PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
IRB. Author manuscript; available in PMC 2012 July 11.
Published in final edited form as:
IRB. 2011 May-Jun; 33(3): 1–5.
PMCID: PMC3394177
NIHMSID: NIHMS378997

Broad Data Sharing in Genetic Research: Views of Institutional Review Board Professionals

Institutional review boards (IRBs) play a key role in ensuring that genetic research with humans meets ethical and regulatory standards. They are involved with issues such as how research participants will be informed about researchers' intent to share genomic data broadly, what genetic information will be included in databases, whether the data are adequately deidentified, and whether consent documents adequately explain data sharing.1 Moreover, IRBs must ensure that investigators conducting genome-wide association studies (GWAS) comply with the National Institutes of Health's (NIH's) policy to share deidentified GWAS data with the wider scientific community.2 GWAS examine associations between genetic variants (genotypes) throughout the human genome and observable traits, such as height, cholesterol levels, or disease (phenotypes).

The GWAS data repository is dbGAP (the database of Genotypes and Phenotypes), which the National Center for Biotechnology created as a centralized, controlled-access database.3 IRBs play a role in implementing the GWAS data-sharing policy, and the NIH has provided specific guidance about this role.4 When investigators plan to submit GWAS data to dbGAP, IRBs are expected to ensure that 1) uses of the research data are consistent with the terms of participants' informed consent; 2) deidentification plans are consistent with the NIH data-sharing policy; 3) potential risks to individuals, families, and groups are commensurate with benefits; and 4) data submission and collection are conducted in a manner consistent with federal regulations.5

Given the central role of IRBs in facilitating data sharing in genetic research and the important ethical and regulatory issues data sharing raises, we conducted a survey to investigate the views, experiences, and attitudes of IRB professionals regarding broad data sharing in genetic research.

Study Methods

Study Population and Recruitment

The study population was selected in consultation with leaders from Public Responsibility in Medicine and Research (PRIM&R),6 a nonprofit organization that promotes the ethical conduct of research through education and outreach to individuals who oversee human subjects protections. Eligible members were defined as those whose PRIM&R registration information indicated either a primary interest in human subjects research or general interest in all research, as well as those who did not report a particular research interest. In April 2009, an introductory e-mail was sent inviting eligible members (n = 2,777) to take the survey. This was followed by a reminder e-mail two weeks later. The survey remained open for one month. Both e-mails noted that personally identifying data would not be collected, thus making the survey anonymous. A parallel survey was sent to genetic researchers, and findings are published elsewhere;7 a comparative analysis of the results of the parallel studies is underway.

Both recruitment e-mails asked recipients to forward the survey invitation to interested colleagues who may not be PRIM&R members. An open invitation to participate in the survey was published in the June 2009 edition of the Northwest Association for Biomedical Research's Fresh Sheet, a quarterly e-newsletter that is sent to IRB members and staff in Washington, Oregon, and Idaho. These additional recruitment methods were intended to increase overall survey participation; however, these strategies—in combination with the anonymous nature of the survey, which makes duplicate responses and multiple invitations undetectable—make it impossible to calculate a true response rate.

Survey Instrument

A 90-item, Web-based survey was developed that included questions about the application and review process for research involving human participants at respondents' institutions, committee functions, design-specific issues in genetic research, and demographic information. The instrument used a variety of question types (multiple-choice, Likert rating scales, and a single, open-comment question). The survey instrument has been published elsewhere.8

Survey design was informed by findings from 31 semistructured, individual interviews with IRB members and staff at several institutions, as well as informal input from representatives of key stakeholder organizations and institutions. Pretesting included internal and external expert review, as well as cognitive interviews (n = 8) with members of the target population.9 Piloting the tool on-line indicated that it took approximately 15–20 minutes to complete. Once the instrument was finalized, it was uploaded into a University of Washington Web tool, Catalyst WebQ, for survey administration. Several members of the research team conducted pilot testing to identify any potential technical difficulties with the survey software.

Data Analysis

Survey data were downloaded and converted into a Stata (version 10.0) dataset for analysis.10 Descriptive summary statistics were used to summarize responses. To facilitate analysis and interpretation, five-point Likert scales were collapsed to three categories, combining the “strongly” and “somewhat” categories at either end of the scale (e.g., collapsing “strongly agree” and “somewhat agree” to a single “agree” category). A neutral response category was also included in the scale, and respondents were provided with a “don't know” option. The survey allowed respondents to skip individual items; the sample size therefore varies by question.

Survey Results

Atotal of 208 participants completed online surveys. No minimum threshold for survey completion was imposed, and the sample size varied by question. The range of missing data for the survey items was 0–10.6%; on average, the 90 items had missing data for 2.4% (n = ~5) of the participants. Self-reported participant characteristics indicate that almost all respondents were PRIM&R members based in the United States (97.1% and 98.5%, respectively). Approximately 77% of respondents were female. Most (59.6%) had served on an IRB for more than five years, and the majority (69.1%) reported that their primary IRB affiliation is with a biomedical committee or panel (rather than a social, behavioral, or educational review body). Table 1 provides additional information.

Table 1
Participant Background Characteristics (N = 208)

Participants were asked a number of questions regarding guidance provided during the IRB application process. In one question, participants were asked to indicate the importance of IRB guidance for investigators on specific aspects of genetic studies as compared with other kinds of research studies. The majority (77.2%) indicated that such guidance is very or somewhat important for developing a data repository or biobank that includes genetic data. Approximately 70% said guidance is very important or somewhat important for researchers who plan to use large-scale data repositories containing genetic information, and 82.6% indicated that it is very or somewhat important for those who wish to share genetic research data with other investigators. In response to a related question, the majority also indicated that different guidance is needed for developing these three elements in genetic research studies, compared to other types of research (Table 2).

Table 2
Need for Different Guidance for Genetic Research Studies Compared to Other Research Studies (N = 208)

Another survey question sought to identify specific issues that prompt multiple rounds of communication and negotiation between IRBs and researchers. Respondents were given a list of possible issues requiring “considerable discussion” with researchers (defined as more than two or three exchanges and/or more than a one-hour conversation). Approximately 58% indicated that procedures for protecting participants' personal information or samples fit this description, and 45.6% reported that the process and documentation of informed consent required extra consultation with researchers.

To learn more about IRB professionals' perceptions of the NIH GWAS data-sharing policy, the survey asked participants to indicate whether the NIH guidelines for data sharing are clear. The survey did not provide details about the substance of the data-sharing policy; it simply noted that such a policy had been published. Approximately one-third (36.3%) of participants agreed that the NIH guidelines for sharing of data from GWAS are clear, while 16.7% disagreed, and an equal number neither agreed nor disagreed. The remaining 32.4% responded that they did not know. Participants were not asked whether they were aware of the NIH GWAS data-sharing policy, which may account for some of the “don't know” responses.

One of the risks associated with broad sharing of deidentified genetic research data is the potential for individual participants' identities to become known, either by accident or by design. Half of survey respondents considered it somewhat or very unlikely that such reidentification would occur, while slightly over one-third believed it likely to occur. If research participants were to be identified individually, 34.8% of respondents thought that harm would somewhat or very likely result, while 45.1% considered this unlikely. Responses were evenly split (35.8% each) regarding the likelihood that a federal agency or other law-enforcement agency might compel investigators to disclose information to participants in genetic studies.

Discussion

Most experts agree that the potential for breach of privacy—and the possible harm that could result from unauthorized release of personal information—is the primary risk to participants in high-throughput genomic research. Each individual's genome is unique and, as recent research demonstrates, even very sophisticated technical approaches to deidentification offer incomplete protection.11 Researchers may request a federally issued Certificate of Confidentiality to protect research data from compelled disclosure, and state laws and the Genetic Information Nondiscrimination Act (GINA) prohibit certain kinds of discrimination based on genetic data, but gaps in protection remain, and these measures are largely untested in the courts.12

At the same time, driven by the need to increase research efficiencies and hasten the translation of basic science findings into clinically beneficial applications, NIH policy has prioritized data sharing within the research community. Researchers, motivated by funders' directives and their own scientific goals, wish to share data; institutions desire to participate in consortia without exposing themselves to undue liability; and funders want to demonstrate a return on the investment of public resources in genetic investigations. With so many powerful forces moving toward wide data sharing, the role of IRBs is vital in ensuring that appropriate participant protections are in place.

One can argue that clear, well-disseminated regulatory guidance is a prerequisite for IRBs to fulfill this charge. It is, therefore, worth noting that 16.7% of survey respondents disagreed with the statement that the NIH policy on data sharing for GWAS is clear, 32.4% said that they did not know, and 14.7% neither agreed nor disagreed. These results suggest that almost half of those who answered this question were not familiar enough with the GWAS data-sharing policy to hold an opinion about its usefulness. This finding is a matter of some concern, particularly in light of the fact that 67.5% of survey respondents indicated that they held positions in either IRB administration or regulatory/compliance functions.

Most of the survey respondents believed that special guidance should be provided to researchers whose plans include developing a data repository or biobank that would include genetic data, secondary use of deidentified genetic data, or sharing their genetic research data with other investigators. Without clear regulatory guidance on these issues at the federal level, however, individual IRBs across the country are likely duplicating efforts, dedicating a portion of limited resources to solve a shared problem. In addition, the fragmented and incompatible nature of locally developed standards has been identified as a drawback to effective collaboration13—a particular problem for high-throughput genomic research, which often demands sample sizes beyond those any single research team or institution could reasonably hope to accumulate. One possible solution might be for IRB professionals and researchers to develop regional or national policy consortia. Such groups could work together to define guiding principles and best practices, while leaving room for appropriate local adaptation.

One issue that could be valuably addressed by such collaboration is the question of harm in population-based genomic research. Respondents identified privacy protection as an issue requiring considerable interaction with researchers, but half believed that reidentification was unlikely to occur. If individual research participants were to be personally identified, about a third of respondents thought that harm would result. These results suggest that respondents were not distinguishing between the potential harms to genetic research participants and the likelihood that such harms would be realized. In a clinical trial in which participants may be asked to accept the risk of potential physical injury or death, it makes sense to describe severe, foreseeable harms, even if they are considered unlikely. In the case of population-based genetic studies, on the other hand, the risk of reidentification is unknown, and the severity of possible harm to participants (or others, including relatives and community members) is difficult to assess. Given the substantial uncertainties involved regarding the risk of reidentification, there is a need for empirical studies with prospective and current research participants in GWAS14 to determine what laypeople wish to know before deciding whether to allow their deidentified data to be shared within the research community.

Our survey was an initial exploration of the perceptions of IRB professionals with respect to genetic research review. More in-depth assessments are needed to better understand their concerns. We also note that the survey's low response rate (less than 8%, estimated conservatively against the total PRIM&R membership), combined with our inability to compare responders and nonresponders, calls for caution in generalizing from these results. One contributing factor to the low response rate may have been the length of the survey. These limitations notwithstanding, the results do suggest a need for increased education for IRB professionals regarding the NIH policy for sharing GWAS data; collaboration among IRB professionals and researchers to define best practices and perhaps harmonize standards; and further empirical research to determine information needs and preferences of prospective research participants in the context of wide data sharing.

Acknowledgments

This study is supported by grants from the National Human Genome Research Institute: U01 HG004609, P50 HG003374, and P50 HG003390. This investigation involved multiple collaborators including the University of Washington Center for Genomics and Healthcare Equality, the Case Western Reserve University Center for Genetic Research Ethics and Law, Public Responsibility in Medicine and Research, the American Society of Human Genetics, and the Genetic Research Review and Issues Project (GRRIP) investigators. Additional GRRIP project investigators include (in alphabetical order) Joann Boughman, Wylie Burke, Lynn Dressler, Karen Edwards, William Freeman, Nancy Gerson, Eric Juengst, Susan Lewis, Patricia Marshall, P. Pearl O'Rourke, Roselle Ponsaran, Nancy Press, Mary Quinn Griffin, Helene Starks, and Georgia Wiesner. We would also like to thank the two anonymous IRB: Ethics & Human Research reviewers for their helpful comments.

This minimal-risk study was approved by the University of Washington IRB; all other study sites were approved to receive anonymous survey data.

References

1. Lowrance WW, Collins FS. Identifiability in genomic research. Science. 2007;317:600–602. McGuire AL, Gibbs RA. Genetics: No longer de-identified. Science 2006;312:370-371. [PubMed]
2. National Institutes of Health Policy for sharing of data obtained in NIH supported or conducted genome-wide association studies (GWAS) Federal Register. 2007;72(166):49290–49297. http://grants. nih.gov/grants/guide/notice-files/NOTOD-07-088.html.
3. Mailman MD, Feolo M, Jin Y, et al. The NCBI dbGAP database of geno-types and phenotypes. Nature Genetics. 2007;39(10):1181–1186. [PMC free article] [PubMed]
4. National Institutes of Health NIH points to consider for IRBs and institutions in their review of data submission plans for institutional certifications under NIH's policy for sharing of data obtained in NIH supported or conducted genome-wide association studies (GWAS) 2007 http://grants.nih. gov/grants/gwas/gwas_ptc.pdf.
5. deWolf VA, Sieber JE, Steel PM, et al. Part III: Meeting the challenge when data sharing is required. IRB: Ethics & Human Research. 2006;18(2):10–15. [PubMed]
6. Public Responsibility in Medicine and Research (PRIM&R) http://www.primr.org/
7. Edwards KL, Lemke AA, Trinidad SB, et al. GRRIP Consortium Attitudes toward genetic research review: Results from a survey of human genetics researchers. Public Health Genomics. 2011 Apr 11; DOI: 10.1159/000324931), http:// content.karger.com/ProdukteDB/produkte. asp?doi=10.1159/000324931. [PMC free article] [PubMed]
8. Lemke AA, Trinidad SB, Edwards K, et al. Attitudes toward genetic research review: Results from a national survey of professionals involved in human subjects protection. Journal of Empirical Research in Human Research Ethics. 2010;5(1):83–91. [PMC free article] [PubMed]
9. Willis GB. Cognitive Interviewing: A Tool for Improving Questionnaire Design. Sage Publications; Thousand Oaks, CA: 2005.
10. Statacorp . Stata Statistical Software: Release 10. StataCorp LP; College Station, TX: 2005.
11. Malin BA. An evaluation of the current state of genomic data privacy protection technology and a roadmap for the future. Journal of the American Medical Informatics Association. 2005;12(1):28–34. Homer N, Szelinger S, Redman M, et al. Resolving individuals contributing trace amounts of DNA to highly complex mixtures using high-density SNP genotyping microarrays. PLoS Genetics 2008;4(8):e1000167; Lin Z, Altman RB, Owen AB. Confidentiality in genome research. Science 2006;313:441-442. [PMC free article] [PubMed]
12. National Human Genome Research Institute (NHGRI) http://www.genome. gov/12513976; Reitz S, Connecticut woman alleges genetic discrimination at work. ABC News, April 28, 2010, http://abcnews. go.com/Business/wireStory?id=10500701.
13. McWilliams R, Hoover-Fong J, Hamosh A, et al. Problematic variation in local institutional review of a multicenter genetic epidemiology study. JAMA. 2003;290(3):360–366. Green LA, Lowery JC, Kowalski CP, et al. Impact of institutional review board practice variation on observational health services research. Health Services Research 2006;41(1):214-230; Mansbach J, Acholonu U, Clark S, et al. Variation in institutional review board responses to a standard, observational, pediatric research protocol. Academic Emergency Medicine 2007;14(4):377-380. [PubMed]
14. Beskow LM, Friedman JY, Hardy NC, et al. Simplifying informed consent for biorepositories: Stakeholder perspectives. Genetics in Medicine. 2010;12(9):567–572. [PMC free article] [PubMed]