Atmospheric risk factors of human mortality in Toronto, Ontario

A logical and alternative approach to the assessment of atmospheric-health relationships would be to search for the existence of tripartite links between health outcomes, air quality, and the spectrum of air mass types or weather situations that may affect any particular location. In Canada, there appears to be a paucity of studies that attempt to explore such links.

The goal of the research is to model the relationship between air pollution and human health in the context of locally occurring synoptic weather patterns. Modelling of air pollution impacts on human health has traditionally treated climatological variables as potential confounders. The methods employed are essentially statistical in nature in which some variable is regressed on a number of covariates representing both air pollution and meteorology, with suitable adjustments to allow for seasonality, long-term trends, and serial correlation.

Synoptic categorisation of various weather elements into air masses is a technique that has also been applied to the modelling of human health impacts. The approach categorises weather patterns using factor analysis and offers categories that represent groupings of meteorological variables as they actually occur at any locale. The advantage of utilising this approach is that it captures the multivariate manifestation of the synergistic behaviour of a range of atmospheric variables. Previous research has shown increases in expected mortality due to 'oppressive' air masses, normally those with high temperatures and relative humidities.

Time series data for human mortality, air pollution concentrations and climatology were collected for a 15 year period, 1980 to 1994. Using locally-weighted smoothers and a generalised additive Poisson-modelling framework, relative risk ratios were calculated for human mortality of various air pollutant types. Air mass calendars were created and then incorporated as a factor variable into the regression in order to explain changes in mortality.

The research is unique because it looks at the total atmospheric exposure experienced by a population by integrating air pollution and climatological data into the assessment of human health impacts. The research also constitutes the basis for the development of a predictive tool which may be used to forecast the effects of thermal stress and high concentrations of air pollutants on human health.