Cislaghi C, Braga M, Biggeri A
Istituto di Statistica Medica e Biometria, Università di Milano.
Epidemiol Prev. 1995 Jun;19(63):142-9.
Methods for the analysis of the spatial aggregation of health events has received growing attention under the pressure of public opinion concern and as tools for the identification of potential risk sources, for monitoring relevant geographical areas and, finally, for public health decisions. The development of statistical methods for the detection and localization of spatial clusters has mainly concerned individual data. This paper is aimed at describing one of the methods proposed for the identification of clusters in the case of information at individual level and at presenting its extension to grouped data. This method, the surface density estimation method using the Kernel approach, offers remarkable advantages in terms of simplicity of implementation and flexibility, this latter being an extremely important characteristic in the case of exploratory analyses. For exemplification purposes, the density estimation method has been applied to individual data concerning the spatial distribution of cerebral tumors in Campi Bisenzio (FI) and to the distribution of gastric cancer mortality in the municipalities around Arezzo and Pesaro.
在公众舆论关注的压力下，作为识别潜在风险源、监测相关地理区域以及最终用于公共卫生决策的工具，健康事件空间聚集性分析方法受到了越来越多的关注。用于检测和定位空间聚集的统计方法的发展主要涉及个体数据。本文旨在描述针对个体层面信息识别聚集提出的方法之一，并展示其对分组数据的扩展。这种方法，即使用核方法的表面密度估计方法，在实现的简易性和灵活性方面具有显著优势，而灵活性在探索性分析中是一个极其重要的特征。为了举例说明，密度估计方法已应用于关于坎皮比森齐奥（佛罗伦萨省）脑肿瘤空间分布的个体数据，以及阿雷佐和佩萨罗周边市镇的胃癌死亡率分布。
