Share this article on:

Epidemiology at the Edge

Greenbaum, Daniel S.


Health Effects Institute, 955 Massachusetts Avenue, Cambridge, Massachusetts 02139.

Address correspondence to: Daniel S. Greenbaum, Health Effects Institute, 955 Massachusetts Avenue, Cambridge, Massachusetts 02139.

The views expressed here are those of the author, and do not necessarily reflect those of the Health Effects Institute, or of its sponsors in the U.S. Environmental Protection Agency or industry.

The worlds of scientific research and public policy could not be more different—an observation I make from the perspective of both sides. I have worked as a policy-maker, first as an advocate for a non-governmental organization and then for 6 years as the Commissioner of the Massachusetts Department of Environmental Protection. I now see these differences from the vantage point of the head of a research organization that attempts to provide scientific evidence relevant to policy decisions. At the edge between these two worlds are five fundamental differences.

First, there is the tension between science’s need, over an extended period of time, to come to a robust understanding of a biological, physical, or other process, and the policy maker’s need (especially when political pressure has built) to make a decision and act in a timely fashion. There are frequently pressures and frustrations when decision makers need “answers” from science, and scientists respond, not surprisingly, with cautiously presented, conditional findings, which may or may not be directly relevant to the policy question.

Second, there is the tension between science’s desire to be objective, driven by the facts of carefully constructed observational studies and experiments, and the adversarial nature of policy making, where advocates see “facts” as useful only if they lend advantage to their position in the debate. Beneath this tension is also the fact that scientists are human beings, with beliefs on larger world issues; in some cases it can be a challenge to maintain a balance between those beliefs and the impartiality (perceived or actual) of scientific research.

Third, there is the tension within science between basic and applied research, with a strong predisposition to value basic research (and the deeper insights it is thought to bring), over research specifically designed to inform public questions. This tension is further amplified by differences between disciplines, with epidemiologic studies often having the potential to be more directly relevant to decisions about public health, but viewed by other disciplines (eg toxicology) as less “scientific.”

Fourth, there is a tension between the questions policy makers want answered and the techniques available to scientists to answer them. Environmental policy makers want to know the risks from exposure to a particular pollutant or diet at levels found in everyday life. Scientists can answer with toxicological studies of small numbers of animals exposed at levels well above everyday levels, or of small numbers of relatively healthy humans exposed at more realistic levels. Or they might respond with epidemiologic studies of exposures in the real world, but often with limited ability to assess exposures and other potentially confounding factors accurately. The difficulty is that no one set of scientific techniques can best answer policy questions. The opportunity is that the pressure to answer these questions can be a positive force for the development of new, more sophisticated scientific techniques. 1

Finally, but by no means least, there are the inherent differences between scientists who would like to see science play a more central role in decision making, and decision makers who must of necessity make decisions based on a complex mixture of scientific, technical, economic, and policy and political considerations. This broad base of decision factors is in part inevitable in a democracy, where decision makers’ authority and tenure are linked directly or indirectly to the public’s satisfaction with their performance. Even in this context, many decision makers will make every reasonable effort to incorporate the best science in decisions. Some frustration for scientists is, however, inevitable.

The question of whether and how scientists should advance particular policy positions must be viewed against this background of cultural and professional differences. Most observers would agree that scientists should be able (some would say encouraged) to take policy positions if their research leads them to conclude that public action is necessary. 2,3 The question, then, is not whether, but when and how.

At one level, and despite these differences, policy makers have often sought scientific input when setting policy. Witness, for example, the provisions in most major regulatory legislation calling for some form of scientific scrutiny to support public agency action. In the 1997 U.S. Congressional hearings on proposed new National Ambient Air Quality Standards, prominent scientists were asked again and again, “What does science tell us is the right level at which to set these standards?”4

Yet this example illustrates the pitfalls. Some scientists answered the members of Congress with carefully constructed, scientifically justifiable answers, attempting to explain both the facts and the uncertainties, with the result being frustration on the part of the legislators. Other scientists chose to take positions in support of or against the proposed standards, resulting in fairly readily being typecast by “the other side” as no longer objective.

There is no question that scientists should have a place at the policy making table, and they should be able to state their views if the science supports it. This is especially true of epidemiologists, whose very training embodies the principles of public health on which policy decisions are often based. But taking these differences and pitfalls into consideration, one can begin to construct some principles by which scientists can effectively participate to the benefit of both policy and science.

First, it seems that, in all but the rare exception, scientists should avoid taking positions based on the results of a single study. Few (if any) studies ever are large enough, or comprehensive enough, to justify a particular policy position. The best science usually comes from a series of studies, with different designs, strengths, and weaknesses, from which an overall understanding can be derived. Relying on the results of a single study has the added risk that a subsequent study will find an opposite result, raising the frustration of decision makers and the public if the first results have been advocated too strongly, and calling into question the objectivity of the investigator who took the position.

It follows that scientific opinions should flow from thoughtful reviews of a body of literature, placing any single study into its context, and fully and fairly acknowledging both supportive and opposing results from other studies. It is conceivable that such reviews could be worked into articles reporting the results of single studies. But it seems unlikely that most journals would be able to provide the space necessary to report the findings and conclusions adequately. More appropriately, review articles, commentaries, and editorials offer the opportunity for scientists to expand their views on how research supports or opposes a particular policy.

Finally, the opinions of scientists on policy related issues will have the greatest weight if they represent the views of a broad group of experts (ideally from diverse disciplines) rather than those of a few individuals. Two recent examples illustrate this. The conclusions of a broad, international group of scientists on the Intergovernmental Panel on Climate Change have had far more impact on the international debate than any single scientist could have had. 5 Similarly the recent report from the National Academy of Sciences on the health effects of methylmercury has markedly changed the debate over whether proposed government actions are appropriate. 6

The experience of the Health Effects Institute (HEI) is also instructive here. Over nearly 2 decades, HEI has funded, reviewed, and published a number of studies on the health effects of exposure to diesel exhaust. Although each study was published along with a commentary by the Institute’s Review Committee, those individual comments did not attempt to take a broader position on the policy implications of the scientific results. In addition to these individual studies, however, HEI has convened two expert panels of scientists to review entire areas of the science. The resulting reports provided far more focused conclusions on what we know, what we do not know, and the implications for policy. 7,8 These reviews have had far more weight in the decision making process than any of the individual studies that preceded them.

In the broadest sense then, there is a very real and necessary place for science at the policy making table. And science, when it attempts to be successful at the edge, can be prodded to higher levels of sophistication in its effort to create scientific findings that truly meet the needs of decision makers. In this context, the issue of whether scientists should take positions in connection with the publication of a single article is something of a distraction. The reality is that a number of prominent journals do allow such statements, but one can readily question whether those statements, uttered in the last paragraph of an article in a scientific journal that few if any policy makers will read, will have a significant impact on specific decisions. Rather, if there is an opinion to be expressed, doing so in an accompanying editorial or even better by enlisting an expert panel of scientists, is far more likely to bridge the differences between these two worlds, and to place the best science at the center of important public decisions.

Back to Top | Article Outline


1. Samet JM, Zeger S, Dominici F, Dockery D, Schwartz J. The National Morbidity, Mortality, and Air Pollution Study (NMMAPS). Part I: Methods and Methodologic Issues. Research Report 94. Cambridge, MA: Health Effects Institute, 2000.
2. Teret S. “So What?” Epidemiology 1993; 4:93–94.
3. Rothman KJ. Policy recommendations in epidemiology research papers. Epidemiology 1993; 4:94–95.
4. Committee on Science, US House of Representatives. The Science Behind the Environmental Protection Agency’s Proposed Revisions to the National Ambient Air Quality Standards. Hearings before the Subcommittee on Energy and Environment, March 12, May 7 and 21, 1997. Washington, DC: US Government Printing Office, 1997.
5. Intergovernmental Panel on Climate Change. Climate Change 2001; The Scientific Basis: WGI Contribution to the IPCC Third Assessment Report. Geneva, Switzerland: IPPC Secretariat C/O World Meteorological Organization, 2001.
6. Committee on the Toxicological Effects of Methylmercury. Toxicological effects of methylmercury. Washington, DC: National Academy Press, 2000.
7. HEI Diesel Working Group. Diesel Exhaust: Critical Analysis of Emissions, Exposure and Health Effects. Cambridge, MA: Health Effects Institute, 1995.
8. HEI Diesel Epidemiology Expert Panel. Diesel Emissions and Lund Cancer: Epidemiology and Quantitative Risk Assessment. Cambridge, MA: Health Effects Institute, 1999.
© 2001 Lippincott Williams & Wilkins, Inc.