This is an archive page. The links are no longer being updated.

Statement on H.R. 88: Research Data Available under the Freedom of Information Act by Harold Varmus, M. D.
Director, National Institutes of Health
U.S. Department of Health and Human Services

Chairman Horn, Mr. Turner and other members of the Subcommittee on Government Management, Information and Technology. I am Harold Varmus, the Director of the National Institutes of Health. I am pleased to be here today to testify about H.R. 88, a bill authored by your colleague George Brown, the Ranking Member of the Science Committee, and a long-time advocate of openness in scientific research.

I too am an advocate for openness in scientific research. Exchanging ideas and sharing data are absolutely vital to the success of biomedical research. They are the hallmarks of the success of the NIH research programs. The true understanding of the breakthroughs we are making in genetics research, medical imaging, clinical research, and all other scientific investigation funded by NIH could not be accomplished without public access to methods and data.

Openness has many virtues. It fosters trust in scientific outcomes as well as trust in the use of federal dollars to conduct biomedical research. It engenders faith that human subjects and animals participating in research trials are adequately protected. And it sparks technological innovations that help us find the answers to public health problems. However, we know that the requirements in different scientific fields vary and that great care must be taken in crafting such strategies. There are many examples of responsible data-sharing at NIH. The remarkable strides we are making in our understanding of the genetic components of various diseases would not be possible without the sharing of data. Also, NIH's firm stance regarding the patenting of early DNA sequencing data has ensured that such data are easily available to the scientific community.

And we require that coordinate data developed by x-ray crystallographers be made available at the time of publication in a peer reviewed journal. In these cases, and many others, we promote openness.

But a word of caution. I think it would be a mistake to open all underlying scientific data to public scrutiny simply because of the concept that all openness is good. There are pitfalls in unrestrained openness, including unwarranted violations of privacy, the potential harassment of scientific investigators and the chilling effect that inappropriate public scrutiny could have on the free exchange of ideas and the willingness to take risks to find answers.

The amendment to OMB circular A-110 contained in Public Law 105-277 could have unintended, but nonetheless grave consequences. The regulatory requirements embodied in the law are far more complex than the apparently simple mandate to share data. The kind of widespread access to data envisioned by the Act could result in unforeseen abuses. Unless otherwise protected, patient privacy rights could be violated. The willingness of scientists to speak openly about new ideas and take experimental risks could be fettered by unrestricted data access. Further, I am concerned that inappropriate sharing of preliminary data could lead to misinterpretation of results. Finally, the new requirements could undercut the ability of researchers to build private sector partnerships that now lead to the eventual marketing of products.

It is because of these concerns that we at NIH have taken the position that, while expanded access to scientific data should be encouraged, the A-110 amendment may be a poor vehicle to achieve this goal. I am particularly concerned about the legislation's requirement that the Freedom of Information Act (FOIA) be the tool of regulatory implementation. FOIA was not designed to accommodate the confidentiality requirements of the most sensitive scientific data. Consequently, without additional protections, it would be possible for the privacy of patients to be compromised or individual scientists to be harassed by selected interests opposed to their work due to moral or financial concerns. Such intrusions could stop promising scientific research in its tracks, and the mere threat of such intrusions could impede the Nation's efforts to recruit its most talented students into publicly-supported research.

For example, imagine what would happen if HIV-infected patients thought their condition might be revealed by someone using the new requirements to examine raw experimental data. Patients would not participate in clinical trials if they believed there was an opportunity for their infected status to be revealed. Progress toward treatment of the disease would be stymied.

Under this new application of FOIA embodied in the A-110 amendment, a request for research data would be directed to federal agencies, such as NIH. We would be required to forward the request to grantee institutions and to the scientists with direct responsibility for the data. Their data often contains information about individuals who entered into the research project under a promise of confidentiality. This information would be forwarded from grantee institutions to federal agencies, who would be responsible for determining what to release and what to exempt.

It may sound simple to make that determination, but it is not. FOIA would allow the government agency to remove obvious identifiers such as name, Social Security number, telephone number, but in a given data set it is quite feasible to identify subjects using other information. If the requestor knew a few items about an individual's history, such as place of birth, education, occupation, marital history, or other general information, an individual could be identified. Such identification would then open up the whole research record, including personal medical information to the requestor.

While there are many aspects of the A-110 amendment that trouble us, there is a particular provision that presents a new challenge to those who would make their data accessible. I am thinking about the multiplicity of partnerships between public sector researchers and private companies, non-profit organizations, and even foreign governments. Some of these partnerships make strict requirements on the researcher not to share data further. Without such agreements, private researchers would not participate in these partnerships.

In regard to proprietary data, the Bayh-Dole Act specifically provides protections for the intellectual property of individual researchers. But the A-110 amendment threatens these protections for our partners. Industry scientists may avoid collaborations with publicly-funded institutions, including universities, if they believe they can no longer protect their data from exposure.

The A-110 amendment will lead to increases in administrative burdens and cost for granting agencies, such as NIH, but also for grantees. Universities and other institutions that receive Federal grants will need to create formalized procedures to respond to FOIA requests. Increased administrative costs are not in themselves a reason not to move forward with policies in the public interest, but we would like to ensure that the benefits are commensurate with the costs. Increased administrative costs will come at the expense of research, in both dollars and in investigator time. The paperwork that requests will generate is enormous and counter to the reduction in paperwork efforts. Entire staffs will have to be recruited to make the decisions that will have to be made, with consequent increases in costs of conducting research.

In addition to the administrative burdens, the A-110 amendment is likely to lead to the filing of lawsuits by individuals or organizations whose requests for data are rejected, which would be costly to the government and private institutions that perform publicly-funded research.