An "Old" Methodological Pitfall: Numbers of Deaths Due to Reducing Air Pollution Cannot Be Identified from Epidemiological Data.

To the Editor:

Although it is important to put numbers to the health benefits of reducing air pollution, we would like to draw attention to a methodological misconception in the focused review by Schraufnagel and coworkers (1). Here are some examples of how the authors described the effects of reducing air pollution in their review: “These benefits, valued at $2.0 trillion in 2020, are primarily attributable to lower mortality—an estimated 230,000 attributable deaths avoided per year—due to lower concentrations of outdoor PM. Other benefits include premature mortality due to lower ozone concentrations (7,100 deaths avoided per year)….” “Secondhand smoke is a form of air pollution that caused an estimated 1.2 million deaths in 2017.” “Home air filtration after a Southern California wildfire prevented … 7–39% of the deaths attributable to wildfire particles.”

Importantly, the number of deaths due to exposure or avoided by a reduction in exposure cannot be identified from epidemiological data, as was done in the publications cited by Schraufnagel and colleagues in their statements above. This was demonstrated by leading epidemiologists in the 1980s (2, 3). The scientific arguments were detailed in a recent paper and applied to environmental burden-of-disease and impact studies (4).

Such analytical problems are widespread (5) and they impact important burden-of-disease projects (6). Ignoring identified pitfalls rather than acknowledging the well-known limitations of epidemiologic data (2, 3, 5) may impede appropriate interpretations of important research regarding environmental epidemiology and public health.