Navigating the Perfect Storm:Ethical Guidance for Conducting Research Involving Participants with Multiple Vulnerabilities
The development of ethical guidelines and regulations regarding human subjects research has focused upon protection of vulnerable populations by relying on a categorical approach to vulnerability. This results in several challenges: First, Institutional Review Boards (IRBs) struggle to interpret and apply the regulations because they are often vague and inconsistent. Second, applying the regulations to subjects who fit within multiple categories of vulnerability can lead to contradictions and the rejection of research that would be permissible if only one category were applicable. Finally, some potential subjects have social and other context-based vulnerabilities that are not described in the federal regulations and therefore not considered in IRB deliberations. IRBs and investigators lack guidance on how to address the problem of multiple vulnerabilities in a way that strikes a balance between protection and respect for persons. In this essay, we evaluate the acceptability of the existing federal regulations with respect to research participants with multiple vulnerabilities, offer strategies for rethinking the concept of vulnerability, and outline a context-based normative framework to account for the compounding effects of multiple vulnerabilities.
VULNERABILITY IN HUMAN SUBJECTS RESEARCH
The development of ethical guidelines and regulations regarding human subjects research has focused upon protection of vulnerable populations by relying on a limited typology of vulnerabilities. This results in several challenges: First, Institutional Review Boards (IRBs) struggle to interpret and apply the regulations because they are often vague and inconsistent. Second, applying the regulations to subjects who fit within multiple categories of vulnerability can lead to contradictions and the rejection of research that would be permissible if only one [End Page 451] category were applicable. Finally, some potential subjects have social and other context-based vulnerabilities that are not described in the federal regulations and therefore not considered in IRB deliberations. Generally, IRBs and investigators lack guidance on how to address the problem of vulnerability, especially multiple vulnerabilities, in a way that strikes a balance between protecting them and respecting them as persons by allowing them or their surrogates to choose to participate. In this essay, we evaluate the acceptability of the existing federal regulations with respect to research participants with multiple vulnerabilities (including where they are vague, inconsistent, or contradictory), offer strategies for rethinking the concept of vulnerability more generally, and outline a context-based normative framework to account for the compounding effects of multiple vulnerabilities.
Under the Federal Policy for the Protection of Human Subjects, 45 C.F.R. Part 46, also known as the "Common Rule," vulnerability is understood as an attribute of a particular population.1 One is considered "vulnerable" simply by being classified within the following categories: "children, prisoners, pregnant women, handicapped, or mentally disabled persons" (45 CFR 46.107(a)). Safeguards meant to protect prisoners may not be applicable to minors, yet would need to be in place to protect minors who are prisoners. In addition, subjects who have mental disorders are generally vulnerable to coercion, yet there is no guidance to be found in the federal regulations on how to provide additional safeguards for this population. For these and other reasons, this "categorical" definition of vulnerability is too limited to be helpful for guiding IRBs and needs to be replaced with a "contextual" approach, which we will outline in the final section.
In order to illustrate the current gaps in the regulations regarding vulnerable populations, consider the following hypothetical research project. The proposed study focuses on a sample population that represents three distinct but related vulnerabilities: incarceration, legal incompetence due to age, and mental illness (Table 1). The study presented is plausible and based upon current research, but does not represent any real investigation, proposed, funded, or underway. The study focuses on exploration of recent findings in epigenetics concerning aggression (Provencal, Suderman, et al. 2014) and addiction (Demers, Bogdan, and Agrawal 2014), particularly for genes related to serotonin and dopamine metabolism (Preuss et al. 2015) as they relate to impulsivity (Iofrida, Palumbo, and Pellegrini 2014). Although studies in adult populations [End Page 452] have provided some insight into potential mechanisms, understanding the epigenetic mechanisms requires longitudinal study (Tremblay and Szyf 2010) and focuses on younger populations to determine the gene x environment interaction and development of behaviors.
The hypothetical study proposes recruiting at least 120 incarcerated youth (ages 13 through 17) for study of methylation patterns in target genes related to aggression and addiction. Comparison groups will be divided into four cells based on history of physical aggression and substance use disorders (non-aggressive/non-substance use, aggressive/non-substance use, non-aggressive/substance use, and aggressive/substance use). Any youth with a diagnosis of psychosis will be excluded from the study. Blood samples are collected from all participants for analysis of methylated DNA patterns and will be stored in a biobank. Additional history regarding childhood behaviors and possible exposures to environmental stresses, such as abuse, are obtained from the youth and parent or guardian. The authors of the proposed research intend to restudy the participants at 5 and 10 years following initial investigation to determine outcomes and possible changes related to the epigenome and target behaviors.
Applying the Federal Regulations to the Hypothetical Case
In order to determine whether and how the current federal regulations fail to provide guidance for investigators who wish to conduct research involving subjects with multiple vulnerabilities, it is necessary to first test the limits of the categorical approach to vulnerability. In this section, we attempt to use the categorical approach to address issues related to the appropriate selection of research subjects, informed assent and parental permission process, and the appropriate use and disclosure of genetic information within the context of the hypothetical research study.
Applying the Regulations Leads to Contradictions and Rejections of Otherwise Permissible Research
According to the federal regulations, the permissibility of enrolling certain subjects in research is to be determined according to different criteria, depending on the population. For populations given special protections by the federal regulations, investigators must justify the inclusion of vulnerable populations to the IRB. For prisoners, permissible research must directly address some aspect of prison life or incarceration. The criterion of relevance is controlling here. More specifically, research [End Page 453]
Click for larger view
View full resolution
must focus on issues that are especially relevant to prisoners (45 CFR 46.306(a-b)). Permissible research may be relevant but not exclusively to prisoners, exclusively relevant to them, or relevant and could not be conducted with a different population. For example, research on the prevalence of substance abuse is an issue that is relevant but not exclusively to prisoners and could be conducted with a different population. Research examining sexual assaults in prison or the processes and effects of incarceration would be considered exclusively relevant to prisoners (45 CFR 46.306(a-b)). Research that is relevant but could be conducted with a different population requires a higher level of oversight than research [End Page 454] exclusively relevant to prisoners. These provisions are meant to guard against the use of prisoners as a population of convenience and to ensure that research remains relevant to this particular population.
Investigators would likely be able to satisfy the criterion of relevance for this study if they can successfully argue that it seeks to understand the causes of incarceration. To do so, they would need to be more explicit in establishing the hypothesized link between incarceration, substance abuse, and aggression. Investigators also would have to satisfy the potential counterclaim that aggressive individuals with substance abuse disorders may be found outside the prison walls. However, recruiting individuals from the general population who have engaged in potentially illegal activities related to aggression and substance use could actually present greater legal risks, as the disclosure and collection of such incriminating evidence could be damaging for potential subjects despite all efforts to maintain confidentiality. While histories of physical aggression and substance abuse are more likely to be found in youths charged with violent and drug-related offenses, they are not exclusive to such a group. These potential challenges aside, the criterion of relevance as it applies to prisoners could theoretically be met by this study. However, there are additional criteria that would also have to be met, because of the multiple vulnerabilities of this population.
Since these participants are also minors with histories of substance use disorder and aggression, additional justifications for including this population would be needed. The National Bioethics Advisory Commission (NBAC) provided 21 recommendations that address ethical issues involving persons with mental disorders that may affect decision-making capacity, some of which may be applicable to these participants. It should be noted that these recommendations are not codified in the federal regulations and it is unclear how widely IRBs have adopted them for use in their deliberations (HHS Working Group 2001). However, some of their recommendations, such as one that would prohibit selecting individuals with mental disorders as a population of convenience, are similar to protections already afforded to other vulnerable populations. In order to justifiably recruit this population, investigators must demonstrate that the burdens of participating will not outweigh the benefits. Considering the extensive psychiatric interviews as study-related benefits to be weighed against legal and physical risks could provide sufficient proof of the intent to satisfy this requirement. In this case, the criterion of relevance would be satisfied because the research is relevant and could not be conducted on [End Page 455] patients without a history of substance use disorder. As for the participants' status as minors, the criteria for approving their involvement in research likely to be regarded as minimal risk is not based on relevance to the child, but would instead be based on a risk/benefit assessment and the provisions for obtaining assent and parental permission. Put together, these present alternative criteria for justification for enrolling minors that we will call the criteria of risk. Thus, there are two competing ethical obligations.
In considering the ethical appropriateness of the criterion of relevance, more evidence is needed to support the claim that relevance should be a controlling factor in deliberations about the use of vulnerable populations. Although we will confine the following remarks to the juvenile population, in regard to the criterion of relevance the same arguments should apply to adult offenders as well. Arguably, juvenile offenders have interests that could be promoted by participating in research that focuses on aspects of their lives other than their incarceration (Pasquerella 2002). Limiting juvenile offenders' participation to either research focused on incarceration or research that offers direct benefits leaves out a range of important research (which may include the present study, depending on how the potential benefits are framed) that may serve these interests. Excluding juvenile offenders from certain types of studies based on the criterion of relevance, which is applied only in one other scenario that also involves pediatric research, could also be seen as an unfair practice that further limits their already constrained range of choices. Although there are historical reasons why research involving prisoners has been limited in this manner, respect for persons and justice require that juvenile offenders be treated fairly and allowed to choose from a less restrictive range of research studies (Hornblum 1999; Advisory Committee on Human Radiation Experiments 1996). Furthermore, it is not clear that relevance does or should determine utility for a particular population. Utility as measured by the value of the knowledge that investigators gain may be weighed very differently by potential participants. Although potential participants should have input into whether the criterion of relevance is a useful and effective safeguard against exploitation, IRBs and parents/guardians will ultimately be responsible for weighing the criterion of relevance against the criteria of risk.
To satisfy the criteria of risk in order for minors to be recruited for this study, investigators would have to show that the risk is minimal and that assent has been solicited and parental permission secured to allow participants to take on these risks. The current study would likely [End Page 456] be categorized as minimal risk by most IRBs since the physical risks are not greater than a routine doctor's visit. However, an argument could be made that it should be considered greater than minimal risk based on the subjects' current incarceration. Incarcerated children would be more susceptible than their non-incarcerated peers to some of the potential legal risks associated with disclosure of the test results. For example, biomarkers that identify participants as higher risks for recidivism may influence the decision of the courts regarding the conditions of parole or other legal matters. Unauthorized disclosure of this information could either harm or benefit the subject, depending on the specificity of the results. Likewise, publication of these results could impact sentencing guidelines, thus placing "at-risk" youth at higher risk for lengthy mandatory minimum sentences. Although the study would address problems related to incarceration (which may or may not benefit the participants), the prospect for direct benefit in a study like this is uncertain. The potential for direct benefits is limited to the psychiatric interviews (which may reveal additional sources of mental illness that can then lead to referrals to psychiatric treatment) and the disclosure of genetic information that may or may not be useful for the treatment of their mental illness. The present study may help investigators understand the possible genetic link between aggression, substance abuse, and incarceration, but may also perpetuate harmful stereotypes about marginalized populations. Comparing incarcerated and non-incarcerated minors in terms of the commensurability of psychosocial and legal risks is necessary for satisfying the criteria of risk for juvenile offenders (45 CFR 46.305(a) (3)). If this study were found to be greater than minimal risk with the potential for direct benefit or increases in knowledge-for example, about aggression, substance abuse, or impulsivity-minors could be justifiably enrolled. However, enrollment of juvenile offenders who are minors would require a much higher level of scrutiny (including DHHS review) that focuses more on relevance than risk and would require investigators to show that the study is focused on "conditions particularly affecting prisoners as a class" (45 CFR 46.306(a)(2)(iii); 45 CFR 46.405-6). This guidance is vague, but could include substance abuse and aggression, though these conditions are not specific to prisoners. Again, the criteria of risk and the criterion of relevance conflict in confusing ways. In this case, context is important as it signals that the intended subject pool is at greater risk of legal consequences of participating in research than a control group of non-incarcerated children with substance abuse and a history of physical aggression. [End Page 457]
Although the criterion of relevance may be satisfied for the vulnerable categories of incarceration and mental disability, the question of whether the criteria of risk can be satisfied is dependent on the IRB's determinations regarding minimal risk. In this case, it appears that although the likelihood of potential inappropriate disclosure to law enforcement and loss of confidentiality would be the roughly the same for prisoners and non-prisoners, the consequences could be more severe for the latter as descriptions of criminal activity related to violence and substance abuse could carry legal implications where none existed before. Thus, the risks would not be commensurate with what non-prisoners might accept.
Taken together, the criteria of risk and relevance provide a framework for IRB deliberations to follow, yet there is no method for determining whether relevance or risk should be the controlling factor in deliberations involving populations with multiple vulnerabilities. There is no criterion of relevance that would guide IRB deliberations for minimal risk research involving children and so no basis for comparison between vulnerabilities (incarceration vs age/maturity). Since the most restrictive interpretation of the regulations would normally prevail in IRB deliberations, this population would potentially be excluded from participating in research that is greater than minimal risk-even though other equally applicable guidelines would allow them to enroll. For research involving persons with mental illness, the regulations are silent and the only NBAC recommendation offered is that these subjects may not be enrolled in research if the study protocol could be conducted with persons who do not have mental disorders.
The current regulations fail because of the incommensurability of these criteria. There is no single standard by which to judge, nor is there a clear method for weighing the competing criteria. Should the least restrictive or the most restrictive standard determine approvability? How should the criteria of risk be evaluated when there are two competing interpretations? These questions may have answers, but they are unlikely to be found in general guidelines. Instead, IRBs should draw on their knowledge of the potential subject pool and what is known about their daily lives and preferences. Knowledge of preferences and what is important in the decision-making process can be discovered by turning away from a categorical approach that leads to contradictions and incommensurability to focus on a contextual approach. Context is particularly important when making determinations regarding institutional and developmental constraints on the voluntariness of assent and parental permission. [End Page 458]
Regulations are Vague and Inconsistent Regarding Multiple Vulnerabilities
As incarcerated juvenile offenders, the ability of these subjects to freely assent to participate in the study would likely be constrained by their immaturity and their incarceration. As minors, they are not considered legally competent and are therefore not able to consent to research, but ordinarily would be asked to give their assent. However, their ability to do so voluntarily as incarcerated subjects with mental illness may be questionable. As incarcerated subjects, they are susceptible to undue inducement and coercion (Pasquerella 2002). This section will attempt to identify and address some of these concerns using the current regulations as a guide.
There are many questions related to assent, parental permission, and potential future consent that this case raises. One particular controversy made more difficult by the confluence of multiple vulnerabilities is the definition of "assent" and how it should function to protect young subjects while allowing them to participate in the decision-making process. This term is defined by the federal regulations as the "affirmative agreement to participate" (45 CFR 46.402). This definition is unhelpful as it does not provide guidelines for determining precisely which capacities an adolescent would need to demonstrate in order for his agreement to be honored or whether some research should require a higher threshold for voluntariness. Bioethicists disagree about what level of understanding adolescents must demonstrate in order for their assent to be taken seriously (Diekema 2003; Tait and Geisser 2017). Some commentators have interpreted "assent" as simply an exercise in moral reasoning designed to help promote the child's moral growth, but not to be taken as a true expression of their nascent capacity for self-determination (Miller and Nelson 2006). Others have argued that, based on empirical data, adolescents should be allowed to consent to research participation (along with their parents or guardians) at the age of 11.2, which would mean that participants in this study would be held to a higher standard of decision-making (Hein et al. 2015). Still others argue assent should not be required until age 14 (Wendler and Shah 2003). It is beyond the scope of this paper to try to solve this problem of multiple conflicting standards. However, it is important to note that the lack of guidance from the federal regulations has contributed to the confusion regarding who should be consulted to provide consent or assent.
IRBs are responsible for ensuring that the proper procedures are followed for obtaining assent from minors. Assent forms and processes [End Page 459] must be reflective of the "ages, maturity, and psychological state of the children involved" (45 CFR 46.408(a)). How IRBs are meant to determine the appropriateness of the proposed processes is left to the individual Boards to decide. However, these processes may be waived at the discretion of the IRB, provided that the subjects cannot "reasonably be consulted" (45 CFR 46.408(a)). This waiver is intended to cover the participation of the very young (newborns, children under 7 years old). However, minors who are developmentally disabled or whose decision-making capacity (already less than adults' due to age and maturity) may be limited by mental illness may also qualify for this waiver.
In this case, potential participants who suffer from a condition that is known to affect impulse control may be more prone to disregard the study-related risks or want to take excessive risks. In addition, as minors, they may be unable to understand the full range of rights to which they are entitled and may be not be able to recognize an incomplete assent process. In this case, they would become overly reliant on the research team, which would result in a diminishment of their nascent autonomy. Although the severity of their cognitive impairments may vary among participants, it is reasonable to assume that the capacity to assent will likely be diminished for many potential participants. In cases where capacity is unclear, an assessment by a clinician will be needed. Even if the requirement to obtain assent is waived, parental permission may still be required.
Permission from either a parent or legal guardian is ordinarily required in order for minors to participate in research. In this hypothetical study, only one parent needs to give their permission for their child to participate. However, this requirement may create some practical problems for investigators and some participants (Lane et al. 2012; Wolbransky et al. 2013). Parents of minors who are incarcerated may not be easily reached or may not wish to participate in decision-making involving their child. In some cases where minors have been abused or neglected, parents may not be permitted to be involved in the decision-making process. They also may choose not to visit their incarcerated children, which would make in-person consent impossible (Wolbransky et al. 2013). In these cases, the court may appoint a guardian to make decisions for the child or the investigator may seek a waiver of informed consent (45 CFR 46.116 (d)). Although the federal regulations require only one parent's permission to enroll a minor subject, some IRBs will not allow a subject to be enrolled if the other parent objects. The federal regulations provide no guidance on how to resolve these conflicts and very limited guidance on the other issues [End Page 460] related to obtaining parental permission within the often complex social context of research involving participants with multiple vulnerabilities. A contextual approach is best suited to address these conflicts, as it would involve evaluating the risk level of the study, involvement of each parent in the child's life, and the potential for secondary gain or loss. Mediation by an IRB member or an outside research ethics consultant might be needed to facilitate communication among the parents and investigator(s) in order to determine whether the subject should be enrolled.
In addition to potential problems related to obtaining parental permission and evaluating participants' capacity to assent to participate, the question remains whether participants' multiple vulnerabilities make them more susceptible to undue influence or coercion. Central to concerns about undue influence is how the term "undue" is defined and to what degree the offer of some social good distorts decision-making. According to Emmanuel et al., there are four necessary components of undue influence: offered good, excessive offer, risk of serious harm, and poor judgement (Emmanuel, 2005a; Emmanuel et al., 2005b). Each of these components must be present for the influence to be undue.
There are several possible sources of coercion or undue influence in the hypothetical study under consideration. In general, adolescents are thought to be susceptible to the influence of authority figures, which may include family members, teachers, or, in the case of institutionalized populations, guards or other prison staff. Institutional staff may, intentionally or not, give the impression that incarcerated minors who participate in the study would earn favor or be kept out of harm's way. Children who give their assent may be uncomfortable with withdrawing from research that they find unpleasant, especially if the research appears to be endorsed by the institution in which they are confined. Parents may also feel pressured (unintentionally or not) by the institutional staff to give their permission in order to secure some perceived advantages that they think may be related to their cooperation. Thus, incarcerated minors and their parents are likely to encounter offers of social goods that affect their ability to give their permission as well as pressure to conform to the authority of the investigator in order to obtain these goods.
Beyond questions of undue inducement, there are concerns that prisoners, as an institutionalized population whose daily activities are constrained by the terms of their incarceration and the carefully ordered social structure in which they live, may be more vulnerable to coercion than their free-world counterparts. In this case, the voluntariness of these [End Page 461] young offenders to provide assent may be more likely to be threatened by "arbitrary intervention by prison authorities or prisoners" (45 CFR 46 Subpart C) than if they were adults. Participants who have mental disorders may be particularly susceptible to these kinds of interventions, which might involve unexpected changes in prison rules or routines that adversely affect participants' ability to complete the study. In addition, immature minors may be susceptible to being coerced by staff and fellow prisoners. For incarcerated subjects, additional protections are deemed necessary because of the longstanding assumption that prison life is inherently coercive. The history of human subjects research involving prisoners has shown that this population has been abused by unscrupulous investigators who wished to take advantage of their vulnerability, especially to financial pressures (Hornblum 1999).
In any study involving populations with multiple vulnerabilities, additional protections are needed to mitigate any potential sources of coercion or undue influence. In this case, there are a number of institutional pressures that have been identified and will need to be addressed. However, the current regulations focus on possible sources of undue inducement in prison and ignore the pressures that adolescents might face to continue with a study that has the appearance of being supported by the institution. Furthermore, they are silent on how possible sources of coercion or undue influence within the prison environment may disproportionately affect minors or persons with mental disorders. In addition, the current regulations are vague with regard to what constitutes "assent," have no process for managing disputes between parents who disagree about whether the adolescent should participate, and provide no guidance for how or when to test the decision-making capacity of adolescents with mental illness.
Some portions of the guidance provided by NBAC may be helpful in devising processes that can solve some of these problems. NBAC recommends that potential participants undergo a capacity assessment and, if found to be incapacitated, be allowed to participate with the approval of a surrogate decision maker or legally authorized representative (LAR). Surrogate decision makers would be required to use the best interests standard for making decisions regarding the participation of a decisionally incapacitated minor. LARs must also be available to monitor recruitment, participation, and withdrawal. Decisionally incapacitated patients must be informed of that determination before permission to participate is sought [End Page 462] from a LAR. Special care should be taken in communicating this to children and adolescents. If permission is granted, the participant must be notified and can object. In general, the objection must be heeded. However, it is unclear whether individuals who were never capacitated, for example, adolescents who have suffered from mental illness for most of their lives, would be able to participate under these guidelines. Furthermore, there may be similar difficulties in locating LARs as may occur when trying to locate parents.
The confluence of multiple vulnerabilities in this population affects their capacity to give assent to participate in research, especially research that involves an extended psychiatric interview and the genetic analysis of their biological samples. Although the physical risks of harm are minimal, the question of whether institutional pressure to enroll or not enroll would be too great for minors with mental disorders to manage remains open. The standard for assent is much lower than for consent, yet the potential for limited judgment due to immaturity or the undue inducement to participate posed by the study remains relatively high. These considerations should affect both the approvability of the initial study procedures and the long-term follow up outlined by the investigators, including biobanking of participant's blood samples.
Regulations Fail to Account for Social and Other Context-Based Vulnerabilities
It is still an open question in the bioethics literature whether subjects who are minors at the time of their enrollment in a study need to be re-contacted in order to be asked to consent to future uses of their blood samples (Samuel, Knoppers, and Avard 2012). If the samples and interviews are made publically available or are de-identified, future studies using these data would not constitute human subjects research. However, it is unlikely that the follow-up study described in the protocol would use publically available data, nor would it be appropriate to make the past medical history of these individuals public. In this case, it would be appropriate for the parental permission and assent forms to contain a separate section indicating that a future study is intended. Such a study would raise logistical difficulties in that it is very likely that many of these individuals will be lost to follow up once they leave the juvenile justice system. In order to be viable as a longitudinal study, investigators would need to develop a plan for maintaining contact and state their intention to do so in the assent and parental permission forms. Investigators would [End Page 463] need to establish means for maintaining contact with participants that is independent of the criminal justice system. Additional questions regarding the return of research results to the subjects or their parents and the ability of subjects to withdraw their samples at a later date will also need to be addressed. These questions are part of a much larger debate in the research ethics literature regarding privacy rights and are made more complex by the presence of multiple vulnerabilities.
Given the heritability of methylation patterns, access to the DNA samples of subjects may have value for future research that extends beyond the 5-10 year timeline mentioned in the protocol. If this research study and others that are built on it succeed in uncovering heritable methylation patterns, disclosure of the results may eventually have an influence on subjects' reproductive decision-making, including decisions about potential germline genetic modifications to avoid the mental disorders mentioned above. Results from this study and others like it may also influence more mundane decisions, such as where to live, how to achieve access to psychological support, etc., based on the risk profile that further research on these methylation patterns might reveal. Although these potential downstream effects may not come into play until much later, the need to consider how research in this field might help participants avoid future vulnerabilities is present in these more nascent stages of development. Potential subjects may find this information useful and could request that individual research results be returned. If they do, would investigators be obligated to provide this information? Would they also be required to provide participants with continued control over their samples? Or are investigators justified in asking participants to relinquish control over their samples to future investigators?
One challenge that emerges when considering these issues is whether children have the right to decide how their genetic information will be used and disclosed and how to protect that right if these decisions will need to have already been made by others in order to conduct the study. This right to control how one's genetic information is used may be unjustifiably denied if children who grow up and decide that they want to have their samples removed from the biobank find that investigators have already used their data and generated additional links to identifiable information about them (Samuël, Knoppers, and Avard 2012). If made publically available, this information may be used to generate a criminality profile or used in other discriminatory ways that would not be covered by existing legislation (e.g. the Genetic Information Non-Discrimination Act). In [End Page 464] genetic counseling, this variant of privacy rights is commonly known as the "right to an open future," i.e. the right to have decisions about what genetic information is disclosed postponed until the child reaches the age of majority (Davis 1997). Advocates for this right argue that privacy rights as applied to decisions regarding the disclosure of genetic information should be considered one of the "rights in trust" that parents should not violate. Due to children's vulnerability and the potential for the loss of a right to an open future, Gurwitz et al. (2009) recommend collecting pediatric samples but withholding them and the associated individual phenotypic data from investigators until the subjects are able to legally consent. They argue that public trust in research may be violated if children are held to a "lower standard" for voluntariness and altruism than adults (2009). However, others have argued that focusing on protecting these rights would result in a delay in access that unjustifiably impedes research when what is needed to protect confidentiality is tougher security measures (Hansson and Maschke 2009; Brothers and Clayton 2009). They also note that the potential for widespread stigmatization as a result of publishing information from studies that provide support for genetic determinism will affect both participants and non-participants alike, rendering the potential for an open future less certain (Hens, Wright, and Dierickx 2009). Another question related to informed consent in longitudinal studies arises when considering that children, especially those with mental disorders, are unlikely to be able to predict whether they would in the future find some kinds of research objectionable.
There may be uses of their DNA sample to which pediatric research subjects would not in principle be able to agree. Depending on the level of severity of their mental disorder, persons suffering from impulse control disorders may not gain this capacity or may continue to have limited ability to make decisions about their participation in research. It is ethically problematic for children to be asked to "give up the right to withdraw at the moment of donation" (Hens et al. 2011, 736). It should also be noted that some of the adolescents in the study may be parents themselves and so would need to consider whether and how sharing their genetic information with investigators might affect their children's right to an open future. The option to allow investigators to make this information available to subjects' children could be included as part of the assent or parental permission form. Biobanking the blood samples of subjects with multiple vulnerabilities such as those described above raises several questions about assent, the right to control one's biological information, [End Page 465] and the right to an open future. Subjects who are limited in their ability to assent may not be cognitively capable of understanding the potential far-reaching consequences of the decisions that are being made on their behalf. In these cases, the irreversibility of the decision plus the potential for harm are relevant to the IRB's decision-making process.
In this section, we have discussed a number of gaps in the federal regulations regarding how IRBs should determine the justifiability of enrolling subjects with multiple vulnerabilities. Among the concerns are how closely the research study should attempt to address the health issues related to incarcerated minors with mental disorders and how investigators should analyze the risks of harm compared to the potential for benefit. With a subject population facing the number of consent-related challenges brought on by the confluence of multiple vulnerabilities, how the assent/ parental permission process is structured will be critical. It is clear that these subjects will need assistance with decision-making. How one might enhance the permission-seeking process in light of these challenges is one of many issues we will discuss in the final section.
RECOMMENDATIONS ON HOW TO CONSTRUCT A CONTEXTUAL APPROACH TO THE EVALUATION OF RESEARCH PROTOCOLS INVOLVING SUBJECTS WITH MULTIPLE VULNERABILITIES
We recommend a contextual, rather than a categorical approach to the concept of vulnerability as it is applied across heterogeneous populations. Such an approach would require that IRBs and investigators identify and evaluate potential vulnerabilities by examining the interface between the investigators, sponsor and host institutions, communities, and potential subject pools to identify the research-related burdens that participation would impose (without corresponding direct benefit) and to determine whether participation in the research study is likely to exacerbate baseline inequalities and discrimination (Hurst 2008; Schroeder and Gefenas 2009). For example, if a research study were found to be likely to exacerbate baseline inequalities and discrimination without corresponding benefit to the subject, it would not be approvable. In addition to this focus on research-related burdens, investigators would also be responsible for drawing inferences about possible sources of structural coercion that may inappropriately influence subject populations that already experience numerous forms of discrimination and marginalization in their daily lives. The term "structural coercion" describes the "ways in which broader social, economic, and political contexts" shape the options available to [End Page 466] potential research participants (Fisher 2013, 355). In the hypothetical case described above, a contextual approach would look at the institutional environment in which the research is to be carried out, including the likely impact of structurally coercive forces that may need to be identified and mitigated by the investigators. Contextual investigation of the potential subject pool would not be based on idle speculation, but on examination of the social, economic, political, historical, and cultural characteristics of the local community using data from the social sciences to draw inferences about ways in which certain subpopulations within the subject pool may be vulnerable. While the criteria of risk would still play a role in evaluating the appropriateness of enrolling a particular subject pool in a study, the contextual approach would deepen and broaden concerns about risk by focusing on characteristics of specific subjects' pools and their research-related burdens while also identifying and addressing multiple sources of coercion and exploitation within the research environment itself. The criteria of relevance and risk would not stand alone as rigid and static thresholds for permissibility, but would be informed by the context of the research in order to form complementary descriptive and normative frameworks. Within this framework no single criterion would be necessarily weighed more heavily than the others. As applied to the hypothetical case mentioned above, a contextual approach would draw out the various forms of power-based asymmetries and social vulnerabilities that affect the ability of children to freely give their assent as well as the distribution of risks and benefits to the subject and the community. This approach is similar to the "layers" of vulnerability approach proposed by Luna and Vanderpoel, though their approach focuses instead on concerns that potential subjects are unfairly or inaccurately labeled as vulnerable and that those "labels" are persistent and reductive (Luna 2009; Luna and Vanderpoel 2013).
Incarcerated children with mental disorders face a number of asymmetrical power relationships that place them at a serious disadvantage. These power asymmetries bear special scrutiny as they offer the opportunity for exploitation, manipulation, and exposure to undue risk. Situationally, these subjects face a number of institutional constraints. First, the juvenile offender resides in a facility with its own rules and conflicting demands. Study participants would face the conflicting demands of the staff and guards who are responsible both for their wellbeing and for maintaining order. In addition, participants reside in an institutional context in which privileges (including shortened placement time) are earned by compliance [End Page 467] with rules. Even if the protocol explicitly precludes that kind of treatment, participants are likely to assume that compliance with the rules of the protocol will earn rewards. Second, the investigator and the research team are interested in uncovering a detailed psychiatric and social history. Subjects may feel compelled to participate, yet uncertain about how the information will be used. Once the interview is completed, they are no longer in control of what is done with the information, how it will be used, and who will have access to it. Third, participants are susceptible to exploitation because of a lack of contact or involvement by their parents. In some cases, there may be no one to look after their best interests, except themselves. These power dynamics are not brought to the fore in order to discourage research, but to point to the three areas in which subjects are most vulnerable. Even if capacitated to give assent, they are still subject to the relational vulnerabilities inherent in their position as incarcerated minors with mental disorders (Henderson, Davis, and King 2004).
Kipnis has developed what he calls an "analytical approach" to address some of the deficiencies of the categorical approach to vulnerability. In particular, this approach outlines a typology of vulnerabilities in pediatric research subjects for the purpose of determining which of these characteristics may undermine consent or the ethical permissibility of the study (Kipnis 2003). Using these typologies as a guide for a contextual approach, there are four varieties of vulnerability that cut across the multiple categories of vulnerability described above. Incarcerated children with mental disorders are incapacitationally, juridically, deferentially, and socially vulnerable. First, their incapacitational vulnerability emerges from what is known about their capacity to make decisions. Although, as a population, adolescents are not considered mature enough to make informed decisions, adolescents with high impulsivity will likely face greater challenges than adolescents with normal impulse control. Brain imaging studies featuring fMRI have shown that adolescents in general are susceptible to peer pressure, have difficulty with planning and monitoring their responses to stimuli, and are more likely to take excessive risks (Pope, Luna, and Thomas 2012). Furthermore, IRBs may have some concerns with enrolling subjects below the age of 14, especially if it may be impractical to obtain parental permission. Studies have shown that most children are not able to understand the purpose of research studies before this age (Wendler 2006). This study involves a number of laboratory tests that will occur without the active participation of the subjects. It is unlikely that even older adolescents would be able to understand the [End Page 468] purpose of these tests, which would potentially undermine their ability to provide consent or assent.
Second, these potential participants would be juridically vulnerable due to their incarceration and reliance on parental permission to enroll. Their incarceration puts them under the supervision of the criminal justice system, which has an interest in studies that might illuminate the means for identifying a genetic predisposition to criminal acts related to substance abuse and violence. Persons in authority may promote participation as a way to further their own interests and use their power over the juvenile offenders to compel their participation. Based on the current state of knowledge about how adolescent brains function, it would be plausible to conclude that adolescents with impulse control disorders and/or substance abuse problems would find themselves especially susceptible to the authority of others.
Third, although adolescents are not known for being deferential to authority as a way to mask their "underlying unwillingness to participate," in situations where deference might serve a useful purpose, it is plausible that these participants might feign interest and give their assent in order to gain access to some other resources (Kipnis 2003, 110). Even careful design by investigators would not necessarily dispel the view that some fantasized indirect or non-medical benefit would result from their participation.
Finally, there are a number of what Kipnis describes as "social vulnerabilities" that are not described in the federal regulations. They may arise from "stigmatizing perceptions" related to historically marginalized populations: racial or ethnic minorities, undocumented immigrants, LGBTQ adolescents, and children with limited English proficiency. Historically, these populations have been the subject of discrimination and oppression by various groups and individuals, including law enforcement. Racial discrimination in the juvenile justice system may exacerbate other research-related vulnerabilities, including the potential for inequitable subject selection and the exacerbation of mental illness due to increase in allostatic load (Green 2010; Snowden 2003). Assumptions about criminality and aggressiveness in some minority populations may lead to misclassification of these juvenile offenders. They may be more likely to be either underdiagnosed or misdiagnosed (Escobar 2012; Prins et al. 2012). In addition, potential subjects with existing mental illness who are also racial or ethnic minorities may face discrimination while incarcerated that further inhibits their ability to assent. Or their ability to assent may be underappreciated due to prejudicial assumptions about their decisional [End Page 469] capacity. The potential for misclassification as mentally ill due to these stigmatizing perceptions or the exacerbation of existing mental illness due to the stress of incarceration may also be present in children and adolescents who are undocumented immigrants.
Children and adolescents who are undocumented immigrants may be less willing to participate in research studies that collect demographic information, even though they may be able to contribute valuable data. They may also face pressure from authority figures who may be tempted to use the potential subjects' tenuous immigration status to encourage participation. Conversely, potential subjects may assume that their participation would improve their candidacy for citizenship. Vulnerability related to immigration status may overlap with language barriers that could deny subjects the opportunity to participate in research due to their inability to reply to the study announcement, understand the assent form, or to complete the enrollment process.
LGBTQ adolescents who are also juvenile offenders face a number of additional challenges due to their sexual orientation or gender identity. According to the Bureau of Justice Statistics, "inmates who reported their sexual orientation as gay, lesbian, bisexual, or other were among those with the highest rates of sexual victimization in 2011-12" (Beck 2014, 7). In particular, LGBTQ inmates who have mental disorders reported the "highest rates of inmate-on-inmate sexual victimization (21.0% of prison inmates and 14.7% of jail inmates)." LGBTQ adolescents may view research participation as an opportunity to spend time away from an environment that is particularly hostile and thus may be unduly influenced. Parental permission to enroll in research may also be challenging, if the sexual orientation of the subject is relevant to the research and is not accepted by the parent or guardian (Miller and Nelson 2006). Since substance abuse disorder is one of the variables being examined in this study and LGBTQ youth are at a higher risk for developing these disorders (McCabe et al. 2013), sexual orientation and gender identity may be part of the demographic data collected. LGBTQ youth may also be reluctant to participate in research that might disclose this information to others, especially their parents, which would result in loss of access to potentially beneficial research (Macapagal et al. 2017). In all of these cases, children with social vulnerabilities need to have those vulnerabilities addressed prior to their participation in research. [End Page 470]
Strategies for Rethinking the Concept of Vulnerability
In the above paragraphs, we follow a contextual model for understanding vulnerability proposed by Kipnis and extend it to minors with mental disorders who are also prisoners. Understood in this way, "vulnerability" becomes a lens through which research-related risks and benefits can be framed, relative to the knowledge that may be produced. This approach avoids some of the incomplete comparisons that occur when tacking between the "ideal" research subject (mentally and physically healthy, not incarcerated, decisionally capacitated, economically and educationally advantaged, etc.) and the subject who experiences multiple vulnerabilities. The question remains as to what extent the convergence and amalgamation of multiple vulnerabilities amplifies the perceived deficits in capacity or power that these subjects possess. The federal regulations do not provide guidance on how investigators should adapt their protocols to account for these effects. It is clear that, while we all may be vulnerable in one way or another, certain subject populations are at a higher risk for manipulation, coercion, and exploitation due to the presence of multiple vulnerabilities. The compounding effects of multiple vulnerabilities are likely most intense as the age of the subject decreases and their isolation from parents or guardian and tendency towards impulsive behavior increases. What is needed is a context-sensitive concept of vulnerability that addresses consent-based as well as power-based vulnerabilities.
The convergence and amalgamation of multiple vulnerabilities has implications for the equitable treatment of research subjects. Under the principle of justice, IRBs and investigators are obligated to protect vulnerable populations from exploitation. However, the concept of justice in research has recently shifted to require investigators to provide vulnerable populations with greater access to research-related resources and benefits (Mastroianni and Kahn 2001; King 2005). From a practical standpoint, adopting a contextual approach that draws more heavily on a justice-based framework than a nonmaleficence-based framework would allow these potential subjects greater access to research but would risk compounding their vulnerabilities without the appropriate safeguards.
Following Martha Nussbaum's concept of a capability as an entitlement that supports a "life worthy of human dignity" (Nussbaum 2006, 70), we note that those who have particular social, psychological, or physical vulnerabilities are already restricted in the range of opportunities they have for enhancing their capabilities or for protecting those that are already diminished (i.e., freedom from non-discrimination).2 Participation [End Page 471] in research may challenge some capabilities, i.e. by posing risks to bodily integrity and health, while enhancing others (such as those related to quality of life). In the hypothetical case above, a subject who is LGBTQ in addition to being an adolescent juvenile offender with mental illness may have his capability for living without discrimination diminished by his treatment while incarcerated, but enhanced by the researchrelated benefit of speaking with a trained mental health professional. On a systematic level, power-based vulnerabilities may be mitigated by involving persons with multiple vulnerabilities in community-based conversations with investigators about which research areas should receive federal funding. In addressing social vulnerabilities, justice requires that IRBs attend to the ways in which subject selection and research design enhance or detract from common capabilities of affiliation and control over one's environment. These capabilities of affiliation and control over one's environment can be promoted by designing research enrollment criteria that avoid discrimination and by establishing fair and transparent procedures for banking biological samples, compensating subjects, and disclosing research results, respectively. This capabilities approach to justice is congruent with the contextual approach we discuss above in that both seek to address common structural problems with organizational processes that unfairly target marginalized populations.
Epigenetic research on the hypothesis that there is a connection between methylation patterns, substance abuse disorders, aggression, and criminality has the potential to revolutionize the judicial system and mental health treatment field. However, in order to benefit the most vulnerable members of our society, they will have to be given the opportunity to participate in research that might place them at greater risk for manipulation, coercion, and exploitation with a limited possibility of direct benefit. These risks are compounded by the presence of multiple vulnerabilities, particularly when two different aspects of vulnerability (e.g., being an incarcerated minor) act synergistically in ways that the categorical approach fails to address. Yet, there is a way forward that takes these effects into account. In order to address the compounding effects of multiple vulnerabilities, a comprehensive assent and permission process would be needed. This process would include a detailed interview prior to enrollment, a debriefing session after the initial information-sharing process, and modes of information delivery that are age and educational level appropriate. Although it may be time-consuming, in cases where the study risks are greater than those to be expected in the everyday lives [End Page 472] of incarcerated youths, it would be ethically supportable to introduce an independent party to advocate on behalf of these children. A sliding-scale approach to oversight—based on localized knowledge about the vulnerabilities of the intended subject pool—would allow those who are the most vulnerable to be protected while allowing for those who have more independence, milder cognitive or emotional limitations, and are less isolated from parents or guardians to exercise more of their growing capacity for autonomous decision-making.
The concept of vulnerability for research participants has evolved greatly over the past fifty years and continues to change. We have argued that the federal regulations have historically treated vulnerability as a separate categorical designation meant to highlight the need for protection of particular populations (children, prisoners, mentally ill), guided by the incorrect assumption that these categories seldom or never overlap and that a categorical approach is the most efficient way to protect vulnerable populations (Coleman 2009). In particular, we have noted that the contradictory guidance provided by Subparts C and D regarding what we have called the criterion of relevance and the criteria of risk as well as the justification for including these populations needs to be clarified and should lean towards a contextual account of vulnerability. We have described how multiple vulnerabilities affect the recruitment, enrollment, and participation of research subjects, and advocated for a contextual approach to vulnerability that incorporates these considerations as part of IRB deliberations. In designing studies involving such populations, the context created by their multiple vulnerabilities should guide the protections put in place. [End Page 473]
Andrew M. Childress, PhD, is an Assistant Professor in the Center for Medical Ethics and Health Policy at Baylor College of Medicine. He also conducts clinical ethics consultations at Houston Methodist Hospital and Baylor St. Luke's Medical Center. His research interests include the integration of the humanities within medical education to prevent physician burnout, ethics of research involving human subjects, and illness narratives within the context of clinical ethics consultations.
Christopher R. Thomas, MD, is the Robert L. Stubblefield Professor of Child Psychiatry in the Department of Psychiatry and Behavioral Sciences at the University of Texas Medical Branch at Galveston. He currently serves as the Assistant Dean for Graduate Medical Education and the Director of the Child and Adolescent Psychiatry Residency Training Program. His research focuses on forensic child psychiatry, gangs and antisocial behavior in youth, pediatric burn survivors, and ethics in mental health care.
1. Hereafter, cited as 45 CFR 46.
2. Nussbaum's list of capabilities includes: life; bodily health; bodily integrity; sense, imagination, and thought; emotions; practical reason; affiliation; other species; play; and control over one's environment.