A Temporal, Multicity Model to Estimate the Effects of Short-Term Exposure to Ambient Air Pollution on Health

BACKGROUND: Countries worldwide are expending significant resources to improve air quality partly to improve the health of their citizens. Are these societal expenditures improving public health? OBJECTIVES: We consider these issues by tracking the risk of death associated with outdoor air pollution over both space and time in Canadian cities. MATERIALS AND METHODS: We propose two multi-year estimators that use current plus several previous years of data to estimate current year risk. The estimators are derived from sequential time series analyses using moving time windows. To evaluate the statistical properties of the proposed methods, a simulation study with three scenarios of changing risk was conducted based on 12 Canadian cities from 1981 to 2000. Then an optimal estimator was applied to 24 of Canada’s largest cities over the 17-year period from 1984 to 2000. RESULTS: The annual average daily concentrations of ozone appeared to be increasing over the time period, whereas those of nitrogen dioxide were decreasing. However, the proposed method returns different time trends in public health risks. Evidence for some monotonic increasing trends in the annual risks is weak for O(3) (p = 0.3870) but somewhat stronger for NO(2) (p = 0.1082). In particular, an increasing time trend becomes apparent when excluding year 1998, which reveals lower risk than proximal years, even though concentrations of NO(2) were decreasing. The simulation results validate our two proposed methods, producing estimates close to the preassigned values. CONCLUSIONS: Despite decreasing ambient concentrations, public health risks related to NO(2) appear to be increasing. Further investigations are necessary to understand why the concentrations and adverse effects of NO(2) show opposite time trends.