Share this article on:

Air Pollution and EPIDEMIOLOGY: “Déjà Vu All Over Again?”

Samet, Jonathan M.


Department of Epidemiology, Bloomberg School of Public Health, Johns Hopkins University, 615 North Wolfe Street, Suite W60141, Baltimore, MD 21205-2179;

For centuries, air pollution has been a public health and aesthetic concern, managed by governments (with varying degrees of success) to protect the public. Although there is still uncertainty about many aspects of air pollution and health, there are now evidence-based regulations in many countries to protect the public from air pollution by motor vehicles and by stationary sources such as power plants and factories. Epidemiologic evidence is a cornerstone of these regulations, providing not just documentation of health effects but the dose-response data necessary for setting rational standards. Air quality has improved in many countries as a result.

Since the early 1990s, the volume of epidemiologic publications on air pollution has grown dramatically (although not at a pace faster than epidemiology generally). This increase has several explanations. First, there were the surprising reports beginning around 1990 that contemporary levels of air pollution were associated with excess mortality. Second, new analytic methods for time-series data, together with advances in hardware and statistical software, made it possible to conduct time-series analyses that exploited easily accessible databases of morbidity, mortality, and air pollution. Third, funding agencies, particularly in the United States, supported an expanded research agenda on airborne particulate matter that focused on key uncertainties. 1 And fourth, the U.S. Environmental Protection Agency (EPA) required findings to be published in a peer-reviewed journal before they could be incorporated into the regulatory process.

The result has been a dramatic rise of epidemiologic publications on the health effects of air pollution. Figure 1 shows results of a MEDLINE search using the terms “air pollution” and “epidemiology.” Publications on this topic increased nearly 140% between 1990 and 2000. The EPA’s 1996 Criteria Document for Particulate Matter included more than 200 pages reviewing epidemiologic evidence. EPA’s draft update of epidemiologic data since 1996 covers 286 pages and cites more than 300 papers.



Epidemiology, which itself dates to 1990, has published 76 articles on air pollution, including 17 in 2000. This reflects not only the trends in the field, but an increasing number of high-quality submissions on environmental topics since our 1994 affiliation with the International Society for Environmental Epidemiology. Methods developed for air pollution research have also been creatively applied to other areas of epidemiology, such as infectious disease. 2,3 Epidemiology has provided a forum for discussion of these new methodologies and for divergent views on the findings and their interpretation. 4–7

Not surprisingly, many of our recently submitted manuscripts on air pollution follow in the footsteps of the earlier studies. We see many time-series studies of daily mortality, or hospitalization, or emergency room and clinic visits, and “panel studies” of susceptible groups. New studies may offer refinements in the form of methodologic enhancements or larger sample sizes, and many add new populations to those previously studied. But are these refinements enough for the new studies to advance knowledge of such a well-studied topic as air pollution?

Replication is critical in epidemiologic research. Still, many of the newer studies add little to an already large literature. By recent count, there have been 110 papers on time-series studies of particulate air pollution and daily mortality. The inevitable 111th paper is not likely to shift the balance of evidence. In addition to individual studies, there is a new generation of analyses combining data from multiple locations to explore heterogeneity across locations. 8,9

In the context of this abundant literature, what can yet another study add? Replication of findings in a new population may be noteworthy if there are unique exposure or susceptibility characteristics. But at this point, the finding of an association between air pollution and morbidity or mortality in a new setting may be most useful to local authorities, as an indication that local conditions are consistent with those elsewhere, and therefore publication in a local journal would be most appropriate. Such publication also helps to avoid bias in meta-analyses or other summaries.

In short, the quality of this research is often higher than research on other topics, but the net contribution may be less certain. Epidemiology must unfortunately reject a high percentage of such manuscripts. We encourage authors to consider carefully the contribution of their manuscript, and whether publication in an international journal is appropriate. Indeed, this applies to papers on any well-studied topic in which epidemiologic methods have reached a high level of excellence and yet the marginal contribution may be small. Such papers will have a better chance of success in our journal if they are submitted as a “Brief Report” or letter. Authors can provide more complete details on their website, or in the electronic version of their publication in Epidemiology.

Publication in a peer-reviewed journal is a mandated requirement for use of research findings by the EPA. To meet its deadlines, the EPA has itself held conferences and then arranged for special issues or supplements to be published. A flood of papers has resulted, challenging the ability of the scientific and regulatory communities to integrate the new information. This strategy has the potential to diminish the quality of peer review as well. Regulatory agencies may need to identify new models for bringing together evidence from different investigators. Peer review should not be set aside, but evidence presented in innumerable scientific papers cannot efficiently be synthesized to determine what has been learned.

We do not discourage the submission of manuscripts on air pollution or other well-studied topics to Epidemiology. In fact, we take pleasure in manuscripts that offer new approaches or provide fresh insights to the understanding of important public health problems such as air pollution. But when the topic is one as widely studied as air pollution, there is a special burden on authors to show readers that their work is not just (in the immortal words of Yogi Berra) “déjà vu all over again.”

Jonathan M. Samet


Back to Top | Article Outline


1. National Research Council, Committee on Research Priorities for Airborne Particulate Matter. Research Priorities for Airborne Particulate Matter. No. 1. Immediate Priorities and a Long-Range Research Portfolio. Washington DC: National Academy Press, 1998.
2. Schwartz J, Levin R, Goldstein R. Drinking water turbidity and gastrointestinal illness in the elderly of Philadelphia. J Epidemiol Community Health 2000; 54: 45–51.
3. Curriero FC, Heiner KS, Samet JM, Zeger SL, Strug L, Patz JA. Temperature and mortality in eleven cities of the eastern United States. Am J Epidemiol 2001; 155: 80–87.
4. Dockery DW, Schwartz J. Particulate air pollution and mortality: more than the Philadelphia story. Epidemiology 1995; 6: 629–632.
5. Moolgavkar SH, Luebeck EG, Hall TA, Anderson EL. Air pollution and daily mortality in Philadelphia. Epidemiology 1995; 6: 476–484.
6. Moolgavkar SH, Luebeck EG. A critical review of the evidence on particulate air pollution and mortality. Epidemiology 1996; 7: 420–428.
7. Samet JM. Particulate air pollution and mortality: the Philadelphia story. Epidemiology 1995; 6: 471–473.
8. Samet JM, Dominici F, Curriero FC, Coursae I, Zeger SL. Fine particulate air pollution and mortality in 20 U.S. cities: 1987–1994. N Engl J Med 2000; 343; 1742–1749.
9. Katsouyanni K, Schwartz J, Spix C, Touloumi G, Zmirou D, Zanobetti A, Wojtniak B, Yank JM, Tobias A, Ponka A, Medina S, Bacharova L, Anderson HR. Short term effects of air pollution on health: A European approach using epidemiologic time series data. The APHEA Protocol. J Epidemiol Community Health 1996; 50 (suppl 1): S12–S18.
© 2002 Lippincott Williams & Wilkins, Inc.