TO: John Callahan, Assistant Secretary for Management and Budget

FROM: Director, NIH

SUBJECT: Response to OMB's Call for Comments on the NPRM to Modify Circular A-110

I am responding to the Office of Management and Budget's (OMB) call for comments on the Notice of Proposed Rulemaking (NPRM) to modify circular A-110 extending the Freedom of Information Act (FOIA) to data produced under Federal grants. I appreciate the effort of the OMB in limiting such access to published data used in Federal regulations or policy, but many questions remain. A desire for access to research data must go hand in hand with protections for the research subjects' privacy and the research investment of investigators in research, and take into account the burdens on agencies, institutions, and investigators that such access might create. Openness in science is very important, and the National Institutes of Health (NIH) already endorses many strategies to make scientific data and research tools accessible. However, we know that the requirements in different scientific fields vary and that care must be taken in crafting such strategies.

I would like to call to your attention several issues. Some of these could be clarified in the rulemaking process, but as we reviewed the proposed rule we came to the conclusion that the FOIA is fundamentally unsuitable as a vehicle for providing access to scientific research data. Nevertheless, I felt it was important to provide a response that addresses specific aspects of the proposed implementation of the existing law. This analysis is based on the OMB's narrowing of the law to include only published data used in Federal policy-making or regulation; it is, therefore, important that this narrowing be sustained. If there were any indication that this would not be sustained, we believe the entire rule should be reopened. The issues identified below are discussed in greater detail in the attached review:

• Scientific data are quite varied and the NPRM offers no definition of data.
• Publications may take many forms, and it is important that early sharing of research findings not trigger release of the underlying data. The scope of "publication" must be clarified.
• The NPRM limits release to data used in Federal regulations or policy making, but it is important to explain what constitutes "policy" for the purposes of this rule.
• Data may be used in the regulation or policy process in different ways; data release should only be triggered when there is an explicit link between the data and the regulation.
• The procedures of the FOIA require that the agency obtain information before redaction to protect private information. The agency's retention of such data constitutes a risk to our research participants and would hurt our ability to carry out our primary mission.<
• Application of the FOIA to our grantees would place undue administrative burdens on the agency, the grantees, and the investigators; procedures to compensate grantees and agencies are unclear and incomplete.
• The application of this rule to data, regardless of the level of Federal funding, will impede our ability to work with private or other governmental partners who find such data access unacceptable. This rule as proposed would thus hinder our ability to support the best research and develop public-private partnerships.
• Complex data sets and sophisticated methodological tools have increased the risk of identification of research subjects. Therefore, far more than the names of subjects must be redacted to protect subjects' privacy. It is not clear that FOIA can offer such protections.
• FOIA clearly protects proprietary data, but this may not be sufficient since the need to provide data to the agency for determination may constitute disclosure, especially in terms of foreign patents.

At the NIH, we require that coordinate data developed by x-ray crystallographers be made available at the time of publication in a peer reviewed journal. Many demographic studies have moved large, valuable data sets into public repositories. Data archives have been supported to capture data on specific topics and make it available. I have taken a stand regarding the patenting of early DNA sequencing data to ensure that such data are easily available to the scientific community. Because these strategies have been constructed to meet the specific demands of the science involved, they have not caused undue administrative burden. The application of the FOIA to data produced under grants would place significant administrative burdens on us and our grantees.

We at the NIH understand the value of broad access to research data, but we also understand how complicated this can be. The complexity stems from such issues as the variability of types of data, the need to truly protect our research subjects' privacy, and the importance of protecting our scientists' ability to build partnerships and often to take their findings on to marketable products. It is not clear from our review that this can be accomplished under the proposed rule. The FOIA was designed for other purposes and is unsuitable as a vehicle for making research data accessible.

The scientific community has been slow to recognize the significance of the change occasioned by this new law. I believe it would be premature to rush to a final regulation when so many questions remain. I request that the comment period be extended and the time used to have further discussions with the scientific community and the agencies that would be affected by this law.

Harold Varmus, M.D.

Attachment

NIH:OD:OER:WBaldwin:4/1/99
496-1096