关键词: 结核, 肺/预防和控制, 登记, 时空聚类分析, 地理信息系统, 湖南省

Abstract: Objective  To analyze the spatial distribution pattern of the active TB patients registered in Hunan province, and find out the gathered area, to provide theoretical basis for further study of the tuberculosis risk factors and prevention and control strategies.  Methods  There are a total of 475 125 cases of active TB registered between 2003 and 2011 in Hunan province, with the prefecture annual average registration rate of 80.1/100 000 (61.3/100 000-122.8/100 000). We collected the active TB registration number and average annual population in each prefecture between 2003 and 2011, analyzed the spatial distribution of average annual registration rate by SPSS 13.0 clustering analysis (hierarchical cluster), the aggregation characteristic of spatial, time and time-space distribution by SaTScan (Version 9.1.1) Possion distribution probability model, and mapping by MAPGIS 67 software.  Results  There was a descending trend of the spatial distribution of active TB patients registration from northwest to southeast in Hunan province between 2003 and 2011. Purely spacial analysis identified 4 clusters, the first cluster included Xiangxi, Zhangjiajie and Huaihua, and 85 065 patients were registered actually, while the expectation number of registration patients were 62 095，RR=1.45 (P<0.001), the second cluster was Loudi, 31 381 patients were registered actually, while the expectation number of registration patients were 27 043，RR=1.17 (P<0.001), the third cluster was Changde, and 47 831 patients were registered actually, while the expectation number of registration patients were 43 366，RR=1.11 (P<0.001), the fourth cluster was Yongzhou, and 41 837 patients were registered actually, while the expectation number of registration patients were 37 656，RR=1.12 (P <0.001). Purely temporal analysis identified there was a cluster between 2004 and 2006 in terms of notification on active tuberculosis patients, and 162 067 patients were registered actually, while the expectation number of registration patients were 151 859，RR=1.10 (P=0.001). The spatial analysis identified 4 clusters, the first cluster included Xiangxi, Zhangjiajie and Huaihua, and 41 695 patients were registered actually, while the expected registration number were 25 306，RR=1.71 (P<0.001), the second cluster was Yongzhou, 9518 patients were registered actually, while the expected number of 6334，RR=1.51 (P<0.001), the third was Loudi, 7223 patients were registered and the expected of 4589，RR=1.58 (P<0.001), the fourth was Changde, 23 058 patients were registered while expected of 19 149，RR=1.21 (P<0.001). Conclusion  There was a descending trend of the spatial distribution of active TB patients registration from northwest to southeast in Hunan province between 2003 and 2011. The notifications are not randomly distributed in space, time and timespace, and clusters did exist in Hunan. The cluster most likely existed in Xiangxi, Zhangjiajie and Huaihua, and gathered between 2004 and 2006.

Key words: Tuberculosis,pulmonary/prevention &, control, Registries, Space-time clustering, Geographic information systems, Hunan province