Abstract

Objectives: To compare methods for defining the population at risk from a point source of air pollution. A major challenge for environmental epidemiology lies in correctly identifying populations at risk from exposure to environmental pollutants. The complexity of today's environment makes it essential that the methods chosen are accurate and sensitive.

Methods: Environmental and mathematical methods were used to identify the population potentially exposed to a point source of airborne pollution emanating from a waste incinerator. Soil sampling was undertaken at 83 sites throughout the city and environs. The concentrations of arsenic and copper were measured at each site. Computer software produced smoothed contour plots of the distribution of arsenic and copper in the soil based on the information derived from the sampling sites. The population at risk was also identified using concentric rings of varying radii, with the source of pollution at the centre. Lastly, we used the sites that had previously been selected and measured the frequency of wind direction, speed and distance from the source of pollution at each site. Theoretical contour plots were constructed using the distance from the source of pollution at each site, with and without incorporating wind frequency as a function of direction.

Results: Each method identified different populations at risk from airborne pollution. The use of circles was a very imprecise way of identifying exposed populations. Mathematical modelling that incorporated wind direction was better. Soil sampling at many sites was accurate, as the method is direct; but it is very costly and the close proximity of high and low concentrations hindered interpretation. The smoothed contour plots derived from the soil sampling sites identified an exposed population that was similar to that derived from the spot sampling.

Conclusions: Using circles as the only means of identifying the exposed population leads to dilution of the potential health effect. The best approach is to use local knowledge about wind direction and speed to estimate the population likely to be at risk; to back up this estimate by judicious use of soil sampling; to use contour mapping to guide the final selection of exposed and non-exposed populations; and finally, to interpret the populations identified as being at risk by incorporating information about other potential sources of pollution (past and present) in the area.