Acute symptoms related to air pollution in urban areas: a study protocol

The harmful effects of urban air pollution on general population in terms of annoying symptoms are not adequately evaluated. This is in contrast to the hospital admissions and short term mortality. The present study protocol is designed to assess the association between the level of exposure to certain ambient air pollutants and a wide range of relevant symptoms. Awareness of the impact of pollution on the population at large will make our estimates of the pertinent covert burden imposed on the society more accurate.

A cross sectional study with spatial analysis for the addresses of the participants was conducted. Data were collected via telephone interviews administered to a representative sample of civilians over age four in the city. Households were selected using random digit dialling procedures and randomization within each household was also performed to select the person to be interviewed. Levels of exposure are quantified by extrapolating the addresses of the study population over the air pollution matrix of the city at the time of the interview and also for different lag times. This information system uses the data from multiple air pollution monitoring stations in conjunction with meteorological data. General linear models are applied for statistical analysis.

The important limitations of cross-sectional studies on acute effects of air pollution are personal confounders and measurement error for exposure. A wide range of confounders in this study are controlled for in the statistical analysis. Exposure error may be minimised by employing a validated geographical information system that provides accurate estimates and getting detailed information on locations of individual participants during the day. The widespread operation of open air conditioning systems in the target urban area which brings about excellent mixing of the outdoor and indoor air increases the validity of outdoor pollutants levels that are taken as exposure levels.