[Time series methods in epidemiological studies on air pollution]

Area of Science:

• Epidemiological studies
• Time series analysis
• Environmental epidemiology

Context:

• Air pollution impacts public health, particularly among the elderly.
• Previous studies have analyzed daily mortality data in relation to air pollution.
• Barcelona experienced significant smog pollution between 1991-1995.

Purpose:

• To review time series methods for epidemiological studies on air pollution.
• To illustrate the application of autoregressive Poisson regression.
• To model the relationship between daily mortality and air pollution, controlling for confounders.

Summary:

• An autoregressive Poisson regression model was used to analyze daily deaths (all causes, 70+ years) in Barcelona (1991-1995) and daily smog levels.
• The model accounted for weather variables, seasonality, trends, day of the year, and flu epidemics.
• It incorporated non-linear relationships, time lags, and residual autocorrelation to refine the analysis.

Impact:

• Provides a standardized method for analyzing air pollution's impact on mortality.
• Enables objective point analysis and facilitates comparison of results across studies.
• Offers a robust approach to controlling for confounding variables in epidemiological research.