Three presenters examined the issue of informed consent through the lens of specialized populations. Each of the special types of subjects examined—groups of family members, people who have experienced disasters or other traumas, and children—present different challenges with respect to informed consent. Margaret Foster Riley, of the University of Virginia, noted in her introductory remarks that the proposed changes to the Common Rule bring up the possibility of a standardized informed consent form, but the variety of issues raised by these distinct populations makes it clear that it will be exceptionally difficult to develop a single form—or a single informed consent process—that can be applied uniformly to these disparate types of research participants.
Sally Powers, of the University of Massachusetts, Amherst, spoke about her 30 years of research on the effects of social stress on depression and the consent issues associated with such studies. Her research is longitudinal, and she follows her child, adolescent, and adult subjects over several years. Her particular research interest is the stress that arises from conflict within families and how that stress influences the emergence and course of depression. She works with multiple family members within a family. Her subjects have included newlywed couples during the first years of their marriages, new parents with infants and young children,
families with teenagers, and young adults just out of high school who are part of dating couples. To observe how the various subjects respond to stress, she asks subjects to discuss heated, unresolved conflicts in their relationships. She measures stress through observations of behavior, self-reports, and analysis of stress hormones in saliva samples taken before the conflict task, during the conflict task, and afterward.
Powers focused on consent issues involving rich data, which she defined as data that can be “recoded, reassayed, and retested to yield new information that was not proposed in the original study.” She explained that she was not referring to the reanalysis of existing coded information from large survey datasets, which can produce new insights and interpretations but which usually does not produce new data. Instead, she was specifically referring to datasets that allow researchers to derive new information through new coding, assaying, or testing of the original behavioral or biospecimen data samples.
In the case of biospecimens, for example, Powers explained, there are many different types of information that can be gained through various analyses, and the original researchers are likely to have obtained only a small portion of everything that is possible. In her research, Powers said, she analyzes the biospecimens from her participants mainly for stress hormones because she is interested in stress levels. The most common biospecimen she collects is saliva, which provides an indication of stress levels around the time of the sample, but she also looks at hair samples, which gives her information on stress levels over the preceding months.
An important fact about these biospecimen samples is that over time researchers are able to get more and more information from them. When she began collecting saliva more than a dozen years ago, Powers said, the purpose was to analyze the levels of the stress hormone cortisol. Today there are dozens of endocrine molecules, neurotransmitters, immune factors, and other molecules whose levels can be detected in saliva samples, each of which provides a different set of information about the subjects. “We have hundreds of those samples frozen from many, many families,” she said, so she can go back and extract a tremendous amount of information about the subjects from their samples. Furthermore, the amount and types of information that can be extracted from the biospecimens are only going to increase further in coming years.
Behavioral data can also be rich data, Powers said. In her experiments she videotapes subjects in various situations, including conflict situations, and the videotapes can be analyzed to obtain many different types of information. When she carried out her earliest studies, she was focused
on submission and dominance behaviors—how much and how often subjects were acquiescing to what other subjects wanted, for instance—and on how those behaviors might be associated with the onset of depression. Later on she became interested in what she calls “secure base behaviors” in married couples and dating couples, and she was able to go back to the videotapes from earlier work and examine them for this new type of behavior. Most recently, she said, she has begun working with computer scientists to amplify color changes in subjects’ faces as a way to observe cardiovascular stress reactions to family conflict. The analysis allows her to see the color of a subject’s face go from normal to flushed with the subject’s pulse, making it possible to get a second-by-second reading of stress, which can be combined with measurements of stress hormones taken at the same time.
The opportunities posed by rich data do, however, come with a number of ethical and consent issues, she explained.
Consent Issues for Rich Biological Data
A major issue with rich biological data such as saliva samples, Powers said, is that the presence of so much information waiting to be unlocked by various analytical techniques means that it will become increasingly possible to identify the subjects who provided the samples. This possibility may blur the line between identified and de-identified data, and make it possible in the future to determine the identity of the people who provided the samples. Thus, it will be increasingly problematic for researchers to assure participants who provide biospecimens that their identities can be protected by such measures as keeping their names and other identifying information confidential.
Given that, she asked how the proposed changes to the Common Rule might affect her research. The current rules allow de-identified biospecimens to be used in other studies without re-consent, while for identified specimens it is necessary to obtain re-consent before carrying out new research. The proposed revision to the Common Rule would require getting prior consent whenever one obtains biospecimens and also asking for blanket consent for open-ended use, meaning that the research participants would agree to have their samples used for any type of future analysis. This consent would be requested whether or not the biospecimens were to be de-identified.
In her own research, Powers said, she has always considered the biospecimens and other data she collects to be identifiable because she is conducting longitudinal research, which by its nature requires that the researcher know which data comes from which subject in order to observe how the data change over time. Furthermore, she said, because it is
unlikely that even de-identified biospecimens will be able to remain anonymous in the face of increasingly powerful analytical tools, she always assumes that biospecimens will become identifiable in the future. Thus, she asks her subjects for prior consent for future use of their specimens, but she does not ask for blanket consent for open-ended use. Instead she asks for consent within broad categories of testing that are designed to anticipate the various future uses to which the saliva samples might be put. For example, she said, “a particular … study may be funded to assay cortisol, but I know that in the future I would like to also explore relations between stress hormones and immune functioning.” In that case, she would ask the participants at the beginning of the study if they would be willing for their biospecimens to be used in such a future study.
The difference between such a broad prior consent and a completely open-ended prior consent might be important, she said, because some participants are very concerned about the open-ended use of their data. In particular, many of them worry that their DNA data will be identifiable. On the other hand, she noted, the vast majority of her participants do not seem to worry about it at all. It is unclear that subjects’ lack of concern is based on a clear understanding of the risk of identification in the future.
Consent Issues with Rich Behavioral Data
The issues concerning consent for rich behavioral data are somewhat different, Powers said. While she feels very cautious about requiring open-ended, all-inclusive prior consent for biospecimens, she feels less cautious about the sort of videotape data that she collects. The reason is that every participant who gives informed consent on videotaped data knows that his or her face is identifiable on the videotapes. Participants have no expectation that the data can be de-identified, as biospecimens can, so any informed consent implicitly includes the acknowledgment that the subject will be identifiable in any future use of the data. Thus, she said, “I suggest that this [blanket, open-ended] standard is acceptable for video data, but not for biospecimen data.”
Because the proposed changes in the Common Rule seem to be focused mainly on biospecimen data, she said, it would be useful to think carefully about those changes as they would apply to behavioral research and, particularly, videotape data. “IRBs are so variable in terms of how they assess the risk of videotaped identifiable data,” she said, that it would be to everyone’s advantage to pay attention to this area and to develop some guidelines for IRBs about “what is risk with videotape data and what is not.”
Prior Consent for Longitudinal Follow Up
Powers also noted the issue of getting consent when doing the follow-up studies in a longitudinal series. Institutional review boards (IRBs) differ, she pointed out, on whether participants who have not already given consent to be re-contacted for future studies can be re-contacted using their contact information from the original study. One particularly tricky situation arises when a researcher wishes to do a follow-up study with subjects who were adolescents in an earlier study and who are now adults. Because they were adolescents at the time of the previous study, they could not have given informed consent for future contact. To address the issue, Powers said, she always asks participants in her studies for permission to re-contact them later. “We ask them to give us contact information; we even ask them if they would give us names of friends who may know where they’ve moved.”
Consent for Interdependent Data
Because her data involve multiple subjects who interact—a videotape with a mother and daughter arguing, for example—Powers must deal with consent issues involving interdependent data. She illustrated with an example of a woman who had taken part in an earlier wave of the study as an adolescent and is now a mother with her own teenager. Following up on this subject could provide insights into, for instance, how her interactions with her own mother have affected the way she interacts with her own child. This sort of research requires analyzing not only new videotapes of the subject with her daughter but reanalyzing videotapes from the earlier study showing the subject at a younger age with her mother. It is clearly necessary to get new informed consent from the woman who was an adolescent in the previous study, Powers noted, but what about re-consent from this woman’s mother? “There is no way to [separate] her data out from her mother’s data.”
Powers argued that it is not necessary to re-contact the mother because the mother’s original permission to use that tape would cover the later use. The one exception, she said, would be if she were to ask the study participants to view their own videotapes in order to get their reactions and their judgments about what was going on in various interactions. That would require re-consent from the mother because her data would be viewed by someone other than members of the research staff (in this case, by her daughter).
Roxane Cohen Silver, of the University of California, Irvine, discussed the issue of obtaining informed consent from people who have experienced traumatic events, such as the death of a child, a terror attack, or a natural disaster. Silver began with a description of the sorts of research studies she carries out. She has studied the sudden unexpected death of an infant, California neighborhoods devastated by firestorms, and the aftermath of the Columbine High School shootings. The events she studies are random, unpredictable, and uncontrollable, she said, which makes the events particularly interesting to psychologists but also makes them particularly difficult to study and raises ethical issues.
One difficulty is that, to be most useful, the research on such an event needs to start almost immediately after the event’s onset. Immediate data are necessary to identify early predictors of long-term difficulties, Silver said, and early identification of at-risk individuals allows mental health professionals to target interventions to those who are most vulnerable. Early data allow educational and intervention efforts to be better planned, more sensitive, and more cost-effective. Furthermore, research shows that people cannot accurately reconstruct emotional experiences long after a traumatic event, so it is important to study people as soon after the event as possible. However, the need to move quickly after a disaster or traumatic event raises a number of ethical issues, Silver said. “First, is it ethical to conduct research immediately after a tragedy?” she asked, “can or should we intrude on individuals during a potentially vulnerable period?” A second issue is whether individuals can provide true informed consent while they are experiencing a life crisis. A related procedural issue is whether participants must provide written consent, especially when the research is seeking representative, population-based samples and needs to move quickly.
Silver described a national study she conducted following the September 11, 2001, terrorist attacks. The study was done in collaboration with Knowledge Networks, a research survey firm that retained the participants’ identities to make longitudinal data collection possible while ensuring that all the data were de-identified. On behalf of Silver’s collaborators, Knowledge Networks also went back to a random subsample of original participants and collected biospecimens—saliva samples—two years after the project ended, which were ultimately linked to the earlier data respondents had provided. The people who gave the samples were re-consented, Silver noted.
To maintain anonymity of participants, Silver got a waiver of written informed consent. The online data collection process provided contact information participants could use to ask questions or express concerns, as well as reminders that the participants could skip questions or quit the
survey at any time. Participants conveyed their consent by clicking on an embedded web link. Silver noted that the study, which covered seven waves of data collection, and more than 10,000 survey completions among a nationally representative panel, had very low dropout rates. There were no complaints from the participants about procedures, the survey questions, or anything else.
How IRBs Can Facilitate Disaster Research
Silver has had generally positive interactions with her IRB, she said. Of the four large disaster studies she discussed, all were either exempted from review or given expedited review. Reflecting on her experiences with IRBs, she offered several ways that the review boards can facilitate this sort of post-disaster survey research.
Most important, she said, is that the IRB needs to be willing to approve the research very quickly. “After the Columbine High School shootings, my IRB convened over e-mail the weekend after the shooting, and I was able to pick up my signed IRB approval on Monday morning.” Because her research team had driven to Littleton, Colorado, over the weekend following the shootings, the team was able to get started within a week of the incident. The IRB must be willing to issue a waiver of written consent in certain cases, particularly for fast-moving research. She has also found it very valuable to have the IRB preapprove a generic proposal that can be activated immediately after a disaster. She has an approved generic disaster trauma protocol at her university that describes the background and rationale of the research, proposed methods, proposed risks and benefits, who the research team is, and sample questionnaires and interview questions. This generic protocol was reviewed and received expedited approval by her IRB.
Furthermore, she said, she has a “contract” with her IRB that it will review and decide on proposals within 24 to 48 hours of submission once she provides them with the specific purpose or event to be studied, the specific research methodology, the specific sample and sample size, and any deviations from the generic proposal in the research team, methods, compensation, or other elements. She has not yet used this generic proposal, but her relationship with the IRB is such that after a national event has occurred, the IRB has called her to see if she is going to study it.
At this point, Silver said, she is proposing a project that will begin even earlier. Working with meteorologists and wind engineers, she will focus on communities being targeted by a hurricane and will identify participants 36 to 48 hours before landfall. The goal is to link pre- and during-storm emotions, risk assessments, and behaviors to post-storm adjustment. Participants will be recruited orally as the storm approaches, and they will
be given Internet-enabled tablets that will be used to collect data before, during, and after the storm. One goal will be to accumulate data that illuminates who evacuates and why, because at this point no such data exist (only anecdotal information has been collected). Speed is essential for this project, Silver said. “We need to recruit 750 individuals within 18 hours, so we will be requesting a waiver of written consent.” The project will also require flexibility of methods, she said, in order to protect both the participants and the research personnel.
The Special Case of Intervention Research
A related type of research that Silver considers even more challenging is intervention research—studies of psychological treatments or procedures that are carried out after a disaster. Examples of such treatments include the psychological debriefing sessions offered to all New York City police officers and fire fighters after the September 11 terrorist attacks, and the counseling offered in classrooms when students and teachers return after there has been a school shooting. These interventions are well intentioned, Silver explained, but are not based on evidence that they are actually helpful. Indeed, there is growing evidence that they may be unhelpful or even harmful because they may disrupt the natural course of social support and recovery. Thus, she said, it is important to evaluate the effectiveness of such post-trauma interventions.
Silver described one such intervention study she carried out in Yogyakarta, Indonesia, of a rural community that is subject to repeated natural disasters, such as earthquakes, volcanoes, tsunamis, and floods. She recruited 500 families with children in six different elementary schools. The parents consented for themselves and their children, and the children provided assent. Silver’s Indonesian collaborators had told her that because of cultural norms the participants would be unwilling to sign their names to documents, so Silver received a waiver of written informed consent from the IRB. The study was explained to the participants orally, research assistants were available to answer questions, and participants were told they could skip questions or terminate their participation without penalty. There was no coercion or pressure to complete questions, and remaining in the room as the surveys were about to be distributed implied consent to participate.
Both parents and children completed surveys before the intervention, and a subsample provided saliva samples for future cortisol and genetic analysis. The parents were randomly assigned, by school, to either a skills-based psychosocial intervention or a waiting list control group. Parents and children completed surveys after the intervention, and the intervention was subsequently provided to the waiting list control group.
Finally, Silver noted that there are several ways IRBs can facilitate disaster and trauma research. They should review and decide on proposals quickly. They should also allow subjects to be randomly assigned to different conditions—an intervention group versus a waiting list or a no-treatment control group, for instance. And they should allow flexible informed consent procedures. The Indonesian project went through a full-committee IRB review, Silver said, and it was successful “because of IRB acceptance of my proposed methodology.”
To obtain representative samples of participants in trauma and disaster research, she said, researchers find that “flexibility in consenting procedures is crucial, including the opportunity to maintain anonymity [for participants].” Such flexibility might mean allowing a waiver of written consent, or allowing a delay in consent for the future use of biospecimens, particularly in the early aftermath of a disaster or a tragedy. “Requirements in the revisions to the Common Rule that prohibit flexibility will impede this kind of research,” Silver observed. Defining or evaluating trauma and disaster research as necessarily “emotionally charged,” and therefore disqualified from being considered exempt or excused, will impede scientific progress, she added. Finally, eliminating the waiting period before exempt or excused data collection can commence would facilitate trauma or disaster research, whereas requiring even a brief waiting period—as little as one week—could impede the research.
Celia Fisher, of the Center for Ethics Education at Fordham University, spoke about a variety of issues that arise in studies involving children and other special populations (see also Fisher et al., 2013). One of the first recommendations made in the advance notice of proposed rulemaking (ANPRM) concerns the length, content, and documentation of consent for informed consent forms. The proposal to shorten the length of informed consents is timely, Fisher said, commenting that many researchers seem to agree with this recommendation. However, she said, the proposal to standardize forms poses problems. It could lead to confusion and misinformation, she observed. “We need the flexibility in format and language to ensure that any informed consent is crafted so it is age appropriate, it is language appropriate, and it is appropriate to the educational level and familiarity with research of the population,” she explained. Informed consent forms should also be flexible enough to take cultural understanding into account, she added. For example, when doing research with American Indian tribal nations, it is very important to obtain the consent of the tribal leader or tribal groups before seeking the consent of individuals.
The issue of oral consent is equally important, she said. Informed consent is a process, not a standardized document, and its goal is to ensure that decisions to participate are informed, rational, and voluntary. Thus, oral consent may be more appropriate than written consent for certain populations. With children, for example, oral consent can be less coercive because of children’s more limited reading skills, their deference to authority, and their lack of experience in signing forms. Furthermore, written consent can jeopardize participant safety in war zones; in research in which partner violence is an issue; and in studies looking at stigmatized or illegal behavior, such as drug use, HIV status, or homosexual activity. Thus, Fisher said, it would be useful to have population-specific guidelines to help IRBs and investigators craft appropriate oral consent.
Other issues raised by the ANPRM concern flexibility and accuracy in the informed consent process. It would certainly be useful to have the flexibility to waive irrelevant informed consent components, she said. With respect to Brian Mustanski’s comments about IRBs requiring participants to be warned about the possible harmful effects of stressful questions, she commented that such statements are often not accurate because there is no evidence that such harm may occur. “We are actually communicating untruths to our participants.” Such unsubstantiated statements should be removed from informed consent forms, she said.
Fisher suggested that the following default statement could be included in forms to describe minimal risk research: “This research presents minimal risks no greater than those of daily life or routine medical, dental, psychological, or educational examinations or tests” (Fisher et al., 2013). Such a statement would not only truthfully describe the risks but would also serve an educational purpose because it is based on the IRB’s evaluation of the risks.
Research Risk Versus Institutional Liability
Fisher also addressed the issue of distinguishing research risk from institutional liability. One reason informed consent forms are so long, she said, is because the institutional liability information is attached to them. However, in Fisher’s view, such institutional liability statements do not belong on an informed consent form, because they refer to risks outside of the research procedures themselves, such as falling in the research facility. “In fact,” she said, “Section 46.116 says that no informed consent ‘may include any exculpatory language through which the subject or representative is made to waive or appear to waive any of the subject’s legal rights or appears to release the investigator, the sponsor, the institution, or its agents from liability for negligence.’”
Furthermore, she said, including this liability language within an
informed consent form is unfair to children or other vulnerable populations, who may not be familiar with what their legal rights are or have access to legal advice. “In some of the research I have done with populations where I ask them about ethical issues, some of them believe that they are signing their rights away when they sign an informed consent or a legal liability form, and I think we have to be very careful about that,” Fisher observed. In her view, it is important to separate institutional liability from the informed consent document. If institutions wish to notify prospective participants or their guardians about limits to the institution’s legal liability, they should do so in a separate document, she added.
Waiver of Guardian Permission
An issue specific to minors is the waiver of guardian participation, Fisher noted. Emancipated minors are adolescents who are supporting themselves and who are, under the laws of their state, considered adults; they may themselves be parents. Mature minors are those adolescents who by state law can independently and without parental permission gain access to health or mental health services. Under federal regulations, both emancipated and mature minors are considered to be adults. However, Fisher said, most states do not include language specific to research participants in their emancipated and mature minor laws, and “this has been incredibly confusing to IRBs because they don’t know whether or not they should require parental permission.”
As a result, she said, IRBs often needlessly require guardian permission for minors’ involvement in research related to treatment and procedures for which they’ve already obtained a legal right to adult status. For example, an adolescent who can go to a clinic and get sexual health treatment or prescriptions independently under the mature minor rule may still be required by an IRB to get parental permission to participate in a survey that asks about his or her experiences. This requirement could deprive adolescents of their rights and of the potential benefits of research participation, Fisher said.
Waiver of guardian permission is also relevant to Section 46.116 of the Common Rule, which discusses procedures to ensure that “waiver or alteration will not adversely affect the rights and welfare of the subjects.” When studies involve children and adolescents—and some other vulnerable populations—IRBs often overreact when considering what the participants actually understand, Fisher said. However, there is a considerable body of developmental research that can be used to determine whether subjects can give an independent consent, so IRBs should use “evidence-based literature to evaluate whether or not an age group has an understanding of their rights and research procedures.”
A provision in Section 46.116c stating that components of informed consent may be waived if “the research could not practicably be carried out without the waiver or alteration” also poses an issue, Fisher noted. The Secretary’s Advisory Committee on Human Research Protections has recommended that guardian permission should never be waived for the convenience of the investigator, or solely for reasons of cost or speed or other expedient measures, if doing so will weaken the protection of subjects’ rights and welfare. This body has also stated that parents’ reluctance to permit their children to participate in research is not a legitimate reason to waive this protection. Researchers who work with marginalized communities often encounter parents who do not want to give permission, perhaps because they don’t trust the research or perhaps because of other concerns. “Their reluctance is legitimate,” Fisher said, and parental permission should not be waived simply because the researcher thinks that the research is important.
There is also plenty of research on how to enhance consent for children and for adults with impaired cognitive capacity, Fisher said. That knowledge should be used, she suggested, to develop enhancement procedures for use in the informed consent process to help members of these groups autonomously consent to research, and to make sure that the language used is age- and participant-appropriate.
Informed Consent for Future Use of Data
Fisher also discussed informed consent for the future use of bio-specimens and archived socially sensitive data, focusing on permission granted by guardians for future research on data collected from children. It is important to ask under what circumstances permission granted by a guardian will be sufficient for future data use even after the participant reaches adulthood, she noted. In her view, it is appropriate not to require re-consent if several conditions are met: (1) appropriate security protections are in place and are updated to reflect evolving information technologies and federal standards; (2) the level of harm associated with the informational risk has not increased with changes in societal attitudes, health coverage, or other policies; and (3) the original informed consent informed the guardians—and minors old enough to understand—that their consent represented a default permission for continuing to use the data after the child reached the age of maturity.
Another important question, Fisher noted, is when it is acceptable to expand the original informed consent commitment to de-identified data for socially sensitive research. Emerging technologies may make obsolete the original de-identification data security protections to which guardians, minors, or vulnerable adult populations originally consented. One
way to address this, Fisher said, is to have the initial informed consent indicate that all investigators who will have access to data in the future will be bound by the best practices in data and confidentiality protections at the time of the data collection as well as any new protections that emerge. Federal regulations should ensure that future investigators honor this commitment. This approach is consistent with proposals in the ANPRM to establish regulatory procedures for the continuous updating of data security procedures.
A final question, Fisher said, pertains to when informed consent is necessary for linking identifiable archival data to the collection of new data. Whenever an original investigator or a new investigator wants to link archival identifiable data with collection of new data, she explained, it is necessary to get re-consent from the original participants, Fisher said. The original informed consent should indicate that any investigators interested in linking new data collection with the archival dataset will be given access to the participants’ contact information in order to request additional permission for its use. In the case when archival data was collected during childhood, once a participant reaches adulthood consent for linking new data to the archival set should be obtained from the original participant, not the guardian.
Summing up, Fisher said that informed consent procedures should be age- and population-sensitive; should be based on the substantial evidence base concerning consent capacities; and should include, where appropriate, consent-enhancing procedures. Any decisions regarding the waiver of informed consent components should provide adequate protections against misunderstanding by and exploitation of participants and such waivers should also ensure that children and vulnerable populations have equal access to the potential benefits of research.
Fisher, C.B., D.J. Brunnquell, D.L. Hughes, V. Maholmes, P. Plattner, S.T. Russell, S. Liben, and E.J. Susman. (2013). Preserving and enhancing the responsible conduct of research involving children and youth: A response to proposed changes in federal regulations. SRCD Social Policy Report 27(1):1, 3–15.
This page intentionally left blank.