By e.stephenson | March 25, 2013
This is from the COSSA Newsletter ([Consortium of Social Science Associations).]{.st} March 25, 2013 Volume 32, Issue 6 Regarding a workshop on proposed changes to the Common Rule. Readers of these blog entries will recall that these proposed changes would require that data identitified in social science research would be required to meet HIPPA standards; potentially rendering many public datasets unuseful for research purposes.
A link to the webcast is here: http://sites.nationalacademies.org/DBASSE/BBCSS/CurrentProjects/DBASSE_080452#Workshop George Alter spoke on the panel on Data Security and Sharing.
Here is a summary of the COSSA report:
On March 21 and 22, the National Academies’ Board on Behavioral, Cognitive, and Sensory Sciences (BBCSS) held a workshop on the “Proposed Revisions to the Common Rule in Relation to the Behavioral and Social Sciences.” In 2011, the Department of Health and Human Services proposed changes to the Common Rule, the regulations governing the protection of human subjects in research, in an Advanced Notice of Proposed Rulemaking (ANPRM). (For more information, see Update, January 28, 2013 and click here for a response to the ANPRM from the social and behavioral science community.) Several COSSA member organizations helped sponsor the workshop. More information about the workshop, including presenters' slides and an archived webcast, is available here. BBCSS will publish a summary report of the workshop. According to Robert Hauser, Executive Director of the Division of Behavioral and Social Sciences and Education (DBASSE), the Academies expect to convene a panel a panel that will produce a consensus report with conclusions and recommendations.
The workshop’s opening session reviewed existing knowledge and evidence about the functioning of the Common Rule and Institutional Review Boards (IRBs). Connie Citro, Director of the Committee on National Statistics at the National Academies, gave an overview of the many National Academies’ reports on human subjects protection published since 1979 and summarized the lessons learned. She pointed to four major takeaways from the existing literature. First, one-size-fits-all approaches often have unanticipated negative consequences. Second, there is no need to reinvent the wheel regarding human subjects’ protection. Third, a balance needs to be struck between leaving subjects vulnerable and handicapping researchers. Finally, the social and behavioral sciences (SBS) are often not given the same consideration as the biomedical sciences in writing regulations and thus need to be constantly vigilant to make sure that new rules are appropriate for a SBS context.
Noting that there is a relatively small evidence base on the efficacy of the Common Rule and IRBs, Jeffrey Rodamar, Department of Education, reviewed some of the existing data. He found that despite popular perception, IRBs function pretty well. They are generally no more of an administrative burden than other grant-related activities; on average, review takes less than three percent of a study’s time; a majority of studies are approved; expedited review takes less than a month on average and full review takes less than two months; and extreme delays are statistically uncommon. Rodamar described data showing that both SBS and biomedical researchers generally approve of the IRB system. He conceded that there are some problems with the Common Rule regulations and IRBs, but, paraphrasing Winston Churchill, suggested that perhaps “IRBs are worst form of governing research except for all those other forms that have been tried from time to time.”
The “Minimal Risk” Standard
The second session, moderated by Celia Fischer, Fordham University, focused on the types of “risks and harms” encountered in SBS research. Richard T. Campbell delved into the concept of “minimal risk,” an important area for researchers dealing with human subjects. The determination of whether participation in a study represents a “minimal risk” dictates the level of IRB review that takes place. Under the Common Rule, a study represents minimal risk if “the probability and magnitude of harm or discomfort 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.” Noting that it is a “cognitively complex” concept, Campbell suggested that risk can be thought of as the relationship between the probability of harm occurring and the severity of potential harm. Thus, the Common Rule provides some flexibility in that it does not dictate that both probability severity must be “minimal,” just that, as probability increases, the severity of possible harm must decrease (and vice versa). Given that other parts of the definition are also thorny (such as what is meant by “daily life”), Campbell suggested that the Office of Human Research Protection (OHRP) could provide guidance to facilitate more consistent application of the minimal risk benchmark.
Brian Mustanski, Northwestern University Feinberg School of Medicine, spoke about his research on risky and sensitive behavior (such as drug use, sexual behavior, and HIV) in youth, which are topics that often make IRBs skittish. He conducted a study that was reviewed by two IRBs. One board approved it immediately, while the other delayed the study for six months because it was felt to be a “slight increase” over minimal risk. However, when Mustanski surveyed his subjects, a large majority felt that their participation was less uncomfortable than a routine medical exam (the minimal risk standard). Mustanski argued that such institutional reluctance to approve research into controversial or sensitive subjects as minimal risk can have a chilling effect, leading to a poor evidence base for interventions with already underserved populations, which is indeed the case regarding HIV prevention in LGBT youth.
Steve Breckler, American Psychological Association and COSSA Board Member, discussed the concept of risk in the SBS context. He reminded the audience that the broad goal of assessing risk is to calibrate the level of review to the level of risk a study poses to participants, in other words, to protect subjects and reduce unnecessary regulatory burden. He argued that the social science community should put greater focus on producing evidence to determine how well regulations are working and that having better guidance and tools for assessing risk would facilitate the work of IRBs.
Charles Plott, California Institute of Technology and a former COSSA Board member, posed the question of whether the entirety of the research endeavors for some fields, like economics, political science, game theory and decision science, could be said to be wholly without risk. In a survey of economics, political science, and judgment and decision making associations, Plott found very low numbers of adverse incidents and reports of harm, all of which were low-magnitude events (such as feelings of stress or frustration). He argued that some research topics– markets, committees and voting, games, processes, and decisions– and some research methods– questionnaires, computer games, etc.– can be said to pose no potential harm to subjects and should thus be exempted from consideration under the Common Rule.
Informed Consent and Special Populations
A session on the consent process and special populations was moderated by Margaret Foster Riley, University of Virginia. Sally Powers, University of Massachusetts, Amherst, discussed how consent operates in her research on depression, which collects “rich” behavioral and biospecimen data (which can be recoded and analyzed as part of future analysis). The proposed changes to the Common Rule would require that prior consent is obtained for re-analysis of biospecimens, but that consent should be given for open-ended use of specimens. However, the changes do not address rich behavioral data; Powers argued that the same standards should be applied.
Roxane Cohen Silver explained how she conducts research on victims of disasters and traumatic experiences (like natural disasters, infant death, and mass shootings) shortly after such events occur. Silver argued that such research can be conducted ethically and sensitively, if participants are given multiple opportunities to opt out, are allowed to refuse to answer questions and researchers and staff are well-trained. Noting that this type of research is most valuable if it is commenced immediately after a traumatic event, Silver described her arrangement with her IRB, which pre-approved a generic post-disaster proposal. In the aftermath of a traumatic event, Silver provides the IRB with specific information and can get full approval within 48 hours.
Celia Fischer, Fordham University, spoke about some of the issues involving obtaining informed consent from children. She argued that simplifying consent forms, as proposed by the ANPRM, would be useful. However, relying on standardized forms can be problematic for certain types of research and subjects of different ages, language skills, and educational backgrounds. Fischer observed that verbal consent can be a better form in certain contexts. She also noted that emancipated minors are often not treated as full adults by IRBs, despite being adults under the law. Fischer pointed out the issue of re-obtaining consent from adults, for whom parental consent had been granted when they were minors.
Data Security and Sharing
David Weir, University of Michigan Survey Research Center, moderated a panel on “Data Use and Sharing and Technological Advancement.” The proposals in the ANPRM would mandate that all studies that collect identifiable or potentially identifiable data to have data security plans. George Alter, University of Michigan Interuniversity Consortium for Political and Social Research (ICPSR), which archives and protects social science data, spoke about some of the ways data can be kept secure. Informational risk can be reduced by improving study design (implementing certain sampling procedures, using multiple sites), having protection plans in place, using data repositories and archives, and training. ICPSR restricts data based on the degree of risk of disclosure and severity of harm from that disclosure, from publically releasing data online to requiring researchers work with data in physical data enclaves.
Taylor Martin, University of Utah, spoke about the data security implications of her research into math learning, which collects rich data from children playing online educational games. This type of research shows promise in terms of providing new information about how different kinds of children learn and how we can teach them better. However, concerns about data security can have a chilling effect on data sharing and reuse among researchers. Martin observed that for-profit companies are collecting data and doing the same kind of research without having to go through the same hurdles as researchers.
Susan Bouregy, Yale University Human Research Protection Program, raised concerns about the ANPRM’s proposal to apply HIPAA standards for deidentification of data (requiring removal of 18 specific identifiers). Bouregy noted such standards may make some data sets unusable while ignoring other ways individuals could be identified. She also argued that some of the mandated HIPAA security elements are not appropriate for certain types of social science research. Furthermore, it ignored that not all identified data is risky. Finally, Boregy suggested that the ANPRM’s requirement that all suspected data breaches be reported should be made more flexible and allow IRBs to tailor reporting to the context of each situation.
Multi-Disciplinary and Multi-Site Studies
Robert Levine, Yale, University, moderated a session focused on multi-disciplinary and multi-site studies. Pearl O’Rourke, Partners Health Care System, discussed the requirement that multi-site studies use a single IRB of record. She noted that having a central IRB does not absolve the individual institutions of fulfilling a number of responsibilities in overseeing and approving research. O’Rourke was concerned that mandating a central IRB would not address the complexity of each situation. Furthermore, the requirement underestimates the costs and time involved in running a central IRB.
Laura Stark, Vanderbilt University Center for Medicine Health and Society, gave an ethnographic perspective on IRB decision-making. As an explanation for why IRBs reach different conclusions regarding the risk level of similar research, Stark suggested the concept of “local precedents,” or allowing past decisions to govern the evaluation of subsequent research. Such precedents may lead to faster decisions and internal consistency, but they can be problematic for researchers working with multiple IRBs. Stark offered three strategies to work around local precedents: 1) study networks (having a central IRB for multiple sites), 2) collegial review (allowing departmental experts to review research), and 3) decision repositories (online archives of approved protocols from many IRBs).
Thomas Coates, University of California, Los Angeles Program in Global Health, shared his experience with multinational studies (which are not addressed by the ANPRM). Some concerns he encountered included whether requiring other countries to adhere to U.S. requirements could be considered paternalistic, how to evaluate minimum risk in different cultural and economic contexts, and how to harmonize U.S., international, and local regulations. Coates also stressed the importance of receiving approval from local bodies in addition to U.S.-based IRBs.
The Scope of Institutional Review Boards
A final session, moderated by Yonette Thomas, Howard University and a COSSA board member, focused on the “Purview and Roles of IRBs.” Lois Brako, University of Michigan, discussed the ANPRM’s proposed changes from the perspective of an IRB that has made strides to become more innovative and flexible. Brako praised the ANPRM’s proposals to reduce the oversight burden for minimal risk studies, eliminate annual review, and harmonize federal regulations (so long as the harmonization does not take the form of a unilateral one-size-fits-all approach). However, she argued that some of the proposals are unnecessarily burdensome, including requiring all institutions that receive Common Rule funding to be subject to federal oversight, some of the information security provisions, requiring reports of all adverse events to be submitted and stored in a central database, and expanding “human subjects” to include deidentified biospeicimens. Brako also suggested that in some cases, clearer guidance from OHRP would be more helpful than changed regulations.
Rena Lederman, Princeton University, observed that the Common Rule regulations were written from a biomedical perspective and are particularly unsuited for certain types of SBS research, such as anthropological fieldwork. Anthropologists establish thick relationships with their subjects, immerse themselves in other cultures, and do not test hypotheses or run controlled experiments. The ANPRM’s requirements for informational security could cripple anthropological research (anthropologists’ detailed fields notes would treated as data with informational risks under the new rules, raising the question of how such notes could be deidentified). Rather than trying to adapt the Common Rule to fit SBS research, Lederman proposed that it be only applied to biomedical research. She proposed the creation of a National Commission to develop an alternative guidance and framework to address SBS research risks.
Cheryl Crawford Watson, National Institute of Justice (NIJ), discussed the Department of Justice’s (DOJ) approach to confidentiality and how it differs from other regulations regarding human subjects protection. Researchers funded by DOJ must submit a Privacy Certificate, which protects researchers and data from subpoena. It also prevents the researcher from violating subjects’ privacy for any reason other than future criminal conduct. The DOJ privacy certificate differs from the certificate of confidentiality mandated by other agencies (like Health and Human Services) in that it prohibits researchers from reporting child abuse, reportable communicable diseases, and threatened harm to self or others. In order to be allowed to report such abuse, researchers must get the subjects to sign a separate consent-to-report form. The certificate is so strict due to concerns that few of the subjects under DOJ’s purview would consent to participate in research otherwise.