March 31, 1999

FROM: Administrator

SUBJECT: OMB Circular A-110 NPRM, AHCPR Response

The Agency for Health Care Policy and Research, Department of Health and Human Services has prepared the attached comments in response to the Notice of Proposed Rule Making (NPRM) published in the Federal Register on February 4, 1999. This NPRM pertains to a proposed revision to OMB Circular A-110, as required by P.L.105-277, Treasury and General Government Appropriations Act for FY 1999. Our understanding is that, essentially, all data produced under a grant award by a Federal awarding agency must be made available to the public in accordance with the Freedom of Information Act (FOIA). The agency may charge a reasonable fee equal to the incremental cost of obtaining the data, according to P.L. 105-277.

One of the most important disadvantages of the proposed revision is that it will undermine or even jeopardize AHCPR's partnerships (for data, data linkages, or shared funding) with private institutions (such as hospitals, health plans, employers, insurance companies, professional associations and others). These private institutions will be reluctant to collaborate or share their data with AHCPR if they perceive that such a partnership will likely give competitors access to their proprietary tools and the characteristics of their customers, or yield information about how they manage their businesses. The added administrative and financial burdens on AHCPR are additional issues to be considered seriously.

John M. Eisenberg, M.D.

Attachment

The proposed revision to OMB Circular A110 states "[I]n response to a Freedom of Information Act (FOIA) request for data relating to published research findings produced under an award that were used by the Federal Government in developing policy or rules, the Federal awarding agency shall, within a reasonable time, obtain the requested data so that they can be made available to the public through the procedures established under the FOIA." The consequences to AHCPR depend on the interpretation of the Congressional intent behind the quoted language. Under a narrow reading, AHCPR grantees will not be affected unless their research data directly informed Federal policy. However, a broader reading would not only implicate all AHCPR grantees, but would also give rise to serious administrative burden, financial costs, and research complications.

AHCPR would require a significant additional allocation of manpower to comply with this directive, for administrative and statistical/research purposes. Depending on the number and scope of requests, the costs could be substantial for AHCPR, the principal investigator, and the principal investigator's institution.

• AHCPR or the Principal Investigator (PI) would have to eliminate or modify (perhaps statistically) all personal identifying information from the data set to be given to the FOIA requester.

• The PI will likely have to give up unique, one-of-a-kind data elements that would preclude future research on the data, e.g., blood or tissue samples that help determine the causes or outcomes of a particular medical condition.

• The PI's sponsoring institution will likely have to hire FOIA advisors to counsel the researcher about his/her rights to the data, how to respond to an AHCPR request after the PI has promised not to release any data based on sensitive-subject interviews.

• If the PI changes universities, the location of the data will have to be tracked by both institutions.

• If the research is collaborative among many researchers, each one might have to yield the part of a project's data on which they worked. If scattered across the country, or the world, the burden of complying with this change in A-110 would be extremely burdensome.

Financial Burden

• While the agency may impose a burden on the PI, pay him/her, and incur its own expenses of redacting the data, the ability to charge the requestor a reasonable fee may not keep the agency from depleting its own resources without replenishment. When AHCPR receives such funds, it must turn them over to the Treasury, without reimbursement. A greater A-110 burden translates into a greater loss of budget funds.

• AHCPR manages a national panel survey each year costing approximately $30 million, obtains hospital discharge and other identifiable health data from states, and frequently supports other surveys. If such data were made available to grantees of AHCPR, even if identities were hidden and the researcher were bound by contract from trying to identify any individual surveyed, the realization by survey responders that some of their data could be obtained by FOIA requesters would likely reduce the survey response rate. This would increase the cost per usable data element and potentially introduce bias in the data obtained. The change in A-110 intended to dramatically increase project data sharing (which is a desirable goal in itself) will reduce the quality and quantity of data obtained from surveys (which is greatly undesirable).

Partnership

• AHCPR undertakes partnership with many national organizations such as the American Medical Association, American Association of Health Plans, American Hospital Association, and large manufacturing companies. In the instances where proprietary data are involved, if these organizations believe that the changes in A-110 will increase the probability that their data will get out into public hands, they will cease partnership. For them, the risk of having to fight to protect their proprietary data from public requestors is averted if they just say no to the requests for their data from AHCPR and its researchers.

• Even if only a small percentage of a cooperative project were funded by a federal research agency, such as AHCPR, the other partners will likely avoid the potential exposure of their data (even if redacted) by refusing to collaborate with AHCPR. This could lead to fewer research findings and inferior methods of research and would diminish the opportunities to make good research investments.

• Research results frequently depend on originally collected data that are merged with proprietary data (data in which someone other than the researcher has ownership). An example would be a researcher merging managed care enrollment data with data about health services use or a researcher merging employment records with health claims data. The owners of the proprietary data will likely refuse to cooperate with the researcher after the implementation of the proposed revision to OMB Circular A-110 because of the heightened fear that the data, even if scrubbed, would be released and might be used against its interests.

• As the proposed regulation is stated, there is no definition of what data are.
• When are data releasable? After the first publication by the PI? After a time limit?