煤矿区大气污染与儿童呼吸道疾病住院人数的时间序列分析

目的评价神木县大气污染与儿童呼吸道感染住院人数的关系。方法收集2009年11月至2012年10月间神木县的气象因素（气温、相对湿度）和环境因素：二氧化硫（SO，）、二氧化氮（NO：）和可吸入颗粒物（PM10），以及16岁及以下儿童的住院资料。利用SAS9．3对气象、环境因素和住院人数资料进行描述性分析，并判断主要污染物。利用R3．0．1统计软件的建立Poisson回归的广义相加模型，引入大气污染物浓度及气象因素拟合线性模型，并考虑其滞后效应，评价主要大气污染物对儿童呼吸道感染住院的相对危险度。结果神木县的首要污染物为PM10，PM10与空气污染指数（API）的相关系数为0．917，P〈0．01。在控制了长期趋势、“星期几效应”、气象因素及其它污染物的影响后发现，首要大气污染物PM10浓度每上升10μg／m3时，每日呼吸道感染住院患儿的上呼吸道相对危险度（RR）及其95％可信区间（95％cx）为1．002（0．996～1．007），t=0．589，P=0．556；下呼吸道为1．000（0．995～1．005），t=-0．007，P=0．995。当日或滞后1～10日内PM10浓度变化均不能显著改变儿童上、下呼吸道感染的住院率。结论神木县首要污染物PM10浓度变化对儿童上、下呼吸道感染的住院情况无明显影响。

Time series analysis of air pollution in coal mining area and number of hospitalized children with respiratory tract illness

Objective To analyze the relationship between air pollution in Shenmu County and the number of hospitalized children with respiratory tract infections. Methods Data regarding meteorological factors （temperature and relative humidity） and environmental factors including SO2, NO2 and particulate matter 10 （PM10） in Shenmu County were collected, and the data regarding hospitalization of children under 16 years of age were also collected during the period of November 2009 to October 2012. SAS 9.3 was used to describe and analyze meteorological and environmental factors and the number of hospitalized patients and to determine the main air pollutants. Using statistical software R 3. 0. 1, a generalized additive Poisson regression model was established, the linear fitting models of the air pollutant concentrations and meteorological factors were introduced considering lag effect, and the relative risk of the main atmospheric pollutants on children hospitalization because of respiratory tract infections was evaluated. Results Primary air pollutant in Shenmu County was PM10 and its Pearson eoiTelation coefficient with air pollution index （API） was 0. 917 （P 〈 0.01 ）. After controlling long-term climate trend, ＂week day effect＂, meteorological factors and impact of other contaminants, it was found that when PM10 concentration increased by 10μg/m3, the RR and 95 % CI of hospitalized children with upper respiratory tract infection was 1. 002 （0. 996-1. 007 ） （ t = 0. 589, P = 0. 556） and 1. 000 （0. 995-1. 005 ） （ t = - 0. 007, P = 0. 995 ） for lower respiratory tract infection. Change of PMIO concentration on the same day or during the lag of 1 to 10 days had no significant effect on children hospitalization rate because of upper and lower respiratory tract infections. Conclusion In Shenmu County the main pollutant PMIO concentration has no significant effect on children hospitalization rate because of upper and lower resoiratorv tract infections.