Abstract

OBJECTIVES
This study was conducted to investigate whether air pollution levels in Seoul have short term effects on respiratory disease outpatients.
METHODS
We analyzed five air pollutants (CO, NOy, Oy, SOp, PMyp) and daily counts of outpatient for respiratory disease divided by age group ((15yrs, 15-64yrs, 65+yrs) during the period of July 1997-June 1998 using Generalized Additive Model(GAM).
RESULTS
After controlling for seasonal trends, day of the week, week of the month, temperature and humidity, CO (RR=1. 041 ~ l. 051, 0-day lag) and NOy (RR=1. 024~1. 050, 3-day lag) were significantly associated with daily counts for respiratory disease in all age groups. 0, and SO, were associated, with a 3-day lag in the age group under 15 years (O RR=1. 047, 95% CI=1. 045- 1. 049 i SOp RR=1. 018, 95% CI=1. 017- l. 020) and with a 0-day lag in the age group 15 to 64 years (0, RR=1. 035, 95% CI=1. 032-1. 037; SO, RR=1. 037, 95% CI=1. 035-1. 039), in the age group over 65 years (O RR=1. 042, 95% CI=1. 033-1. 051; SO RR=1.029, 95% CI=1.022-1.036). PM was weakly associated, with a 1-day lag in the age group under 15 years (RR=1.011, 95% CI=1.010-1.013) and with a 0-day lag in the age group 15 to 64 years (RR=1. 021, 95% CI=1. 019-1. 023), in the age group over 65 years (RR=1. 025, 95% CI=1. 018-1. 032). In the two pollutant model, the magnitudes of the relative risk for CO, NOy and Oq were maintained, but SOy was not associated with daily counts for respiratory disease in the age group under 15 years after controlling for NO,. PMyp was not associated in all age groups after controlling for CO, so the association with PMyp in the single pollutant model might be confounded by CO. In this study, the impact Of PMyp on the respiratory disease was not large as compared with other pollutants.
CONCLUSIONS
Air pollution in Seoul may increase the incidence of the respiratory disease, so more positive attention for the control of air pollution should be paid.

• Keywords: Air pollution;Respiratory disease outpatient;Generalized additive model(GAM);Time series analysis