The trusted source for
healthcare information and
Sharing data requires policies and procedures
Here’s what IRBs need to know
The longtime research tradition of sharing data has been challenged in recent years with the implementation of the Health Insurance Portability and Accountability Act (HIPAA) and its privacy requirements. However, the National Institutes of Health (NIH) has confirmed its commitment to the tradition with its mandate requiring certain grant proposals to describe how data will be prepared for public use.
NIH issued a mandate, effective Oct. 1, 2003, requiring research grant proposals of more than $500,000 to describe how data will be shared for public and other research use without breaching confidentiality, or researchers can explain why this would not be possible.
IRBs may consider updating their policies and procedures regarding data sharing in order to ensure that researchers follow the mandate.
"The researcher needs to have a plan in the proposal, and the IRB has a major responsibility in making sure the plan is adequate," says Joan Sieber, PhD, professor of psychology emerita at California State Hayward. She spoke about confidentiality and data sharing at the PRIM&R 2003 annual IRB conference.
In updating policies and procedures regarding data sharing, IRBs may desire to consider these questions and issues:
1. How should informed consent be adjusted to reflect the possibility of data sharing?
IRBs may need to consider requiring investigators to obtain consent for the sharing of a subject’s data for future studies, Sieber says.
For example, suppose research involves a videotape of a family, and the standard informed consent has been obtained before the family was videotaped and involved in the study. Then it might be a good idea for the researcher to return to the family after the videotape is made and ask them for informed consent for archiving the tape in a place where it might be shared with other qualified scientists, she explains.
"That way, the person knows what it is you are going to share," Sieber says. "If you put that information in the informed consent before the data are gathered, then the information wouldn’t be as useful to the subject."
It might be useful to change the name of the second informed consent to something that more closely reflects its purpose, such as "permission to release data," she adds.
"The agreement might go like this: Your data will be archived in a secure place and would be shared only with a qualified scientist who submits a proposal — for research use of the data — that is approved by the archive," Sieber explains. "The protocol would comply with the confidentiality agreement initially set forth, and the borrower and his or her institution agree to comply with the confidentiality agreement."
2. How should researchers reduce risk of the data becoming identifiable?
An IRB will need to know the sensitivity of the data, because if the data are not sensitive, then the risks and IRB’s concerns are fewer.
"But when you gather data about things that would stigmatize or incriminate someone or make them financially vulnerable, such as in health data or genetic data, then I think great care needs to be taken in the storage of those data, and it has to be considered on a case-by-case basis," Sieber says.
Preventing re-identification of specific subjects is a new process, involving new technology, so it might be beyond the expertise of some IRBs, although investigators also will need to become educated in this area, she notes.
"Of course, not every institution is going to have a lot of researchers who will share data, but certainly major institutions need to take steps to support the IRB by providing statistical consulting expertise," Sieber says. "These statisticians could help advise investigators to come up with appropriate plans that could be set forth in the proposal."
3. Defining data.
IRBs that address policies regarding data sharing might use the guidelines and definitions, including the definition of data, provided by NIH at this web site: http://grants2.nih.gov/grants/policy/data_sharing/data_sharing_guidance.htm.
"IRBs should make sure they understand that information and go by the NIH definitions," Sieber advises.
For example, NIH defines Final Research Data this way:
"Recorded factual material commonly accepted in the scientific community as necessary to document and support research findings. This does not mean summary statistics or tables; rather, it means the data on which summary statistics and tables are based. For the purposes of this policy, final research data do not include laboratory notebooks, partial data sets, preliminary analyses, drafts of scientific papers, plans for future research, peer review reports, communications with colleagues, or physical objects, such as gels or laboratory specimens. NIH has separate guidance on the sharing of research resources, which can be found at http://grants.nih.gov/grants/policy/nihgps_2003/NIHGPS_Part7.htm#_Toc54600131."
4. What additional resources are needed?
Besides having access to a statistician or similar expert, an IRB and institution may desire to consider the costs of sharing data and how these will be distributed.
"The marginal costs of sharing with a particular person can be shared, but the cost of preparing data for sharing can become part of a research proposal," Sieber says. "The cost of preparing data for sharing are primarily the cost of explaining the data so well that someone who is not on the initial research team truly understands what each variable means, how the information was gathered, and what limitations there are to interpreting it."
This information requires a great deal of documentation, she adds. "The cost of preparing data for sharing is considerable, and many little studies that are not carefully scrutinized by a funding agency’s review panel probably aren’t worth sharing," Sieber says. "On the other hand, many journals and scientific societies tell people that if they want to publish their research, they need to hold on to the data for at least five years, so that if anyone questions the data they may reanalyze the data."