基于时空重排扫描统计量的深圳市流行性腮腺炎聚集性研究

目的探讨深圳市流行性腮腺炎时空聚集性特征和变化趋势。方法应用回顾性时空重排扫描统计量和空间自相关系数分析深圳市2009年1月1日～2011年12月31日流行性腮腺炎时空聚集性,利用地理信息系统（geographic information system,GIS）技术表达相应的聚集区域。结果回顾性时空扫描分析确定了23个可能的发病聚集区域,一类聚集区域发生在2011年11月27日～12月26日,聚集范围包括沙头角和海山街道（P〈1.0×10^-17）,二类聚集区域包括22个,2009年探测到6个,包括4个聚集区域和2个聚集点,2010年探测到7个,包括1个聚集区域和6个聚集点,2011年探测到9个聚集点。呈现大范围聚集区域减少,小范围的聚集点逐渐增多的趋势。23个聚集区主要出现在4～7月份和11～2月份,与深圳市和广东省腮腺炎疫情高峰出现时段相一致。局部自相关分析发现的高-高聚集区域为：2009年海山、沙头角、莲塘、南头和龙岗街道;2011年盐田、海山、平湖和福永街道,结果与时空扫描结果基本一致。结论时空重排扫描统计量结合地理信息系统,能够更加直观、全面地展示了发病聚集区域,为以后开展针对性的预防控制措施,提供了科学参考依据。

A clustering study on mumps in Shenzhen based on space-time permutation scanning statistics

Objective To explore the spatiotemporal clusters of mumps in Shenzhen. Methods Retrospective space-time permutation scan statistics and Spatial autocorrelation analysis were applied to analyze mumps reporting data of Shenzhen from January 1,2009 to December 31,2011,and geographic information systems（ GIS） technology was used to show mumps clusters. Results A total of 23 significant mumps clusters were detected,including one most likely cluster and 22 secondary clusters. The most likely cluster was occurred in Shatoujiao and Haishan streets from November 27 to December 26,2011（ P 〈1. 0 × 10^-17）. Secondary clusters included four clustering areas and two clustering points in 2009,one clustering area and six clustering points in 2010,one clustering area and nine clustering points in 2011. The trendency showed that large-scale clustering areas were reducing,while small-scale clustering points were gradually increasing. All clusters mainly occurred on April to July and November to February,which were also the mumps epidemic peaks in Shenzhen and Guangdong Province. Local autocorrelation analysis indicated that high-value clusters were noted in Haishan,Shatoujiao,Liantang,Nantou,Longgang Streets in 2009,and Yantian,Haishan,Pinghu,Fuyong Streets in 2011,consistent with the results of space-time scan analysis. Conclusions Space-time permutation scan statistics together with geographic information system could be used to clearly and fully evaluate spatiotemporal clustering characteristics of infectious disease,and provide more valuable information for the disease control and prevention.