调查点不足偏倚对疾病空间分布趋势面分析结果的影响  

The Bias of Surveying Points Shortage in Trend-surfce Models and its Impaction to the Results of Trend-surface Analysis

作　　者：王发银 薛付忠[2] 王洁贞[2] 张玉军 

机构地区：[1]山东省平邑县卫生防疫站,山东平邑273300 [2]山东大学公共卫生学院,山东济南250012 [3]山东省临沂市卫生防疫站,山东临沂276001

出　　处：《预防医学文献信息》2004年第1期3-6,75,共5页Liferatue and Information On Preventine Medicine

基　　金：国家自然科学基金资助项目 (30 0 70 6 77)

摘　　要：[目的 ]探讨调查点不足偏倚对趋势面分析结果的影响。 [方法 ]以基准面为基础 ,在保证调查点分布均匀化的前提下 ,依次减少调查点个数 ,并拟合同阶趋势面模型 ,观察模型的趋势等值线图和剩余等值线图 ,评价随调查点的递减 ,趋势面形态的变化情况。 [结果 ]随调查点数的递减 ,趋势等值线图的偏倚较小 ,但剩余等值线图的偏倚严重。 [结论 ]当用趋势面模型反映疾病发病水平在大尺度范围内的空间变化趋势时 ,用较少的调查点就能反映出来 ,不至于发生严重偏倚 ,但调查点数必须大于多项式的项数 ;当用趋势面模型寻找有地理流行病学意义的异常区域时 ,调查点太少会发生严重偏倚。[Objective]To explore the bias of surveying points shortage in trend surface models and its impaction to the results of trend surface analysis [Methods] By compare with the trend contour plot and residual contour plot of the same order trend surface models between the benchmark surface and the surfaces, prove the impaction of the bias of surveying points shortage to the results of trend surface analysis [Results]With the number of surveying points being descended, the bias of surveying points shortage is little in the trend contour plot , but it is large in the residual contour plot [Conclusions] Using the trend surface models to survey the spatial trend of disease in a large scale , the number of surveying point may be few; but the number of surveying point should be enough to survey the geographical abnormal area of disease

关 键 词：调查 偏倚 趋势面模型 疾病 趋势面分析 地理流行病学 影响 等值线图 个数 结论 

分 类 号：R181.35[医药卫生—流行病学] G633[医药卫生—公共卫生与预防医学]