The American College of Epidemiology recently proposed 10 Principles concerning health data control, access and confidentiality. The Principles are well-intentioned and, for the most part, sensibly protect medical privacy.
But proposed Principles Nos. 9 and 10 are contrary to the scientific method and threaten the scientific credibility of epidemiology. They give researchers absolute discretion to withhold data.
Proposed Principle #9 states "Federal mechanisms are [needed] to protect investigators and research institutions from the forced disclosure of confidential data created as part of the research process. Researchers should not be subject top pressures from commercial and special interests to release data collected under conditions of confidentiality. Such protections are needed to ensure the independence of the process of scientific discovery and the confidentiality of individual health-related information.
Proposed Principle #10 states "Federal law should preempt states laws on the subject of data access and confidentiality. This is needed to ensure consistent nationwide governance of access to individually identifiable health data. Many large epidemiologic and health services research studies are organized either as multi-center studies or in multiple states or are performed by health care organizations responsible for the care of individuals in multiple states.
The scientific method is the process by which scientists conduct studies and collect data. By following the scientific method, scientists are able to conduct studies with replicable results. If the scientific method is not followed, it is not possible to know whether results were due to what was tested or merely the result of chance.
To do an epidemiologic study according to the scientific method, epidemiologic data must be available for independent researchers to analyze. Otherwise, reported results cannot be confirmed. While confirmation of a single study is not always important, consider the recent controversy over the Environmental Protection Agency's new air quality standards.
The EPA claimed two specific epidemiologic studies showed the new standards would prevent 15,000 premature deaths annually from fine particulate air pollution. Opponents claimed these studies were of questionable validity and did not justify what would be the most expensive environmental regulation ever--estimated to cost in excess of $100 billion annually.
Both studies reported very weak statistical associations (i.e., relative risks of 1.26 and 1.17) between fine particulate pollution and mortality. These weak associations were based ecologic exposure data. Questions arose about whether confounding risk factors were adequately considered--especially important since mortality is a complex event and rarely due to a discrete cause.
When the researchers were asked by Congress to make available the data underlying the studies, they refused. They offered excuses ranging from medical privacy to proprietary concerns.
No one was interested in violating anyone's medical privacy or interfering with proprietary interests in the data (some of which was collected at taxpayer expense!). Data access was requested only to confirm epidemiologic results that were claimed to support significant regulatory action. It would have been a simple matter to mask the identity of study subjects and to restrict use of the data for the purposes of independent review.
Proposed Principles 9 and 10 would officially countenance this anti-scientific behavior. If researchers want to be considered "scientists," they must comply with the scientific method, including reasonable data access. Science is supposed to be an open, not a secret process.
Steven Milloy is executive director of The Advancement of Sound Science Coalition, 1155 Connecticut Ave., N.W., Suite 300, Washington, D.C. 20036. Tel: 202-467-8586. Fax 202-467- 0768. E-mail: milloy@cais.com
Material presented on this home page constitutes opinion of the author.
Copyright © 1998 Steven J. Milloy. All rights reserved. Site developed and hosted by WestLake Solutions, Inc.