The EPA's Fine Particulate Matter (PM2.5) Standards, Lung Disease, and Mortality

The EPA's Fine Particulate Matter (PM2.5) Standards, Lung Disease, and Mortality

A Failure of Epidemiology
September 07, 2006

Full Document Available in PDF

Congress passed the Clean Air Act of 1970 based on the belief that reducing air pollution levels saves lives and improves health.  The Act mandated the Environmental Protection Agency (EPA) to base its regulatory policies on good science.  In 1997, EPA promulgated standards for fine particulate matter that were the most stringent and expensive in the agency’s 35-year history.  The standards were widely criticized, and even EPA’s own science advisory committee did not endorse them.  Instead of preventing 20,000 deaths and saving $69 to $144 billion a year at a cost of $6.3 billion (for partial attainment), as claimed, the standards have cost at least $70 billion a year to implement, eliminated hundreds of thousands of jobs a year, and likely have cost lives (because of the huge cost) without providing any public health benefit.

One reason for this failure of public policy lies with the epidemiological environmental studies used.  Two large studies served as the scientific basis for the standards promulgated—the 1993 Harvard Six Cities Study and the 1995 American Cancer Society Study.  These and other studies showed only a weak association between exposure and disease or death—an increased relative risk of 1.26 and 1.17 respectively—and yielded several discrepant results.

Epidemiology is the study of health in populations.  With air pollution studies, as a practical matter, exposure is estimated at the group level, while health outcomes are measured at the individual level.  In addition, observational epidemiological studies, unless they show overwhelmingly strong associations—on the order of an increased relative risk of 3.0 or 4.0—do not indicate causation because of the inherent systematic errors that can overwhelm the weak associations found.  These errors include confounding factors, methodological weaknesses, statistical model inconsistencies, and at least 56 different biases. 

In order to show causation, environmental epidemiological studies showing strong associations must be accompanied by experimental animal toxicologic studies that provide evidence for a plausible biological mechanism.  But dozens of such animal studies performed over the past 30 years all have been negative. 

The introduction of causal assumptions into observational epidemiological studies that show only weak statistical associations is a problem that has been recognized for many years, and has been well documented in the literature.  Since this process bypasses the scientific method, it has been labeled “statistical malpractice.”  In addition, the improper use of science to promote political agendas is rightly considered unethical.

EPA’s PM2.5 regulations are a tragic failure of public policy that are shown to have no basis in science and thus are not saving lives or preventing illness.  Instead they are imposing billions of dollars of net cost each year on the American people.