Study On Statistical Analysis System Of Epidemiological Dot Map

The epidemiological dot map is widely used in spatial epidemiology. Therefore it is essential to make full use of epidemiological dot map in geographic epidemiology and spatial epidemiology. The epidemiological dot map has much significance in the analysis of spatial distribution of diseases and indication of the epidemiological features. Additionally, it can be used to explore the spatial risk factors according to the spatial heterogeneity of diseases, thereby provide scientific strategies for the government to prevent diseases in certain regions. However, the previous studies on spatial analysis of disease were only based on spatial point pattern analytical methods used in geology or landscape ecology, and there is not a complete spatial point process analytical system to be used in spatial epidemiology. Some of the spatial point analytical methods currently used in analysis of diseases do not combine the essence of spatial epidemiology and they are not suitable to use in spatial analysis of diseases. As a result, this study establish a statistical analysis system to be used in spatial point pattern analysis of diseases according to principles of geographical epidemiology and the geographical epidemiology features of diseases, following the basis classical idea of the statistical analysis. It aims to enrich and develop the geographic theory and epidemiological method, and to provide a system idea for the statistical analysis of epidemiological dot map.The statistical analysis system of epidemiological dot map is as follows:1. Statistical description of the epidemiological dot mapBecause of the specialty of spatial data, traditional statistical description indexes and charts are no longer applicable in the statistical description of epidemiological dot map. The statistical description of the epidemiological dot map includes centrographic statistics, density index, indices of spatial autocorrelation, Convex Hull and the standard deviation ellipse. Similar to traditional statistical methods, the centrographic statistics contains center analysis (such as mean center, geometric mean, and median center), deviation analysis (such as standard distance and relative standard distance), convex hull and standard deviation ellipse which are similar to the classical statistical graph and used to directly show the distribution of cases. The same as the application of descriptive indices in classical statistical analysis, standard distance and mean center require the distribution of x and y, the location of the cases, is normal, and median center and quartile distance are applicable in the skewed distribution of x and y of the cases.2.Statistical inference of the epidemiological dot mapThe density-based methods of statistical analysis, also called analysis of the first property of spatial distribution, include statistical inference of the distribution of the cases, the probability function and kernel estimation of the spatial distribution of the cases.The distance-based methods of statistical analysis could indicate whether the distribution of the cases is random, aggregated or regular, as well as reveal the spatial characteristics at different spatial scales, describe the relationship between the distribution pattern of two different types of cases and the temporal dynamic distribution. These methods include nearest neighbor distance method, Ripley's K function method and the expansion of Ripley's K function methods (the dual variable model and space-time model).Spatial "hot spot" analysis, also called spatial cluster analysis, is used to detect highly risk areas of the cases. It is able to directly display the hotspot of the cases combined with the GIS software. Therefore it is very important to establish effective strategies and measures for the prevention and control of diseases in epidemiology. The major methods in detecting the spatial hot spots of diseases include nearest neighbor hierarchical clustering, risk-adjusted nearest neighbor hierarchical clustering.Conclusion:1. Statistical description of epidemiological dot map is very useful for the initial description of the diseases (cases) and their spatial distribution. We should use different statistical indices according to different research purposes. By this course, we can evaluate the effectiveness of disease prevention and health services, and optimize the allocation of health resources. Similar to classical statistical analysis, spatial description should also be the first step in the course of spatial data analysis. It can be used to reflect the whole characteristics of the spatial data, and find the overall spatial heterogeneity. Therefore it is the foundation and prerequisite of statistical inference.2. The density-based methods of statistical analysis use the characteristics of the density and frequency of cases in defined regions to study the spatial distribution of point patterns of cases. These methods, also called analysis of first-order property, can indicate the overall trend at the large scale of spatial data. However, the spatial information is gathered into the area units, so a lot of information is lost. Unlike the density-based methods of statistical analysis, the distance-based methods of statistical analysis eliminate the defect of gathering the spatial information into area units and have more significance in the analysis of the spatial data. These methods could analyze the spatial clusters of point pattern data in any scales in order to finely reveal the spatial distribution property of the cases. They are used to illustrate the second-order property. Spatial "hot spot" analysis reveals the relative high risk regions of the cases. It can be used to detect the geographical risk factors and provide guidance for prevention and control of diseases.3. This study makes use of analytical system of epidemiological dot map above to analyze the spatial structure of incidence of tuberculosis cases recruited in 2006 and 2007 of Qingdao, Shandong province using the from statistical description and inference. It could provide foundations to detect geographical risk factors of TB and furthermore, it can be used to set up effective strategies and measures for prevention and control of TB.4. The statistical analysis system established in this study enriches and develops the methods in spatial epidemiology, especially in the analysis of the epidemiological dot map. Therefore it is a useful complement in the quantity analysis of geographical epidemiology.