Share this article on:

Policy and Science: Should Epidemiologists Comment on the Policy Implications of Their Research?

Teret, Stephen


Professor of Health Policy and Management, Director, Center for Law and the Public’s Health, Johns Hopkins University Bloomberg School of Public Health, Baltimore, MD

Address correspondence to: Stephen Teret, JD, MPH, Johns Hopkins University Bloomberg School of Public Health, 624 N. Broadway, Baltimore, Maryland 21205.

Several years ago, an exchange took place in the pages of this journal between the editor, Kenneth Rothman, and this writer. 1,2 At issue was an editorial policy stating that “Opinions or recommendations about public-health policy should be reserved for editorials, letters, or commentaries, and not presented as the conclusions of scientific research.”3 This position was consistent with earlier opinions expressed by Rothman, such as his 1985 editorial in the American Journal of Public Health arguing that “Having focused on a research area,...scientists should ignore policy questions to persevere in pursuit of their objective, which is knowledge.... The time for a scientist to be a political and social mover is after hours.”4

My argument (only one in a series of scholarly differences of opinion with Rothman going back to the seventh grade) was that health policy needs to be informed by scientists, and scientists should not be proscribed from discussing the policy implications of their research. While the scientist need not feel an obligation to discuss policy, on the other hand, “If the scientist feels equipped to state the policy implications of his or her work, then why prohibit the creation of that most important bridge between research and policy?”1

Now, 8 years later, the editors of Epidemiology have chosen to revisit this issue. On reflection, whether an epidemiologist is allowed to mention policy in his or her manuscript, rather than in a letter, commentary, or editorial, is of less importance than the more general question of whether epidemiologists should be encouraged or discouraged from discussing anywhere the policy implications of their work. Should epidemiologists eschew policy because it interferes with their quest for knowledge?

It might be instructive to focus on areas of probable agreement regarding the relationships between science and policy. If one were to speculate on these possible areas of concurrence, they could include the following:

  • Policy will be made and needs to be made to protect the health and safety of the public.
  • Policy is often made by those who are inadequately informed by and untrained in science.
  • While policy will probably never be determined solely by scientific findings, policies that are based on sound scientific information are preferable to those that are not.
  • In order for policy to be informed by science, there must be some exchange of information, either directly or indirectly, between scientists and policy makers.

How, then, shall we attempt to bridge the gulf between science and policy, for the benefit of the public? Perhaps it is useful to start with a specific example, such as guns. Much of the descriptive epidemiology of firearm-related morbidity and mortality had not been developed 20 years ago. Health information on firearms was presented, if at all, in a compartmentalized manner, with gun-related suicides being seen as a mental health problem, homicides being relegated principally to the fields of law enforcement and criminology, and unintentional injuries being ignored. In 1987, Dr. Garen Wintemute published an article that aggregated gun death data; to the surprise of many in the field of public health, it was revealed that guns, in the 50-year period of 1933 to 1982, killed nearly 1 million people in the United States. 5 In addition to reporting on the epidemiology of gun deaths, Wintemute commented that changes to the design of firearms, making them less lethal and less concealable, might reduce the number of deaths. That same year, Wintemute and colleagues published an epidemiologic analysis of a case series of unintended gun deaths involving children shooting themselves or other children. 6 Again, the authors discussed the need for designing guns so that they would be safer consumer products.

Certainly, any comments on gun policy are controversial, given the strong opinions Americans hold regarding firearms. But the suggestion, coming from scientists, that a public health approach to gun violence reduction includes changing the product as well as trying to change the behavior of those using the product, launched a sea-change in how we, as a nation, view gun policy. Surveys show that the vast majority of the general public, and even the majority of gun owners, now favor laws that would mandate safer designs of handguns. 7 In this author’s opinion, changing social norms so that we now view guns as consumer products that should be regulated was materially assisted by the inclusion of policy comments in papers reporting on scientific findings. If scientific experts back away from discussions of the policy issues, or relegate such discussions to “after hours” only, they would be leaving an important matter to be resolved by those with far less understanding of the problem.

Epidemiologists in the public health field of injury prevention have been exemplary in their willingness to make a bridge between their scientific findings and the policies that would best reduce the incidence of injuries. (It was an article on injury prevention that stimulated the first editorial in Epidemiology on the subject of policy.) As another example, more than 2 decades ago, injury epidemiologist Susan Baker published the first data on the high death rate of infants involved in motor vehicle crashes. Whereas the vehicle occupant death rate for children ages 6 to 12 years was approximately 3 per 100,000, for children younger than 6 months the vehicle occupant death rate was 9 per 100,000. In her article published in Pediatrics, it was stated that “...the high (death) rate for infants calls for emphasis on their need for protection. Only 7% of children in the United States are restrained when they travel, and almost all countries with seat belt laws have exempted young children from required use.”8 In this paper and in others, 9 the authors made a call for greater use of infant and child restraints, which could be accomplished through legislation. Articles and opinions such as these were instrumental in the passage of child restraint device legislation, which was passed in every state of the nation in the several years following the publication of these articles. More recently, Chen and colleagues have published epidemiologic data demonstrating that “[T]he incidence of motor vehicle crashes fatal to 16- and 17-year-old drivers increased with the number of passengers.”10 Following the description of the results of the research, the authors comment that “...restrictions on carrying passengers should be considered for inclusion in graduated licensing systems for young drivers.”10 Comments on the policy implications of the research, such as these, seem to enhance the public health value of the research.

One of the warnings sometimes heard regarding scientists who engage themselves in or comment on the formulation of policy is that their credibility will be impaired because they are no longer objective. Rothman and Poole again offer an opinion here: “It is important ... for scientists to safeguard their scientific objectives as much as they can from secular influences. The conduct of science should be guided by the pursuit of explanations for natural phenomena, not the attainment of political or social objectives.”4 There appears to be a way to test the hypothesis of impaired credibility. Survey research, based upon carefully drawn scenarios and the solicitation of opinion responses from one’s peer scientific group, can be performed. Why not posit situations in which researchers comment on the policy implications of their findings in a published article? Then ask other scientists how they assess the credibility of the author, compared to his or her credibility if the article had been devoid of policy discussion. Answers based upon data rather than rank speculation might make for sounder editorial policies.

In the final analysis, most people who devote their careers to public health are probably interested in seeing the reduction of diseases or injuries, and the enhancement of the health of populations. Health policy plays a crucial role in accomplishing these goals. If epidemiologists can inform policy makers of the comparative effectiveness of policies or of the need for formulation of new policies that would reduce the incidence of disease and injury, it seems counterproductive to silence the voice of epidemiologists in this regard, or to instruct them that their voices should only be heard by policy makers “after hours.”

Back to Top | Article Outline


1. Teret S. So what? Epidemiology 1993; 4:93–94.
2. Rothman KJ. Policy recommendations in epidemiology research papers. Epidemiology 1993; 4:94–95.
3. Guidelines for Contributors. Epidemiology January, 1993.
4. Rothman KJ, Poole C. Science and policy making (Editorial). Am J Public Health 1985; 75:340–341.
5. Wintemute GJ. Firearms as a cause of death in the United States, 1920–1982. J Trauma 1987: 27:532–536.
6. Wintemute GJ, Teret SP, Kraus JF, Wright MA, Bradfield G. When children shoot children: 88 unintended deaths in California. JAMA 1987; 257:3107–3109.
7. Teret SP, Webster DW, Vernick JS, Smith TW, Leff D, Wintemute GJ, Cook PJ, Hawkins DF, Kellermann AL, Sorenson SB, DeFrancesco S. Support for new policies to regulate firearms. Results of two national surveys. N Engl J Med 1998; 339:813–818.
8. Baker SP. Motor vehicle occupant deaths in young children. Pediatrics 1979; 64:860–861.
9. Karwacki JJ, Baker SP. Children in motor vehicles: never too young to die. JAMA 1979; 242:2848–2851.
10. Chen L, Baker SP, Braver ER, Li G. Carrying passengers as a risk factor for crashes fatal to 16- and 17-year-old drivers. JAMA 2000; 283:1578–1582.
© 2001 Lippincott Williams & Wilkins, Inc.