Traditional standards of informed consent in research require “that subjects enter into research voluntarily and with sufficient information about the research activity to make a comprehending decision about participation.”Reference McGuire and Beskow1 Federal regulations state that the basic elements of informed consent include information about the study purpose; procedures (i.e., biospecimen and information collection, storage and access, re-contact and data-sharing); risks and benefits; protections for privacy and confidentiality; plans for costs, payment and commercialization; plans for sharing results and disclosure; and procedures for withdrawal.2 Biobanking and the creation of other multidimensional data repositories introduce new challenges to this standard of informed consent.Reference Faden, Beauchamp and Kass3 In biobanking, biospecimens are stored and shared to investigate research questions that are not determined at the time of data collection, and plans for commercialization and data sharing may be difficult to articulate in advance. Large-scale research infrastructure projects, like the All of Us Research Program (AoURP) (a government supported research initiative that aims to enroll 1 million people) and Project Baseline (a corporate sponsored project to study a cohort of 10K), also aim to store multiple types of data (e.g., genetics, physical measurements, self-reported behaviors, electronic health records) collected from a large number of participants for future unspecified research uses.Reference Sankar and Parker4 Potential participants may be asked to grant permission to have their data stored at a time when study recruiters will find it “impossible to describe in detail or even to foresee all of the future research for which they might be used.”5
Recently, bioethicists have raised fundamental questions about whether and through what mechanisms current informed consent procedures can accommodate new research models that feature long-term storage and use of data. Reference Budin-Ljøsne, Teare, Kaye, Burke, Beskow, Trinidad, Fullerton, Brelsford, Sabatello, Appelbaum, Murphy, Scott, Kaufman, Geller, LeRoy, Hudson, D'Abramo, Sheehan, Khodyakov, Mendoza-Graf, Berry, Nebecker and Bromley6 In a recent mixed-method examination of researchers' views about the importance of understanding during consent for biobanking, Beskow and WeinfurtReference Beskow and Weinfurt7 identify disagreement among experts about whether participants with gaps in understanding should be allowed to consent for biobanking. Given that many of the experts did not see poor understanding as barring participation, Beskow and Weinfurt conclude that informed consent may need to be re-conceptualized, with understanding considered an aspiration rather than a requirement.Reference Wieten, Blythe and Magnus8 Many of the commentaries accompanying Beskow and Weinfurt's article appear to reconsider the role of information exchange in consent. As Ellen Wright Clayton says, “It is time to think more clearly about what we hope to achieve with informed consent.”Reference Clayton9
This paper contributes to this reconsideration by focusing on research stakeholders' views of a model of consent appropriate to biobanking and research infrastructure initiatives like the AoURP and Project Baseline (hereafter referred to as health data repositories) that entail long-term storage and undetermined future research use of multiple types of health data. We conducted semi-structured interviews with a variety of research stakeholders to elicit their views of a model of consent appropriate to health data repositories. We explored their views of the function of consent for participation in health data repositories; and we examined their views of the overlap and divergence between health data repository and traditional biomedical research consent. In particular, we explored research stakeholders' views of the role of information exchange in consent and challenges to consent that arise when there is limited certainty about the future purposes, procedures, and risks of a health data repository study at the time of enrollment.
This paper contributes to this reconsideration by focusing on research stakeholders' views of a model of consent appropriate to biobanking and research infrastructure initiatives like the AoURP and Project Baseline (hereafter referred to as health data repositories) that entail long-term storage and undetermined future research use of multiple types of health data. We conducted semi-structured interviews with a variety of research stakeholders to elicit their views of a model of consent appropriate to health data repositories. We explored their views of the function of consent for participation in health data repositories; and we examined their views of the overlap and divergence between health data repository and traditional biomedical research consent.
We demonstrate that, when considering health data repositories, research stakeholders refine and reframe their concept of consent. Rather than describe consent as informed, they describe a concept in which consent is engaged. Stakeholders' model of engaged consent emphasizes the importance of the participant's ongoing relationship with the study as a substitute or adjunct to information exchange at the time of enrollment. Moreover, we demonstrate the ways in which research stakeholders' views of the need for re-consent are shaped by their concerns about limitations of information exchange at enrollment. By focusing on stakeholders' understandings of risks and ethical obligations in consent for participation in health data repositories, we outline some contemporary uncertainties about the functions and requirements of consent for initiatives that aim to store data for unspecified research purposes.
Background
Informed consent is the essence of autonomyReference Capron10 or respect for persons, a key ethical principle in the Belmont Report that remains a foundation for contemporary research ethics.11 A large empirical literature describes the extent to which individuals understand information presented during the informed consent process,Reference Sugarman, Paasche-Orlow, Flory and Emanuel12 as well as strategies for improving comprehension.Reference Del Carmen, Joffe, Goldberger, Kruse, Kadish, Passman, Bergner, Palmer, Cassidy, Dunn, Spira and Sheikh13 As procedures for long-term storage of data for unspecified research purposes have developed over the past few years,Reference Patil, Majumdar, Awan, De Souza and Greenspan14 bioethicists have raised questions about whether and in what way individuals should consent for initial donation of data or biospecimens to repositories and for re-use of data. A large national survey of adults in the U.S. found that a majority of respondents were supportive of large-scale research on genes and environment, with the opportunity to learn more about one's own health appearing to be a strong motivator for participation.Reference Kaufman, Baker, Milner, Devaney and Hudson15 Yet bioethicists have raised difficult questions about which informational features must be included in consent for health data repositories and how those features should be implemented.Reference Lemke, Halverson, Ross, Burke, Antommaria, Bennett, Sanderson, Brothers, Mercaldo, Beskow, Dombeck, Thompson, Watson-Ormond and Weinfurt16
The revised Common Rule for the protection of human research participants incorporates into regulations a model of broad consent. 17 Broad consent provides “consent for an unspecified range of future research subject to a few content and/or process restrictions.”Reference Grady, Eckstein and Berkman18 The U.S. Health and Human Services Secretary's Advisory Committee on Human Research Protections (SACHRP) states that broad consent “may be the most suitable pathway for research involving identifiable private information or identifiable bio-specimens” and can function as a “substitute for traditional informed consent in certain circumstances.”19 In this broad consent framework, participants receive sufficient information about potential study purposes and procedures, foreseeable risks and benefits, and plans for use of identifiable information or specimen and information sharing.20
For a variety of reasons, policy guidelines may not settle questions about information exchange during consent to health data repositories. First, the regulatory version of broad consent differs in certain ways from how broad consent has been discussed and used in practice, and other models of consent have been proposed to ensure that the rights of donors are upheld when donating data to a health data repository.Reference Nanibaa'A, Sathe, Antommaria, Garrett, Dohan and Koenig21 Various models of consent — blanket consent, layered consent,Reference Bunnik, Janssens and Schermer22 tiered consent,Reference Lentz, Kennett, Perlmutter and Forrest23 dynamic consent,Reference Kaye, Whitley, Lund, Morrison, Teare, Melham, Wee, Henaghan and Winship24 and partnered consent25 are just some examples — provide a plan for information exchange while highlighting its limits. For instance, in a blanket consent model, participants donate their data and samples with no ability to limit their uses for future research,Reference Wendler26 an approach that some say undermines participant autonomyReference Caulfield27 and under-communicates study risks.Reference Greely28
Moreover, the Common Rule's broad consent requirement that participants receive sufficient information about purposes and procedures, risks and benefits, and plans for use and sharing of data may not be achievable at the time of consent to participation. The necessary components of a broad consent process remain unsettled, and SACHRP “favored allowing institutions to create their own broad consent forms that could be tailored to a variety of circumstances.”29 Barriers to fully informing participants at the time of consent for health data repository studies have been long-recognized and include overestimation of personal benefit and underestimation of risk.Reference Klima, Fitzgerald-Butt, Kelleher, Barr, McCarty, Nair, Austin, Giampietro, Ormond, Cirino, Helenowski, Chisholm and Wolf30 Moreover, data-banking studies often involve multiple researchers, sites, and projects, complicating dialogue with participants as new uses of the data or new plans for return of results emerge.
Recently, functions of consent beyond informing have received attention from bioethicists. Recognition of limitations of information exchange during consent has brought other key functions of consent to the fore.Reference Dickert, Fehr, Llanos, Scicluna, Samady, Mandava, Pace, Campbell, Emanuel and Grady31 For instance, Dickert et al. emphasize the fact that, even where understanding remains limited, consent processes promote “transparency, integrity, and trustworthiness” between researcher and participant; these functions of consent, “each have independent ethical value.”Reference Dickert, Eyal and Goldkind32 Brudney emphasizes that consent is best thought of not as a means of promoting autonomy but of contributing to authentic decision-making.Reference Brudney33 And, Faden et al., arguing that consent processes can over-protect patients from research that could benefit them, suggest using strategies that prioritize transparency and trust instead of traditional consent procedures.34 In this context of change and reconsideration, we conducted a broad exploration of the functions that research stakeholders ascribe to consent, the aspects of consent they see as critical, and the aspects of health data repository consent that raise concerns about protection of participant autonomy.
Methods
We conducted a series of semi-structured telephone interviews with a diverse group of 44 experienced research stakeholders about their perspectives on consent for large-scale health data repositories. Our goal was to solicit the perspectives of stakeholders with previous experiences with biobanks, longitudinal studies, human subjects' protection, data privacy and security, and stakeholder-engaged projects. We sought interviewees representing universities, health provider organizations, community organizations, non-profit research institutions, and private companies. We aimed to enroll a purposive, variant sample of experts in the field to capture a wide range of contemporary views about consent for health data repositories. Such a sample is useful to expand the range of differences of opinion that may emerge in the data.
To assemble a diverse yet information-rich sample, we reached out to (1) authors of recent publications on ethics of genetic research, (2) individuals involved with commercial and research biobanks, (3) the members of relevant AoURP working groups, and (4) individuals representing community-based organizations attending stakeholder engagement conferences about the AoURP. Because it is one of the largest and most diverse health data repository programs — both in terms of involved institutions and participants — we assumed individuals involved with AoURP would have highly salient perspectives. We tracked enrollment as it proceeded and adjusted sampling to ensure that we included perspectives from all of these groups.
A team of three experienced qualitative researchers conducted all interviews by telephone using a semi-structured interview guide. It included open-ended questions about the enrollment and consent of participants, approaches to consent for a project like AoURP including broad consent, pros and cons of different consenting approaches, and the need for re-consent. Questions included, “What may motivate a large and diverse group of individuals to enroll in a health data repository? What does broad consent mean in the context of AoURP or related initiatives? That is, what are people consenting for? What are the benefits and drawbacks related to the reliance on broad consent? What are the situations where you think additional consent should be required for studies that plan to use existing de-identified data? How might the use of broad consent affect participants' interest in, enrollment in, satisfaction with, and retention in the repository?” All interviews were audio recorded and transcribed verbatim. Institutional Review Board approval for this study was obtained from the RAND Human Subjects Protection Committee.
All researchers jointly developed the codebook, which included codes derived deductively from interview topics and inductive categories describing different views of consent that emerged during data collection. The codebook included codes for the goals, meanings, and implications of consent; the need for modifications to consent procedures to ensure ethical study procedures; the role for updates to consent or follow up consenting; participants' motivations for consenting; ethical concerns about different consent approaches and solutions to address them; and implementation challenges of different consent approaches and solutions to address them. After the main codes were developed, two transcripts were coded separately by the two coders and compared. Disagreements were discussed until consensus was achieved. Next, one coder coded the remaining interview transcripts. A second coder then checked text marked with categorical content codes (i.e., need for re-consent), found no disagreements with content codes, and tallied counts. Both coders re-reviewed coded text, identified the most salient themes, looked for cross cutting themes across interviewees, linked related concepts, and searched for stakeholder differences and similarities in perspectives. All three researchers reviewed data summaries, offered revisions and refinements to thematic descriptions, and finalized selection of illustrative quotes.
Results
The sample is described in Tables 1 and 2. We approached 77 individuals to yield a sample of 44 (57%) who completed a phone interview. Females were more likely than males to respond to our interview invitation (65% vs. 37%, p=.019), and we interviewed more project administrators than Ph.D.-level researchers. Roughly one-third of our participants were employed at academic institutions, and one quarter were employed by a health provider organization. Interviewees are identified by a study ID as well as a code for employment sector (Health Provider Organization=H; University=U; Non-Profit Research Organization=NP; Private Research Organization=PR; Community-Based Organization=C); interviewees highly-involved in the AoURP are identified with a plus (+) symbol.
Table 1. Interviewees' Place of Employment
* Includes organizations that offer specialized research expertise in areas such as infrastructure support, marketing and communication, or data management
Table 2. Sample Characteristics
* Defined as participating in 2 or more AoURP working groups
I. Functions of Consent to Participate in a Health Data Repository
Interviewees described consent to a health data repository as a process eliciting participant agreement to evolving uses of their data as overseen by an institutional infrastructure that stewards that data over a long period of time. Instead of emphasizing the informing function of consent, interviewees suggested consent could be informed enough: it could provide a general overview that lacks many details. These aspects of consent as described by interviewees are detailed below.
Evolving Uses of Data:
Interviewees contrasted traditional informed consent, in which participants agree to a specific study purpose, to consent in a health data repository context, in which participants consent for use of their data for evolving purposes. While broad consent has been described as allowing future, unspecified, or indeterminant uses of data, we call these uses of data ‘evolving’ to reflect interviewees' emphases on data being relinquished to researchers' emergent choices about study questions. Interviewee 14 (NP) said, “Essentially, an individual consents at the outset that their data can be used for purposes that are not yet established.” Another (ID23; U+) agreed that participants must know “that their data is going to be used for research and available to researchers for all kinds of research.” According to interviewee 26 (H+), participants are giving up their right to disallow particular uses of their data: “we make it clear if you're not comfortable with having your samples and data used for [a] specific kind of research, then you should consider not participating because literally it could be used for any and all research.” Interviewee 41 (U+) agreed that participants needed to know that they were giving up rights: “let's just really be absolutely clear that you are donating your data and your blood and your information, and we do not know what researchers are going to do with it.” These interviewees emphasized that participants relinquish their right to prevent data from being used. Interviewee 50 (U+) said that the consent “needs to cover … the future assay that can be done on the specimen.”
Other interviewees emphasized that consent to participation in a health data repository entails an agreement that their data will be put to use. Interviewee 43 (NP+) said that consent is an agreement to share data with “someone out there [who] will hopefully be using it…to help us understand illnesses and diseases better.” Interviewee 19 (U+) said that the purpose of a repository is discovery, which necessitates an exploratory use of the data: “You are telling them that the data will be used for every discovery that will happen in the future.” Interviewee 17 (PR) called the lack of specificity of the study purpose “part of the transparency” of a data repository study since “we're moving into a discovery phase, so we won't know everything that we're going to learn.” These interviewees framed consent not so much as an agreement to any subsequent use of data but to a purpose dictated by the imperative of scientific discovery.
Institution and Infrastructure:
Rather than describing consent as an agreement to complete specific study procedures, interviewees described participants as agreeing to be a part of a broader research process. They described consent as conferring decision-making authority on the institution or infrastructure that manages the health data repository. As interviewee 26 (H+) explained, “They're consenting to be part of a large longitudinal study that will engage with them multiple times” over an extended period. When asked to clarify what participants are consenting to, interviewee 25 (U+) said they “are volunteering to be a part of [a study] and contribute information…that could be used for any type of research for years and years to come.” Interviewee 13 (U) said participants agree to a method of collecting and storing data, including “how the study will operate,” or “the infrastructure of the study and how the study is set up,” which comprises “a process by which their samples and data are going to be collected, how they are going to be stored, what sort of confidentiality protections…and then this process by which decisions will be made about who can use data and for what purpose.” Interviewee 31 (H+) summarized that the consent “has to accurately describe… the rules of engagement” of the study. Interviewee 27 (H+) described a requirement that researchers “be up front with folks that … these are the way your materials will be used, these are the protections that are in place, and we're going to do our best to make sure that it's being used as wisely as possible.” Interviewee 14 (NP) agreed that consent processes entail “making sure that at the outset participants understand that if they participate, this is the framework that they are participating in.”
Informed Enough:
Interviewees consistently deemphasized the importance of information exchange and understanding at the time of consent to a data repository. As interviewee 26 (H+) said, for example, the recruitment process “is actually about getting people jazzed up and excited” about a long-term relationship with the study. Interviewee 50 (U+) said that a “constant information loop with participants” is a goal, rather than one-time conveyance of information, a procedure this interviewee called “really looking at consent in a different light.” Interviewee 50 (U+) said that while traditional consent consists of “’I'm explaining something to you, I'm going to ask you some questions, making sure that you understand, and I'm going to ask you to sign this piece of paper,’” in consent for a health data repository “the engagement of participants … has to be continuous.” Interviewee 50 (U+) continued that, “engagement is part of the consent, you know, a more modern, and better at that, consent process.”
Interviewees considered this kind of consent to be informed enough: they acknowledged that initial information might be incomplete yet sufficient to protect participant autonomy. Interviewee 14 (NP) felt that complete information about a study was not needed to achieve an adequate consent process: “the idea of knowing everything in order to be able to make an informed decision is a naive concept. We never know everything when we make a decision and that doesn't impede us in making autonomous decisions.” Interviewee 13 (U) said that study purpose, procedures, and risks and benefits could be communicated “in enough detail that people can make an informed decision about whether they are willing to agree to that.” Interviewee 38 (U+) described a reasonable compromise between feasibility and the need to inform. The consent process must “weigh the balances between protecting human subjects, but also the balance between making a consent form that's like 40 pages long that no one's going to sign or be a part of.” Interviewee 38 (U+) said, “I don't think there's anything wrong with a consent that…is probably more feasible and practical” than a comprehensive informed process. While these comments reflect a ubiquitous need to balance feasibility and information provision during consent, we note interviewees' consistent dismissal of the need to completely inform potential participants. What we here call an informed enough consent is, according to interviewee 38 (U+), “still protecting a human subject, rather than something that is so detailed or so far into the weeds.”
Despite overall endorsement of the informed enough approach, the most common concern expressed by interviewees was the risk of inadequately informing individuals during consent. Interviewee 21 (H+) agreed that “just conceptually it's hard for people to understand,” and as a result “our biggest hurdle is…feeling like our patients are truly consenting to the process.” Interviewee 54 (H+) called “the lack of specificity” in the consent a major source of “negative feedback.” Five interviewees said that inadequate information provided at the time of consent decreased participation rates. Each aspect of the broad consent requirement for sufficient information-exchange — study purpose, procedures, risks and benefits, plans for use of data, and information sharing — was mentioned as a potential source of inadequate comprehension.
For instance, regarding study purpose and procedures, interviewee 46 (H+) said that “people really don't know potentially what could be done with their data.” Interviewee 31 (H+) described that a consent for research on ‘health related’ topics is not well-understood by participants: “a lot of them…have [sic] very narrow understanding,” and may think that “health-related has to do with disease and not wellness, for example.” Interviewee 18 (U), noting that “it's one of the few things I am struggling with,” was worried that “the data can get sent out and studies that could be done on things that may not be compatible with someone's beliefs.” Interviewee 15 (NP) agreed that “I could see people not wanting the data to be used for a topic that they felt very strongly about.” Interviewee 13 (U) anticipated that if studies “are not in keeping with what people thought was going to happen, that that will be a major problem with trust in research and with people's willingness to continue.”
Interviewee 54 (H+) empathized with participants' wish for information: “I think they literally just want to know more specifics. They want more specificity of exactly, okay, well, you're taking six tubes of blood from me, what are you doing with six tubes of blood? What tests are you running?” Citing concerns that procedures were inadequately explained, interviewee 53 (U+) recommended “[r]ather than say, ‘Oh, we don't know what yet we can use your information for,’” the consent process should express “possible potential uses; list them all out and categorize them,” and afterward return to the participant to ask, “’Did I make myself clear?’”
II. Engagement as Protection Against Inadequate Information
Most interviewees identified participants' commitment to an ongoing relationship with the study as a key component of consent for health data repository research. As interviewee 27 (H+) explained, participants agree to “a long-term commitment, that people will … stay with the project.” In this sense, “consenting is not a one-time step, single piece of paper;” it is “an ongoing process.” Interviewee 27 continued that consent and engagement are synonymous: “whether you call it engagement, or you really understand the term consent … I don't think it makes a huge difference” because “consent should be an ongoing activity. It shouldn't be something you do once, and you're in.” While interviewees did not provide a consistent definition of engagement, the most common descriptions were of an enduring relationship in which communication deepened and developed over time.
Interviewees consistently described the purpose of engagement as mitigating limitations of information exchange at the time of enrollment. Interviewee 14 (NP) used the analogy of a marriage: individuals enter with incomplete information “but still feel like [they] made an informed decision … people can choose to participate in a research study that allows for broad reuse of their data and so long as in a healthy marriage, there is clear communication and engagement over time and the ability to check in and also to step out any time a person feels their needs are not being met then it's still ethically an okay arrangement.”
Interviewees' descriptions of engagement often included options to receive information that might shape participants' willingness to continue participation. For example, interviewee 50 (U+) said that, “the [health data repository] is going to be in contact with participants so that participants know what is going on in that [health data repository].” Within this engagement, “they have access to their own data. And they may decide at any time what they want, what their level of participation may be.” Follow up could remind the participant they can opt out: “You can revoke your authorization to be part of the program, you can change your mind, you can … opt out at any time.” Interviewee 36 (NP+) endorsed an approach in which participants can “have a voice about what we're studying or if they can initiate their own research.”
Re-consent:
For interviewees, options for re-consent presented ambiguous solutions to the informational deficiencies of consent for health data repositories. On the one hand, re-consent procedures might mitigate limitations in information exchange at the time of consent, strengthening participants' link to and trust in the project. On the other hand, re-consent procedures could undermine participants' commitment to a study by suggesting initial consent was insufficient. In this way, interviewees had mixed views of whether re-consent was a form of engagement or an impediment to it.
For example, a minority of interviewees were confident that no additional permissions for use of data would need to be gathered from participants. The eight interviewees who said that there would be no situations where additional consent would be necessary described re-consent as undermining participants' ongoing relationship with the study. For instance, one said that a requirement for re-consent “would mean that there's a huge barrier to doing research … with data from communities that are already underserved by biomedical research” (ID36; NP+); that “you've consented people under this broad framework”; and that any re-consent “changes the terms that you agreed to initially” (ID14; NP). Three interviewees said that a re-consent process could heighten participant anxiety. It may, interviewee 17 (PR) said, attenuate engagement because it is a form of “bothering them about consenting and re-consenting” and “makes them feel like somehow everyone else is trying to protect themselves from something.” Interviewee 16 (U) agreed that “it's going to prompt people to say, ‘wait, what are they doing with my medical records?’” By going back to participants, interviewee 32 (U+) said, “you really run the risk of having people not want to continue” to participate.
Yet 23 of 44 of interviewees felt that some additional process of consent should be a part of participant follow up. Nine interviewees said that further consent would be necessary if new types of data were to be collected or analyzed in the future, for instance for study procedures that are “more high risk” (ID38; U+), in which “a different type of procedure … carries more risk than a blood draw” (ID25; U+), or that use a “brand new technology” (ID18; U) or “an additional questionnaire” (ID39; C). Some of these nine interviewees suggested re-consent should be used as the data analytic plan becomes clearer even if no additional data is collected. For instance, interviewee 34 (PR+) mentioned circumstances that should trigger “some sort of re-consent” including “anytime there's a major programmatic change” or “halfway through at the five-year mark, just to remind people,” or, as interviewee 40 (PR+) described it, as a study “gets more in-depth and things change.”
Three other interviewees considered additional consent to be important when topics for study were identified. Interviewee 15 (NP) recommended “more specific consents down the road” when the study is “easier to understand.” For interviewee 15, additional consent would be needed because, “when there is a specific topic, I think participants should be allowed to have” more information. Interviewee 24 (U+) said any topic should require re-consent “if the scope is not included in what the participant consent,” and interviewee 53 (U+) assumed “that down the road” a health data repository should include “very specific…follow-up consent process[es]…when other studies that are very specific then take place.” Two interviewees mentioned that using data for profit should require additional consent, and two others mentioned the need for special procedures for Native American participants whose “samples [have] to be completely segregated from everyone else's so that they have greater control over what can be done with it” (ID30; U+) or for participants with brain injuries or reduced capacity to consent (ID47; H+).
Four other interviewees said consent would be important if the data were to be used to study sensitive topics. As interviewee 13 (U) said, “if…a researcher comes along in and wants to collect a very particular information [sic] on a topic that is -- could be considered sensitive…then I think that we either have to make the criteria for wavier or exemption or you know, get consent, hopefully simple. Hopefully simple consent.” Interviewee 43 (NP+) agreed that the question of “when to go back and ask people if they're okay with certain kinds of things” would be important as researchers “are getting information about what they want to study [and] say, ‘Oh, this is kind of a sketchy topic or maybe we need to go back and ask.’” Interviewee 43 (NP+) mentioned the difference between “making a good drug…for…an illness that I care about that I want to be solved” or a “lifesaving drug” versus “if it's for, like, Viagra” or something “more cosmetic” in which going back to the participant for re-consent may be important.
Interviewees who felt confident that re-consent was needed could not always name what should trigger it: as interviewee 22 (U+) said, “I don't have a good example in mind, but I just can foresee if something is more sensitive.” Others described re-consent as a form of engagement: a means of “asking people if they're okay” (ID43; NP+) with study procedures or a strategy “just to remind people” (ID34; PR+) of their participation in the study.
Discussion
In this paper, we draw on semi-structured interview data from a diverse group of 44 research stakeholders to examine the ways in which understandings of consent to a health data repository overlap and diverge from traditional consent principles. Interviewees described new challenges to the concept and practice of informed consent when there is limited certainty about the future purposes, procedures, and risks of a data repository at the time of participant enrollment. Interviewees indicated that building repositories for unspecified uses conflicts with the traditional view that the research participant be fully informed at the time of consent. Instead, interviewees understood health data repository consent not as informed but as engaged. Deemphasizing the importance of full information at the time of consent, they stressed the value of the ongoing relationship with the study as a mechanism for periodic information-sharing.
In this paper, we draw on semi-structured interview data from a diverse group of 44 research stakeholders to examine the ways in which understandings of consent to a health data repository overlap and diverge from traditional consent principles.
About half of our interviewees viewed some type of re-consent as an important means of mitigating limitations in information exchange at initial consent. The researcher stakeholders who endorsed a requirement for re-consent directly linked this endorsement to the inadequacy of information exchange at initial consent. As well, these research stakeholders suggested that re-consent provided one avenue by which participants could grow in their understanding of the repository's purposes, procedures, and risks and benefits. They described this deepening understanding as likely to increase commitment to the repository. At the same time, those who did not prefer re-consent expressed concerns it would undermine engagement with the study. Overall, research stakeholders saw engaged consent as an adequate means of mitigating informational gaps, gaps they felt should be closed as long as the participant's relationship with the study can be preserved. Research stakeholders had no shared notion of which informational gaps needed to be closed, what events should require re-consent, or what emergent information should be shared over time.
In these ways, our findings are consistent with the conclusions of Beskow and Weinfurt that nationally recognized experts disagree about the importance of comprehension in consent.35 Many in their study did not believe poor comprehension should bar enrollment. Our findings confirm that research experts are questioning “whether adequate comprehension should be viewed as an ethical requirement,”36 and that some view understanding as not essential to participation.Reference Kim, Miller, McKinney, Beskow, Ford, Kraft, Garrison, Wilfond, Solberg and Ursin37 Our interviewees seemed to suggest that rejecting understanding is “an ethical proposal whose time has come,” as Solomon Cargill says.Reference Cargill38
A few years ago, bioethicist Erik ParensReference Parens39 noted a concern that contemporary research norms are drifting away from the need to fully elicit and respect participants' specific preferences. Parens describes traditional informed consent practices as recognizing the wide diversity of viewpoints about the preferability of learning genetic information about oneself.Reference Lakes, Vaughan, Jones, Burke, Baker and Swanson40 Given this diversity, traditional informed consent procedures provide a means to “show respect to persons by asking them if they want access to a given piece of genetic information.”Reference Patil, Majumdar, Awan, De Souza and Greenspan41 That is, the traditional model of informed consent “was partly meant to ensure that individuals weren't coerced into decisions that someone else thought were in their — or the state's — best interest.”42 As evidence of a drift from this idea, Parens notes that proponents of genetic research justify their practices by citing population-based survey data indicating that most people are interested in their genetic information. “Drifting away from consent,” in his view, is a move that undercuts respect for the fact that “different people will reasonably want different sorts of information.”43 Our results provide an empirical suggestion that a similar drift away from informed consent can be seen in the concepts and practices emerging around health data repositories.
Yet our findings suggest that pronouncements of the collapse of the role for information exchange in consent are incomplete. Our interviewees appeared to agree with Appelbaum's view that ignorance is not bliss.Reference Appelbaum44 In different ways, most all of the expert research stakeholders we interviewed looked for means to improve and augment information exchange, including through engagement or re-consent. They emphasized a drift toward more thorough and long-term engagement with participants. They were reluctant to jettison traditional informed consent activities, such as approval for study purposes, procedures, and risks. While they did not agree on the events that should trigger re-consent, most preferred to retain the possibility that it would occur. They vested considerable responsibilities within the institution or infrastructure of the study to manage data responsibly. Most consistently, they stressed the importance of an ongoing relationship between participants and health data repository representatives, and indicated that informational shortcomings at the time of enrollment could be overcome through intentional engagement over time. These findings are consistent with calls for embedding the functions of consent in communication over timeReference Zawati and Knoppers45 and the recognition that “communication and decisions about research participation may not be contained within a discrete consent encounter.”Reference Dickert46
Importantly, Henderson et al.Reference Henderson, Edwards and Cadigan47 found in a survey of 456 U.S. biobanks that most employ robust strategies for managing specimens and decision-making about data use, a finding they saw as an indication of a developing “ethics of stewardship” among health data repositories. While these institutional responsibilities were rarely elaborated by our interviewees, they recognized the place of institutional reputation and infrastructure to the consent process. They suggested that responsible long-term data management can stand alongside consent as a mechanism for ensuring the protection of participants.Reference Jeffers, Laurie, Dove and Ganguli-Mitra48 And, both implicitly and explicitly, our interviewees identified the importance of strategies to build and maintain trust in the individuals and institutions managing the repository.Reference Kraft, Cho, Gillespie, Khodyakov, Mikesell, Bromley, Yarborough, Fryer-Edwards, Geller, Sharp, Moreno-John, Fleming and Ford49 Trustworthiness in physicians, researchers, health care systems, governments, and corporations involved in research has specifically been identified as important to enrollment in precision medicine initiatives.Reference Kraft, Cho, Gillespie, Burgess and O’Doherty50 Our interviewees described engagement as a strategy that could help to maintain and enhance trust by increasing transparency, confirming ongoing interest in participating, and attending to unique or historical challenges to trust (e.g., commercialization, controversial research topics, research misconduct experienced by some subpopulations). In these ways, fully informing participants about study purpose, procedures, and risks and benefits seemed important as a means of protecting participants, but interviewees could imagine new timelines and novel means of achieving these and related goals.
Nonetheless, our interviewees neither presented a clear definition of engagement nor delineated its component responsibilities. In addition, the plan for implementing engagement as a mechanism for mitigating limitations of information exchange at enrollment remained imprecise. In fact, Henderson et al.61 found that most biobanks do not create ongoing relationships with participants. Interviewees described several different and mutually-exclusive roles for engagement, including: (1) keeping participants advised of the option to opt out, providing access to their data, or otherwise giving them a voice in the study choices but not asking for further consent; (2) periodically updating them on what is being done with their specimens or providing them with some access to their own data; (3) re-consenting around new uses of data or around new specific topics, particularly those that may be sensitive; and (4) introducing additional consent procedures for vulnerable populations.
One final area of uncertainty can be noted. A traditional informed consent includes the obligation that the participant be fully informed prior to opting in, while interviewees described engaged consent as a means of re-affirming over time the participant's right to opt out. If fully embraced, this shift in emphasis from opt in to out out could alter the incentives for researchers, increasing the importance of participant retention and inviting investment in transparency, return of results, and information exchange during a study. Interviewees suggested frequently that an engaged approach should remind participants that they can opt out, but they seemed uncertain about which information should be provided simultaneous with this reminder. Taken as a whole, interviewees did not demonstrate consensus about best practices for long-term engagement with participants in a health data repository. Future research will need to clarify the necessary elements of engagement as well as the implications of using engaged consent as a means to mitigate limitations in information exchange at the time of consent to a health data repository.
Limitations
We note several limitations of our study. Our findings are limited by the non-representative sampling of the interviewees. While research stakeholders who participated reflect a diverse array of sectors and experiences, the findings cannot be said to generalize to the entire health data repository community. A second limitation is that our sample included many individuals familiar with the AoURP, an initiative that foregrounds the importance of engagement.
Conclusion
With the development of large-scale health data repositories, the research stakeholders we interviewed described a procedure for engaged rather than informed consent. Results indicate the need for further research on engagement as a means of improving participants' comprehension of a health data repository and of ensuring the affordance of appropriate protections from research risk.
Acknowledgements
This work was supported by the National Institutes of Health under Grant 1U24OD023176-01. results
