Developing a Model to Estimate Individual Long-Term Exposure to AIr Pollutants

ISEE-516 Introduction: Previous studies of the health effects of chronic exposure to air pollution often relied on poor or proxy estimates of long-term exposure to air pollutants. Historical air quality data are often not available, data from monitoring sites are averaged over wide areas, and residential histories of participants are not taken. The aim of this study is to develop models for estimating cumulative exposure of individuals to air pollution. The initial scoping of this model is applied to a cohort of 1,142 participants born in a metropolitan area in the United Kingdom (UK) in 1947, for whom detailed health records are available until age 15 and at age 50. Methods: Historical air quality data were used to estimate individual monthly pollution levels. Data were available for soluble and insoluble particles (collected by deposit gauges), SO2, and black smoke reported as monthly averages. Using the geostatistical technique of kriging, pollution levels were calculated for each of approximately 12,000 grid squares of 0.01 km2. The influence of meteorological conditions, known point sources of pollution, household density and indoor fuel usage were also tested in the model. The feasibility of using calculated levels of pollution was tested against prospectively collected residential histories of cohort participants during the period 1947-1962. Results: Air quality has been monitored in the UK since the early part of the 20th century. As the number of monitoring sites increased, so too did the frequency of reporting data. By 1947, seven monitors covered the conurbation of interest, increasing to 26 in 1962. Monthly averages were available from each station (active for 1 to 15 years). There was a three-fold variation between monitoring sites, but no consistent trend for individual stations emerged over the 15-year period. Full residential histories of cohort participants were obtained from written and electronic archives. Postcodes assigned to each place of residence were used to locate participants on the modelled monthly pollution grid. In this way, cumulative exposure to pollutants could be calculated for each participant. In future, these data will be used to assess the relationship between exposure and known health outcomes. Discussion: This model describes a novel way of estimating an individual’s long-term exposure to air pollutants. Such a model depends on the availability of air quality data and residential history. Where such information is available this model adds a valuable environmental component to lifecourse epidemiology studies.