socw 6301

Psychological and Social Risks of Behavioral Research BY SUSAN M . LABOTT AND TIMOTHY P. JOHNSON

isks to human subjects asso- k x ciated with participation in

JL \S.research are an important consideration in the design of stud- ies and in the informed consent process. Yet, there are significant gaps in our knowledge about the psychological and social risks of behavioral research. Our review of federal regulations governing research with humans and empirical work pertaining to behavioral research risks shows that we have little data on the frequency of specif- ic psychological and social risks and on the implications of these risks for research subjects who experience them. We make suggestions for fur- ther research in this area to address the current ambiguities.

Risk Assessment

n p h e Belmont Report,^ issued by JL the National Commission for

the Protection of Human Subjects of Biomedical and Behavioral Research, described the need for Institutional Review Boards (IRBs) to evaluate the risk-benefit ratio of research studies. IRBs were urged to consider the probability and severity of potential risks, the justifiability of risks, the likelihood of anticipated benefits, and alternative ways to obtain the benefits of the proposed research. The federal regulations for research with humans expand upon this idea by providing guidelines for the evaluation of risks and benefits. The regulations require that all risks

Susan M. Labott and Timothy P. Johnson, "Psychological and Social Risks of Behavioral Research," IRB: Ethics & Human Research z6 No. 3 (1004): 11-15.

to subjects should be minimized to the extent possible. The importance of the potential knowledge gained through the research can be consid- ered in this determination, but the potential long-range risks and bene- fits of applying the knowledge should not. Further, the regulations require that "reasonably foresee- able" risks be disclosed to subjects during the process of consent, along with any anticipated benefits.^

The controversial regulatory defi- nition of minimal risk involves a sit- uation in which "the probability and magnitude of harm or discom- fort anticipated in the research are not greater in and of themselves than those ordinarily encountered in daily life or during the performance of routine physical or psychological examinations or tests."' One prob- lem in assessing minimal risk involves issues of the definition of "daily life." Whose daily life? That of an investigator or of a potential subject who is currently homeless? Further, is a study with a potentially significant harm that has a low like- lihood of occurrence the same as a study with relatively lesser harm but a greater likelihood that many sub- jects will experience it?'' Lastly, is risk level determined a priori, or after the investigator's safeguards to protect subjects are in place? Regardless, research that does not meet the criteria of minimal risk is defined as greater than minimal risk and warrants additional safeguards.

The National Bioethics Advisory Commission provides suggestions for risk assessment that include data on the nature, likelihood, and

acceptability of risks, but notes that these assessments are difficult to make because of unknown informa- tion about many of the risks associ- ated with human subjects research.5 In spite of the inherent problems, researchers and IRBs regularly make risk estimates.

Regardless of the category of risk, researchers and ethicists agree that risks need to be minimized to the extent possible. That is, a study should be designed in the safest manner and procedures should be in place to manage any problems that do occur. Further, the process of informed consent requires that the risks (and benefits) be clearly described so that individuals can consider and weigh these along with other aspects of the research as they make their decisions about partici- pation.

Behavioral versus Biomedical Research Risk

A t this point the perspectives on . i \ b e h a v i o r a l versus biomedical research risk diverge, partly because behavioral research is generally min- imal risk while biomedical research is more likely to be greater than minimal risk; the clarity with which risks can be specified also differs. In the area of biomedical research, the nature and acceptability of potential risks can be determined to a great extent, at least qualitatively, and can be evaluated in the context of poten- tial benefits. For example, consider research on two new medical proce- dures designed to slow the progres- sion of cancer. The most significant risk of one procedure is infection.

IRB: ETHICS &. HUMAN RESEARCH MAY-JUNE 2004

for the second procedure it is death. Clearly these risks are qualitatively different in that an infection can be treated with antibiotics, is likely to be reversible, and will probably have no long-term consequences. For many potential subjects, this risk may be acceptable, especially if there is a corresponding potential for med- ical benefit. The risk of death is like- ly to be unacceptable to most peo- ple, unless there is a high likelihood that the procedure could result in significant benefit, i.e., a cure for the cancer. In this context, the risks are clearly different; their impact is eval- uated accordingly, and is weighed against the potential benefit (which may be dramatic). For many bio- medical protocols, risk frequency can also be quantified. For example, negative reaction rates and other side effects for many medications are well established (i.e., one person in a thousand or a million may have a life-threatening reaction to a given drug), and this information is com- monly used to assist potential research subjects in evaluating the risks of participation.

In the area of behavioral research, how to assess risks is less clear. First, psychological and social risks are less tangible than physical risks, and we do not know their nature in an exact way. For example, what exact- ly is emotional discomfort and how reversible is it? Second, there is often no anticipated benefit to consider in relation to the risks of behavioral research. Much behavioral research offers no individual benefit for par- ticipants at all, providing no com- pelling reason to become involved in research that offers any (even mini- mal) risk. Third, although it is not difficult to identify potential social and psychological risks, estimating the frequency or degree of Ukelihood associated with each currently amounts to little more than guess- work. Finally, some would argue that because psychological and social risks do not result in physical harm.

they are not actual "risks" at all.

Psychological and Social Risks

What exactly are thi; psychologi-cal and social risks of research? The National Research Council (NRC), in its thorough review of issues related to human subject protection, has designated six categories of potential subject harms: physical, psychological, social, eco- nomic, legal, and dignitary.* Prentice has also indicated that social and psychological risks may occur, not only to the research participant, but also to family members who may experience negative consequences as a function of their family member's participation.'' The psychological risks that have been discussed in the literature include depression, altered self-concept, increased anxiety, decreased confidence in others, guilt, shame, fear, embarrassment, bore- dom, frustration, receiving informa- tion about oneself that is unpleasant, and inconvenience.^ Social risks gen- erally involve stigma, decreased opportunities, or negative changes in relationships.'

• Confidentiality. The area of risk that has been studied most deals with the issue of confidentiality, like- ly the most significant social risk of research. Singer and colleagues have studied this issue extensively, provid- ing information on the effects of confidentiality assurances on survey research. In one set of studies, assur- ances of confidentiality decreased willingness to participate in a survey because they elicited subjects' expec- tations that the information request- ed would be sensitive."' In a meta- analysis of studies on this issue, con- fidentiality assurances affected response, but only when data requested were sensitive; when data were not sensitive, negligible effects of confidentiality assurances were obtained.^^ When several researchers reviewed responses to the 1990 mail census, they discovered that privacy and confidentiality concerns had a

significant effect on the census mail return. ̂ ^ Further, the effect of priva- cy and confidentiahty concerns dif- fered for black and white respon- dents. It is also important to note that these studies raise the issue of psychological discomfort (e.g., anxi- ety, apprehension, embarrassment), which msiy result from the confiden- tiality assurance itself.^'

• Emotional Distress. The risk of emotional distress has been con- sidered in the area of trauma research. Here, the concern is that recall and discussion of a traumatic event may result in re-traumatization of the individual. Newman and col- leagues examined the risks involved and the potential vulnerability of this populatic'n.^'t They found little empirical evidence to specifically address the risks, and conclude that there is a need to generate data on the prevalence of adverse occur- rences and to determine if individu- als with c:ertain trauma-related symptoms are more vulnerable than others. They also suggest that studies on violence and trauma may provide some benefit to participants. Such studies might be a way for individu- als to beg;in to seek help from profes- sionals, or as a way to begin to use a negative experience to potentially help others.

In an empirical study addressing this issue, women were asked to respond 1:0 a survey that included items about childhood abuse and maltreatment; a subset of partici- pants was also interviewed.^' Only 10% of the survey respondents reported unexpected upset, and these were individuals with higher levels of post-traumatic stress disorder symp- toms. Many participants (23% ques- tionnaire, 86% interview) reported benefit and only 5% reported regret for having participated. Finally, the appraisals of benefit and regret about the; interview were generally stable over the 48 hours post-inter- view. The authors suggest that the follow-u]5 contact contributed to the satisfaction that participants report-

MAY-JUNE 2 0 0 4 IRB: ETHICS & HUMAN RESEARCH

ed, and proper informed consent procedures helped to minimize upset and regret. However, they also noted that more extensive informed con- sent procedures, especially for those with a history of maltreatment, might further decrease the unexpect- ed upset reported by some partici- pants.

• Social Risks. In one of the few studies of social risks associated with biomedical research, an attempt was made to delineate the social harms associated with participation in a Phase II trial of an HFV vaccine. ̂ ^ Because participation in a study of this type can result in positive HIV tests for uninfected individuals, the potential for negative social conse- quences occurs. These authors ques- tioned 247 participants in a vaccine trial about adverse social conse- quences (job- health- or life insur- ance-related problems, travel prob- lems, disclosure of participation, and the presumption by others that the volunteer has AIDS). Forty-five respondents (18.2%) reported they had experienced at least one of these adverse social consequences in rela- tion to trial participation. Individuals also rated how upset they were by the event; the overall mean was 4.0 on a 10-point scale (i=not at all to io=extremely upset). Despite this, 9 4 % of the participants said they would join the study again. Therefore, while social harms occurred fairly frequently, these problems were less upsetting to par- ticipants than might have been expected. However, other subject samples that differ culturally may have different perceptions of these risks.

• Other Risks. With respect to the other psychological and social risks, we have little information to guide our evaluations of their impacts. For example, consider research on emotion and mood. Many subjects (including vulnerable populations) have been subjected to inductions specifically designed to alter mood, either in a positive or

Table 1 . Process of Risk Evaluation

Investigators

1. Provide information on potential risks and benefits

2. Provide information on probability, magnitude, and harm associated with each risk

IRBs

1. Evaluate validity of investigator's assessment of the above, considering alternate

procedures with less risk

2. Ensure risks are clear to potential subjects

3. Evaluate risks against benefits

4. Do not consider potential long-term effects of the information gained from the

research

5. Be aware of the difficulties and biases in estimating potential risks

negative way. The studies are gener- ally construed as acceptable research, because the changes in mood are seen as "minimal risk" and transito- ry, and because appropriate safe- guards are in place for individuals who could be harmed, e.g., coun- selors are available. Yet, we do not actually know the consequences of these manipulations on human sub- jects, the subjects' perceptions of acceptability, or under what condi- tions individuals may be at greater risk.

Embarrassment is another poten- tial risk. Everyone has suffered embarrassment when others in the social world are witness to events or obtain personal information. We assume that, in most cases, this expe- rience results in minimal impact on the individual. However, some indi- viduals, those most sensitive to social evaluation, may not be resilient with respect to experiences of this sort. Group data collection in which indi- viduals share personal information may produce significant risk for embarrassment of individuals, but we have no data on the potential long-term consequences for individ- ual participants.

Oakes presented anecdotal reports described by IRB members in which survey questions seemed to cause harm to survey respondents.^'' In one case a man completing a questionnaire about cancer commit-

ted suicide following his participa- tion. In another, several individuals completing surveys on rape became upset and asked for help during the survey. In a third, an adolescent responding to questions about sexual preference misinterpreted the ques- tions as suggesting that he was homosexual; he attacked the researcher and had to be restrained.

It is, of course, important to acknowledge that the evaluation of risk is subjective, and that individu- als may vary dramatically in their views on the unpleasantness of a particular risk. Further, as noted by Sieber in a discussion of risk associ- ated with sexuality surveys, some subject perceptions of risk are associ- ated with imagined outcomes (rather than actual ones).*^ As researchers, then, we may need to address the perceptions of both actual and per- ceived risks. This is a topic that remains largely unexplored.

Guidance for IRBs on Risk Assessment

"Tpable 1 provides a summary of JL the principles described above

for the evaluation of research risk. While this appears to be a fairly straightforward process, the NRC noted that it is difficult to estimate the severity, likelihood, and duration of specific risks, although it is fairly easy to imagine specific harms occur-

IRB: ETHICS & HUMAN RESEARCH MAY-JUNE 2 0 0 4

ring to study participants. The NRC also pointed out that we have no data on how much IRBs over- or underestimate risks. Singer and Levine note that IRB members gener- ally make assessments of risk based on their own experience, with almost no information on subjects' percep- tions of the probability, magnitude, duration, and consequences of these r i s k s / '

Several studies have examined dif- ferences in risk perception across various groups. In a study of risk in industrial psychology research, Ilgen and Bell asked IRB members, univer- sity faculty, human resource man- agers, and job applicants to evaluate the risks associated with study proto- cols.^" Of note was the finding that IRB members perceived significantly greater risk than human resource managers when standard consent procedures were not used. The IRB members, university faculty, and human resource managers all per- ceived greater risk than the job applicants when the survey items were sensitive. Another study by Schreier and Stadler compared the evaluations of student research par- ticipants, psychologist investigators, and IRB members when evaluating a research protocol.^^ The IRB mem- bers and psychologists were similar in that they reported less comfort endorsing the study and saw less benefit of it than did the student par- ticipants. Finally, Kimmel has pro- vided evidence of bias in the esti- mates of risk made by psychologists, reporting that gender, specialty area, and work setting all influenced risk estimates.^^

The NRC, as well as Singer and Levine, noted that IRBs might be applying greater scrutiny to behav- ioral protocols, using provisions (on minimal risk research) that were developed for higher risk biomedical protocols. The effect of this is more cumbersome procedures for behav- ioral researchers, without any addi- tional protection for research partici-

pants. The NRC noted i.hat IRBs are quite variable in the extent to which they use exempt and expedited pro- cedures; greater use of these cate- gories adds flexibility to the review process.

Future Directions

We are not arguing that researchinvolving psychological or social risks is inappropriate or uneth- ical. Rather, there is a need for empirically-based understanding of these risks so that they can be evalu- ated more clearly and communicated to human subjects where relevant. We agree vk̂ ith Oakes that there is a need to understand "the risks posed and literally experienced in social sci- entific investigations. "̂ 3 Methods are needed to evaluate the frequency and impact of social and psychologi- cal risks, as well as differences in risk perception in various populations. As the situation stands now, human subjects are frequently presented with "minimal risk" studies involv- ing the possibility of psychological and social harms, and are left to determine the risks for themselves.

Although it seems ironic to call for more research on the nature of research risks, there are some low- cost actions that could be immedi- ately taken to help addiess our cur- rent lack of knowledge. For exam- ple, as the NRC recommended, researchers could be encouraged to build debriefing questions into their research studies in order to begin collecting information regarding per- ceptions of risk and harm in behav- ioral research.̂ "* Simple probes designed to learn about subject per- ceptions and reactions to specific behavioral research procedures would be particularly valuable. Fendrich and colleagues employed debriefing probes to evaluate subject reactions to a request to provide hair samples as part of a survey interview concerned with substance use behav- ior.̂ 5

The development and widespread

adoption of common sets of proce- dures for evaluating subject percep- tions of research risks could provide importani: empirical information regarding the nature, severity, and frequency of research risks as per- ceived by participants. Of course, development of feedback mecha- nisms to ensure this information is routinely shared with IRBs would also be important. Such information would pnjve useful to protocol reviewers., enabling them to consider more directly subject reactions to potential risks when evaluating new behavioral protocols. This informa- tion would also prove invaluable to future research subjects, potentially providing them with risk assessments similar to those now available in bio- medical research.

• Susan M. Labott, PhD is Director, Health Phychology Service, Department of Psychiatry at the University of Illinois at Chicago; Timothy P. Johnson, PhD is Director, Survey Research Lj.boratory, College of Urban Planning and Public Affairs at the University of Illinois at Chicago; Both are Vice-Chairs of the behavioral IRB at UIC.

Referenctis 1. National Commission for the Protection

of Human Subjects of Biomedical and Behavioral lilesearch. The Belmont Report: Ethical Principles and Guidelines for the Protection of Human Subjects of Research. Washington, D.C.: Government Printing Office, 1979.

2. 45 Cl-R 46. Protection of Human Subjects.

3. 45 CFR 46.102. 4. Barnbaum D. Making more sense of

"minimal risk." IRB: Ethics & Human Research i o o 2 ; 2 4 : i o - i 3 .

5. National Bioethics Advisory Commission. Research Involving Persons with Mental Disorders That May Affect Decisionmaking Capacity. Rockville, MD, 1998.

6. National Research Council. Protecting Participant! and Facilitating Social and Behavioral Sciences Research. National Academies Press: Washington, D.C., 2003.

7. Prentice ED, Gordon BG. Institutional review board assessment of risks and benefits associated with research. In: National Bioethics Advisory Commission. Ethical and Policy Issues in Research Involving Human Participants, Vol. II, Bethesda, MD, 2001.

8. Hermeren G. Human and social conse- quences of research. Research Ethics

MAY-JUNE 2004 IRB: ETHICS &. HUMAN RESEARCH

6o-369; Sieber JE. Planning research: Basic ethical decision-making. In: Sales BD, Folkman S, eds. Ethics in Research with Human Participants. American Psychological Association: Washington, D.C., 2000:13-26.

9. See ref. 8, Sieber 2000. 10. Singer E, Hippler H, Schwarz N.

Confidentiality assurances in surveys: Reassurance or threat? International Journal ofPuhlic Opinion Research i^^z;4:z$6-i68.

11. Singer E, Von Thurn DR, Miller ER. Confidentiality assurances and response. Public Opinion Quarterly 199$-,59:66-77.

12. Singer E, Mathiowetz NA, Couper MP. The impact of privacy and confidentiality concerns on survey participation. Public Opinion Quarterly i993;57:465-482.

13. See ref. 10, Singer 1992. 14. Newman E, Kaloupek DG, Keane

TM, Folstein SE Ethical issues in trauma research. In: Kantor GK, Jasinski JL, eds. Out of the Darkness. London: Sage, 1997, p. 271- 281.

15. Newman E, Walker EA, Gefland A. Assessing the ethical costs and benefits of

trauma-focused research. General Hospital Psychiatry I 9 9 9 ; 2 i : i 8 7 - i 9 6 .

16. Sheon AR, Wagner L, McElrath MJ, Keefer MC, Zimmerman E, Israel H, Berger D, Fast P. Preventing discrimination against volunteers in prophylactic HIV vaccine trials. Journal of Acquired Immune Deficiency Syndromes and Human Retrovirology

work organizations. Ethics & Behavior

17. Oakes JM. Survey research. In: Amdur RJ, Bankert EA, eds. Institutional Review Board: Management and function. Jones and Bartlett: Sudbury, MA, 2002, p. 428-433.

18. Sieber JE. The ethics and politics of sensitive research. In: Renzetti CM, Lee RM, eds. Researching Sensitive Topics. London: Sage, 1993, p.14-26.

19. Singer E, Levine EJ. Protection of human subjects of research: Recent develop- ments and future prospects for the social sci- ences. Public Opinion Quarterly

20. Ilgen DR, Bell BS. Conducting indus- trial and organizational psychological research: Institutional review of research in

3 5 21. Schreier BA, Stadler HA. Perspectives

of research participants, psychologist investi- gators, and institutional review boards. Perceptual and Motor Skills i99x;74:1067- 1072.

22. Kimmel AJ. Predictable biases in the ethical decision making of American psychol- ogists. American Psychologist \99i;t\6:7S6- 788.

23. Oakes JM. Risks and wrongs in social science research: An evaluator's guide to the IRB. Evaluation Review 2oo2;26:443-479, p. 467.

24. See ref. 6, National Research Council 2003.

25. Eendrich M, Johnson T, Wislar JS, Sudman S. The feasibility of hair testing in a household survey on drug abuse. In: Mieczkowski T, ed. Drug Testing Technology: Assessment of Field Applications. Boca Raton, FL: CRC Press, 1999, p. 235-253.

NEWS & NOTES

Food and Drug Administration. "Innovation Stagnation: Challenge and Opportunity on the Critical Path to New Medical Products."

March Z004. FDA white paper that analyzes the factors contributing to the slowdown of new drug and biological medical products reaching patients. http://www.fda.gov/oc/initiatives/ criticalpath/whitepaper.pdf

Institute of Medicine. "The Ethical Conduct of Clinical Research Involving Children." March 2004. Under the Best Pharmaceuticals For Children Act of 200Z (RL. 107-109), Congress charged the IOM to review the current regula- tory context involving research with children and to make recommenda- tions regarding the regulations' appro- priateness in relation to child partici- pants' ages; IRB interpretation of regu- latory criteria in approving research; the parental permission and child assent processes; child and parental expectations and comprehension about research participation; payment for research participation; regulatory com- pliance and enforcement issues; and

IRB roles and responsibilities. http://www.iom.edu/report.asp?id=i94 2Z

Department of Health and Human Services. "Clinical Research and the HIPAA Privacy Rule."

Detailed information that includes a Frequently Asked Questions and Answers section explaining how the Privacy rule affects researchers and IRBs. http://privacyruleandresearch.nih.gov/ clin_research.asp

National Institutes of Health. Office of Biotechnology Activities. Recombinant DNA Advisory Committee (RAC) Informed Consent Working Croup. "Cuidance on Informed Consent For Cene Transfer Research." February 2004. Provides researchers, IRBs, and others with guidance in applying Appendix M of the NIH Guidelines for Research Involving Recombinant DNA Molecules (gene transfer research). http://www4.0d.nih.gov/0ba/rac/ic/

National Institutes of Health. Office of Biotechnology Activities. NIH Genetic Modification Clinical Research Information System (CeMCRIS).

A comprehensive resource that pro- vides information about NIH-regis- tered human gene transfer trials, e.g., investigators and institutions conduct- ing these trials, types of gene products in use, and study protocol summaries. http://www.gemcris.od.nih.gov/

United Kingdom Department of Health. "Proposals for New Legislation on Human Organs and Tissue."

A summary of the proposed Human Tissue Bill and justification for its pas- sage. The bill provides a consistent legal framework for the storage and use of human organs and tissue obtained from adults and children. http://www.dh.g0v.uk/assetR00t/04/07/

United Kingdom Parliament. Human Tissue Bill (text). http://wvvrw.parliament.the-stationery- 0ffice.c0.uk/pa/cm200304/cmbills/009/ 2004009.htm

IRB: ETHICS & HUMAN RESEARCH MAY-JUNE 2004