Synoptic approach to evaluate the effect of temperature on pediatric respiratory disease-related hospitalization in Seoul, Korea

Abstract

BackgroundThe spatial synoptic classification (SSC) scheme is a daily weather-type classification scheme that is widely used in climatological and epidemiological investigations. We investigated the modification effect of temperature on respiratory disease-related hospital admissions (RAs) in children based on the SSC type. MethodsData on RAs among children aged ≤19 years were collected from the National Inpatient Sample (NIS) database between 2011 and 2015 in Seoul, Korea. In this time-series study, we used Poisson generalized linear models allowing for over-dispersion, with stratification by season and each of the seven distinctive SSC types (Dry moderate [DM], Dry polar [DP], Dry tropical [DT], Moist moderate [MM], Moist polar [MP], Moist tropical [MT], and Transition [T]). We investigated the modification effects of temperature by SSC type through seasonal stratification, adjusting for other meteorological conditions, air pollution, and time trends, and considered the delayed effect for up to four consecutive days. We reported the relative risk percentage change with 95% confidence intervals (CIs) of RAs for every 1 °C increase in ambient temperature, by season and SSC type. ResultsOf the 8346 RAs overall investigated between 2011 and 2015, there were 5271 [63.2%] RAs in children aged 0–4 years. Overall, on majority of the days, the SSC types were DM (665 days [36.4%]), and DP (371 days [20.3%]). With regard to seasons, the different SSC types were distributed as follows: spring, DM (193 days [42.0%]); summer, MT (193 days [42.0%]); fall, DM (241 days [53.0%]); and winter, DP (228 days [50.6%]). A stronger association between RAs and temperature was observed overall (1.28% [95% CI; 0.04%, 2.53%]), and in spring (2.10% [0.62%, 3.60%]). According to SSC type, increased associations between RAs and temperature were detected with DM (2.20% [0.47%, 3.97%]) condition during spring and DT (2.41% [0.22%, 4.64%]) during fall. In addition, on re-categorizing the SSC types from a temperature or humidity standpoint, a modified association was observed especially in children aged ≤4 years and 5–9 years during spring. ConclusionsUsing temperature variability to distinguish seasonal characteristics, we found that the relationship between temperature and RAs in children varied by SSC type. Application of the SSC scheme as an integrative approach may assist in gaining an understanding of seasonal characteristics and health effects due to temperature change.