Complex mixtures and indoor air pollution: overview of epidemiologic methods.

The likelihood of an epidemiologic study correctly identifying an adverse health outcome associated with exposure to indoor air pollutants is increased if a) substantial variation exists in the frequency or level of exposure among study subjects otherwise at similar risk of the health outcome; b) the number of study subjects or study communities is large; c) the health outcome can be assessed with accuracy; d) relevant exposure levels can be measured with accuracy; e) an unbiased sample of exposed and nonexposed subjects is selected for study; and f) other determinants of the adverse health outcome can be measured. Nonetheless, given a strong enough impact of exposure to one pollutant or a mixture of pollutants on the risk of illness, it is possible for epidemiologic studies to discern a relation even if only some of the above circumstances are present.