Consideration of the Principle of Justice under 45 CFR part 46
Secretary’s Advisory Committee on Human Research Protections
July 22, 2021
Injustice has no place in human subjects research and undermines public trust in science. One consequence of injustice, whether it manifests as inappropriate exclusion from participation or as exploitation of “populations of convenience,” is the belief that human subjects research serves the interests of the privileged and powerful and therefore perpetuates economic, racial, religious, sexual, gender, and cultural biases. The history of research on human subjects reflects numerous examples of the relationship between blindness to, or disregard of, issues of justice and consequent justified mistrust. Most recently, the ongoing disparate economic and public health impacts of the COVID-19 pandemic, including issues related to vaccine equity and hesitancy, illustrate some of the sources and consequences of distrust in science and its social goals. Researchers and many others recognize human subjects research as a primary human activity dedicated to objectivity and empiricism; however, it continues to be marred by unjust policies, practices, beliefs, and systems of power. It is time to reconsider and reestablish justice as a core principle in biomedical and social-behavioral research, reflecting the reality that the science of people must be accountable to people to be legitimate.
SACHRP acknowledges that achieving this goal requires a commitment of others beyond OHRP and the regulated community. Objectivity will always be an aspiration, and decisions about which research questions and methods are selected and funded will reflect the experience and backgrounds of the members of the researcher and sponsor communities. Equity (and the quality of research) will suffer until these communities reflect the experiences and backgrounds of the people they are meant to serve (https://extramural-diversity.nih.gov/diversity-matters). Human Research Protection Programs (HRPPs) and Institutional Review Boards (IRBs) have a limited but important role to play, and the recommendations in this document are intended as a starting point from which to develop more detailed policies and practices to help ensure fair access to opportunities for research participation and reasonable assurance that the potential benefits from research are available and meaningfully applicable to all.
Framing of Justice in the Belmont Report
It is customary to look to the wisdom of the Belmont Report for guidance on the principles that allow ethical research on humans. Belmont articulates “three principles, or general prescriptive judgements, that are relevant to research involving human subjects”: Respect for Persons, Beneficence and Justice. The applications to the research enterprise of Respect for Persons and Beneficence are clearly described. Respect for Persons dictates treatment of individuals as autonomous agents, and mandates additional protections for those with diminished autonomy. Beneficence expands on the Hippocratic mandate to “do no harm,” with the qualification that the nature of research means that harms are often unknown or potential and can only be minimized, not eliminated. Applying these two principles to a specific research study requires understanding the study and the science behind it, but the meaning of the principles themselves is generally clear.
The understanding of all three principles has evolved, but application of an expanding appreciation of Justice and its broad scope and impact has not been adequately reflected in interpretation of Belmont or the research regulations to date. Belmont framed Justice as a question of fairly distributing the burdens and benefits of research, but largely limited its discussion to subject selection within the context of an individual trial or institution. The Report justified this limitation by observing:
“Injustice may appear in the selection of subjects, even if individual subjects are selected fairly by investigators and treated fairly in the course of research. Thus, injustice arises from social, racial, sexual and cultural biases institutionalized in society. Thus, even if individual researchers are treating their research subjects fairly, and even if IRBs are taking care to assure that subjects are selected fairly within a particular institution, unjust social patterns may nevertheless appear in the overall distribution of the burdens and benefits of research. Although individual institutions or investigators may not be able to resolve a problem that is pervasive in their social setting, they can consider distributive justice in selecting research subjects.”
In this way, the Report anticipated current concerns about structural injustice, but explicitly chose to limit such concerns to the selection of research subjects.
Both the Belmont Report and the Common Rule were primarily reactions to egregious examples of inappropriate inclusion of research subjects (as populations of convenience and/or members of groups vulnerable to exploitation). The result of this focus is an ethically necessary tradition of protections from inappropriate inclusion, but these protections have been imperfectly applied and have not been balanced by equally necessary attention to participation in research and its subsequent benefits.
Distributive Justice in Research
Consideration of distributive justice, in the sense of fair distribution of burdens and benefits, requires an unambiguous identification of those burdens and benefits. In biomedical research, the burdens are (1) the specific risks of harm consequent to participation in a study, and (2) economic and social burdens of participation, including opportunity costs; the latter may be experienced by both individuals and groups. By definition, the benefits of research, defined as an activity intended “to develop or contribute to generalizable knowledge” (45 CFR 46.102(l)), are safe and effective therapies or preventive strategies for disease, but specific biomedical studies may also hold out the possibility of a direct benefit to the participating individuals.
While the nature of their primary benefit is what differentiates research, which is designed to teach us something generalizable so that we can learn how to better treat patients in the future, and medical care, which is the treatment of the individual patient, in practice treatment may contribute to research (e.g., “Learning Healthcare Systems,” retrospective research on clinical records, patient registries) and research may contribute to treatment. Justice requires that potentially therapeutic research opportunities be available to those who suffer from the condition under study, and both individuals and communities should have a place in assessing just access.
Where there is no such possibility of benefit to participating individuals, whether from the therapeutic effect of the study product or from better access to care, fairness suggests that those exposed to the risk or burdens of the research be members of the population likely to benefit from the knowledge gained, i.e., those who suffer from the condition being studied. In this case, the demands of fairness align with those of scientific generalizability, in that it may not be valid to extend research results to a population affected by a disease but not directly studied. The application of these principles requires that a study population adequately reflect those who may benefit from future risk reduction or therapeutic knowledge (no inappropriate exclusions), and also that the population be limited to those who may so benefit (no inappropriate inclusions). Sensitivity to injustice will be highest when communities have been unjustly excluded from both potentially beneficial research and the ability to contribute to future knowledge and the generalizability of results.
SACHRP’s recommendations are often focused on biomedical research, perhaps because the nature of risk in this setting has made it the focus of both regulation and guidance. But considerations of justice apply equally to social behavioral research. In particular, such research can still lead to harms, economic and social burdens and opportunity costs, and may or may not offer potential direct benefit. The same arguments above apply: research should be open to members of populations that will be affected by the application of the knowledge gained (e.g., policy and program development).
Inappropriate Inclusion Versus Inappropriate Exclusion
Although Belmont acknowledged the concern that research should not be limited to populations convenient for the researcher, its consideration of justice mainly examined the exploitation of vulnerable subjects. This attention to inappropriate inclusion was in response to real incidents, and was intended to help prevent those who would not benefit, those who did not represent a population that would benefit, or those who represented only a small portion of those who might benefit from being chosen as research subjects simply because they were “readily available.” The regulations address this concern in their consideration of vulnerable populations at 45 CFR 46.107, 111 and 116, as well as explicitly in subparts B, C and D. While this original concern remains, one-sided attention to exploitation has led to concern that individuals will be deprived of potential direct benefit and groups will be deprived of future benefits if they are inappropriately, unnecessarily or inadvertently prevented from participating in research, perhaps because of restrictive study design, operational concerns, social circumstance, or conscious or unconscious bias on the part of the researchers. Such inappropriate exclusions are unjust.
There are a number of populations that have been systematically excluded from research participation, including women, pregnant persons, children, the elderly, individuals with limited English proficiency, persons with mental or physical disabilities, and people from particular racial and ethnic groups. The historic reasons given for exclusion have differed, but in all cases these reasons are now suspect. In the past, it was considered scientifically justifiable to restrict the population of subjects (e.g., to adult white males) to control variability or reduce risk, without regard to whether the results would be generalizable to populations not studied. This assumption has reinforced biases and resulted in material harms for the excluded populations, including a failure to recognize disease and the use of inappropriate treatments. Concerns about risks to pregnant persons, fetuses, children and the elderly have too often allowed the research community to avoid the more challenging calculation of whether potential harm is outweighed by potential benefit. More egregiously, the exclusion of individuals from certain racial and ethnic groups has been blamed on the reluctance (perceived or real) of such groups to participate. Where such reluctance is real, it is justified by both the history of exploitation in research and an ongoing blindness to the differential burdens of study design on communities with different circumstances, needs and resources. Justice and good science require that the research community develop strategies to address that history, make participation practical, and earn and rebuild trust. In reality, exclusion from research, like exclusions from other areas of social participation, often results from long-standing structural biases.
The continued failure to include racial and ethnic minorities in clinical research has additional insidious consequences. Racism is grounded in a fallacy, in that it disenfranchises populations from economic and social participation and then uses the predictable consequences of that disenfranchisement to justify its continuance. Science practiced knowingly or naïvely within the context of such a tautology can give it a patina of legitimacy. Assuming that race and ethnicity are medically relevant variables because they are related to an individual’s genetic heritage can hide more problematic and actionable causal relationships with the social determinants of health. Race and ethnicity are strongly correlated with access to care, environmental exposures, income, employment, and other social determinants of health, which, by definition, affect health outcomes. Thus, an enrollment restriction or analysis based on race or ethnicity without considering these deeper causes will be misleading and will confirm pre-existing bias.
Further, equity and inclusiveness in a pluralistic, democratic society are a measure of that society’s fidelity to its core principles. At a time of widespread distrust, objective science well executed will eschew bias and lead to better understanding and better public health and serve as a model of how we should treat one another as human beings.
The charge to SACHRP is explicitly to consider how institutions, investigators, IRBs, sponsors and the research enterprise as a whole can help address injustice that “arises from social, racial, sexual and cultural biases institutionalized in society” and that represents “a problem that is pervasive in [our] social setting.” In responding to that charge, the first question to ask is whether Belmont’s narrow interpretation of the responsibility of the researcher and IRB is still appropriate. Belmont is explicitly about balance; the question to SACHRP is how to balance existing protections against exploitation with equally important protections against inappropriate exclusion.
Our response to the charge includes several high-level recommendations. Sources of injustice are structural and embedded in long-standing practice; they will take time and commitment to reduce. Some of our recommendations refer to earlier SACHRP recommendations, which, when mentioned, are incorporated by reference and not fully discussed herein. In addition, several issues discussed briefly are sufficiently complicated that their full examination is beyond the scope of this document. Those issues could be topics for future recommendations.
Charge to SACHRP and SACHRP’s Response
§46.111(a)(3) requires the IRB to take into account “the setting in which the research will be conducted.” For much biomedical research, particularly research with the potential for direct benefit to participants (i.e., studying potentially therapeutic interventions) the setting in which research is conducted is the healthcare delivery system. There is ample data demonstrating that the current healthcare delivery system has structural inequities related to race, income and socioeconomic status. What is the role of the IRB in ensuring that research conducted in this setting appropriately protects the rights and welfare of research participants? How can research be conducted so that it does not implicitly inherit the injustices of the healthcare delivery system (e.g., restricting participation to individuals who are insured and able to afford co-pays)?
As Belmont implies, it would not be realistic to expect the IRB to unilaterally change the circumstances of the current healthcare system. Problems of insurance, access to care, socioeconomic status and systemic injustice are interrelated and cannot be adequately addressed within the authority of the IRB, nor can they be addressed on a study-by-study basis. On the other hand, there are some areas in which the IRB can have an impact, particularly in reducing burdens of participation that may fall disproportionately on some groups. Accordingly, IRBs should consider:
- Individual study design and the burdens it places on participants. The impact of these burdens may vary depending on social and economic circumstances. Within the limits allowed by scientific integrity, the IRB should expect investigators, sponsors and institutions to explicitly address and minimize such burdens. Approaches could include limiting the number and duration of visits to the minimum necessary to achieve the study aims; considering remote access, home visits, or the use of other alternative locations whenever feasible; paying for transportation/housing costs that have been or will be incurred by subjects to participate; and considering how to enable participation by individuals with inflexible employment hours. Absent such considerations, subjects in a study are likely to represent a population of convenience, even if not one traditionally considered disadvantaged.
- Compensation, incentives and reimbursement. Belmont and the research regulations are primarily focused on avoiding exploitation, not on ensuring fair opportunity to participate (i.e., access). Accordingly, IRBs have traditionally eschewed financial incentives, and even compensation, as representing possible “undue influence.” SACHRP has previously released recommendations that question this reason for reluctance to pay research participants. If the risks and potential benefits are appropriately balanced and the rights, interests and safety of subjects are protected, it is not clear what standard would establish a financial influence as “undue.” SACHRP recommends that the framework of incentives, compensation and reimbursement described in its earlier recommendation be embraced, and every individual protocol should describe how and why it is paying participants in all three categories.
- Compensation for injury. IRBs should ensure that provisions for compensation for research injury are weighted to participant protection, not to minimizing institutional and sponsor liability. In particular, the cost and complexity of seeking legal redress makes that avenue more available to some than others, and relying on the right of a participant to individually seek redress through the courts should not be relied upon as a fallback for protections. Similarly, the informed consent document should not be designed to establish a legal defense against claims.
- Study recruitment plans. IRBs should assess whether the study recruitment plans match the goals of the study and are likely to reach the intended study populations. The selection of the study location, method of recruitment (e.g., online vs. phone vs. in-person), and whether study documents are available in languages other than English are just a few of the study design decisions that can impact the equitable selection of subject. Researchers should be able to explain why their choices are appropriate for the study.
- Study inclusion/exclusion criteria. IRBs should evaluate study inclusion and exclusion criteria critically. Criteria that explicitly exclude certain groups need strong justification, and criteria that may result in over- or under-inclusion of certain populations, intentionally or not, should similarly be justified.
§46.111(b) asks the IRB to ensure that “additional safeguards” are in place to protect research subjects “vulnerable to coercion or undue influence.” What measures constitute adequate safeguards in these circumstances, and how should their adequacy be assessed? Should the requirement for such safeguards be limited to those “vulnerable to coercion or undue influence”, or are there concerns of social justice, as that term is used in the Belmont Report, that should lead to a more expansive interpretation of vulnerability to exploitation?
IRBs have historically been concerned with exploitation, which reflects the regulatory focus on protecting research subjects regarded as vulnerable. This protection has been seen primarily as prevention of the inappropriate or not entirely voluntary inclusion of subjects. The demands of certain groups, beginning with those living with HIV/AIDS, for access to research and its potential benefits turns this concern “on its head,” and forces us to ask if there are subjects who are inappropriately excluded.
The history of scientifically unjustified exclusion from research of women and racial and ethnic groups, for example, suggests that this is a valid concern. The exclusion of such individuals often lowers the costs and burdens to the research community, while denying to the excluded populations the potential benefits of participation and the longer-term potential benefits to be gained by knowledge about their condition. Institutions and sponsors are already responding to this concern by actively seeking broader participation, but this effort needs to be systematically embraced.
Inappropriate exclusions have two major consequences: they can limit the scientific generalizability and applicability of results; and they disenfranchise communities from contributing to the common good. It is tempting to justify the need for broader inclusion primarily as a matter of scientific generalizability, but broad inclusion should not be justified based on science alone. For many trials, sufficient inclusion to guarantee the generalizability of results to all possible groups or categories of patients is impractical or impossible, but that impracticality cannot be used to justify exclusions that really reflect implicit bias, paternalism, or financial considerations. Disenfranchisement is unjustifiable and contributes to loss of trust in, and support for, research and science in general.
The language and current interpretation of the regulations is problematic in another way. While there are populations, like children and those lacking capacity to fully consider whether to participate in research, that are intrinsically vulnerable, there are many more populations that are made vulnerable by their environment and social circumstances. Such extrinsic vulnerabilities may be manifestations of economic need, societal power structures, and structural inequities in the healthcare delivery system. To accept such extrinsic sources of vulnerability as fixed is to use reasoning akin to the logic of racism, in that the resulting exclusions only reinforce the original vulnerabilities, and exclusion is implicitly blamed on the excluded.
SACHRP recommends that IRBs avoid taking the easy route of simply excluding potentially vulnerable populations from research participation in response to circumstances that could lead to coercion or undue influence. This recommendation may raise practical challenges: for example, individuals may be unduly influenced if participation in research is the only way to gain access to medical care. In this case, the proper solution is to ensure that adequate care is available outside of research participation. While individual IRBs cannot change the inequities in the current healthcare delivery system, they can suggest ways to minimize the burdens of research to enable broader, more equitable participation, and institutions conducting research could address inequities of access in their communities as an aspect of reciprocity for seeking their participation in the research enterprise.
§46.107(a) requires that the IRB “consider inclusion” of individuals who are “knowledgeable about and experienced in working with” disadvantaged populations if it regularly reviews research involving such groups. This language is often interpreted to apply when research targets a specific group, but given the diversity of the U.S. population, should the language be interpreted more broadly to require that inclusion of such members be the rule, rather than the exception? And, given the awareness of ubiquitous structural racism, is it sufficient to rely on expertise and experience rather than representation to “promote respect for (the IRBs) advice and counsel?”
Knowledge about, and experience in working with, disadvantaged populations cannot substitute for the lived experience of those populations. At the same time, the requirement that IRBs be of a manageable size and also include substantial expertise related to scientific disciplines, research methodologies and ethical analysis makes it unrealistic to expect representation of all potentially involved communities at all IRBs during the review of all protocols. On the other hand, the mandate for single IRB review in multi-site studies suggests that there could be heightened requirements for such reviewing IRBs. Some of the administrative efficiencies that were the justification for the single IRB requirement could be invested in broader community representation on those boards. Accordingly, SACHRP recommends that, as a start, OHRP provide guidance that IRBs serving as the single IRB in multi-site research of federally funded projects have robust community representation. In 2001, the National Bioethics Advisory Committee recommended that non-scientists and non-affiliated members should constitute 25% of an IRB’s membership to ensure adequate representation.
SACHRP also recognizes that simple representation may not be sufficient to guarantee a community voice. The voices of community members may not be fully valued, and there may be power imbalances between non-scientist community members and other members of the committee. Ensuring that voices are heard is the responsibility of the meeting Chair, and SACHRP recommends that OHRP develop guidance on what constitutes adequate training for IRB Chairs and standards for their evaluation, with particular attention to eliciting input from all members of the committee.
General considerations and the role of the IRB
An evolving understanding of the application of Justice to research involving human subjects raises practical concerns of how the research enterprise should fulfill its ethical duties. The IRB is not the sole or even primary seat of ethics in research; IRBs function as oversight bodies that detect and sanction ethical lapses and thus help motivate a primary concern for ethical investment, study design and execution among sponsors, institutions and investigators. While the IRB can be attentive to issues of Justice, it alone cannot create a just research enterprise.
If society demands, or institutions proactively commit to, Justice in research, the IRB is a natural entity to take on responsibility for oversight. IRBs are already charged with ethical oversight, and Justice is within their mission. They are the only entities established by regulation and are themselves subject to compliance requirements. On the other hand, IRBs are already criticized for “mission creep,” and must deal with an increasingly complex research environment using rapidly evolving technology in the context of social norms that are evolving equally quickly. They must compete with other institutional and population needs for resources at a time when such resources are stretched. But the alternative, creating yet another institutional entity that would not have the well-developed community that supports IRBs, would be an inefficient use of resources and would risk a fragmented response to Justice concerns. Further, research ethics is a matter of balancing overlapping and sometimes inconsistent principles (e.g., Respect for Persons, Beneficence, and Justice, or protection from inappropriate inclusion versus disenfranchisement through inappropriate exclusion), and balance would be hard to find if different groups were tasked with addressing different principles.
Therefore, IRBs should be tasked with seeking to ensure that duties arising from Justice are discharged. To this end, SACHRP recommends that those duties be articulated as plainly and as unambiguously as possible, extending the recommendations of the Belmont Report if necessary, and that federal agencies develop formal guidance or regulation that gives them force. It is beyond SACHRP's scope and expertise to develop the specifics of such recommendations, which should be based on a national conversation that involves the research community, the public (as actual and potential research subjects and as the funders of research), and traditionally underserved or excluded communities. Nonetheless, until such a conversation takes place, the research community cannot be complacent. While it may not be appropriate to articulate standards, as a start, SACHRP recommends that IRBs encourage attention to Justice by requiring that research proposals include a discussion of Justice and access. Such discussions should be informed at the institutional level by formal consultation with affected or potentially affected communities.
Many investigators and IRBs have made progress in the use of study-specific and standing community advisory boards, community review boards, and community-engaged research designs that incorporate the perspectives of racial and ethnic minorities, persons who use a language other than English, and underserved groups. Knowledge-sharing and best practices in this area should be encouraged, so that served communities have a voice in the development and execution of the research agenda. Industry sponsors, if they are not already doing so, should consider ways to elicit involvement of affected communities (representing potential customers) in both investment strategy and study design.
SACHRP recognizes that the reasons for the systematic underrepresentation of some populations in research are complex. Many are related to the social determinants of health, and others may be cultural. IRBs, which operate in the context of individual research studies or research programs, do not have the reach or the tools to dismantle these entrenched social and structural barriers. Yet, both Justice and Beneficence demand that research be open to all members of society affected and potentially affected by the conditions, problems or diseases being studied.
Issues of justice inspire demands for quick action, but it is far easier to identify problems than to craft meaningful and sustainable solutions. In addition to the general recommendations above, SACHRP recommends that DHHS establish equity and justice to be a primary research program, not only an add-on consideration to existing research. Research into health equity, the social determinants of health, and the effects of social and structural injustice on health disparities would fall in both the biomedical and social behavioral domains. Such research is methodologically feasible, may be less expensive than traditional interventional biomedical research, and should help to clarify the complex relationships between structural injustices, the social determinants of health, and their biological effects. Policy informed by such research could have more immediate and profound impacts on America’s overall health than any individual therapeutic discovery, even a “breakthrough.”
- Bierer BE, White SA, Meloney LG, Ahmed HR, Strauss DH, Clark LT. Achieving Diversity, Inclusion, And Equity in Clinical Research. Guidance document from the MRCT Center of Brigham and Women’s Hospital and Harvard. 2020. https://mrctcenter.org/diversity-in-clinical-trials/
- National Institutes of Health, UNITE, https://www.nih.gov/ending-structural-racism/unite.
- National Institutes of Health (NIH) National Institute on Minority Health and Health Disparities (NIMHD) released a new funding opportunity announcement, Understanding and Addressing the Impact of Structural Racism and Discrimination on Minority Health and Health Disparities. This initiative will support (1) observational research to understand the role of structural racism and discrimination (SRD) in causing and sustaining health disparities, and (2) intervention research that addresses SRD in order to improve minority health or reduce health disparities. https://grants.nih.gov/grants/guide/rfa-files/RFA-MD-21-004.html
- Walensky RP, Centers for Disease Control, Media Statement. https://www.cdc.gov/media/releases/2021/s0408-racism-health.html
- NIH stands against structural racism in biomedical research. https://www.nih.gov/about-nih/who-we-are/nih-director/statements/nih-stands-against-structural-racism-biomedical-research
- King, PA (2005) Justice Beyond Belmont. In JF Childress, EM Meslin (Ed.), Belmont Revisited: Ethical Principles for Research with Human Subjects (pp.136-147). Georgetown University Press.
 Language is important. SACHRP represents the scientific and academic establishment, and inherits that enterprise's assumptions and biases. Justice in the practice of research on human subjects cannot be fully realized without trying to make these assumptions and biases explicit, but we are doing so with language that has become charged with meaning. There are several terms, in particular, that we have used that may carry unintended meanings that are contrary to the intent of our recommendations.
Science — when we write about science, we are talking about systematic learning from observation or experiment. We are not talking about the academic, government and commercial entities that embody institutional science. We mean science as a method that recognizes that truth is best approximated empirically, and that we must not accept the truth of our perceptions or beliefs without testing.
 Objectivity — objectivity and science are inseparable, but objectivity is aspirational. As people, we can only try to distance ourselves from our assumptions and beliefs, but true objectivity is impossible. There are two reactions to that impossibility — we can declare objectivity unattainable and therefore without value, but such a reaction condemns us to knowingly embrace one set of assumptions as better than another, a strategy that cannot but lead to conflict and division. Or we can keep trying to identify and eschew our assumptions to try to build a common understanding where the only assumptions come from our collective humanity.
 Access — Facilitating access does not necessarily lead to participation; it is a core ethical tenet that research participation must be voluntary. Thus, opportunities must be available, but individuals must also have reason to participate. Our recommendations largely address access, which is a necessary prerequisite, but articulating the goals of research in a way that is compelling to historically excluded communities is likely to require time and the rebuilding of trust, which will only happen through practice.
 Phase 1 biomedical research on healthy volunteers represents a special case that does not fit easily into this framework. Such research enrolls healthy individuals to assess the safety of an intervention while trying to avoid risks that might arise from existing health conditions. It is difficult to apply considerations of distributive justice in this context, although the considerations discussed in the following section, “Inappropriate inclusion versus inappropriate exclusion,” still apply.