地理研究 ›› 2010, Vol. 29 ›› Issue (1): 181-187.doi: 10.11821/yj2010010019

• 地球信息科学 • 上一篇    下一篇

应用地理信息系统探测消化道癌症死亡率空间聚集性

戚晓鹏1,2, 周脉耕1, 胡以松1, 王黎君1, 葛辉1, 庄大方2, 杨功焕1   

  1. 1. 中国疾病预防控制中心, 北京 100050;
    2. 中国科学院地理科学与资源研究所, 北京 100101
  • 收稿日期:2009-03-20 修回日期:2009-07-06 出版日期:2010-01-20 发布日期:2010-01-20
  • 通讯作者: 杨功焕(1949-), 女,研究员,教授,博士生导师,中国疾病预防控制中心副主任。主要研究方向为慢性病和危险因素的流行病学。E-mail:yangghuan@vip.sina.com E-mail:yangghuan@vip.sina.com
  • 作者简介:戚晓鹏(1975-), 女, 黑龙江省哈尔滨人,助理研究员,博士生。主要从事应用系统维护和疾病监测。 E-mail:caroline_qi@163.com
  • 基金资助:

    十一五科技支撑项目"淮河流域水污染与肿瘤的相关性评估研究"(2006BAI19B03);中国疾病预防控制中心青年科研基金项目(2009A202)

Spatial hotspot exploration on digestive tract cancer mortality with geographic information system

QI Xiao-peng1,2, ZHOU Mai-geng1, HU Yi-song1, WANG Li-jun1, GE Hui1, ZHUANG Da-fang2, YANG Gong-huan1   

  1. 1. Chinese Center for Disease Control and Prevention, Beijing 100050, China;
    2. Institute of Geographic Sciences and Natural Resources Research, CAS, Beijing 100101,China
  • Received:2009-03-20 Revised:2009-07-06 Online:2010-01-20 Published:2010-01-20

摘要:

根据研究区分村4种消化道癌症死亡监测数据和人口数据,描述癌症死亡率的空间分布规律,探测病例在空间上存在的聚集性热点,为进一步开展环境和人群监测提供参考依据。本研究采用基本的图层Voronoi处理技术以及全局空间自相关和空间热点探测的方法,通过绘制空间自相关系数图,描述不同空间尺度与消化道癌症死亡率自相关系数之间的关系。结合空间探测技术和癌症死亡率分布特点,确定空间探测的合理参数,应用地理信息系统,对消化道癌症在研究区的空间聚集热点进行探测。结果发现,研究区在4300m尺度存在有意义的显著空间正自相关,探测到3个消化道癌症高值聚集区,共58个村,每个聚集区平均人口在3万左右。3个聚集区癌症粗死亡率明显高于非聚集区和该县平均粗死亡率。空间热点的探测与分析,引入空间权重矩阵的概念,弥补传统统计学缺乏空间信息和空间关联的缺陷,为引起消化道癌症高发的危险因素探寻提供线索,是传统统计学的必要补充。

关键词: 消化道癌症, 空间分析, 地理信息系统

Abstract:

This paper describes the spatial distribution of cancer mortality and explores the spatial hotspot of death cases in the study area, based on the 4 kinds of digestive tract cancer death surveillance data and the population data. According to it, the environment and public surveillance will be held in the next step. With basic layer Voronoi technique, global Moran's Index method and spatial hotspot exploration, the spatial autocorrelation index graph was drawn using automatic multi-dimension exploration, which describes the relationship between the Moran's I and the distance. The accurate parameter was identified under the spatial analysis technique and the distribution character of cancer mortality, which was used to observe the spatial cluster in this county with GIS. It had a remarkable positive autocorrelation in the 4300 meters in space. At the same time, three hotspots were confirmed as high value cluster, including 58 villages and a population of about 30,000 in each cluster. The crude death rate in the hotspots is significantly higher than that in other areas and the average level of the county. The spatial hotspot exploration and analysis, which imported the spatial weight matrix, made up for the deficiency of traditional statistical method in spatial information and spatial correlation. It offered the evidence for making the risk factor of high cancer incidence much clearer. And it is the necessary makeup for the traditional statistics.

Key words: digestive tract cancer, spatial analysis, GIS