焦虑障碍的认知偏向矫正元分析：效果量、影响因素及中介检验

Meta-analysis on CBM for anxiety disorder: effect sizes, moderators and mediation

焦虑障碍的认知偏向矫正效果及作用机制是当前一个研究热点且有争议的主题。采用元分析、元回归分析并且组合两步结构方程元分析技术,探索认知偏向矫正对焦虑障碍干预的临床和亚临床有效性,分析可能的影响因素并考察认知偏向改变在认知偏向矫正(Cognitive Bias Modification,CBM)和焦虑症状之间的中介作用。元分析共纳入46篇文献,CBM对焦虑症状干预具有虽小但显著的效果量(包含非标准化测量:g=0.29;标准化测量:g=0.26,p<0.001;压力任务后焦虑症状g=0.34,p<0.001),对认知偏向有中等效果量(g=0.48,p<0.001);干预范式、出版特征、实验环境是焦虑症状CBM效果量的调节因素,焦虑障碍类型和实验环境对CBM认知偏向效果量有调节作用;认知偏向改变在CBM干预与焦虑障碍改变之间起显著的中介效应(z=3.13,SE=0.014,p<0.001),为部分中介。将来研究可关注提升CBM症状测量或者认知偏向测量工具的信效度、发展新干预范式,尝试CBM同其他干预方式相结合,进一步探索CBM理论基础和干预机制,提升干预效果。

The current update meta-analysis focused on the effect sizes of Cognitive Bias Modification （CBM） for anxiety disorder, and explored the potential moderators by subgroup analysis and hierarchical meta-regression. To understand the mechanism of CBM, mediation analysis by two-stage structural equation method （TSSEM） integrating meta-analysis and structural equation method is applied. The results demonstrate that （1） CBM may have small effects on anxiety symptom reduction at the risk of publication bias; however, CBM can modify cognitive bias better with a medium effect size. （2） The type of intervention, training environment, publication year and impact factor moderate the effect size of CBM on post-training anxiety. Training environment and the type of anxiety disorder are moderators of the effect size on cognitive bias. （3） Cognitive bias is the partial mediator by which CBM may have an influence on anxiety symptom.