When “Minimal Risk” Research Yields Clinically-Significant Data, Maybe the Risks Aren’t So Minimal
by Helen M. Sharp, Robert D. Orr
2004. The American Journal of Bioethics 4(2):W32-W36

Surveys and routine clinical procedures applied in research protocols are typically considered only minimally risky to participants. The apparent benign nature of "minimal risk" tasks increases the chance that investigators and Institutional Review Boards (IRBs) will overlook the probability that clinical tools will identify signs, symptoms, or definitive test results that are clinically-relevant to subjects' welfare. "Minimal risk" procedures may also pose a particular hazard to participants in clinical research by increasing the therapeutic misconception because the tasks mimic clinical care and are often conducted in clinical settings. Investigators should anticipate which measures could yield clinically-important findings and should describe explicit plans for data monitoring, disclosure, and follow-up. Protocols that include reliable and valid clinical measures should prompt a more detailed risk assessment by the IRB, even when the tasks meet the regulatory criteria for minimal physical, psychological, or emotional risk.

Key Words: research ethics, behavioral research, informed consent, disclosure, depression, nontherapeutic human experimentation, institutional review board (IRB).

Jane is about to begin treatment for breast cancer in an academic medical center. Her primary oncologist is the principal investigator for a longitudinal study of patient-reported functional status and quality of life among women being treated for breast cancer. He asks Jane if she is interested in learning more about the study. She agrees to meet with the research nurse to review the consent forms.

Participants in this study complete several questionnaires before they begin treatment, at monthly intervals during treatment, and for a year after treatment is completed. Usually, subjects complete the questionnaires on their own during their clinic visits. The data forms are filed by subject identification number, kept in locked files, and the data are entered at the completion of the study. A depression screening survey was included in the protocol because the investigators think there may be a relationship between depression and subjects' reports of quality of life.

Jane completes the initial surveys and returns them to the receptionist when she leaves the clinic. As the research assistant files Jane's questionnaires, she notices that Jane indicates she has recently thought about suicide. The assistant notifies the research team. Some members of the team feel strongly that they should contact Jane and intervene, but others are unsure whether it is appropriate to contact her because this information was gained through her participation as a research subject.

The Problem
The conflict among the members of the research team raises a primary question: are investigators obligated to notify adult participants of findings, referral, or treatment options when research data have implications for the subject's well-being or future health? This question applies across all health services, epidemiologic, and genetic research. With disclosure and intervention, Jane's participation in research could benefit her. On the other hand, Jane agreed to participate in the study knowing that the study would not provide direct benefit to her and with assurances that her responses would be kept confidential and would not influence her clinical care. If the presumption is non-disclosure, the danger in this case is that Jane may believe she has communicated her suicidal thoughts in a clinical context, particularly because her treating oncologist is also the principal investigator.

This case also raises secondary questions about data monitoring. Data are often handled by research assistants or technicians who may not recognize the clinical significance of a particular lab value or response on a survey. It is unusual to explain that the primary investigator may not see the raw data until the study is complete, if ever. While subjects are usually informed about how and where their data will be stored, it is rare to inform people that their data may not be examined by anyone for weeks, months, or even years.

Discussion
"Minimal Risk" Research
The National Bioethics Advisory Commission (NBAC) (2001) commented that a system for classifying research as "minimal risk" is useful because it allows IRBs to focus on research that is more likely to cause harm to subjects. The Code of Federal Regulations (CFR) defines minimal risk as "the probability and magnitude of harm or discomfort anticipated in the research are not greater in and of themselves than those encountered in daily life or during the performance of routine physical or psychological examinations or tests" (45 CFR 46.102 i 1991). Included in the NBAC analysis is discussion about whether the subject's characteristics (e.g., having cancer) should influence interpretations of risk encountered in daily life. The report concludes that IRBs should interpret risk relative to "common risks - for example driving to work, crossing the street, getting a blood test, or answering questions over the telephone" (NBAC 2001). The study of quality of life in patients with breast cancer that we describe fits both the CFR and NBAC descriptions of minimal risk.

When a study is identified as "minimal risk," the chair of the IRB is authorized to review and approve the study using an "expedited review" process, without a full IRB review of the protocol and consent forms (46 CFR 110 a & b 1991). Some "minimal risk" research allows for IRBs to waive written consent from participants (46 CFR 117 c.2 1991). If new data are collected, such as in the case described, consent forms are typically used. Consent forms acknowledge the risks of participation including breach of confidentiality and the possibility that subjects could be distressed by the research tasks, such as answering personal or emotional items on a survey. Protections such as removing identifiers, using subject numbers not names, and secure data storage are offered. Subjects are reminded that they may skip any questions or withdraw from participation at any time during the study (21 CFR 50.25 8 1998; World Medical Association 2000).
The label "minimal risk" and the expedited review process tend to focus the reviewer on the risk of the tasks, rather than the implications of using standardized clinical measures to identify conditions such as depression, diabetes, or hypertension. While individual IRBs may consider these implications, the inherent variability across IRBs coupled with the regulatory focus on the task-associated risk, increase the chance that both investigators and IRBs will overlook the risks embedded in apparently innocuous research. Studies that undergo full IRB review also have the potential to overlook the implications of "minimal risk" procedures when these are included in trials or studies that involve clear physical risk to participants.

The Therapeutic Misconception
One purpose of written informed consent for participation in research is to uphold the distinction between research and therapy for both subjects and investigators. Differentiating research from therapy is important so that clinician-investigators and patient-subjects are clear about their obligations and expectations (Appelbaum 2002; Miller and Brody 2002). Miller and Brody (2002; 2003) argue that the distinction between the ethical obligations of the investigator and the duties of the clinician become blurred when a clinician-investigator takes on a dual role. This boundary is blurred further when reliable and valid clinical measures are used for research purposes by the primary treating physician. Despite efforts to establish clear boundaries between clinical care and research through informed consent, subjects often maintain a "therapeutic misconception" or belief that there will be some personal health benefit from participating in research (Appelbaum et al. 1987). Even when subjects are able to explain concepts such as randomization to treatment or placebo arms of a trial, many seem to retain a belief that their physician maintains some control and interest in the therapy they receive (Appelbaum et al. 1987). Sugarman and colleagues (1998) report that patient-subjects also infer differing degrees of risk based on the words used to describe the research. People tend to assign greater risk and uncertainty to "medical experiment," when compared to "medical research," while "medical study" is considered the most "benign" and most likely to benefit the participant.

Survey research or the use of routine clinical tests, such as measures of blood pressure, may present a particularly high risk for advancing the therapeutic misconception, especially when subjects complete research tasks during routine clinic visits. In the case we describe, Jane is likely to believe she has conveyed her suicidal thoughts directly to her physician. If Jane is at all confused about the research-therapy distinction, the lack of response from her clinical-research team could worsen her isolation and depression.

The therapeutic misconception, coupled with tasks that seem like clinical procedures (e.g., hearing screenings or measures of blood pressure), often conducted in clinical settings, are likely to disadvantage or even endanger research participants if investigators isolate clinically-relevant data as "research" and distance these from their clinical obligations to promote the patient's welfare. The therapeutic misconception is sufficiently prevalent that when research data yield clinically-concerning results it is most reasonable to assume that the patient-subject does not differentiate the role of the investigator as substantively detached from that of physician.

Disclosure of Research Findings to Participants
With the exception of genetic research, there is little guidance in the literature regarding disclosure of research findings to participants. The NBAC report, Research Involving Human Biological Materials: Ethical Issues and Policy Guidance, addresses the issue of disclosure and concludes that:

IRBs should develop general guidelines for the disclosure of the results of research to subjects and require investigators to address these issues explicitly in their research plans. In general these guidelines should reflect the presumption that the disclosure of research results to subjects represents an exceptional circumstance. Such disclosure should occur only when all of the following apply:
a) the findings are scientifically valid and confirmed,
b) the findings have significant implications for the subject's health concerns, and
c) a course of action to ameliorate or treat these concerns is readily available. (NBAC 1999)

NBAC also recommends that investigators specify in the protocol the findings that would prompt disclosure and that "appropriate medical advice or referral should be provided" (NBAC 1999). To our knowledge, these recommendations for handling clinically-relevant findings in genetic and tissue research have not been widely adopted across other biomedical or behavioral research.

The primary obligation to uphold the welfare of the individual is stated in the Declaration of Helsinki "in medical research on human subjects, considerations related to the well-being of the human subject should take precedence over the interests of science and society" (World Medical Association 2000), and is echoed in many other writings. (e.g., American Association of Universities 2000; Miller et al. 1998; Shalala 2000). Non-disclosure of treatment options to subjects when standard clinical treatment is available has been critiqued in observational studies, such as the Tuskegee Syphilis Study (Caplan 1992; Rothman 1982). Researchers' obligations to disclose or act on research findings in order to benefit the subject of the research remain somewhat unclear in the regulations, yet seem intuitive from the perspective of maximizing the subject's welfare.

At present, there is no mandate for disclosure or referral for participants in clinical, behavioral, or epidemiologic research. Clinical measures are routinely employed in these forms of research and will identify people with abnormal clinical signs, symptoms, or disorders. Given the likelihood that the participants are at least slightly unclear about the boundary between clinical practice and research, the current presumption for non-disclosure provides inadequate protection to participants.
Based on the literature and current guidelines for the conduct of research involving human subjects, we offer six recommendations for investigators and IRBs when "minimal risk" research protocols include reliable and valid clinical tools.

Recommendations
1. Reconsider the Assessment of Risk When Clinical Measures or Survey Tools Could Detect Clinical Conditions
Current guidelines that govern risk assessment appear to limit the evaluation to the potential harms (physical, emotional, or psychosocial) associated with the research tasks, deception about the purpose of the research, and the potential for breaches of confidentiality. The Office for Protection from Research Risks (OPRR), now the Office of Human Research Protections (OHRP), recommends in its Institutional Review Board Guidebook (1993) that IRBs should consider the potential for social or psychological harm related to the findings of genetic research when evaluating risk to participants. The guidebook specifies that answering interview questions may appear to qualify as "minimal risk" research, but may present a higher risk when the implications of the data are considered. This same logic should extend to the assessment of risk for any protocol in which researchers use clinical tools that appear to meet minimal risk standards, but yield reliable and valid measures of any condition relevant to the subjects' welfare.

2. Identify Relevant Clinical Information in Every Protocol
When research yields information about potentially treatable and/or life threatening conditions, there should be little debate that the finding and information about referral and treatment options should be discussed with the participant. Investigators should be able to identify the research tasks that could yield reliable and valid information that is clinically-relevant to the participants. When such tasks will be employed, investigators should establish particular responses (e.g., participant is suicidal) and threshold scores a priori (e.g., diastolic blood pressure greater than 105) that will trigger reassessment or disclosure to clinical personnel and the participant.
The obligation to notify participants of clinical findings is distinct from the duty to notify participants of harms caused by the research, particularly with respect to the investigators' or institutions' financial responsibilities to treat conditions that result from participation.

3. Develop a Plan for Disclosure to Participants
The plan for handling these findings should be detailed in the protocol and justified in the IRB application (NBAC 1999). The consent process should inform prospective subjects that particular findings will initiate a clinical response that will override anonymity protections. Some researchers inform participants of the general findings of their research through a written research summary or newsletter (Beskow et al. 2001). This approach is an effective, anonymous method for informing and involving participants, especially with longitudinal research, but should be differentiated from a requirement to disclose abnormal findings or clinically-concerning results directly to the participant.

Disclosure of clinically-relevant findings should be conducted in-person whenever possible and should adhere to clinical standards, including confidentiality. Ideally, a member of the clinical team should be assigned to disclose such findings to a participant. If non-clinical investigators are responsible for conveying test results, making clinical referrals, or suggesting treatment options they should be trained in appropriate counseling techniques.

4. Withholding Clinical Information from Participants Should be An Exception
If the research yields data that are preliminary or cannot be interpreted clinically, then disclosure of findings would be premature or inadvisable in most cases (NBAC 1999). When treatment options are unclear or the finding is relatively minor, it may be justifiable to wait until the completion of the study before disclosure or recommendations for treatment are made. Participants should be informed that no information will be offered until the study is complete. While this option may be viable for benign conditions, waiting for months or even years to recognize suicidal ideation, major depression, or other significant findings cannot be justified, even with consent from participants.

Some researchers might argue that influencing a subject's clinical care based on research findings could have a detrimental effect on the purity of the data. In Jane's case, data regarding depression are being collected as an adjunct to overall quality of life measures and not as a primary end-point of the research. Therefore, a decision not to intervene based on the integrity of the study is unsupportable.

In some cases, clinical tests can be expected to reveal sensitive information such as non-paternity or that an individual is affected with a late-onset degenerative condition, such as Huntington chorea (Lucassen and Parker 2001). Participants should be informed when protocols could reveal sensitive or potentially unwanted information. Prospective participants should decide whether to enroll with a full understanding of what the tests could reveal. While findings such as pregnancy require disclosure, some protocols could allow the participant to elect not to be informed of sensitive information.
If investigators believe it is justifiable to withhold clinically-relevant findings from subjects, the IRB should be aware of this decision. Potential subjects should be informed that they will not be told of negative test results (Beskow et al. 2001), offered referrals, or treatment based on the findings of the research.

5. Monitor Data from Clinical Measures Throughout the Study
The IRB should encourage timely and ongoing review of potentially sensitive data, so findings that could have bearing on a patient-subject's welfare do not go unread for months (Emanuel et al. 2000). This review could be conducted by the researchers themselves or, if blinding is important, data could be entered and reviewed by an independent person or group. This recommendation is not currently required for protocols that meet the standard for "minimal risk" research, but is consistent with existing mechanisms for interim analyses and independent review by Data and Safety Monitoring Boards for clinical trials (Emanuel et al. 2000; NBAC 2001; NIH 1998).

6. Report Unanticipated Clinical Findings to the Participant and the IRB
While advance planning and anticipation is always preferable to post-hoc protocol modifications, planning and protocol design fail occasionally. In circumstances in which research tests reveal unanticipated or incidental information that relates to subject safety or welfare, the welfare of the patient-participant should always be the priority and in most cases, the participant and the IRB should be notified.

Conclusion
The IRB has an important role in educating researchers about the implications of including "routine" clinical assessment tools in research protocols. Routine clinical procedures and diagnostic tests may be a particular hazard in clinical outcomes research because they are difficult to differentiate from clinical care, are often conducted in clinical settings, and qualify as "minimal risk" procedures, yet they yield meaningful data with implications for the welfare of participants.
We recognize that investigators and IRBs are inundated with requirements, regulations, and policy. However, the IRB has the opportunity to review protocols with fresh eyes and to identify clinical tests, tools, and procedures that are likely to yield positive findings and to assist researchers in revising their protocol to provide appropriate follow-up for subjects. The IRB should ensure that investigators have (a) identified any tasks that might lead to clinically-significant information, (b) stated a priori what type or range of data will trigger disclosure, (c) identified a process for appropriate notification, disclosure, or intervention, and (d) planned for timely and ongoing review of sensitive data. We believe that these recommendations lend better protections to participants of clinical research and are supported by existing policy, ethical standards, and guidelines for the conduct of "minimal risk" research involving human subjects.

Acknowledgements
The authors thank two anonymous reviewers for their helpful suggestions in refining this paper.

References
American Association of Universities. 2000. Report on University Protections of Human Beings Who are Subjects of Research. Washington, D.C.: American Association of Universities.

Appelbaum, P. S. 2002. Clarifying the ethics of clinical research: A path toward avoiding the therapeutic misconception. The American Journal of Bioethics 2(2): 22-23.

Appelbaum, P. S., L. H. Roth, C. W. Lidz, P. Benson, and W. Winslade. 1987. False hopes and best data: Consent to research and the therapeutic misconception. Hastings Center Report 17(2): 20-24.

Beskow, L. M., W. Burke, J. F. Merz et al. 2001. Informed consent for population-based research involving genetics. JAMA 286(18): 2315-2321.

Caplan, A. L. 1992. Twenty years after. The legacy of the Tuskegee Syphilis Study. When evil intrudes. Hastings Center Report 22(6): 29-32.

Emanuel, E.J., D. Wendler, and C. Grady. 2000. What makes clinical research ethical? JAMA 283(20): 2701-2711.

Lucassen, A., and M. Parker. 2001. Revealing false paternity: Some ethical considerations. Lancet 357:1033-1035.

National Bioethics Advisory Commission (NBAC). 1999. Research Involving Human Biological Materials: Ethical Issues and Policy Guidance. Volume I. Rockville, MD: National Bioethics Advisory Commission.

-----. 2001. Ethical and Policy Issues in Research Involving Human Subjects: Volume I. Bethesda, MD: National Bioethics Advisory Commission.

National Institutes of Health (NIH). 1998. Policy for Data and Safety Monitoring. Available from: http://grants1.nih.gov/grants/guide/notice-files/not98-084.html.

Office for Protections from Research Risks (OPRR). 1993. Chapter 5. Biomedical and Behavioral Research. Protecting Human Subjects: Institutional Review Board Guidebook. Washington, DC: US Government Printing Office. Available from: http://ohrp.osophs.dhhs.gov/irb/irb_chapter5.htm.

Miller, F. G., and H. Brody. 2002. What makes placebo-controlled trials unethical? The American Journal of Bioethics 2(2): 3-9.

-----. 2003. A critique of clinical equipoise: Therapeutic misconception in the ethics of clinical trials. Hastings Center Report 3:19-28.

Miller, F. G., D. L. Rosenstein, and E. G. DeRenzo. 1998. Professional integrity in clinical research. JAMA 280(16): 1449-1454.

Rothman, D. J. 1982. Were Tuskegee and Willowbrook 'Studies in Nature'? Hastings Center Report 12(2): 5-7.

Shalala, D. 2000. Protecting research subjects: What must be done. New England Journal of Medicine 343(11): 808-810.

Sugarman, J., N. E. Kass, S.N. Goodman et al. 1998. What patients say about medical research. IRB 20(4): 1-7.

World Medical Association. 2000. Declaration of Helsinki: Ethical principles for medical research involving human subjects (reprinted). JAMA 284(23): 3043-3045.