The Role of Participants in a Medical Information Commons

Abstract

Meaningful participant engagement has been identified as a key contributor to the success of efforts to share data via a “Medical Information Commons” (MIC). We present findings from expert stakeholder interviews aimed at understanding barriers to engagement and the appropriate role of MIC participants. Although most interviewees supported engagement, they distinguished between individual versus collective forms. They also noted challenges including representation and perceived inefficiency, prompting reflection on political aspects of engagement and efficiency concerns.

In 2015, Francis Collins and Harold Varmus highlighted the importance of “new models for doing science that emphasize engaged participants and open, responsible data sharing.” ¹ Others have explicitly tied the success of efforts to increase data sharing and create a large-scale, longitudinal, multi-purpose informational resource or medical information commons to participant engagement and participant-centricity.² As Charlotte Hess and Elinor Ostrom have observed, efforts to create common-pool or shared resources are most likely to endure and thrive if governance is “organized in a nested structure with multiple layers.”³ Thus, the term “medical information commons” or “MIC” may appropriately be used to describe both the networked space or ecosystem (singular) in which data sharing occurs and also the initiatives (plural) collecting and broadly sharing diverse kinds of data for research and other purposes, and attempts to engage participants and attend to their interests, values, and concerns may be assessed at multiple levels.⁴

Supporters of participant engagement in biomedical research cite a range of intrinsic and instrumental goals. They justify engagement in terms of respect for persons and autonomy, democratic norms, and considerations of social justice. They also mention benefits such as improved recruitment and retention and increased public buy-in.⁵ Since an MIC involves longitudinal collection of sensitive personal information from large numbers of people, the potential for engagement to advance these goals may be particularly salient to MIC sponsors. However, some scholars have drawn attention to the potentially problematic implications of public engagement rhetoric, such as the contestable implication that there is a civic duty to participate in government-sponsored research initiatives and other initiatives promoted as serving the common good.⁶

While there is no consensus definition of “participant engagement” in the literature, there is broad agreement that the existence of channels for bidirectional communication between researchers and participant representatives is a necessary feature of participant engagement.⁷ Conceptions of participant (or patient or community or public) engagement as a continuum from weak, limited forms of interaction to robust, comprehensive forms of interaction — such as community-based participatory research — are common.⁸ Although there is typically an implicit normative assumption that movement along the continuum toward greater engagement is desirable, in some cases determination of the optimal degree of engagement is specified as requiring the input of those affected. The term “partnership” is often used to indicate the point on the continuum where leadership is truly shared between participant representatives and researchers or other experts.⁹ Advocates and scholars such as Barbara J. Evans have identified a point beyond partnership where participants assume control over their own data, while perhaps enlisting experts and research institutions as consultants and allies.¹⁰

While interviews were framed in terms of an MIC, much of this content is relevant to the participant role in biomedical research more generally. Building on our findings, we conclude by reflecting on the political aspects of participant engagement and efficiency concerns that we believe are worthy of further consideration by the bioethics and policy communities.

We present findings from interviews with diverse expert stakeholders involved in existing and emerging public and private data-sharing initiatives or the creation of relevant ethical and legal frameworks. Interviewees were selected to represent six sectors contributing to the creation of an MIC: academia (including investigators conducting biomedical research, law and policy research, and research on participant engagement), non-governmental organizations (including leaders of patient advocacy organizations and research foundations), technology companies, government, laboratories, and healthcare systems. The interviews began broadly, offering interviewees an opportunity to articulate their own vision of an MIC and assess the current landscape. The interviews then focused on specific areas of interest, including exploring interviewees' perspectives on the role of participants, meaning the people from whom data derive. While interviews were framed in terms of an MIC, much of this content is relevant to the participant role in biomedical research more generally. Building on our findings, we conclude by reflecting on the political aspects of participant engagement and efficiency concerns that we believe are worthy of further consideration by the bioethics and policy communities.

Methods

The research presented here is a sub-analysis of semi-structured interviews conducted with 41 expert stake-holders. Two interviews involved two interviewees; the remainder involved one interviewee. The strategies for sample selection and data collection are detailed in full elsewhere.¹¹ Interviews began with a question soliciting a definition of “medical information commons.” Input was then sought on the following working definition: Medical information commons are networked environments in which diverse sources of health, medical, and genomic data on large populations become broadly available for research use and clinical applications. Interviewees were also asked for their views on the current landscape and barriers to MIC creation. The interviews then shifted to particular areas of interest. This manuscript presents a thematic content analysis of responses to the interview question, What role should the people whose data populate the medical information commons play? (and any reflection on participant role in responses to other questions). This open-ended question was sufficient to prompt an expansive response from many interviewees. The interview guide did include a series of probes to draw out interviewees where necessary, including: How, if at all, should they be involved? What are the challenges to engaging individuals or communities to carry out the role you envision for them?

The initial question used neutral language such as “people” (the word “participant” is currently in favor in the research ethics literature precisely because it implies an active rather than passive role) and intentionally omitted the use of the term “participant engagement” in order to avoid steering interviewees to traditional conceptions of engagement. We use the terms “participant” and “participant engagement” in this paper because “people whose data populate the MIC” is an awkward construction and because the majority of interviewees envisioned an active role for the people whose data populate an MIC and themselves employed the terms “participant” and “participant engagement.”

Coding and Analysis

Three of the authors (JB, MM, and AV) were responsible for coding and analysis of interview transcripts. The development of the codebook drew on a preliminary analysis of 11 transcripts. This initial review and assessment of the codebook led to the creation of additional codes that captured themes that emerged from holistic analysis of the fuller set of transcripts. Responses to the central question, plus additional related commentary on participant engagement stated in response to other questions, were coded by two coders using the NVivo 11 Pro software program. Coding differences were discussed by the coders until consensus was reached. Code reports were generated and reviewed for further analysis.

Results

The distribution of interviewees by sector is shown in Table 1. All interviewees favored expanding the role of participants beyond simply providing one-time informed consent, except for one interviewee who stated, “I think that once somebody agrees to participate, they want to be left alone.” (P25)

Table 1 Sectors Represented by Respondents

Here we explore further findings in two areas. First, while some interviewees focused on ways to increase engagement of participants as individuals, others focused on ways to increase engagement of participants as a collective. Second, many interviewees described challenges they had experienced or anticipated in seeking to expand the role of participants.

I. Individual Focus or Collective Focus

When prompted to reflect on the role of participants, some interviewees first or only discussed interactions with participants as individuals, while others first or only discussed the role of participants as a collective, particularly through representation in governance structures. The distrinction between participants as individuals versus a group is important to explore, as there are potential tensions between the two orientations. Illustrative quotes related to individual and collective focus are displayed in Table 2 and summarized in the following sections.

Table 2 Individual Focus and Collective Focus

As shown in Table 2, interviewees who emphasized the role of participants as individuals identified a number of measures to enhance this role in the context of an MIC. A relationship between participants and an MIC was stated or assumed as a basis for MIC interactions with participants; interviewees did not delve into details such as the precise role of intermediaries where data is flowing into an MIC through a number of primary research studies. Many interviewees noted opportunities for an MIC to return information to individuals, in some cases framing return of information to participants as an obligation. Several interviewees mentioned empowering individuals to remove data, or change their consent status over time consistent with a dynamic, granular consent paradigm (with the Platform for Engaging Everyone Responsibly as one example). One interviewee referenced an instrumental justification: the anticipated positive impact this would have on individuals' willingness to share data. A few interviewees talked about individual entry of phenotypic information or involvement in data curation, and one described how technological advances would allow individuals to take control over an increasingly large and important set of health-relevant data. Finally, several interviewees spoke of individuals having a right to access information about them and/or the desirability of having pathways for participants to suggest research questions or engage in citizen science using de-identified data. However, one interviewee objected to creating such pathways because of the additional work involved and the fact that participants may lack knowledge and expertise.

Other interviewees described measures to enhance the role of participants in collective terms. Several interviewees felt that it was important that participants as a collective be involved in governance, with some specifying a leadership role on the steering committee of an MIC. One interviewee went on to state that in the context of an MIC, this collective approach would be the most appropriate way of realizing respect for autonomy. A few interviewees outlined more modest forms of collective engagement as potentially acceptable, such as focus groups or opinion surveys, or stated that they preferred community advisory boards (with clear delineation of their non-decision-making role) to participant involvement in a steering committee or board. A few interviewees advocated for an involved-in-all-aspects role for participants, beginning with the involvement of representatives of potential participants as a collective (or particular communities) in shaping research priorities and research design and extending to representation of participants as a collective in governance and in the conduct of research, including opportunities to engage in citizen science (i.e., contribute research ideas or directly query databases), but also encompassing greater empowerment of participants as individuals.

II. Challenges

Interviewees explicitly or implicitly referenced challenges they experienced or foresaw in bringing reality into line with their view of the role participants should play in an MIC. Here we focus on five of those challenges and proposed strategies for addressing those challenges (where offered).

COST

Interviewees with experience in participant engagement noted increased recognition of the resources required. They cited initiatives such as the Patient-Centered Outcomes Research Institute (PCORI) as helping to normalize research expenditures on participant engagement. Yet they reported continuing resistance related to cost, for example: “I can have a budget that's approved by the funder and when I submit the invoice for my community partner to be paid I still get the ‘Well, why are you paying them so much?’ kind of response.” (P24) One strategy was simply to draw attention to the disjunction between rhetoric and budget: “How much are we spending on engagement versus technology systems and research protocols and that sort of stuff? I think we're going to find out 1% of the budget goes to these things and that is unconscionable and won't give us the results we want.” (P13)

REPRESENTATION

Interviewees who considered or advocated for some representation of participants on governance bodies or other limited-membership groups described challenges related to selecting individuals who are truly representative of a cohort or population, and meaningfully engaging those representatives. Some linked the challenge of representative selection to scale: “The challenge is when you have 100,000 research participants… [, and] it's not going to be too far where it's going to be 10 million. How do three people…on a committee actually represent that?” (P19) Regarding representative selection, several interviewees noted that participants who volunteer or are selected as representatives are often atypical because they must have the time to participate and a likely unusual level of interest in research relative to other priorities. Related to this, some contrasted professional advocates with the “common (wo)man” or “person on the street.” One interviewee considered selection of representatives from and for groups with schizophrenia or severe depression or severe cognitive impairment especially challenging. Another contrasted Native American nations, which have sovereignty and clear authority structures, with more fragmented communities lacking clear authority structures. A few interviewees mentioned the possibility of experimenting with some kind of election process.

Regarding meaningful engagement, institutional review boards (IRBs) were twice mentioned as examples of engagement that failed the meaningfulness test. One interviewee talked about the limited number of community members on IRBs, implying that critical mass in governance structures is necessary for meaningful engagement. The other brought up aspects of what has been called “epistemic inequity”¹²:

In my experience, community members on IRBs are often overwhelmed by the other members of the IRBs with their greater expertise. And community members or representatives are not often given much credence. They're not listened to. They're just sort of tolerated.

(P32)

Two interviewees reported success with a strategy that amounts to (in our words) multiple kinds of participant voices, adequately supported. This strategy embraces outreach to individuals who have been selected to lead community organizations and to those who put themselves forward as community leaders. It would also involve engagement with constituents outside the control (and presence) of formal and informal leaders. As to “adequately supported,” these interviewees believed that meaningful engagement must include education, e.g.: “The typical individual doesn't walk in off the street…able to provide oversight in some sort of commons that also includes researchers and providers and whomever else is going to be a part of it. They need to be prepared and trained and sometimes they need to be compensated, especially if it's a significant amount of their time that's going to be involved.” (P24)

PERCEIVED INEFFICIENCY

One interviewee reported running up against the view that it would be “unmanageably complicated” to give individuals more say in how their data are used, alongside complacency about the adequacy of traditional research consent: “There's sort of a puzzled look on people's [project leaders'] faces, and I don't know… they'll hold up a piece of paper and say, ‘Well, we have the consent form right here.’” (P10) Another interviewee cited pressures to launch programs and do so under budget as working against a more robustly participant-centric vision. At the same time, several suggested that efficiency considerations had some legitimacy, expressing concerns about the feasibility of a framework that offers individuals more control. For example, an interviewee who expressed reservations about general or broad consent added: “On the other hand, it is a logistical nightmare to have all kinds of levels [of consent]… Not that the software can't handle it…, but it adds substantially to the information management challenge.” (P31)

Some of the expert stakeholders we interviewed have written and spoken publicly about data sharing, the design of data-sharing initiatives, and MIC policy development, often touching on participant role; many of the themes we highlight here are already present in the literature and are also reflected in the emphasis on participant or community engagement in major grant-making initiatives like PCORI. Nonetheless, we believe political aspects of differing approaches to participant engagement and efficiency concerns have not received the attention they merit. In making this case, we build on the results reported above, but we also draw on our own experience with participant and public engagement.

Desire/Ability to Engage

Several interviewees observed that people vary in both their desire and their ability to engage. For some, this was not a significant problem, so long as this variation was acknowledged and accommodated. For example:

[T]here's like the concerned cohort that are just hyper on everything, right? They're your big wearables people, that kind of thing. Retirees who have a lot of time to answer emails, I don't know. If somebody started pinging me all the time they wanted to access my data at 23andMe, I would get on the phone and say, “Take me off your list. I don't want to be bothered anymore.”

(P36)

Others saw the potential to increase health disparities, insofar as only the relatively privileged would benefit (both directly and downstream) from a system designed to promote inclusion and deliver respect and rewards for participation via interactive mechanisms: “If we are starting a system in which it takes a high degree of sophistication and wherewithal [i.e., resources, including unencumbered time] and education and those sorts of other things to make full use of it, then we're going to have an aggravated chasm between the haves and the have-nots.” (P32) Also, while one interviewee pointed to the success of patient advocacy organization-driven efforts as important and inspiring (e.g., multiple myeloma, cystic fibrosis), another suggested that the success of rare disease groups in particular could be misleading: “I think it's a model that works very well in the rare disease community because they're very savvy, very active, and rightly so, very concerned parents, and families, and extremely democratic in sharing their information… [But] to paint [MCI] participants with the same brush as having all those tools?…” (P1) In short, this interviewee questioned the generalizability of engagement strategies based on single-disease focused initiatives.

INAUTHENTIC ENGAGEMENT

A few interviewees were especially critical of efforts to engage participants as a means to an outcome unrelated to respect for individuals and communities. One interviewee contrasted conversations about engagement as a means to increase recruitment and retention (using words like “convincing,” “selling,” and getting “buy-in”) versus conversations focused on how to achieve true “partnership” or true involvement through leadership roles or participation in governance. Several interviewees described programs that used desirable language, at least in their public-facing communications, but in practice failed to live up to it. Finally, as noted above, several interviewees held up community-based participatory research as an ideal, with community needs and concerns a primary consideration even to the point of leading to a rethinking of research priorities, or to policy changes that could add costs or introduce delays.

Discussion

Some of the expert stakeholders we interviewed have written and spoken publicly about data sharing, the design of data-sharing initiatives, and MIC policy development, often touching on participant role; many of the themes we highlight here are already present in the literature and are also reflected in the emphasis on participant or community engagement in major grant-making initiatives like PCORI. Nonetheless, we believe political aspects of differing approaches to participant engagement and efficiency concerns have not received the attention they merit. In making this case, we build on the results reported above, but we also draw on our own experience with participant and public engagement.

Political Aspects of Participant Engagement

Some accounts of participant engagement focus on advances in technology and on participants as individuals.¹³ Equal attention should be given to relationships and participants as a collective. While technology can support health-related “e-communities,”¹⁴ our interviewees commented on the importance of off-line, face-to-face interactions among participants and between participants and researchers. Further, they recognized the importance of engaging participants as members of a collective despite the difficulties presented. We note that keeping participants isolated as individuals may serve to enhance the control of researchers and minimize participant voice and power — even if that is not the intention of those developing individual-focused engagement strategies. Yet, as Blassime and Vayena note, for large-scale “precision medicine” cohorts “the kind of communitarian bonds that a participatory ethos is supposed to capture and to promote are either yet to be formed or bound to compete for the opportunity to have a say regarding the governance of the cohort.”¹⁵ They suggest offering participants opportunities to become actively involved in governance structures at enrollment and leaving the composition and agendas of governance bodies open to reconfiguration by participants.

Blassime and Vayena's work is immensely valuable but stops short of confronting all the challenges associated with selecting and then empowering participant representatives in the context of an MIC, especially in the U.S. One challenge is the problem of representation. Sankar and Parker mention expertise, diversity, equity, inclusiveness, and convenience as possible criteria in the selection of participant representatives for cohorts such as All of Us. They then pose a question about the moral basis for decisions regarding selection criteria.¹⁶ No pat answers emerged from our interviews, but we did hear from experts who were cognizant of the complexities and had practical experience confronting them. One clear message was rejection of what might be called a politics of scarcity (e.g., fixation on how to fill one or two designated “participant” slots on a steering committee). Multiple channels for twoway communication should be created from the planning stage onward. Intrinsic and instrumental ethical considerations underlie the importance of both diversity and equity in selecting individuals to represent collectives of participants or potential participants.¹⁷

As for expertise, we note that it is itself diverse and dynamic. A person who has learned important lessons from her experience advocating for the interests of her community in the face of numerous obstacles has expertise, as does a person who has learned as much as most clinicians or scientists about the particular health condition affecting him or his child. Lay people, especially those from groups that have experienced and continue to experience discrimination and exclusion, need support to develop and convey their expertise in a manner that will be recognized and valued by other stakeholders. For that matter, the ability and willingness of minority and underserved populations to engage at all will be contingent upon efforts to understand and address barriers to active participation (e.g., lack of resources), and to shape engagement to meet their needs rather than using them solely to advance researcher agendas.¹⁸ Reliance on solicitations of interest for candidate identification, and convenience as a core selection principle, works against diversity and equity and narrows the kinds of expertise brought to the table.

Other pitfalls not explored in our interviews but relevant to engagement as a political project include ignoring intra- and inter-group differences and papering over tensions. For example, individuals rightly resent being asked to give voice to the perspective and concerns of an entire racial/ethnic group, and from the other side, question the ability of someone very unlike them in most respects to represent them based on a single commonality.¹⁹ Some are skeptical about motivations and question whether anyone can truly represent them.²⁰ In addition, individuals who are directly affected by a health condition may have a different perspective than family members or caregivers. And patient advocates may have a different perspective than patients “on the street.” Patient advocates may differ in the degree to which they prioritize privacy versus open science, or developing new treatments versus comparative effectiveness research, or support services versus discovery research. Those who identify as “healthy” (or are recruited because others characterize them that way) may have concerns that differ from those affected by a specific condition or identifying as in “poor health.” Strategies such as building in repeated opportunities for face-to-face communication over time can help to create the bonds of trust within and across groups of participants and between participants and other stakeholders, including researchers and funders, that make uncomfortable but important conversations possible and productive. We are hopeful that some of those uncomfortable conversations will address the ways in which traditional research agendas and priorities have failed certain groups, such as American Indian and Alaskan Native communities and African Americans (especially as these identities intersect with sources of disadvantage such as poverty and powerful forces such as structural racism).²¹

Indeed, participant engagement cannot fulfill its quotidian or transformative promise without significant attention to process and to the promotion of epistemic equity. Ideally processes of engagement are designed to promote mutual understanding of the reasons or stories that underlie and explain positions, as in deliberative democracy approaches.²² In studies in the IRB context, non-institutional members often report feeling intimidated and attribute this feeling to their unfamiliarity with technical terms and jargon. There is evidence that professionals have a tendency to discount the input of others (lay people and professionals from other disciplines), which may be based in part on differing linguistic habits or privileging scientific paradigms of explanation. Early expressions of opinion from IRB chairs or other professionals ascribed special authority may preempt alternate points of view. As Wenner notes, “[t]he upshot is that a process…to bring multiple, divergent perspectives to bear on difficult and nuanced problems can very easily become one that instead overlooks or marginalizes precisely those outlier views to which it is intended to give voice.”²³ Broadly, failed engagement reflects lack of communication and collaboration to address power differentials, not simply or primarily knowledge deficits. Nonetheless, education for all stakeholders may have value within a portfolio of strategies. Participant representatives would benefit from technical knowledge that reduces self- and third party-appraisals of incompetence.²⁴ In addition, knowledge about crucial conversations and enhancing storytelling as a mode of communication could equip participants to navigate tensions and relate their real-world knowledge to issues, while also providing them with skills of general value. Researchers and other stakeholder groups would benefit from education about team science as well as epistemic inequity and steps for minimizing it. The goal would be a long-term project of mutual learning.

We have highlighted complexities that may seem daunting to institutions and investigators who are receptive to the case for participant involvement but lack experience in this area. One possibility is to look for collaborators in schools of public health or community organizations with a health mission and a track record of successful engagement of constituents. Consultation with leaders from initiatives such as PCORI (or relevant component projects) or the Health Care Systems Research Network (HCSRN) Patient Engagement in Research Scientific Interest Group may also be valuable, especially as there is now a growing literature reporting “lessons learned” from these efforts.²⁵ Champions should be prepared to advocate with institutional leaders and funders for the resources necessary to do engagement well. As suggested by one of our interviewees, if budgets are a reflection of priorities, then these stakeholders should acknowledge that a financial investment in participant engagement is one marker for authentic commitment.

Efficiency Concerns

It is quixotic to expect researchers in the commercial sector to embrace engagement without considering its impact on efficiency (i.e., the amount of time and effort required to accomplish objectives), and efficiency is also a legitimate concern when scarce public resources are involved. Still, given the intrinsic and instrumental goals of participant engagement, efficiency as an excuse to avoid putting serious thought and effort into how best to engage participants should be resisted. It is certainly not a bad thing that attention is finally being paid to the long-term efficiency gains from meaningful participant engagement, including improved retention, better quality data, and a focus on the outcomes that are most important to patients/participants.²⁶ And yet to focus on such gains is to risk favoring the instrumental over the intrinsic and the pedestrian (i.e., business as usual, only a bit better) over the transformative. Adopting a more inclusive and experimental perspective, Kelty and Panofsky point out that even as engagement brings new risks such as “loss of control and rise of contention in the research process,” it offers “many intriguing possibilities for new forms of collaboration, data, knowledge production, funding, and serving public needs.”²⁷

Our results, including efficiency-focused reservations about dynamic, granular consent, do help make the case for beginning with experimentation rather than detailed mandates. (Innovative approaches can be tested before being promoted as best practices or required of all the initiatives that make up an MIC.) No one we interviewed articulated a goal of rapidly creating new layers of regulation and bureaucracy around participant engagement. Although models exist, there are no legal requirements for engagement, beyond requirements from funders, and rules that each IRB include a community representative — an ill-defined and problematic role (as noted by several interviewees).²⁸ There is also tension between the IRB mandate to protect individual participants and procedures that endorse attention to community concerns. For experimentation to yield maximal value, it will be important to develop appropriately nuanced approaches to evaluating the effectiveness of participant engagement efforts in achieving intrinsic and instrumental goals and systematically assess benefits and costs.²⁹ In the literature it is still uncommon to find evaluations of participant engagement, and published evaluations are usually limited to measurements of levels of engagement (such as the number of people initiating communication with researchers).³⁰ Given the importance of engagement and the resources at stake, it is time to build a more robust evidence base.³¹

Limitations of our research include our recruitment of individuals involved in data sharing, either directly through involvement in data-sharing initiatives or through their work on relevant ethical and legal frameworks. While a few of our interviewees harbor significant reservations about aspects of the data-sharing enterprise, most are committed to the goal of creating an MIC. Also, although we framed an MIC as a resource available for purposes other than biomedical research (e.g., clinical care), most interviewees discussed an MIC within the context of research rather than exploring connections to learning health care systems or public health.

Conclusion

While the challenges described by expert stakeholders are significant, enthusiasm for expanding the participant role in the context of an MIC remains. Each initiative within an MIC ecosystem is a potential laboratory of democracy and space for deliberation about how best to align medical, scientific, and social justice goals. Openness about successes and struggles will be critical to mutual learning and development of best practices for participant engagement.