EFFECTS OF TEMPERATURE ON MORTALITY IN SEOUL: A TIME SERIES ANALYSIS.

초록

Background The r elationship between temper atur e and mor tality is essential to estimate the burden of climate change. Results from other countries in the same climate zone and some organization are used in estimation of national burden and attribution. However, it can be more reasonable using regional data because it depends on its socioeconomic environment and acclimation. Methods Time-series regression analysis was used to evaluate temperature-mortality relationships in Seoul. Mortality data and meteorological data from 1991 to 2011 were used. The qualitative relationship between daily mean temperature and mortality was evaluated using a generalized additive model with Poisson distribution. Further analysis was performed using piecewise linear regression if graph showed non-linearity with threshold. The threshold and effect of temperature on non-accidental and cardiovascular death were compared to those in previous research. Population attributable fraction (PAF) was calculated using regional data. Results A significantly raised risk of heat-related mortality was observed in Seoul. Non-accidental mortality increased by 0.37% (95%CI: 0.28-0.45) and 10.1% (95%CI: 8.54-11.7) with the threshold of 28.2℃. Above the threshold, cardiovascular mortality increased more than non-accidental mortality, which was by 15.1% (95%CI: 12.0-18.3) with the threshold of 28.3℃. These were quite higher than 2.6% increase in cardiovascular mortality from other studies of the temperate zone (World Health Organization). As for PAF, 9.17% and 13.1% of non-accidental and cardiovascular mortality could attribute to hot temperature. Conclusions The current research showed quite big difference in the influence of temperature on mortality with the previous literature. Regional data could be used for more accurate estimation of environmental burden and attributable fraction.