4
Informed Consent

Although the informed consent process necessarily takes place before the collection, storage, and analysis of biological specimens and biodata, it requires giving potential participants an honest explanation of exactly what the research entails and what risks they face by participating. The consent process, therefore, hinges on the information and observations offered in earlier chapters of this report.

While clinicians and biomedical researchers have contended with issues of informed consent related to biological data for many decades, consensus on how to address these issues remains elusive. Active debate continues among clinicians, researchers, and ethicists about exactly what forms of consent are necessary in various situations. The debate has been particularly heated with respect to large-scale repositories of human biological specimens that are linked to clinical or socioeconomic data and are being used to assess interactions between genes and the environment similar to those discussed in this report. Thus, social scientists considering adding the collection of biospecimens to their surveys will find that doing so makes informed consent a much more complex process than is the case for survey research that involves collecting only self-reported psychological, social, or economic data. This is true in particular if the biospecimens and the data derived therefrom will form a large collection to be used by numerous researchers for an indeterminate period of time for research purposes that may all not be well defined at the time consent is obtained, especially if genomic data are involved. The controversies do not make this kind of research impossible, but they do add substantial complications that social science researchers need to anticipate. This chapter begins with an overview of the



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 55
4 Informed Consent A lthough the informed consent process necessarily takes place before the collection, storage, and analysis of biological specimens and biodata, it requires giving potential participants an honest explanation of exactly what the research entails and what risks they face by participating. The consent process, therefore, hinges on the information and observations offered in earlier chapters of this report. While clinicians and biomedical researchers have contended with issues of informed consent related to biological data for many decades, consensus on how to address these issues remains elusive. Active debate continues among clinicians, researchers, and ethicists about exactly what forms of consent are necessary in various situations. The debate has been particularly heated with respect to large-scale repositories of human biological specimens that are linked to clinical or socioeconomic data and are being used to assess interactions between genes and the environment similar to those discussed in this report. Thus, social scientists considering adding the collection of biospecimens to their surveys will find that doing so makes informed consent a much more complex process than is the case for survey research that involves collecting only self- reported psychological, social, or economic data. This is true in particular if the biospecimens and the data derived therefrom will form a large collection to be used by numerous researchers for an indeterminate period of time for research purposes that may all not be well defined at the time consent is obtained, espe - cially if genomic data are involved. The controversies do not make this kind of research impossible, but they do add substantial complications that social sci - ence researchers need to anticipate. This chapter begins with an overview of the 

OCR for page 55
 CONDUCTING BIOSOCIAL SURVEYS informed consent process. It then examines how collecting biospecimens alters the informed consent process with which social scientists are familiar. The third section reviews a number of unresolved issues surrounding informed consent in social science survey research that includes the collection of biospecimens. The chapter ends with a discussion of the role of Institutional Review Boards (IRBs) in the informed consent process for such research. OvERvIEW OF THE INFORMED CONSENT PROCESS This section summarizes the purpose and the basic elements of the informed consent process. The Purpose of Informed Consent Informed consent as it is understood today did not exist until about mid - way through the twentieth century (see, for example, Faden and Beauchamp, 1986). In the medical arena, physicians historically were obligated to act in the best interests of their patients as they understood those interests, but they felt no obligation to inform a patient completely about a situation and then follow the patient’s decision. A scattering of early cases established that physicians had to provide patients with information about proposed procedures and that the terms of the patient’s consent could place legal limits on the procedures that could be performed. It was not until a series of cases in the 1950s and 1960s that clinical informed consent became well established as a legal right with full legal redress, equivalent to the penalties for battery, if it was not obtained. Battery theory was gradually replaced by negligence theory in the 1950s and 1960s (Levine, 1988). Similarly, there was historically no generally recognized obligation of those performing medical research to inform their subjects about the conduct, purpose, and risks of the research and to obtain their consent for participation, although there were some rudimentary versions of informed consent as early as 1900 (Vollman and Winau, 1996). A famous case in which informed consent was obtained was the Walter Reed yellow fever studies in the late nineteenth century. The first internationally recognized code of research ethics, the Nuremberg Code of 1947, was developed in response to revelations concerning abhorrent Nazi human experimentation before and during World War II. It empha- sized the importance of informed consent and the centrality of subjects’ self- determination to participate in medical research. Some years later, in 1964, at the 18th World Medical Assembly in Helsinki, Finland, the World Medical Association adopted the Declaration of Helsinki, which provided a set of guid - ing principles for ethical considerations related to biomedical research. Until the 1960s, informed consent to research generally was obtained only from vol -

OCR for page 55
 INFORMED CONSENT unteers. Then, as a result of U.S. Food and Drug Administration (FDA) statute and the Declaration of Helsinki, the practice was extended to patients. In the United States, the importance of obtaining informed consent for medical research achieved national prominence in 1972 following revelations of unethical practices by the U.S. Public Health Service in the Tuskegee syphi - lis study. In 1974, the U.S. Congress passed the National Research Act, which created the National Commission for the Protection of Human Subjects of Bio- medical and Behavioral Research. In 1979, the commission issued the Belmont Report, which set forth three guiding principles underlying the ethical conduct of research: (1) respect for persons, (2) beneficence, and (3) justice. The purpose of informed consent and the reasons behind it have changed little since the Declaration of Helsinki was issued: no one should be subjected to experimentation—or other research activities—without giving permission, and that permission is meaningful only if the person understands what he or she is giving permission for. Thus the process of informed consent is designed to provide human subjects with sufficient information about the nature and purpose of the research, the procedures to be employed, and the potential risks and benefits of their participation so they can make an informed decision about whether to participate. As the types of studies have become more complex, the issues surrounding informed consent have become more complex and difficult to resolve, but the basic underlying principles have not changed. Basic Elements of Informed Consent The basic elements of informed consent are set forth in the U.S. Code of Federal Regulations (45 CFR 46 Part 116). First, the language of the informed consent document should be reasonably clear and specific. Some commentators have recommended that the document be written at the fifth-grade or some other low level of reading capability, but the precise appropriate reading level is somewhat controversial (see, for example, Levine, 1982). Among other topics, the document should address the following: • he fact that research is being performed, the reason why it is being t performed, and what the participant will be asked to do; • he risks and potential benefits (including financial and nonfinancial) to t the participant; • he level of confidentiality that will be maintained, which can include t such information as whether the data resulting from collected biospeci - mens will be deidentified or coded and whether the data or study results will be released to the participant, his/her family, or his/her health care provider; and • the participant’s right to withdraw consent without penalty.

OCR for page 55
 CONDUCTING BIOSOCIAL SURVEYS Adding biological specimens to social science surveys changes the nature of the informed consent required. Because of the unique risks involved, social science surveys that collect biospecimens should include several additional ele - ments in the informed consent document:1 • process for dealing with blanket consent (see below), including what a process will be used when the participant’s specimens or data are to be used in future studies; • process for dealing with any significant medical information that a is uncovered by the analysis (in general, any information that is not released to the subject will not be released to treating physicians. On the other hand, the subject may choose to have some data released to him/her without also being released to treating physicians); • he process governing withdrawal and what it means with respect to t what can be done with the specimens and data after withdrawal; and • ow third-party issues (see below) will be handled, including questions h about third-party privacy and the furnishing of medically relevant infor- mation that may affect third parties. Generally, although informed consent should be obtained before biospeci - mens are collected, in some instances postcollection consent is appropriate. This is the case, for instance, for the use of leftover specimens from earlier surgery or cases in which it was not initially possible to obtain consent because of the participant’s illness, undue stress, or inability to comprehend the consent procedure.2 DIFFERENCES IN CONSENT PRACTICES BETWEEN SOCIAL SCIENCE SuRvEy AND BIOMEDICAL RESEARCH When social science surveys include the collection of biospecimens, two hitherto separate informed consent processes coincide. Social scientists must deal with requirements and issues different from those to which they are accustomed. 1An alternative approach, used in the Personal Genome Project (PGP) is to make volunteers score 100 percent on a consent exam. This is intended to ensure that participants understand that there is no real way to predict what can be done with genomic data in the future, nor any way to protect confidentiality or protect against possible misuses of the data in the future with 100 percent certainty. The panel does not believe that this is an appropriate or practical strategy in this context. 2Note that even the most sincere investigator cannot anticipate all of the ways in which a biologi - cal specimen may be used in the future. This is a major issue, but without a clear solution.

OCR for page 55
 INFORMED CONSENT Traditional Consent Practices for Social Science Surveys Traditionally, the informed consent process in social science survey research has been relatively straightforward. Because the risk of a breach is understood to be fairly small, the informed consent forms tend to be minimal as well; obtaining consent over the telephone at the beginning of a survey call is a com - mon practice. At many institutions, the informed consent process is overseen by a nonmedical IRB, which generally has a very different idea of appropriate informed consent from that of a biomedical IRB (Martin and Marker, 2007). Traditional Consent Practices for Biomedical Research Biomedical informed consent has typically been a much more rigorous pro- cess than the traditional process in the social sciences. This is the case in part because in certain clinical trials and in other potentially invasive biomedical research, there can be a very real risk of physical harm. The informed consent process is generally seen as the responsibility of principal investigators and is usually overseen by medical IRBs. It involves extensive written consent forms containing detailed accounts of the potential risks, which are typically explained in face-to-face meetings with informed-consent personnel (Doyal and Tobais, 2001). This closer scrutiny of biomedical research has also influenced the treat - ment of informed consent for medical research that poses not physical risks but financial or psychological risks to participants. Informed Consent for Social Science Surveys That Include the Collection of Biospecimens Social scientists who wish to add the collection of biospecimens to their surveys will likely need to adopt standards of informed consent similar to those appropriate for minimal-risk biomedical research. (Minimal risk, as in the Com- mon Rule [see Chapter 3], is taken to mean that the probability and magnitude of possible harm or discomfort resulting from participating in the study is no greater than that encountered ordinarily in daily life or during the performance of routine physical or psychological examinations or tests.) In some cases, as discussed in Chapter 3, the research may involve more than minimal concerns and risks. The addition of data derived from biospecimens to survey data may increase the risk of a breach of confidentiality and the amount of potentially damaging information that could be revealed in such a breach. To the extent that social science surveys that include the collection of biospecimens pose risks commensurate with those posed by biomedical research—including the risk, real or perceived, of disclosure—the informed consent process should be of parallel rigor. IRBs dealing with biosocial surveys should benefit from the standards and practices of both social and behavioral science IRBs and medi -

OCR for page 55
0 CONDUCTING BIOSOCIAL SURVEYS cal IRBs in light of a careful reading of the Common Rule (see also, National Research Council, 2003). uNRESOLvED ISSuES A number of issues remain unresolved with respect to informed consent for social science survey research that involves the collection of biospecimens. They include consent for potential future uses of the specimens and the data derived therefrom, return of significant findings, withdrawal of consent, third-party issues, and the nature and extent of communication with participants. Consent for Potential Future uses The concept of informed consent assumes that participants can be truly informed about the various aspects of a study, including its benefits and risks. Yet with today’s increasingly rapid changes in science and technology, obtaining (and retaining) informed consent becomes a moving target. This is particularly true with respect to whole-genome sequencing, genomics, and information technology, in which changes occur so rapidly that it is difficult to imagine what might be possible in just a few years’ time. As a result, it may be impossible to guarantee a participant’s privacy and confidentiality or to discuss comprehen - sively what risks the participant might face by taking part in a survey in which biospecimens will be collected and stored for future use. Researchers who are collecting biological specimens (or establishing collec- tion protocols for such specimens) as part of a social science survey therefore face a difficult issue concerning consent for potential future research uses of the specimens they collect and the data derived therefrom. Because it is impossible to know what types of research might be performed on the specimens and data in the future, researchers cannot fully inform study participants about those potential future uses when the original consent is sought. There are two basic approaches to dealing with this issue. One is to return to the participants to obtain their consent for each new kind of study, explaining its purpose, con - duct, benefits, and risks. The other is to ask the participants at the time of the original survey to provide a blanket consent that permits use of the specimens and data for various studies in the future without the need to recontact them to request consent. There are also various combinations of these two approaches that entail varying degrees of contact with and commitment from participants, but the focus here is only on the two basic options. In addition, in some cases deidentification or anonymization of the data may obviate the need to obtain informed consent for future research using the data (see Chapter 3), although this means of avoiding informed consent poses ethical issues.

OCR for page 55
 INFORMED CONSENT Recontact Perhaps the most straightforward approach to ensuring that participants are comfortable with how their contributed specimens will be used is to recon - tact them to obtain informed consent whenever a new study is proposed. This approach guarantees that no specimens will be used without participants’ explicit consent. However, it is a process many investigators find too costly, labor-intensive, and burdensome, and the consensus in the field is that it is economically and logistically infeasible. Moreover, it is impossible to reach every person involved in the original study after a significant amount of time has passed, as many of participants will have moved, and some will have died. 3 Recontact procedures raise ethical questions as well, since it can be argued that the act of seeking to contact participants may violate privacy, especially if they asked not to be contacted again in their original consent document. Recon- tact also could be annoying to participants, especially if done frequently, and could reduce their willingness to continue to take part in the research. As a result, a variety of organizations and researchers have concluded that recontact is not a workable option and that insisting on it would prevent potentially valuable research from being carried out. A working group of the UNESCO International Bioethics Committee found, for example, that a “sys - tem which required fresh consent would be extremely cumbersome and could seriously inhibit research” (United Nations Educational, Scientific and Cultural Organization, 2002, p. 16). Similarly, Strouse (2005, pp. 142–143) suggests that “going back to subjects to seek specific consent for each later study undertaken with the initial sample may seriously inhibit potentially important research.” There is one report of a new approach that may make recontact easier and more feasible (Shickle, 2006). First Genetic Trust, a U.S. biotech company, has developed “dynamic consent procedures in which subjects are contacted by e-mail each time their DNA is used.” Study participants offer informed consent for the initial study, and later, if the same or a different group of researchers wishes to use the biospecimens or the data generated therefrom in a new study, the participants can say yes or no to the new research (Shickle, 2006, p. 507). One limitation of this approach is that it depends on the study participants having regular access to the Internet, which may be problematic, particularly for older individuals. 3The experience of the competition for the follow-up to the National Long-Term Care Survey may be instructive here. During the rebidding process, no applicant organization proposed retain - ing the existing longitudinal sample, partly because doing so would have required that the Census Bureau recontact each participant and obtain an additional consent before any new data could be collected.

OCR for page 55
 CONDUCTING BIOSOCIAL SURVEYS Blanket Consent Because of the difficulties involved in recontact, many organizations and researchers have used or have called for the use of blanket consent to the general use of specimens or data in future, unspecified projects. This type of consent can be advantageous both for participants, as less is demanded of them, and for researchers, who avoid costly and time-consuming recontact procedures. “If you are creating a biobank, the idea is that you want to build a resource that is going to have all kinds of uses going forward, many yet undefined. So blanket consent becomes something that, presumably, is more appealing” (Malinowski, 2005, p. 8). Among the organizations that have recommended the use of blanket con - sent for data in biorepositories are UNESCO and the World Health Organiza - tion groups working on these issues (World Health Organization, 1998; United Nations Educational, Scientific and Cultural Organization, 2002). The UK Bio- bank, one of the largest biobanks in the world, is not planning to seek consent prior to each new research study using stored biospecimens once the initial consent has been obtained (Shickle, 2006). The biobank does plan, however, to send out frequent newsletters to remind subjects of their participation and to inform them of what studies are being conducted and planned for the future. The biobank also uses e-mail as a form of contact, which opens up the pos - sibility of using this means of communication each time participants’ DNA is used, giving them sufficient information and time to make an informed decision about whether to extend their consent to the newer projects. Again, however, any approach that relies on e-mail and access to the Internet may leave out some participants, particularly the elderly. Approaches to Mitigating Problems Raised by Blanket Consent One major problem with blanket consent is that it is inherently vague, and thus study participants can never know exactly what their biospecimens and other information will be used for in the future. This lack of knowledge presents a serious ethical issue: blanket consent arguably fails to meet the basic requirement for informed consent that it describe the possible risks of a study to potential participants. If the future studies are unknown, the risks of those studies cannot be described. For example, study participants may agree to enroll in a study that includes collecting biospecimens and to allow the sharing of their samples as long as their confidentiality is protected. But what happens if a future study results in diagnosing a particular participant with a disease? Should the participant be contacted with that information? What if the par- ticipant would prefer not to know? (See the discussion of this issue below.) In short, it is always possible to envision scenarios that neither the participant nor the researcher could have anticipated at the time the informed consent was

OCR for page 55
 INFORMED CONSENT given. Thus, it can be argued that blanket consent cannot truly be “informed” consent (Deschênes et al., 2001). Some researchers have argued that such issues are not serious concerns and that a simple blanket consent is sufficient to safeguard the rights of participants. For example, Rothstein argues that “as long as the potential research subjects are clearly apprised of the range of possible future uses of their sample, they should be permitted to give one-time blanket consent to such uses” (Rothstein, 2005, pp. 92–93). Others, however, have suggested that these issues need to be addressed. Moreover, as noted in Chapter 2, IRBs are increasingly conservative and therefore unlikely to grant approval for the use of blanket consent. Various approaches to addressing this issue have been suggested. For instance, Greely (1999) proposes a variation on blanket consent that adds extra protections. His approach involves telling participants that their biological samples will be stored and explaining the possibility that they might be used for future research projects. At that point, participants could choose whether to consent to such future uses of their samples. Should researchers wish to undertake a future project using these samples, that project would have to be approved by an IRB or other appropriate body, which in certain circumstances could insist that the researchers return to the participants to obtain additional consent (Greely, 1999). Weir and Olick (2004) suggest a hybrid approach: participants would offer their consent for the specific study for which their biospecimens were collected and would also offer consent for future, unspecified studies that fell within certain categories or satisfied certain conditions (see also Kapp, 2008). In effect they would be offered a menu of research options from which to choose, and they could consent to as many as they liked, depending on their values and beliefs. They might, for instance, consent to the use of their specimens for any research on diabetes or cancer but not for studies aimed at understanding the genetic components of behavior. Another approach is to use a system of tiered consent, whereby partici - pants are allowed to decide to what degree their specimens and data can be used. For example, participants might be given the option of having their specimens and data used only for the original study and not for any future research; in this case, the informed consent document should explain what will be done with the data and, especially, the specimens, when the original study is complete.4 A second option might be to allow participants to specify the types of research for which their specimens could be used in the future. For example, a participant could agree for the specimens to be used only for cancer 4It is possible that a tiered consent procedure could introduce biases in research data, but there is a growing array of methods for dealing with missing data, including selectively missing data. A discussion of such methods is outside the scope of this report, but see Little and Rubin (2002) for more discussion on this point.

OCR for page 55
 CONDUCTING BIOSOCIAL SURVEYS research or for research into Alzheimer’s disease and diabetes, or perhaps for any sort of medical research. Care must be taken when considering such a tiered system of informed consent, however, since it may be inappropriate for a broad range of inter- and multidisciplinary research. A tiered approach should be used only if the biorepository (or other appropriate agent) has suf - ficient experience and a tracking system sophisticated enough to monitor the various levels of informed consent. Deidentification or Anonymization and Consent In certain cases when data have been deidentified—that is, meeting the requirements of the Health Insurance Portability and Accountability Act (HIPAA)5 but not rendered completely and irreversibly anonymous—participants are not considered “human subjects” under the Common Rule (45 CFR Part 46, Subpart A), so informed consent is not required (see Chapter 3 for a detailed discussion of the Common Rule and the HIPAA Privacy Rule). This would be the case, for example, if a researcher were using coded data or specimens from a repository that had been collected by other researchers and if the researcher could not readily determine the identity of the individuals involved. (In this context “coded data” refers to records in which identifying information about individuals has been replaced by a number, letter, symbol, or combination thereof.) For example, the researcher may have signed an agreement prohibit - ing him or her from receiving the key to the code—the procedure under which data contributed to dbGaP have been declared not to be human subjects data (see Box 3-3 in Chapter 3) (Office for Human Research Protections, 2008). In short, if the data and specimens have been deidentified or anonymized, if they were collected by someone else for a different study, and if the researcher agrees to use the data without the possibility of learning the identity of the subjects, then in general it is not necessary to obtain informed consent because the 5The HIPAA Privacy Rule makes two methods available for deidentifying health information: 1. Remove 18 specific identifiers listed in the Privacy Rule and determine that there is no other information that may identify the individual. The identifiers are names; geo- graphic subdivisions smaller than a state; all elements of dates (except year) related to an individual (including dates of admission, discharge, birth, death and, for individuals over 89 years old, the year of birth must not be used); telephone numbers; FAX num- bers; e-mail addresses; Social Security numbers; medical record numbers; health plan beneficiary numbers; account numbers; certificate/license numbers; vehicle identifiers and serial numbers including license plates; device identifiers and serial numbers; web URLs; Internet protocol addresses; biometric identifiers (including finger and voice prints); full face photos and comparable images; and, any unique identifying number, characteristic or code. 2. Obtain an opinion from a qualified statistical expert that the risk of identifying an individual is very small under the circumstances; the methods and justification for the opinion should be documented. (See 45 CFR Subpart E 164.514 (b).)

OCR for page 55
 INFORMED CONSENT participants who supplied the data and specimens are not considered human subjects. Aside from such special cases, however, for someone to be considered no longer a human subject for research purposes, complete and irreversible anonymization is generally required; deidentification is generally not enough. The difficulty with this approach is that, as discussed in Chapter 3, using deidentification or even anonymization is no guarantee that participants cannot be identified at some point. This problem is even more pressing in databases that combine social data with biological, and particularly genomic, data because the combination can sharply increase the risk of reidentification. Thus avoiding informed consent by using deidentified or anonymized specimens or data raises a serious ethical issue: although the participants are still exposed to risks from breaches of privacy, they are no longer afforded the usual protection given to human subjects under federal regulations (Greely, 2007). Moreover, this pro - cedure explicitly creates a double standard for what constitutes ethically and legally appropriate practice. The original investigators (and anyone who obtains access to the specimens or data through them or their storage facility) must play by the usual rules governing research with human subjects, while those using the deidentified or anonymized data are deemed no longer subject to the Com - mon Rule. Finally, as long as the original specimens or data exist in identifiable form, it is relatively easy for anyone who has the deidentified version of the data to reidentify them through a merger with the identifiable versions. Return of Significant Findings Investigators need to consider what they will do if their research yields information with implications for a participant’s health. In the course of ana - lyzing the biospecimens collected for a study, for example, researchers may learn of a significantly increased risk of a life-threatening disease linked to a characteristic present in some of the samples, frequently, though not always, a genetic characteristic. Participants often expect that researchers have a moral obligation to share such information with them. Although in many cases the accuracy of genetic testing and its application in disease prediction are subject to debate (Zimmern and Kroese, 2007; Offit, 2008; see also Chapter 1), in other cases the connections are quite clear. It can be argued that if this information could affect a reasonable individual’s health care options, it should be disclosed, at least under certain conditions (Greely, 2007)—for example, if the informa - tion clearly indicates the presence of an illness, and the illness is treatable. On the other hand, preliminary diagnostic data that reveal little about specificity or sensitivity should not be disclosed. Generally speaking, there should be differ- ent thresholds for disclosure depending on the severity of the illness in question and on whether an effective treatment exists. A further complication is that only laboratories covered by the Clinical Laboratory Improvement Amendments (CLIA) are allowed to disclose diagnostic information. Moreover, a participant

OCR for page 55
 CONDUCTING BIOSOCIAL SURVEYS may not wish to be contacted with any information about the study or its results, in which case any kind of disclosure, even that deemed to be a moral obligation, could be considered an invasion of the participant’s privacy. It is now customary to advise participants to contact a CLIA-approved laboratory independently themselves if they are interested in obtaining information from biodata that might have significant implications for their health. Currently, IRBs and ethics oversight committees are encouraged to review disclosure statements on a case-by-case basis. One of the factors that can influence their decision is the expected validity of the prediction based on the presence of the risk factor, which can change over time as a result of new sci- entific discovery (McGuire et al., 2008). In any case, the protocols for sharing study results (if any) with the participants should be clearly explained in the informed consent document, subject to review by the IRB. The issue can be expected to arise with increasing frequency as more becomes known about the human genome, leading to a growing number of incidental findings that reveal genetic information with known implications for health that are not part of the research hypotheses of the study. Moreover, the issue extends beyond genetic and other data derived from biospecimens. A social science study that includes measuring blood pressure will, for example, reveal the occasional participant with dangerous hypertension. One possibility is to use the consent process to inform participants that no diagnostic information will be disclosed, but that they should have regular medical care and consider asking a physician to obtain genetic or other potential diagnostic indicators such as those collected in the study. This may be a tempting option for studies in which the investigators cannot claim competence to deal with diagnostic issues. Withdrawal of Consent There is widespread agreement that a basic component of any system of informed consent must be the ability to withdraw consent at any time. Initially, this concept was applied to human subjects research, as can be seen in the Declaration of Helsinki: “In any research on human beings, each potential subject . . . should be informed that he or she is at liberty to abstain from par- ticipation in the study and that he or she is free to withdraw his or her consent to participation at any time” (World Medical Association, 1964). This concept actually originated much earlier, in the Nuremburg Code. Later, with the development of biobanks, the principle of withdrawal of consent was generally included in ethical guidelines for dealing with subjects who provide specimens (Eriksson and Helgesson, 2005). The Common Rule specifies that “participa - tion is voluntary . . . and the subject may discontinue participation at any time without penalty or loss of benefits to which the subject is otherwise entitled” (45 CFR 46 Part 116(a)(8)). The reasons for withdrawing consent to use stored specimens and data are

OCR for page 55
 INFORMED CONSENT usually different from the reasons for withdrawing consent for participating in medical research.6 In the latter case, participants are generally concerned about some sort of direct physical harm, but this is not an issue for stored specimens and data. Instead, participants are concerned about such things as sensitive information ending up in the wrong hands, leading in turn to discrimination, stigma, or personal distress. For example, prior to the passage of the Genetic Information Nondiscrimination Act of 2008, some participants might have worried that their genetic information would end up in the possession of an insurance company and possibly make it more difficult for them to obtain insurance. It is difficult to estimate the risks of such occurrences, although some researchers believe them to be relatively small (e.g., Eriksson and Helgesson, 2005). Nonetheless, in most cases a study participant may withdraw from the study at any time, and this right should be clearly stated in the informed consent document. When consent is withdrawn, the question arises of whether and to what extent the data gleaned from specimens can be used in research. As Shickle (2006, p. 515) notes: withdrawal raises some difficult questions. It is clear that new data collection should stop. However, the participant may still be content for their DNA and previously collected data to be used. Their decision may depend on whether withdrawal has been triggered by a desire not to share information about a new disease, perhaps that is associated with stigma, or by fundamental con - cerns about the way that the biobank is being maintained and used. In the case of previous disclosure or publication of the data, it may, in fact, be impossible not to use the data. As Shickle (2006, p. 515) comments, “It would be unreasonable and impractical for a participant to require that data previously collected be removed from any analyses previously conducted and potentially published.” The more difficult question is whether data derived from a participant’s biospecimens can be used in studies after consent is withdrawn. One possible solution to the continuing use of such data is anonymization. As with the use of anonymization to avoid the need for informed consent to uses of specimens and data beyond the original research, however, there are a variety of problems with using the procedure to deal with withdrawal of consent. As noted earlier, for example, anonymization renders the data less useful scientifically, especially when it comes to social survey data that are linked to administrative records and other health data (see Chapter 3). And as discussed earlier, anonymization may not completely cut the link to a given individual; it may still be possible to 6The regulations do not require that patients wishing to withdraw consent give a reason for doing so.

OCR for page 55
 CONDUCTING BIOSOCIAL SURVEYS identify the individual through various means, for example, by using identifying characteristics such as age, sex, and location. Stripping the specimen of all such information may render it useless for research purposes. Moreover, anonymiza - tion may not truly satisfy a participant’s wishes. If the participant desires, for whatever reasons, that his or her data not be used for any purposes at all, strip - ping the data of identifying information but still using them does not comply with this desire (Eriksson and Helgesson, 2005). Arguing that biobank research produces a public good, Eriksson and Helgesson (2005, p. 1075) suggest that withdrawal of consent should require more than simply a change of heart: “During the course of the research, a par- ticipant should be at liberty to withdraw his consent if he can present sufficient reasons why it is no longer reasonable to ask for his continued participation.” If the participant has sufficient reasons—something left to the researchers, bio - bank managers, or an ethics committee to decide—the data will be anonymized or destroyed, as the participant chooses. Likewise, samples may be destroyed, anonymized, or returned. The panel is not comfortable with Eriksson and Helgesson’s approach but does believe that there must be some practical limitations on the right to withdraw consent. In particular, the panel agrees that once data have already been published or relied upon in publications, it is not practical to delete them. (For a discussion of practical limitations on withdrawal of consent, see Levine [1988]). The larger message is that researchers should be aware of the issue, make plans for what they will do in cases of withdrawal of consent, and spell out those plans in the consent form. Third-Party Issues Because relatives of the participants in a study share some of the partici - pants’ DNA, many of the same risks participants face exist for their relatives as well, albeit in attenuated form. Any sensitive information, such as a genetic predisposition to a disease that is obtained from the genetic sample of a study participant may also apply to his or her biological relatives. Thus important questions arise concerning whether the researchers have any obligations to the participant’s family. Under federal regulations, third-party relatives are not generally consid - ered to be research subjects, and researchers are not required to obtain their consent. However, a case involving Virginia Commonwealth University (VCU) researchers about a decade ago indicates the difficulties that can arise when a study accumulates information about third parties. A VCU study question - naire had included questions about the psychiatric history of the participants’ parents, and the father of one of the participants objected and complained to federal officials that the questions invaded his privacy (Amber, 2000). When the Office of Protection from Research Risks (OPRR, now the Office of Human

OCR for page 55
 INFORMED CONSENT Research Protections, or OHRP) reviewed the minutes pertaining to the VCU IRB’s review of the study, it found that the IRB had not considered whether the father or other family members should be regarded as research subjects whose consent to participate would be required. OPRR agreed that the father was a subject and faulted the IRB for its failure to consider whether, in this case, family members were or were not research subjects. In the case of genome-wide association studies, it is impossible to avoid accumulating some information about third parties, since sequencing the genome of a study participant inevitably provides genetic information about that individual’s biological relatives. Although it appears inappropriate to extend autonomy-based consent procedures to third-party relatives, it is impor- tant to be aware that these risks exist, and that researchers’ ethical obligations to these individuals increase with the increasing risk of breaches of confidential- ity and autonomy (Greely, 1997). Thus, it can be argued that study participants should be fully informed about these types of risks during the initial consent stage, even to the point of encouraging them to discuss their participation and possible consequences with their families before enrolling in a study (McGuire et al., 2008). If information from a sample reveals potential health issues for a partici - pant, those health issues could affect relatives as well. As noted above, there are no clear answers about what to do if analysis of data in a study indicates a link between a particular characteristic and the likelihood of developing a disease. One can argue that a moral obligation exists to inform the participants in a study of such a risk, but one can also make the opposing argument that at least some participants may not want to know and that recontacting them with the information is an invasion of privacy. The ethical issues are even muddier for third parties, and the question of whether researchers are morally obligated to inform third-party relatives of the results of a biomarker association or similar study remains unanswered. Effective Communication with Participants In obtaining informed consent, it is not sufficient simply to decide what information should be provided to participants; one must also take into account the participants’ ability to understand and process that information. In other words, one must communicate the information effectively. One of the most obvious factors to take into account is language needs. The necessary information should be provided in a language in which the par- ticipant is fluent. Those whose English skills are questionable should be given the information in their native language. Even among native English speakers, it is important to consider exactly how the information is presented. Communications should be at an appropriate level. Furthermore, those preparing the consent form should be familiar with

OCR for page 55
0 CONDUCTING BIOSOCIAL SURVEYS research on health literacy. Studies among those with different levels of edu - cational attainment imply that a good deal of variation exists in health literacy even among populations with more than a high school education. To the extent possible, consent forms should be tailored to fit different populations. One population of special concern is the elderly, for whom cogni - tive impairment may make informed consent problematic. This is a particular issue among the oldest adults—those aged 85 and above. Another issue to consider is how much information to provide. A balance must be achieved between presenting too little information, so participants cannot truly understand what they are consenting to, and presenting too much information, which may interfere with obtaining meaningful consent. A good approach is to consider the materiality of the information—a legal term of art—and provide only that information which is material to the participant’s understanding of the benefits and risks of participation and therefore to his or her capability to make a decision.7 The Role of Institutional Review Boards The Code of Federal Regulations (45 CFR Part 46) requires the use of IRBs for all research funded by the U.S. Department of Health and Human Services (HHS) that involves human subjects, and the Common Rule extends this requirement to most other federal agencies that fund research involving human subjects (see Chapter 3). Thus almost all federally funded research proj- ects with human subjects must use IRBs, and they must follow the requirements concerning IRB use that are set forth in the federal statutes. 8 Review and prior approval by research IRBs serves as the major protective oversight mechanism for all human subjects research in the United States, but the role of IRBs in protecting the rights of research participants—and in par- ticular, their role in the process of informed consent—has been the subject of debate. Among other things, IRBs weigh in on the possible risks and benefits of participation in a study, the proper selection of participants, and whether informed consent should be obtained. If informed consent is required, an IRB must approve the informed consent document before the beginning of the study and then review it periodically. IRBs are also permitted to waive or alter the informed consent requirement if a study meets the conditions described in the Common Rule, the basic federal rule governing the role of IRBs in over- seeing human subjects research. However, this kind of waiver is not always fitting and can sometimes be difficult for researchers to obtain (Littenberg and MacLean, 2006), although appropriate requests are rarely denied. 7With medical studies, patients are sometimes encouraged to take the document home and give consent only after “sleeping” on it. 8See 45 CFR 46, 101 for exemptions.

OCR for page 55
 INFORMED CONSENT The Role of IRBs in Reviewing Informed Consent Procedures There is growing concern about the ability of IRBs to be effective in their role of reviewing the informed consent process. Over the past decade, a number of calls have been made to examine closely how IRBs carry out this role. One editorial in a medical journal commented, for example, that the “tissue issue” has become increasingly vexatious for IRBs, which struggle with the question of how to phrase informed consent forms to explain the genetic studies that might be performed on the tissue in the future (Levine, 2001). Furthermore, IRBs are becoming increasingly overburdened, with det- rimental effects on their productivity and usefulness. In addition, persisting confusion about the federal guidelines that regulate human subjects research and the sometimes conservative interpretations of these guidelines by IRBs have raised concern in the research community as it strives for excellence that is both scientific and ethical (Levine, 2006; Fost and Levine, 2007). The Role of IRBs in Biorepositories and Biobanks The role IRBs should play in research involving biospecimens and bio- data (especially genetic data) from a repository or biobank is unclear, and this subject is not considered in the federal regulations. The current IRB structure was designed primarily for direct human research, and it works especially well in such cases as clinical drug trials because the IRB paradigm assumes that researchers will be working with live human subjects to whom they disclose the probability and magnitude of potential harms that may be associated with the research (Meslin and Quaid, 2004). The current IRB structure is not as well suited to research involving biorepositories or biobanks in which the tissues for the studies are not obtained directly from subjects; the principal harms are nonphysical; the studies may involve someone other than the person from whom the sample was obtained, such as a family member; and the studies do not require the person to be present or even alive to participate (Meslin and Quaid, 2004). This is one good reason why biosocial research should not be delegated completely to medical IRBs. The problem is especially difficult for social surveys that collect and store biospecimens and biodata; there are no clear and immediate benefits of par- ticipation for those who donate the specimens, yet risks are still present. The issues that arise from the storage of specimens and data from social science research were not anticipated when the federal regulations concerning IRBs were developed, and it may be necessary at some point to revisit these regula - tions and update them for use with biorepositories and biobanks. Such issues have long been a concern of social scientists (and their IRBs) because they have a history of sharing data, and for the most part no such tradition exists in biomedical research.

OCR for page 55
 CONDUCTING BIOSOCIAL SURVEYS Issues Related to the Involvement of Multiple IRBs When researchers from different institutions are involved in a single proj - ect or when researchers from one institution share biospecimens or biodata with researchers from another, multiple IRBs are likely to become involved. This involvement of multiple IRBs can cause a variety of problems that have no easy solutions (Burman et al., 2001). As Meslin (2006, p. 279) notes: “All researchers have experienced the frustration of submitting a protocol for a multi-center study only to have one IRB approve it, another decline to approve it, and a third require extensive modifications that the other two did not.” This is a particularly difficult problem for multicenter biomedical research, and in November 2005 a national conference was held to discuss solutions. The conference was sponsored by the National Institutes of Health, OHRP, the Association of American Medical Colleges, and the American Society of Clini - cal Oncology. The workshop report details a number of the issues involved and some approaches to addressing them (National Institutes of Health, Office for Human Research Protections, Association of American Medical Colleges, and American Society of Clinical Oncology, 2005). The issues to which Meslin refers are perhaps the most obvious ones that stem from the presence of multiple IRBs: How does an investigator proceed if one IRB approves the research plan and a second does not, or if one requires modifications that another does not? And what happens if two IRBs impose contradictory sets of conditions that both cannot be met? There are also con- cerns related to the IRBs’ continuing oversight of a project, as disagreements between them can be even more problematic once a project is under way, par- ticipants have been recruited, and specimens and data are being analyzed. As biosocial surveys and studies become more common, it is likely that the number of projects requiring multiple IRBs will increase. To be sure, this issue is not limited to biosocial surveys, but it will become increasingly important to find new ways to provide the oversight that IRBs offer without the complica - tions entailed in having to answer to independent and uncoordinated entities at multiple institutions. Two possible approaches are to develop procedures for centralized review or to find ways for IRBs at different institutions to cooperate in their reviews of multi-institution research (Wolf, Croughan, and Lo, 2002). CONCLuSION Social scientists planning to add the collection of biospecimens to their surveys are likely to find that the informed consent process is quite different from their past practice. The best way for them to deal with these new chal - lenges is to consult with investigators at their institutions or others with prior experience with biosocial or biomedical research, who can provide practical information about developing an informed consent document and about what to expect from their IRBs.