基于GIS的青海省结核病时空分布特征研究

Temporal and spatial distribution of tuberculosis in Qinghai Province based on GIS

目的了解青海省结核病的空间分布特征和趋势,探究其分布的聚集性和发病热点,为青海省结核病的防治提供依据。方法采用空间流行病学研究方法及Arc GIS 10.2,对青海省2011-2015年结核发病资料进行专题地图输出、空间自相关分析、高/低聚类分析、热点分析和Kriging插值预测。结果青海省结核病5年平均登记发病率地理分布图显示:发病率高的县是曲麻莱、玛多县、杂多县、久治县,低发病率主要分布在格尔木市、乌兰县、平安县、湟中县、互助县。青海省结核病高发病率在东西方向和南北方向都有趋势存在,结果与结核地区分布图相符合。全局空间自相关分析结果显示:Moran’s I指数为0.32,Z=7.22,P=0.001;高/低聚类分析显示:Getis Gi系数为-0.13,Z=-2.00,P=0.045;热点分析显示:青海省结核病发病的热点集中在杂多县、曲麻莱县、称多县、达日县、玛沁县、班玛县。结论 2011-2015年青海省结核病地理分布呈现不均衡分布,全省结核病疫情分布呈现正空间自相关,主要以"低—低"模型聚集,并存在发病热点。

Objective To understand the tuberculosis（TB） spatial distribution characteristics and trends of Qing- hai Province and to explore the distribution of hot type and the onset of aggregation, so as to provide evidence for the prevention and control of tuberculosis in Qinghai Province. Methods Data of tuberculosis from 2011 to 2015 in Qinghai were collected and spatial correlation analysis, high / low clustering analysis, hotspot analysis and Kriging interpolation prediction were conducted by spatial epidemiological techniques. Results The results based on five-year average morbidity showed that geographical areas with high incidence of TB in Qinghai were counties of Qumalai, Maduo, Zaduo, Jiuzhi. Low incidence areas were Golmu City, Wulan County, Ping＇ an, Huangzhong and Huzhu. Trend figures showed that high inci- dences of TB in Qinghai existed in both east-west direction and the north-south direction, which were consistent with the ge- ographical distribution. Spatial autoeorrelation analysis showed that： Moran＇ s I index was 0. 32, Z = 7.22, P = 0. 001 ; High/Low clustering analysis showed that the Getis Gi coefficient was 4）. 13, Z = - 2. 00, P = 0. 045 ; hot spot analysis showed that most hot spots of TB in Qinghai were concentrated in counties of Zaduo, Qumalai, Chengduo, Dari, Maqing and Banma. Conclusions The geographical distribution of TB in period of 2011 to 2015 in Qinghai was uneven and ap- peared a positive spatial autocorrelation and displayed a low-low model and a hot spot.