The Exposure Paradox in Particulate Matter Community Time-Series Epidemiology: Can Ambient Concentrations of PM Be Used as a Surrogate for Personal Exposure to PM ?
Objective: Explain why epidemiologic studies find a statistically significant relationship between ambient concentrations of PM and health effects even though only a near-zero correlation is found between ambient concentrations of PM and personal exposures to PM. Method: Consider three aspects of personal exposure to PM: nonambient exposure (due to indoor sources and personal activity sources), ambient exposure (due to exposure to ambient PM while outdoors and exposure while indoors to ambient PM that has infiltrated indoors), and total exposure (sum of ambient and nonambient). Consider the three types of exposure data: longitudinal, exposure data for an individual; daily-average, data for a group of people on the same day for several days; pooled, an arbitrary selection of one person on one day and a different person on another day. Results: The paradox arises from an inappropriate selection of data for the correlation of exposure with ambient concentration and use of an incorrect aspect of personal exposure. Daily community averages of total personal exposure should be correlated with daily ambient PM concentrations. Ambient, not total, exposure should be used in the correlation of pooled exposure data with ambient concentrations. Both theory and experiment indicate that either approach will yield higher correlations. The correlation of daily community average personal exposures to ambient PM with daily ambient PM concentrations should yield a very high correlation. Conclusions: This analysis resolves the paradox and indicates that community time-series epidemiology yields information on the health effects of ambient PM but that other approaches must be used to investigate the health effects of nonambient PM.