Objective: To estimate the prevalence rates of neural tube defects (NTDs) in Heshun County, Shanxi Province, China by Bayesian smoothing technique. Methods: A total of 80 infants in the study area who were diagnosed with NTDs were analyzed. Two mapping techniques were then used. Firstly, the GIS software ArcGIS was used to map the crude prevalence rates. Secondly, the data were smoothed by the method of empirical Bayes estimation. Results: The classical statistical approach produced an extremely dishomogeneous map, while the Bayesian map was much smoother and more interpretable. The maps produced by the Bayesian technique indicate the tendency of villages in the southeastern region to produce higher prevalence or risk values. Conclusions: The Bayesian smoothing technique addresses the issue of heterogeneity in the population at risk and it is therefore recommended for use in explorative mapping of birth defects. This approach provides procedures to identify spatial health risk levels and assists in generating hypothesis that will be investigated in further detail.
CHI Wen-xueWANG Jin-fengLI Xin-huZHENG Xiao-yingLIAO Yi-lan
This study is to assess the prevalence rates spatial pattern of neural tube defects with geographic information system and spatial filtering technique. A total of 80 infants who diagnosed from neural tube defects in the area being studied between 1998 and 2001 were analyzed. Firstly, the geographic information system (GIS) software ArcGIS was used to map the crude prevalence rates. Secondly, the data were smoothed by the method of spatial filtering. We evaluated that the effect of changes in spa-tial filtering radius size was assessed by creating maps based on various filtering radius sizes. The 3 miles or larger filtering radius gives better sec-tion variability than the 2 and 2.5 miles or smaller ones. The maps pro-duced by the spatial filtering technique indicate that prevalence rates in the villages in the southeastern region are to produce higher prevalence than that in the other regions. The smoothed maps based on Heshun County display a more adequate data representation than the raw prevalence rate map.
CHI WenxueWANG JinfengLI XinhuZHENG XiaoyingLIAO Yilan
The estimation of geographical attributes is a crucial matter for many real-world problems,and the issue of accuracy stands out when the estimation is used for between-regions comparison.In this work,our concern is area attribute estimation in a GIS environment.We estimate the area attribute value with a mean Kriging technique,and the probability distribution of the estimate is derived.This is the best linear unbiased observed spatial population mean estimate and can be used in more relaxed situations than the block Kriging technique.Both theoretical analysis and empirical study show that the mean Kriging technique outperforms the ordinary Kriging,spatial random sampling,and simple random sampling techniques in estimating the observable spatial population mean across space.
