基于灵敏度和特异度任意赋权的加权比数积  

Weighted Odds Product Based on Weighted Sensitivity and Specificity

作　　者：李丹玲[1] 段重阳[1] 邓居敏[1] 陈平雁[1] 

机构地区：[1]南方医科大学生物统计学系,510515

出　　处：《中国卫生统计》2016年第2期201-204,共4页Chinese Journal of Health Statistics

基　　金：国家自然科学基金项目(81273191;81402759);全国统计科学研究计划项目(2013LY059);广东省医学科研基金(B2013222)资助;广东省自然科学基金(2015A030310279)

摘　　要：目的建立一种基于灵敏度(SEN)和特异度(SPE)任意赋权的加权比数积统计方法。方法加权比数积(φ_w)的构造满足以下三个原则:灵敏度和特异度的权重(w)之和为1,0≤w≤1;满足特殊性:当灵敏度和特异度等权时,即w=0.5时,加权比数积φ_w等于比数积φ;加权比数积φ_w的取值范围与比数积φ的取值范围相同,为[0,+∞]。所构造的加权比数积为:φ_w=(SEN/(1-SEN))~2w·(SPE/(1-SPE))^(2(1-w)),(0≤w≤1)。根据中心极限定理,推导出lnφ_w的标准误和两个lnφ_w比较的Z统计推断方法,进一步推导出权重w的变化对检验统计量Z的影响。结果所构造的加权比数积满足上述构造三个原则。结论本研究所建立的加权比数积方法解决了应用中对灵敏度和特异度有不同赋权要求的问题,为诊断试验评价提供了新的工具。Objective To develop a weighted odds product（ φw） method for evaluating and comparing diagnostic tests based on weighted sensitivity and specificity. Methods Three principles of constructing weighted odds product φw are as follows：firstly,the sum of two weights which are attached to the sensitivity and specificity should equal to 1; secondly,φw equals to φ when the sensitivity and specificity have the same weights. finally the range of possible values of φw is within [0,＋ ∞ ],which is the same as the odds product φ. Then,the φw is defined by φw =（SEN/（1- SEN）） ^2w ·（ SPE/（1- SPE））^2（ 1- w）（ 0≤w≤1）. According to the central- limit theorem,we obtain the standard error of lnφw and propose a statistical inference method to compare two weighted indexes. Furthermore,we also deduce the test statistics Z can be either a monotonously increasing /decreasing function or non- monotone function of the weight w under different conditions. Results The proposed φw satisfied the above-mentioned three principles. Conclusion For different weights attached to the sensitivity and specificity,φw can be used to deal with such kinds of problems as provide a new and practical tool to evaluate diagnostic tests.

关 键 词：加权比数积 权重 灵敏度 特异度 

分 类 号：R195[医药卫生—卫生统计学]