaboratory for Genomics and Society Research, and we are especially indebted to Jane Kaye, Kazuto Kato, Patricia Marshall, Bridget Pratt, and Jantina DeVries for their insights. All misinterpretated facts and misguided value judgments are, of course, attributable solely to the authors.
* The views expressed by Eric M. Meslin in this paper are his alone and may not reflect those of the Council of Canadian Academies (CCA).