边缘效应偏倚对疾病空间分布趋势面分析结果的影响  

The margin effect bias in trend-surface models and its impact on the results of trend-surface analysis

作　　者：薛付忠[1] 吴晓云 王洁贞[1] 张际文[1] 

机构地区：[1]山东大学公共卫生学院,山东济南250012 [2]山东省平邑县卫生防疫站,山东平邑273300

出　　处：《实用医药杂志》2003年第10期766-769,共4页Practical Journal of Medicine & Pharmacy

摘　　要：目的探讨边缘效应偏倚对趋势面分析结果的影响。方法通过比较基准面和基准面内特定监测区的同阶趋势面模型趋势等值线图的变化趋势,展现边缘效应偏倚对趋势面分析结果的影响,进而应用缘冲区对比法控制特定监测区内趋势面模型的边缘效应偏倚。结果当趋势面模型阶次较高时边缘效应偏倚也较严重,加入缓冲区可有效地控制边缘效应偏倚。结论当趋势面的阶次很低时,可不考虑“边缘效应偏倚”。但当趋势面模型的阶次较高时应当用缓冲区处理;不便于设立缓冲区时,解释研究区域边缘处的结果时要谨慎下结论。Objective To explore the margin effect bias in trend-surface models and its impact on the results of trend-surface analysis.Methods By comparing with the trend contour plot of the same order trend -surface models between the benchmark surface and the special area of surveillance of disease in the benchmark surface,the impaction of margin effect bias to the result of trend-surface analysis was proved;and by setting up the buffer area around the special area of surveillance,the margin effect bias of trend-surface models in the special area of surveillance was controlled.Results When the order of the trend-surface model is bigger,the impaction of margin effect bias to the results of trend-surface analysis is stronger,setting up the buffer area around the special area of surveillance can control the impaction of margin effect bias effectively.Conclusion In order to controlling the margin effect bias,the buffer area around the special area of surveillance of disease would be set up,but when the buffer area could not be set up,the modeling of trend-surface analysis in the margin area should be carefully concluded.

关 键 词：地理流行病学 趋势面模型 边缘效应偏倚 

分 类 号：R181.2[医药卫生—流行病学]