Evaluation of the Air Quality Index As a Risk Communication Tool

Air quality conditions in the U.S. are reported to the general public via the regulatory-based Air Quality Index (AQI). The accuracy of AQI as a risk communication tool is dependent, in part, on an assumption of equivalent health risks for each of the index pollutants. Time-series analyses of 858,030 emergency department visits from 2005- 2010 for respiratory diseases in two New York counties (Bronx and Queens) were completed using a Poisson generalized linear model in order to assess the equivalency of respiratory morbidity risk for four index pollutants. Excess respiratory risk per 1-AQI unit was approximately twice as high for ozone (0.16%, 95% confidence interval [CI] [0.08, 0.24]) as compared with sulfur dioxide (0.09%, 95% CI [0.01, 0.16], nitrogen dioxide (0.07%, 95% CI [0.01, 0.15]), and fine particulate matter (0.07%, 95% CI [0.02, 0.12]). Unequal respiratory risks on a per-AQI-unit basis resulted in inconsistencies between reported AQI values and public health risks, especially during the ozone season. While still useful in reporting general air quality conditions to the public, AQI may be insufficiently precise to inform optimal daily behavior modification decisions.