Exposure to traffic related air pollutants: self reported traffic intensity versus GIS modelled exposure
- 1GSF-National Research Center for Environment and Health, Institute of Epidemiology, Ingolstädter Landstraße 1, D-85764 Neuherberg, Germany
- 2Ludwig-Maximilians-University of Munich, Institute of Medical Data Management, Biometrics and Epidemiology, Department of Epidemiology, Munich, Germany
- 3University of British Columbia, School of Occupational and Environmental Hygiene, 2206 East Mall, Vancouver BC V6T1Z3, Canada
- 4Institute for Risk Assessment Sciences, Yalelaan 2, PO Box 80176, NL-3508 TD Utrecht, Netherlands
- 5Centre for Environmental Health Research, RIVM, Bilthoven, Netherlands
- 6Department of Environmental Health, Norrbacka III, Karolinska Hospital, 171 76 Stockholm, Sweden
- Correspondence to: Dr J Heinrich GSF-Institute of Epidemiology, POB 1129, D-85758 Neuherberg, Germany;
- Accepted 1 February 2005
Background: In epidemiological studies of the potential health effects of traffic related air pollution, self reported traffic intensity is a commonly used, but rarely validated, exposure variable.
Methods: As part of a study on the impact of Traffic Related Air Pollution on Childhood Asthma (TRAPCA), data from 2633 and 673 infants from the Dutch and the German-Munich cohorts, respectively, were available. Parents subjectively assessed traffic intensity at the home address. Objective exposures were estimated by a combination of spatial air pollution measurements and geographic information system (GIS) based modelling using an identical method for both cohorts.
Results: The agreement rates between self reported and GIS modelled exposure—accumulated over the three strata of self assessed traffic intensity—were 55–58% for PM2.5, filter absorbance (PM2.5 abs), and nitrogen dioxide in Munich and 39–40% in the Netherlands. Of the self reported low traffic exposed group, 71–73% in Munich and 45–47% in the Netherlands had low modelled exposure to these three air pollutants. Of the self assessed high exposed subgroups in Munich (15% of the total population) and the Netherlands (22% of the total population), only 22–33% and 30–32% respectively had high modelled exposure to the three air pollutants. The subjective assessments tend to overestimate the modelled estimates for PM2.5 and NO2 in both study areas. When analysis was restricted to the portion of the Dutch cohort living in non-urban areas, the agreement rates were even lower.
Conclusions: Self reported and modelled assessment of exposure to air pollutants are only weakly associated.
- air pollution
- vehicle emission
- exposure assessment
- subjective assessment
- geographic information system
Funding: the work was supported by European Union Environment contract ENV4 CT97-0506 and QLRT 2000-00073
Competing interests: none