Developing methodology for exposure assessment of air pollutants in schools

Abstract

Increasing evidence suggests that exposures to air pollutants present in indoor environments are contributing factors to the recently observed increase in respiratory symptoms among young children. The SchoolAir pilot study aimed to assess the hypothesis that poor indoor and outdoor air quality in schools is associated with increased prevalence of asthma, respiratory and allergic symptoms among primary school children, and assess the feasibility of a bigger full-scale research project in the future. The main aim of this MPhil project was to develop and test a methodology for exposure assessment of indoor and outdoor pollutants in primary schools. A secondary aim was to assess the prevalence rate or respiratory symptoms and their relationship to air pollutant exposure in different schools. The following pollutants were measured indoors and outdoors: carbon monoxide (CO), carbon dioxide (CO2), nitrogen dioxide (NO2), total volatile organic compounds (TVOC) formaldehyde (HCHO), and particulate matter of 0.5–5.0 micrometers in diameter (PM0.5-5.0). A questionnaire was used to assess respiratory health effects in children. Air quality monitoring was conducted in three rounds in four primary schools in England. Real time measurements were performed simultaneously in three indoor locations and one outdoor location within each school, for one week during usual school hours. Personal exposure (PE) to each pollutant was estimated combining time-activity patterns of children and measured concentrations. Findings showed important temporal and spatial variations in concentrations of certain air components. The most prominent variability was observed for PM0.5-5.0 and CO2. Weekly means for PE to PM0.5-5.0, NO2 and TVOC were higher than concentrations measured in classrooms (ME) in the majority of cases, whereas for CO, HCHO and CO2 the opposite effect was observed. The calculated coefficients of variations for ME and PE revealed that variability of modelled PE was higher than that of relevant ME. Thus modelled PE seems to reflect more of the actual variability of exposures that children had during their days at school than exposures measured by fixed monitors in a classroom. The results of linear regression of PE to ME showed that for the three of the six investigated indoor air components – PM0.5-5.0, NO2, and CO2 – less than 50% of the variation of PE could be explained by the variation of relevant ME. For the other three pollutants – CO, HCHO and TVOC the results of linear regression were inconclusive, as half of the calculated coefficients of determinations (R2) were above 0.5 and the other half were below 0.5. Preliminary analysis of the health survey results revealed variations of respiratory and allergic symptoms prevalence between the investigated schools. It was shown that the children in one of the suburban schools, where the modelled yearly mean PEs were in the upper end of the inter-school yearly means range had the highest proportion of respiratory and allergic symptoms, whereas in the rural school the modelled yearly mean PEs were overall in the lower end of the inter-school yearly means range, and the children of the rural school had the lowest prevalence of symptoms. The methodology used in this study for the assessment of children’s personal exposure to air pollutants during a school day employed a combination of measurements by stationary monitors in school microenvironments and children’s time-activity-location patterns. This study revealed important differences between concentrations measured with fixed monitors and estimated personal exposures for all measured pollutants. This methodology is efficient and potentially less expensive than individual personal monitoring.