Analysis of the Relationship between Ozone Pollution, Temperature, and Human Health

EPA Grant Number: F07D30747
Title: Analysis of the Relationship between Ozone Pollution, Temperature, and Human Health
Investigators: Anderson, Brooke Williams
Institution: Yale University
EPA Project Officer: Hahn, Intaek
Project Period: September 1, 2007 through September 1, 2010
RFA: STAR Graduate Fellowships (2007) RFA Text |  Recipients Lists
Research Category: Academic Fellowships , Fellowship - Public Health , Health Effects


Recent research has linked both extreme temperatures and elevated ozone pollution to increased risk of mortality. However, several critical questions regarding these relationships still remain. There is little understanding regarding how temperature-mortality varies by region and by factors such as air conditioning use and sensitive subpopulations. Further, potential confounding by pollution has not been addressed on a national scale. Ozone levels are higher during days with high temperatures, so some of the excess mortality currently attributed to high temperatures could be related to ozone.

Additionally, while ozone pollution and extreme temperatures are both related to excess mortality, it is not yet known how much life expectancy is lost as a result of these factors. Analysis of the importance of short-term mortality displacement for ozone and temperature will enable decision-makers to more effectively address the public health burden of weather and ozone pollution. Also, since both extreme temperatures and ozone pollution are anticipated to rise, on average, with climatic change, an understanding of these relationships will benefit research on the potential health consequences of climate change.


A national database of 108 U.S. communities over a 14-year period will be used to study temperature-mortality and ozone-mortality relationships. Community data (race, socioeconomic factors, air conditioning use) will be obtained from the U.S. Census Bureau. Poisson generalized linear models will be used to estimate mortality relationships. Temperature will be included in models as a natural cubic spline, to allow for non-linear relationships between temperature and mortality. Distributed lag models will be used to examine short-term mortality displacement in both temperature and ozone pollution associations with mortality. Community estimates will be combined using a hierarchical Bayesian model to generate national estimates. The statistical methods developed for this project will be applicable to the study of other pollutants and health outcomes, as well.

Expected Results:

This project will yield several different results. The first stage of research will estimate the effect of extreme heat and cold on human mortality in communities across the United States, as well as the importance of controlling for ozone and particulate matter when estimating these effects. The second stage of research will determine the degree (e.g., number of days or years) by which ozone pollution advances mortality. The third stage of research will estimate the degree by which extreme temperatures advance human mortality. All three projects will investigate whether results differ by region as well as how results are impacted by community-specific factors (e.g., racial composition, air conditioning use).

Supplemental Keywords:

temperature, ambient ozone, mortality displacement, heat-related mortality, cold-related mortality, biostatistics, mortality, particulate matter, air pollution, human health, extreme temperatures,, RFA, Health, Scientific Discipline, Air, Health Risk Assessment, climate change, Risk Assessments, environmental monitoring, heat related health effects, ozone, climate related morbidity, air pollution, mortality, Bayesian statistical theory

Progress and Final Reports:

  • 2008
  • 2009
  • Final