“游程总数”法诊断Meta分析中发表偏倚的准确度评价

Diagnostic Accuracy of "the Total Number of Runs" Method for Diagnosis of Publication Bias in Meta-analysis

目的评价利用文献的发表时间信息以"游程总数"构建的发表偏倚诊断方法("游程总数"法)的准确度,为其推广应用提供理论依据。方法运用Matlab 7.0软件产生有偏和无偏Meta分析研究模拟数据,计算并比较"游程总数"法与Egger法的灵敏度和特异度,对"游程总数"法的诊断准确度进行评价。结果 "游程总数"法与Egger法的灵敏度较低(分别为:0.3302±0.0574和0.360 7±0.1856),而特异度(分别为:0.7062±0.030 8和0.7896±0.0108)较好;"游程总数"法的灵敏度和特异度均低于Egger法,但相差幅度小;"游程总数"法的灵敏度与Meta分析中的子研究个数有关。结论新构建的"游程总数"法与Egger法具有相似的灵敏度和特异度,由于其特殊的理论基础,与基于漏斗图的常用方法联合使用对发表偏倚进行诊断,有利于改善目前灵敏度较低的现状。

Objective To evaluate the diagnostic accuracy of the new method making use of the information of publication date in which ＂the total number of runs＂ was used to diagnose the publication bias in meta-analysis,and then to provide a theoretical basis for the promotion and application of this new method.Methods The sensitivity and specificity of ＂the total number of runs＂ method and Egger method were computed and compared to evaluate the accuracy of the new method from the bias and no-bias simulated meta-analyses generated by the Matlab7.0 software.Results The sensitivities of ＂the total number of runs＂ method and Egger method were low（0.330 2±0.057 4 and 0.360 7±0.185 6,respectively）,but specificities of the two methods（0.706 2±0.030 8 and 0.789 6±0.010 8,respectively） were superior to sensitivities.The sensitivity and specificity of the new method was lower than Egger method,but the difference was small.The sensitivity of the new method was related to the number of sub-studies in meta-analysis.Conclusion The sensitivity and specificity of ＂the total number of runs＂ method are similar to those of Egger method.Because of its particular theoretical basis,if the new method is combined wtih common methods based on the funnel plot to diagnose publication bias in meta-analysis,the current status of low sensitivity for diagnosing publication bias can be changed.