On the interpretation of epidemiological studies of ambient air pollution.

Studies of the effects of ambient air pollution on human health are complicated by the fact that individuals are exposed to a complex mixture of toxic and non-toxic substances that vary in their make-up in space and in time. It is impossible using standard epidemiological designs to uniquely identify any individual component of air pollution as a causal agent of a health effect because of simultaneous exposure to all or a subset of ambient pollutants. Nevertheless, associations observed for a particular pollutant can be interpreted as a marker of toxicity of the mixture, and regulatory action on judiciously selected pollutants can benefit health, as such actions can modify the complex mixture. This paper argues that we may be able to further our understanding of the effects of air pollution on human health by investigating whether particular mixtures are more toxic through (1) conducting and reporting results for all measured components; (2) possibly characterizing the complex mixture by developing synoptic categorizations of daily air pollution, through cluster or latent profile analysis; (3) a careful analysis of differences in the mixtures between locations; and (4) toxicological studies and controlled studies on humans that make use of factorial designs for various components of the air pollution mixture.