Many studies have reported associations between ambient particulate matter (PM) and adverse health effects, focused on either short-term (acute) or long-term (chronic) PM exposures. For chronic effects, the studied cohorts have rarely been representative of the population. We present a novel exposure model combining satellite aerosol optical depth and land-use data to investigate both the long- and short-term effects of PM2.5 exposures on population mortality in Massachusetts, United States, for the years 2000–2008.
All deaths were geocoded. We performed two separate analyses: a time-series analysis (for short-term exposure) where counts in each geographic grid cell were regressed against cell-specific short-term PM2.5 exposure, temperature, socioeconomic data, lung cancer rates (as a surrogate for smoking), and a spline of time (to control for season and trends). In addition, for long-term exposure, we performed a relative incidence analysis using two long-term exposure metrics: regional 10 × 10 km PM2.5 predictions and local deviations from the cell average based on land use within 50 m of the residence. We tested whether these predicted the proportion of deaths from PM-related causes (cardiovascular and respiratory diseases).
For short-term exposure, we found that for every 10-µg/m3 increase in PM 2.5 exposure there was a 2.8% increase in PM-related mortality (95% confidence interval [CI] = 2.0–3.5). For the long-term exposure at the grid cell level, we found an odds ratio (OR) for every 10-µg/m3 increase in long-term PM2.5 exposure of 1.6 (CI = 1.5–1.8) for particle-related diseases. Local PM2.5 had an OR of 1.4 (CI = 1.3–1.5), which was independent of and additive to the grid cell effect.
We have developed a novel PM2.5 exposure model based on remote sensing data to assess both short- and long-term human exposures. Our approach allows us to gain spatial resolution in acute effects and an assessment of long-term effects in the entire population rather than a selective sample from urban locations.
From the aDepartment of Environmental Health - Exposure, Epidemiology and Risk Program, Harvard School of Public Health, Boston, MA; bScience Systems and Applications, Inc, Lanham, MD; and cDepartment of Biostatistics, Harvard School of Public Health, Boston, MA.
Supported by The Harvard Environmental Protection Agency Center Grant RD83479801, NIH grant ES012044, and the Environment and Health Fund Israel.
Correspondence: Itai Kloog, Department of Environmental Health - Exposure, Epidemiology and Risk Program, Harvard School of Public Health, Landmark Center 401 Park Dr. West, Boston, MA 02215. E-mail: email@example.com.
Received May 29, 2012
Accepted January 8, 2013