The Ethics of Consent in a Shifting Genomic Ecosystem

Abstract

The collection and use of human genetic data raise important ethical questions about how to balance individual autonomy and privacy with the potential for public good. The proliferation of local, national, and international efforts to collect genetic data and create linkages to support large-scale initiatives in precision medicine and the learning health system creates new demands for broad data sharing that involve managing competing interests and careful consideration of what constitutes appropriate ethical trade-offs. This review describes these emerging ethical issues with a focus on approaches to consent and issues related to justice in the shifting genomic research ecosystem.

1. INTRODUCTION

The ethical collection and use of human genetic data raise important policy questions about how to balance individual autonomy and privacy with the potential for public good. In the context of increasing demands for broad data sharing, managing competing ethical principles requires careful consideration of what constitutes ethical trade-offs. In the last several decades, there has been a proliferation of local, national, and international efforts to collect genetic data and create linkages to support large-scale initiatives in precision medicine and the learning health system. Marshaling technology to harvest ever more detailed and expansive data has been described as the gold rush of our era: “Data in the 21st century is like oil in the eighteenth century: an immensely, untapped valuable asset. Like oil, for those who see data’s fundamental value and learn to extract and use it, there will be huge rewards” (1). The datafication of health or the conversion of personal information into quantified data has contributed to the convergence of genomic sequencing technologies, electronic health record (EHR) systems, and a shifting regulatory and commercial landscape that creates new ethical challenges (2, 3).

Ethical frameworks initially created to govern clinical research did not anticipate the range of potential uses and users that have emerged in data science, which are increasingly integral to the genomic ecosystem. Institutional actors such as companies, healthcare systems, insurers, universities, government agencies, and others have diverse and, at times, conflicting interests that bear on ethical concerns of control, privacy, transparency, and profit. Capital investment in genomic technologies has fueled increasingly efficient genetic sequencing and testing and has lowered the bar to entry into the genomic marketplace. Public consumption of genomic information has been transformed over the last several decades, and never before have genetics and its nomenclature been so integrated into public discourse. New social models such as the learning health system, in which data are used in real time to improve patient care, challenge conventional approaches to consent, underscoring the limits of ethical oversight.

The development of a sophisticated data sharing infrastructure and its convergence with the growing public appetite for genomic information have contributed to the changing landscape. The shifting boundaries of health versus nonhealth data, research versus clinical care, and human versus nonhuman subjects research demand careful consideration of how well current ethical frameworks and oversight approaches address emerging uses of genetic data. There is growing concern that invisible data collection and integration may compromise privacy and civil liberties by incentivizing the sharing of large amounts of data, which individuals and groups may deem private. Whereas ethical principles identified for research involving human subjects have served to guide biomedical studies, the integration of data science, engineering systems, and tools creates novel scenarios that suggest the need for new approaches.

Recent efforts of targeted recruitment of historically underrepresented populations in genetic research raise further questions about the extent to which consent can address goals of equity. Several analyses have demonstrated that participation in genetic research is significantly biased toward individuals of European ancestry, and typically includes only a small minority of those who identify as African, Asian, Latinx, or Indigenous (4–6). In response, national genome sequencing and precision medicine initiatives have incorporated calls for recruiting diverse populations. A stated aim of such initiatives is to address health disparities by including historically underrepresented and underserved populations (7). An example is the All of Us Research Program, which aims to enroll one million Americans into a national cohort study that will integrate biospecimens, EHR data, and other personal data for future research (https://allofus.nih.gov/). Such national initiatives attempt to redress a history of exploitation, deception, and racism in biomedical research that has contributed to distrust among the racial and ethnic groups that are now targeted for recruitment (8, 9).

The shifting genomic landscape and the focus on collecting data from underserved and historically marginalized groups create ethical challenges for current approaches to consent. In the datafication of human life (2, 10, 11), questions about whether research using genetic data aligns with public values loom large. Is there an imperative to analyze data because such data exist? Are there questions that should never be asked regardless of the researcher’s good intentions and research’s potential value? Are there ethical considerations for determining valid questions, and if so, which principles should guide these decisions? And how should patient values be incorporated into the process of consent? This article describes ethical principles for protecting individuals in research involving genetic data. Drawing on empirical studies of the ethical, legal, and social implications of genomic research, I describe the rationales and approaches to consent and identify challenges that emerge with changes in the genomic ecosystem.

2. ETHICAL PRINCIPLES FOR HUMAN RESEARCH

Just as science is a sociocultural process, so, too, is bioethics. History and social context have informed our ethical infrastructure of principles, guidelines, and regulations (12, 13). Contemporary ethical discourse about genomics in the United States is, to a significant extent, a response to the long history of eugenics in the United States that stoked public concerns about the potential social harms of genetics research. For example, in 1927, the US Supreme Court ruled in Buck v. Bell to permit the sterilization of individuals identified as intellectually disabled (14). Famously, Justice Oliver Holmes declared that “three generations of imbeciles are enough” (274 U.S. 200, p. 207; see also 15). During this same period, American educator and sociologist Harry Laughlin, who led the Eugenics Record Office at Cold Spring Harbor Laboratory, argued that the “American” gene pool was being polluted by a rising tide of “intellectually and morally defective immigrants—primarily from eastern and southern Europe” (15; see also 16, 17). Based on Laughlin’s testimony, Congress voted in support of the Immigration Restriction Act of 1924, the first to use national quotas to delimit the proportion of immigrants from countries due to their assumed inferiority and cost to taxpayers (15). With his declaration that America must remain American, President Calvin Coolidge fueled intense anti-immigration sentiment that kept in place the quota system until it was revised with the passage of the Immigration and Nationality Act of 1965 (18, 19).

It was not until 1974 that the United States passed legislation with the National Research Act (P.L. 93–348, 88 Stat. 342) to regulate ethical protections of humans recruited into research. News stories of the treatment of African American men enrolled in the US Public Health Service Syphilis Tuskegee Study catalyzed governmental action (20). The 40-year nontherapeutic study began in 1932 in Alabama by enrolling 400 men with syphilis and 200 uninfected men who served as controls to determine the “natural course” of the disease in black men (21). Study recruitment materials misrepresented the research as an invitation for clinical care; in order to preserve the research design, researchers did not give penicillin to patients even after it was found effective in treating syphilis in the 1940s (22, 23). The Tuskegee Study is but one example of research based on racist assumptions of population differences during the early twentieth century.

The Nuremberg Code that emerged in the aftermath of the Nuremberg Tribunals at the end of World War II predated the National Research Act (24). Outlining ten requirements for the ethical conduct of human experimentation, the Code recognized voluntary consent as essential. For the first time, the Code articulated a standard for consent in requiring “sufficient knowledge and comprehension of the elements of the subject matter involved as to enable him [a person] to make an understanding and enlightened decision” (24, p. 181). The Declaration of Helsinki in 1964 stated more explicitly that consent should contain clear information on research aims, methods, sources of funding, conflicts of interest, institutional affiliations of the researcher(s), anticipated benefits and potential risks of the study, poststudy provisions, and any other relevant aspects of the study (25). It reiterated that the potential research participants must be informed of their right to refuse participation in research and to withdraw consent to participate at any time without reprisal.

Despite this codification of ethical conduct, many researchers failed to implement these requirements, including the investigators in the Tuskegee Study, which continued for decades (26). In his 1976 article, historian Allen Brandt concludes, “There can be little doubt that the Tuskegee researchers regarded their subjects as less than human. As a result, the ethical canons of experimenting of human subjects were completely disregarded” (27, p. 27). Ethical principles were articulated, but practitioners failed to incorporate them into their clinical practice and research. Ethical oversight and governance of research were lacking.

2.1. The Belmont Report

In recognition of the need for ethical principles to guide researchers and their institutions in the conduct of science, the National Research Act of 1974 established the National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research. The Commission was charged with identifying ethical principles that should inform biomedical and behavioral research involving human subjects, as well as developing a framework for ensuring that researchers would conduct such research in line with these principles. In particular, the Commission considered (a) the boundaries between biomedical and behavioral research clinical medicine, (b) the risk–benefit criteria for determining ethical research involving human subjects, (c) ethical guidelines for the selection of human subjects for participation in research, and (d) the nature and definition of informed consent for participation in research (28). To this end, the US government commissioned the Belmont Report in 1979, which identified three core bioethical principles for human research: respect for persons, beneficence, and justice (Figure 1) (28).

Figure 1.

Bioethical principles for human research, as outlined in the Belmont Report (1979) by the National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research (28).

The principle of respect for persons is founded on the moral requirement that one acknowledge persons’ autonomy and protect those with diminished capacities. To ensure autonomy, researchers must obtain informed consent through a process that demonstrates that individuals are free of coercion. The principle of beneficence and maleficence requires that research promotes public good and prevents or minimizes harm. According to the Commission, the principle of justice focuses on the fair selection of human subjects recruited into research and whether these individuals bear disproportionate burdens (29). In identifying the importance of these principles, the Commission took a nonutilitarian position that potential social benefits from research could not justify disregard for risks to individuals and rejected the notion that such risks could be offset by “the greater good of the greater number” (30). In subscribing to the preceding principles, the Commission reflected public attitudes toward science and the sentiment that while scientific research must be encouraged, it must go forward with minimal possible risk to research subjects (28).

The Common Rule, which was instituted in 1981, is the regulation that operationalizes the principles outlined in the Belmont Report, providing oversight to federally funded research involving human subjects (31). The implementation of these guidelines is overseen by institutional review boards (IRBs), which are responsible for providing an independent ethics review that assesses how researchers address harms to individuals recruited into their studies (32, 33). These harms can include negative psychological impact and discrimination that may result in information disclosure.

Recognizing the potential for social harms associated with genetic information disclosure, the US Congress passed the Genetic Information Nondiscrimination Act (GINA) in 2008 (P.L. 110–233, 122 Stat. 881). The legislation prohibits health insurance companies and employers from using genetic information in decisions. It took over a decade to pass GINA, and the law represented a significant step for protecting genetic information. However, its purview is limited, not extending to life, disability, and long-term care insurance. It also excludes individuals who serve in the military or receive healthcare through the Veterans Administration or the Indian Health Service. Other protections such as the Health Insurance Portability and Accountability Act of 1996 (HIPAA) (P.L. 104–191, 110 Stat. 1936) prevent health insurers from using genetic information to determine eligibility or set premiums and from treating genetic information as a preexisting condition. However, GINA includes a research exception, which allows health plans to request but not require a person to undergo genetic testing for research purposes. It is also important to note that GINA does not penalize insurers for incidentally acquiring genetic information.

2.2. Defining Human Subjects Research

In considering consent in the context of genomics, an essential stipulation for applying ethical guidelines is whether a request for genetic data is deemed research. A question for the Commission was determining what human subjects research is. They concluded that not all activities that produce knowledge or intervene in human lives constitute research, and not all research about humans should be deemed human subjects research. What constitutes research becomes murky when genetic data and biospecimens are collected to build a library, repository, or set of references for future genomic research. As defined by federal regulations in the United States, research is “a systematic investigation, including development, testing, and evaluation, designed to develop or contribute to generalizable knowledge” [Protection of Human Subjects 45 C.F.R. §46.102 (2013)]. Without a specific use in mind, some may argue that biobanking does meet the requirements of research; however, if future use is intended and indeed the raison d’être of the biobank, consent from participants would be expected. How to explain these future uses is a challenge.

In 1951 Henrietta Lacks was a young African American woman whose cervical cells were collected during her care for cervical cancer at Johns Hopkins Hospital. Researchers used her biospecimens to develop the HeLa cell that has proved instrumental in significant biomedical research developments. As was the case at the time, her physician did not administer consent, and the biospecimens were made available to researchers without her permission or knowledge. Over a half-century later, Lacks’ connection to the HeLa cell became widely known through the 2010 book The Immortal Life of Henrietta Lacks by Rebecca Skloot (34). Recognition of the HeLa’s origin raised several significant ethical issues that bear on consent and broader principles of beneficence and justice. Unbeknownst to Lacks, her tissues had been taken in the context of clinical care and then used for medical research. Without explicit consent, Lacks became both patient and research subject without her permission. More recently, the NIH (National Institutes of Health) have put in place several oversight measures (35). However, the story of the HeLa cells reflects the historical disregard for autonomy that was further compounded by a legacy of racism in biomedicine.

With the passage of the National Research Act, informed consent has become the primary process to ensure individuals make voluntary decisions about participation. However, informed consent is only required when research activities are deemed human subjects research. According to the Common Rule, a human subject is defined as a living individual about whom a researcher obtains (a) data through intervention or interaction with the individual or (b) identifiable private information (34). When researchers collect biospecimens from an individual for research, informed consent is required. However, when a researcher uses data or biospecimens that have already been collected and then subsequently have been deidentified, the activities may no longer be considered human subjects research. For example, biospecimens and data collected and stored in a biobank, in which all identifiers have been stripped and replaced with a code and researchers have no access to identifying information, would no longer qualify as human subjects research and future research activities would not require consent.

By deeming anonymized data used in secondary research as not meeting the definition of human subjects research, the Common Rule creates a more efficient yet potentially fraught regulatory pathway toward building big data infrastructure. A point of debate is the increasingly tenuous assumption that genetic data and biospecimens can truly be anonymized. Over two decades ago, computer scientist Latanya Sweeney conducted a study in which she used zip code, date of birth, and gender information from US Census data to reidentify individuals (36). Sweeney demonstrated early that publicly available databases made it increasingly challenging to anonymize individuals. Sweeney and her colleagues confirmed this further by identifying the names of participants in the Personal Genome Project by linking individual profiles to public records (37). Demonstrations of reidentification through seemingly deidentified data challenge the premise of secondary data and jeopardize public trust with a potentially false promise of privacy. This raises concerns that short-term expediency through data collection activities deemed nonhuman subjects research may present longer-term challenges due to the lack of ongoing consent.

3. CONSENT APPROACHES

The linchpin of using genetic data and biospecimens for research is informed consent—the mechanism used to exercise autonomy (38). Consent is a process intended to allow individuals to make decisions for themselves. A bedrock tool for fulfilling the principle of respect for persons as outlined in the Belmont Report, informed consent requires explaining the purpose, procedures, foreseeable risks, and benefits of the research. It should also include information about compensation and a statement that participation is voluntary and that refusal will not result in penalty or loss of benefits, such as healthcare. Different approaches to consent address the varying contexts in which genetic research is identified and used. These approaches can include a combination of opt-in, opt-out, broad, specific, tiered, and dynamic consent (Figure 2). Each differs in the action required of individuals and the potential scope of activities for researchers to use genetic data and biospecimens without a participant’s permission.

Figure 2.

Approaches to obtaining informed consent in human subjects research may be broad, specific, tiered, or dynamic.

In offering individuals a choice to participate or refuse participation in research, researchers may ask individuals to actively accept or decline including their genetic data in research activities. If a study enacts an opt-in approach, individuals will be asked to affirm that they are willing to be included actively; if they do not, they will not be enrolled in the proposed research. Alternatively, an opt-out approach assumes that individuals will be included in a set of research activities unless they actively decline and refuse participation. There are significant trade-offs associated with opt-in versus opt-out approaches. Requiring active affirmation of willingness to participate adheres more closely with the principle of autonomy. However, opt-in may demand more time and resources and decrease sample sizes in comparison to research using an opt-out approach (39).

Broad consent attempts to address the spectrum of potential uses of biospecimens and data by describing this in general terms and asking participants at the time that genetic data are collected. By not stipulating specific future uses, broad consent allows researchers greater latitude in the use and sharing of participants’ genetic data, often for indefinite periods and without limits on research questions that researchers themselves may not have anticipated at the time of recruitment (40, 41). This approach allows for greater flexibility for investigators and biospecimens, and researchers can use data as long it passes IRB approval (40, 42, 43).

In contrast to this approach, specific consent describes in detail the purposes for which collected biospecimens and genetic data will be used, often stipulating with whom the data will be shared and the time frame of the intended research. To use biospecimens or data for purposes beyond what has been outlined in the specific consent would require recontact with the participants and a separate consent process. Besides the added time and resources needed for recontact, preserving links between data and contact information may present privacy risks. Recontacting participants for consent would also almost certainly decrease the dataset due to the challenges of recontacting participants.

The approach of tiered consent attempts to provide more control to participants by providing a palette of options from which a participant can select research uses that align with individual values and comfort. Participants are asked for consent once for the specific study in which they are being recruited, and they can also choose one or more broad fields of research that would be allowable in the future. In addition, participants may be asked for permission to have their biospecimens and identified data used in research that may be ultimately commercialized. Tiered consent may also allow participants the flexibility to offer specific levels of consent for various activities (44). This information may include research questions for which the genetic data may be used, who has access to the genetic data, and permission to recontact participants in future research projects (45).

Dynamic consent attempts to bolster study participants’ autonomy in real time by allowing them to decide whether their biospecimens and data can be used as studies are proposed. By taking this approach, investigators recognize that participants’ support will vary for different types of research by different actors for different purposes. By allowing participants to review uses of genetic data as they emerge in real time, investigators that offer dynamic consent acknowledge that some participants may want to have more control over their data than broad consent would provide. In contrast to specific consent, dynamic consent offers investigators the flexibility to approach participants about research uses not anticipated at the time of recruitment (46).

Dynamic consent requires interactive relationships between investigators and participants in which communication is essential. This information may include ongoing updates on who is using the resources, for what purpose, and whether there are findings that have resulted from the studies. This approach departs from a single-point-of-contact model of consent; instead, consent is an ongoing process in which initial recruitment activity is but one step in building a long-term relationship. As such, dynamic consent requires greater oversight for the donor or their representative and assures the researcher that appropriate consent has been provided, particularly if the research protocol is amended over time. The key to dynamic consent is leveraging online tools and platforms to support ongoing decision-making by study participants. This approach not only involves supporting the ability of study investigators to maintain contact information with participants, but also requires creating an infrastructure that ensures participant biospecimens and data will not be used when consent is not given for future research, i.e., the ability to exit from a study.

4. EMERGING APPROACHES TO CONSENT

Sociotechnical developments, such as the growth of social media and crowdsourced data collected through smartphones and other health apps, have converged to fuel new approaches that offer greater fluidity in the consent process. These innovations aim to address challenges for providing understandable, timely information that allows individuals to fully exercise decision-making authority over their level of participation in research. Groups such as Sage Bionetworks have developed digital e-consents that can be administered through a smartphone or other digital device (47). This approach enables prospective participants to self-navigate through a series of landing pages at their own pace. Information is offered through e-consent screens that contain visual icons, short videos, and concise, highly structured text blocks that simplify key features of program participation (48).

Consent relies on comprehension and informed decision-making. Because genomic research is complex and likely to be unfamiliar to many people, researchers have developed methods for more effectively conveying key concepts, such as heritability, mutation, and risk. The use of visuals in addition to plain language text have been shown to improve participant understanding (Figure 3). For example, in a study of four content-equivalent informational aids (animated videos, slideshows with voice-over, comics, and text) and one no-intervention control, it was found that multimedia aids, such as voice-over with visual reinforcement, can improve participant knowledge more effectively than text alone (49). Other innovations are user friendly digital interfaces that allow for explanatory interactions and for users to review information when they have questions (50). To ensure that individuals understand the nature of participation, researchers have integrated formative evaluation as part of the consent process to reinforce essential concepts and target anticipated common misconceptions (51).

Figure 3.

The learning health system, in which healthcare organizations collect and analyze patient, provider, and other data in real time to improve patient care. Image courtesy of “The Values Project” video series; copyright 2015 Booster Shot Media.

Metaphors are often used as a valuable communication tool because they can link abstract, unfamiliar concepts to concrete, familiar experiences. However, the choice of metaphors requires careful consideration to engender trust, especially among populations underrepresented in genomic research. Ideally, metaphors would convey the most salient information, be unbiased normatively, and describe equivalent meanings across culturally diverse groups.

In our study of public attitudes toward genomic research, our team investigated the meaning and valence of metaphors to convey information about biobanking (52). In support of this research, we developed informational trigger videos about biobank research (http://boostershot.net/values/) to examine how individuals across diverse cultural backgrounds (including native English, Spanish, and Mandarin speakers, who we engaged in these languages) perceive related metaphors. When we asked patients to tell us what the word “biobank” made them think of, they offered several different terms that emphasized different attributes of biobanks. These terms included “financial bank,” “gold mine,” “organ or blood bank,” “cryobank, database,” and “large computer,” which vary in how well they represent or emphasize important concepts of a modern biobank (conceptual accuracy); the terms also vary in whether and to what degree they carry positive or negative associations. This work underscores that the choice of metaphorical terms requires careful consideration to engender trust, especially among populations underrepresented in genomic research.

5. GROUPS AS A UNIT OF ETHICAL ANALYSIS

Although the Belmont Report and its instantiation of ethical principles have been adopted in institutional oversight of federally funded research, it is essential to note that consent approaches have differed across disciplines and fields. These different orientations bear on developments in human genetics and data linkages in important ways. Answers to questions about who should consent, what constitutes consent, and when consent is given may depend on disciplinary background. Whereas consent was first formalized in the biomedical domain with an explicit focus on the individual as the fundamental unit of analysis, researchers in social sciences, such as anthropology and sociology, have long considered issues of consent in terms of groups and communities that are the focus of their research. The notion of group consent is based on recognizing that research on groups can harm individuals and the groups with which they identify. Current ethical infrastructure fails to address this critical social unit of analysis. Gaps include GINA and HIPAA, which apply to discrimination against individual members of a group but do little to protect groups as a category (53).

History provides many cases where individuals have been recruited because of their identity or association with groups that were the focus of research and practice. In addition to the Tuskegee Syphilis Study, other racially and ethnically identified groups involved in biomedical research have been exposed to the potential for group harm. For example, sickle cell testing of African Americans by airline companies has resulted in the firing of African American employees (54), and research on alcoholism among Indigenous populations has created the potential for stigmatization (55).

The Havasupai Tribe case and their lawsuit against Arizona State University in 2010 serve as a cautionary event that reveals the stakes for researchers, communities, institutions, and science to recognize the importance of going beyond the individual in governing the ethical conduct of research. In this case, researchers at Arizona State University had engaged the Havasupai Tribe about a study of the genetic contribution to diabetes in 1990. The study involved collecting 200 blood samples from tribal members as well as other health information (56). Whereas participating tribal members understood the purpose of the study to be focused on diabetes, the consent form used the broader language of “causes of behavioral/medical disorder” (57, 58). The conflict between the tribe and the researchers emerged when tribal members discovered that their samples were used in research on schizophrenia, inbreeding, and evolutionary and migration studies. Among the allegations, the tribe sued the researchers for breach of fiduciary duty, asserting that the researchers had not protected the interests of the tribal members.

Despite case law set by rulings such as Moore v. Regents of California that stipulate that individuals from whom biospecimens are collected do not retain property interests in these tissues (59), regulations are murky in providing clear guidance on the limits of unspecified future uses that broad consent suggests. In the Havasupai case, the university agreed to return biospecimens to the tribe and paid $700,000 to settle legal claims made by the tribal members (60). Scholars have noted the long-term implications of the case include mistrust between researchers and Indigenous populations that has a palpable impact on genomic research and goals of diversity (58).

Bioethics scholars have raised concerns over the shortcoming of current governance structures to recognize and address the potential for group harm in genetics research (61–63). Genetic information impacts the individual from which the biospecimens and data are derived but has implications for family members and populations linked genetically. This risk was evident when two scientific teams published the genome of the HeLa cell in 2013 after the cell line had been identified with Henrietta Lacks. The scientists did not seek permission from the Lacks family to disclose results from the genome sequence—an act that could also reveal information about the Lacks family members (35, 64–66). In an unusual step, the NIH in consultation with the Lacks family created the HeLa Genome Data Access Working Group made up of scientists, bioethicists, and others to review requests to use the genomic data. In addition, the NIH required that researchers that use the data include specific acknowledgment of Henrietta Lacks and her family in their publications. This novel approach attempts to honor the source of the genomic data and to provide some degree of oversight on how the data are ultimately used (64, 67, 68).

An ongoing challenge in genomic research is how to recognize the interests of groups. Innovations in informed consent have focused on methods for fulfilling the principle of autonomy and less on the impact of genomic research on larger social units. Group consent in the context of Indigenous groups has been a source of debate in several research initiatives. These include those involving the collection of genetic data and biospecimens from populations, such as the Human Genetic Diversity Project (69–71) and the All of Us Research Program (72, 73). Governed by international law, Indigenous groups in the United States are recognized as sovereign nations that have inherent rights to self-determination. Researchers who seek to recruit Indigenous populations should address in detail the types of biospecimens and data that will be collected, research uses, data access and security protocols, and risks and benefits to tribal members and create processes for ongoing review of studies and their products (74–76).

Genetic data have implications for groups, such as families and communities, that extend beyond the individual. However, in the absence of recognized legal standing, scholars have pointed out the difficulty in determining who has the right to represent a group, much less consent on its behalf (77). Engaging communities with a stake in the use and potential misuse of data is important and requires innovative approaches to the consent process (78).

6. AUTONOMY AND PUBLIC GOOD

In 1998, Helen Nissenbaum noted that “[t]here is growing awareness as well as resentment of the routine practice of recording, analyzing and communicating information about individuals as they act and transact in the normal course of their commercial and public lives” (79, p. 559). Over two decades later, Nissenbaum’s insight is as pertinent as ever. The public is bombarded with social media apps that collect a wide range of personal data, including purchase history, motion and geolocation data, food intake, and emotional states (80). A growing range of actors in government, business, healthcare, and research collect information on individuals’ public and private activities for an expanding scope of uses that include public health surveillance, product development, quality improvement, and basic science research. Increasingly, these data are linked in unexpected ways and often invisibly, unbeknownst to individuals (81).

In the research context, the Commission in the Belmont Report drew a sharp line between biomedical or behavioral research involving human subjects and routine clinical medicine. The Belmont Report defined research as “an activity designed to test a hypothesis, permit conclusions to be drawn, and thereby to develop or contribute to generalizable knowledge” (28, p. 3). Clinical practice was defined as “interventions that are designed solely to enhance the well-being of an individual patient or client and that have a reasonable expectation of success” (p. 2). In the current landscape, what constitutes research can be murky, as data are linked in ways that are unexpected and, at times, invisible to the donors of the data. Data collection in what has been referred to as the learning health system reflects the troubled boundaries between research and clinical care and presents challenges to consent. As a case study, data collection in the learning health system surfaces many questions about the extent to which institutions can use genetic data and biospecimens collected in usual care in research activities.

The learning health system refers to healthcare organizations that continuously collect and analyze patient, provider, and other data to impact outcomes. The Institute of Medicine describes it as an approach by which science, informatics, incentives, and culture are aligned for continuous improvement and innovation, with best practices seamlessly embedded in the delivery process and new knowledge captured as an integral by-product of the delivery experience (82, 83). In 2011, the Institute of Medicine published a roadmap to guide the development of the learning health system (82). Data collection activities in the healthcare system have expanded in many ways. In this iterative cycle, the learning health system will use a range of activities, including intelligent automation, comparative effectiveness research, positive deviance, surveillance, predictive modeling, and clinical decision support. Some view the learning health system as a new approach to improved public health. Khoury et al. underscore this in their emphasis on a “radical transformation” on a new population-level surveillance approach that “will jump start accelerated use of tools to improve availability, quality, and timeliness of data, and linking public health data more effectively with clinical systems” (84, p. 401).

At the same time, consent presents a challenge for the learning health system. Some predict that requiring patients to actively consent or opt-in to data collection in routine clinical care for research purposes would decrease and create biases in the data. Arguing that the systematic collection of data and biospecimens in the learning health system would reduce mortality due to medical mistakes and inefficiencies in healthcare spending, some claim that consent may not be warranted. Suggesting that conceptions of clinical and research ethics are obsolete in relying on a sharp distinction between research and practice as outlined by the Belmont Report, these individuals argue that requirements for consent in the learning health system demand a reconsideration of what is deemed acceptable measures of protection and must be weighed against the moral imperative to integrate research and practice for the collective good (85, 86).

Ruth Faden et al. write,

the dominant ethical paradigm from the 1970s to the present time is antithetical to and problematic for the learning health care system, at a time when clinical practice is far from optimal and learning to improve care is sorely needed. Several hundred thousand people die needlessly each year from medical mistakes. There is reason to believe that adult patients receive only approximately 50% of recommended therapies and that up to 30% of health care spending is wasted.

(87, p. S16)

They argue that the current ethics paradigm that requires explicit consent for data may hinder improvement.

As such, Faden et al. suggest that ethical protections should be realigned around what they term the “moral priority on learning.” This realignment necessitates a set of new obligations for all actors in the learning health system. A presumption of responsibility extends not only to health professionals and healthcare institutions but also to patients, who they contend have a duty to contribute to the ongoing learning that is integrated with the healthcare they receive. They underscore this realignment as one that is needed to address unjust inequalities in healthcare that only ongoing, systematic data collection can address. The implications of this reframing are that patients become de facto research participants obligated to contribute to knowledge as beneficiaries of the healthcare system. They write:

Traditional codes, declarations, and government reports in research ethics and clinical ethics have never emphasized the obligations of patients to contribute to knowledge as research subjects. These traditional presumptions need to change. Just as health professionals and organizations have an obligation to learn, patients have an obligation to contribute to, participate in, and otherwise facilitate learning.

(87, p. S23)

In defending this “obligation to learn,” they call upon John Rawls’s principle of the common good in asserting a norm of “common purpose” and argue that

The common interest of members of a society in the health care system is that it be positioned to provide each person in the society with quality health care at a cost compatible with individual and societal economic well-being. We also have a common interest in supporting just institutions, including activities that reduce the unjust inequalities …Securing these common interests is a shared social purpose that we cannot as individuals achieve. Our goals cannot be reached efficiently without near-universal participation in learning activities, through which patients benefit from the past contributions of other patients whose information has helped advance knowledge and improve care.

(87, p. S23)

While Faden et al. acknowledge that not all data collection activities are the same (for example, they affirm that randomized, controlled trials of an investigational new device should only include patients who have consented), their proposal suggests an erosion of requirements of consent when explicit hypotheses are absent and data are collected for their potential to answer questions in the future—as nonspecific and continually evolving as this process can be. Instead, patients are not only expected but also obligated to participate because of a common interest in the general improvement of health. Citing philosopher David Hume, they write that the moral underpinning for this duty of beneficence is based on the assumption that

All our obligations to do good to society seem to imply something reciprocal. I receive the benefits of society, and therefore ought to promote its interest…the discharge of obligations of reciprocity occurs through an established practice of making an appropriate and proportional return—returning benefit with proportional benefit, with all alike sharing, as a matter of moral obligation, the burdens necessary to produce these benefits.

(87, p. S23)

It is unclear the extent to which the normative claim of a moral obligation to share data for the good of society is embraced. In the context of the global COVID-19 (coronavirus disease 2019) pandemic, resistance against public health interventions, such as masking, quarantining, and vaccinating, as necessary measures to mitigate mortality, challenges a shared vision of the common good. The lack of transparency and leadership that has resulted in misinformation about the pandemic has eroded public trust in scientific expertise and has resulted in an unwillingness to comply with public health authorities. For example, anti-quarantine groups who protested the governmental requirements to socially distance and remain at home claimed that these public health measures violate their individual right to congregate and work. The learning health system model relies on a strong sense of group identification based on shared values and interests. Yet public behavior in the context of COVID-19, a disease that has resulted in hundreds of thousands of deaths in the United States, challenges notions of public willingness to trade individual autonomy for the health of others and suggests that common interests should not be assumed.

Furthermore, the emphasis on benefit-sharing as a justification for duty and obligation is problematic given what we know about the distribution of benefits in our healthcare system and the outsized burden and risk of data collection for marginalized populations. As Faden et al. recognize, the compartmentalization of research and clinical care is increasingly difficult to maintain; so, too, is siloing and regulating of genetic data and biospecimens within strict domains of health and nonhealth. With the blurring of these domains and the travel of data in and out of healthcare systems, fundamental questions of who gets access to datasets, for what purposes, in what contexts, and with what constraints are left unanswered. Furthermore, regardless of shared common interests, it is unclear whether there is equitable distribution of the common good to all would-be individuals and communities. For example, algorithms used to make distinctions in risk ratios are by and large covered by an opaque veil of secrecy backed by corporate claims of trade secrecy and intellectual property. In the genomic and big data environment, algorithms deliver practically incontestable results.

7. RELATIONAL ETHICS AND PUBLIC VALUES

Understanding participant perspectives of institutional efforts to collect genetic, medical, behavioral, and environmental information for research is critical to creating consent and oversight mechanisms that address public values. Although empirical research of public attitudes is limited, several studies of participants’ willingness to provide health information that includes genetic data indicate a desire and expectation for transparency about the research: What is the purpose? Who are the investigators? Where will it be conducted? Who has access to the data, and when and how would the data be stored? Not surprisingly, studies suggest that those with less familiarity and knowledge of the biomedical research process and more questions about data safety are less likely to participate.

Studies also suggest that patients have a strong preference for control over how their data and biospecimens will be used and by whom. In their systematic literature review of attitudes towards biobanking, broad consent, and data sharing, Garrison et al. found that many people do not favor broad consent for research and subsequent wide data sharing (41). They found that most people preferred tiered or specific consent over broad consent when choices were offered. Furthermore, while individuals were generally willing for data or biospecimens to be shared with other academic researchers, individuals were less inclined to share their data in federal databases or with commercial entities.

In our recent study of attitudes on the collection of genetic data and EHR data for research among individuals who identified as African American, Hispanic, Chinese, South Asian, and non-Hispanic white, our findings underscore the importance of trust for patients contemplating participation in genomic research and the need for public discussion about the collection of genetic data as compared with other patient information (88). Specifically, our findings suggest a range of attitudes that require further exploration to develop effective approaches toward building trust and institutional trustworthiness in genomic research. For example, identifying how institutional policies on storage and distribution of biospecimens affect trust and perceptions of protection of group identity and values is essential to productively engaging with diverse populations.

The potential for genetic data collected in the context of health research to be used for non-health-related purposes challenges principles of autonomy and threatens individuals’ right to exercise control over the use of their genetic data and biospecimens. This risk can be a particular concern for genetic data collected under the general rubric of health research and nevertheless used in what some would identify as nonhealth areas of research. An example is the use of the UK Biobank’s large-scale dataset in genome-wide association discovery research to identify genetic variants associated with same-sex sexual behavior (89). The researchers of this study engaged LGBTQ advocacy groups to discuss the implications of the research; however, some have questioned the ethics of conducting such research. According to the Common Rule, the deidentified data used for the research entails that the study qualifies as nonhuman subjects research and is exempt from regulatory oversight. However, scholars have pointed out the limitations of broad consent, suggesting that the research fails to meet the principle of beneficence and may even present harm to members of a group who identify as LGBTQ (90).

Given social inequities in power across US groups, these activities may have a differential impact across subpopulations, with harmful consequences that are difficult to predict in advance. The history of ethical violations related to protocols for inclusion in biomedical research and the continued misuse of research results [such as white nationalists looking to genetic ancestry to support claims of racial superiority (91–93)] continue to engender mistrust. The COVID-19 pandemic puts in stark relief enduring disparities in disease burden and access to healthcare among Black, Indigenous, and other people of color and the potentially devastating impact of a lack of transparency on public trust in science (94, 95).

Approaching genetic research as a form of gift-giving frames research in terms of relationships guided by moral responsibility. To accept a gift and to acknowledge it as such is to enter into commitments of a moral relationship that depends on fulfilling ethical principles of autonomy, beneficence, and justice (96, 97). Greater transparency that enables inspection of research uses and actors must be part of a larger public conversation about what constitutes fairness and equity. Ensuring mechanisms for autonomy and the right to consent and refuse can bolster individual choice and liberty (98).

8. RECOMMENDATIONS

Current consent models aim to exercise bioethical principles that have emerged from a particular sociopolitical context and history of biomedical research involving individuals and groups. This article describes consent for collecting genetic data and approaches for addressing ethical considerations in balancing privacy protections, risks, benefits, and fairness and equity in genetic research. The shifting ecosystem in which genetic data are linked to a variety of other personal information creates new challenges for meeting these ethical obligations. Of particular concern is the reliance on the individual as the unit of analysis for ethical regulation, which threatens to obscure the implications of genetic research on groups. The shifting genomic ecosystem demands careful consideration of consent and how this process can address how, when, and why genetic data are collected and shared by various stakeholders. The recommendations below aim to support these goals.

Cultivating long-term, trusting relationships with participants underrepresented in biomedical research requires that researchers: (a) address the role of history and experience on trust, (b) engage concerns about potential group harm, (c) address cultural values and communication barriers, and (d) integrate patient values and expectations into oversight and governance structures (99, 100). Creating a responsive, ethical architecture that focuses on groups, inequalities, and the distribution of benefits is critical to research that aims to collect genetic data and biospecimens. It requires a calculus that balances privacy protections with the potential benefits to those who agree to participate in research. To achieve this, researchers and institutions must create innovative mechanisms that integrate individuals’ preferences and consider the implications of research for groups (Figure 4).

Figure 4.

Recommendations for obtaining consent in a shifting genomic ecosystem.

FUTURE ISSUES.

1. How can the balance of power between research subjects and researchers shift to allow for greater trust and equity?

2. What do stakeholders view as appropriate trade-offs between individual privacy and public benefit, and how should these interests be balanced?

3. How can the consent process allow participants to exercise their values by opting in or out of research in real time?

4. How will the erosion of boundaries between health and nonhealth research impact participants’ control over how their data are shared?

5. How can potential benefits of research be addressed on balance with risks and potential harms?