Mortality risks of air pollution: the role of exposure-response functions

A cost-benefit analysis of air pollution control presupposes a policy decision about what is to be valued: number of premature deaths or years of life lost (YOLL)? The difference is more than an order of magnitude. We argue for a YOLL valuation on the grounds of economic rationality. For the implementation there is a difficulty: the relation between E-R (exposure-response) functions for mortality and YOLL is not clear. Whereas, a YOLL calculation needs variations in life expectancy (inverse of population mortality rate), acute mortality E-R functions (based on time-series analysis) report variations in daily death counts, and chronic mortality E-R functions (based on cohort studies) report variations in age-specific mortality. Acute mortality E-R functions carry no information on YOLL, but we try to estimate a typical value based on plausible upper and lower bounds. Chronic mortality E-R functions, by contrast, allow a determination of YOLL at least in principle, although in practice there are large uncertainties in the extrapolation from the study population to the general population. We apply our YOLL estimates to a comparison of mortality costs with morbidity costs, for particles and for ozone. The costs of acute mortality turn out to be small compared to morbidity. For particles all is dominated by chronic mortality. (C) 1998 Elsevier Science B.V. All rights reserved.