广义相加模型在创伤急诊人次与气象要素时间序列相关分析中的应用

Application of generalized additive models in the analysis of the relationship between daily trauma emergency department visit number and meteorological factors:a time-series analysis

探讨广义相加模型在分析气温、相对湿度等气象要素与医院日创伤疾病急诊人次相关性中的应用.收集北京3家综合性医院2008-2012年日创伤患者资料和同期北京市的气象和环境数据资料,根据赤池信息量准则,在控制长期趋势、短期波动、节假日效应以及污染条件等混杂因素的基础上,构建日创伤急诊就诊人次与气象要素之间的广义相加模型,并绘制暴露-反应关系曲线.经广义相加模型分析,医院日创伤急诊就诊人次与日均气温和气压表现出一定的线性正相关和负相关趋势,尤其是气温低于24.5℃时,气压介于1 007~1020hPa之间时,线性变化显著.与日均相对湿度和日照时数表现为复杂的非线性相关,当相对湿度为50%~70%时,相对危险度最高,当日照时数达到9 h时相对危险度最低.针对创伤疾病,利用广义相加模型进行气象要素的作用分析和拟合比较合理.

To explore the application of generalized additive models in the analysis of the relationship between air temperature,relative humidity and other meteorological elements and daily trauma emergency department visit number,daily trauma emergency department visits data of three comprehensive hospitals for the years 2008 to2012,and the same-period meteorological and environmental data in Beijing were collected.Then according to the principle of minimum Akaike information criterion and based on the control of long-term trend,short-term fluctuations,holiday effect and pollution conditions and other confounding factors,a the generalized additive model was constructed between the trauma emergency visits and meteorological factors,and the exposure curve reaction drawn.The analysis showed that the hospital daily trauma emergency visit number had a positive correlation with the daily average temperature and humidity and a negative correlation with the daily average pressure.Especially when the temperature was below 24.5 ℃ or the pressure was between 1007~1020 hPa,there would be a significant linear change.The association between the relative humidity and sunshine hours and daily trauma emergency visit number presented a complex nonlinear correlation when the relative humidity was 50%~70%,and the relative risk was the highest;when the sunshine duration lasted to 9 h the relative risk would be minimum.For traumatic diseases,it was more reasonable to use the generalized additive models to analyze and fit the effects of meteorological elements.