摘要目的 研究蒙哥马利-艾森贝格抑郁量表(MADRS)在重性抑郁障碍患者中应用的信效度和疗效敏感性.方法 由3位经过培训的精神科主治医师对122例目前重性抑郁障碍(DSM-Ⅳ)患者进行MADRS、Hamilton抑郁量表-17项版本(HAMD)和临床总体印象量表-疾病严重程度分量表(CGI-S)评定;12例患者进行MADRS评定者一致性测试;47例患者在抗抑郁剂治疗第2,4,6,8周时进行上述量表评定;采用相关分析、信度分析和计算效应值(ES)评价MADRS的信效度和疗效敏感性.结果 MADRS评定者间信度为0.954;基线10个项目与总分之间的相关系数在0.445～0.770之间(P＜0.01),平均相关系数为0.629;量表Cronbach α系数为0.847;与HAMD总分和CGI-S分校标关联效度分别为0.853和0.672(均P＜0.01);治疗第2、4、6、8周MADRS重测信度为0.737,0.651,0.543,0.524(均P＜0.01);以MADRS为临床终点指标的ES高于HAMD(第2,4,6和8周分别为0.41,0.40,0.87,0.72,1.14,0.88,1.20,0.96).结论 MADRS在重性抑郁障碍患者中使用具有较好的信度和效度,作为疗效评估指标比HAMD敏感.

abstractsObjective To examine the reliability,validity and sensitivity of Montgomery-Asberg Depression Rating Scale (MADRS) for patients with current major depression disorder (MDD). Methods One hundred and twenty-two current MDD (DSM-Ⅳ) patients were administered with MADRS, Hamilton Rating Scale for Depression-17 item version (HAMD) and Clinical Global Impression of Severity (CGI-S) at baseline, 12 patients were selected to complete rater agreement test,and 47 patients receiving antidepressant treatment were followed up at 2,4,6 and 8 week and administered with MADRS and HAMD. Correlation analysis, reliability analysis and effect size (ES) calculation were used to determine the reliability,validity and sensitivity to changes during drug treatment. Results Intra rater reliability for MADRS was 0. 954. Baseline item-total score correlations were between 0. 445 and 0. 770 (P ＜ 0. 01 ), and the average correlation was 0. 629. The Cronbach α coefficient was 0. 847. The criterion related validity with HAMD and CGI-S was 0. 853 and 0. 672 (P＜0.01) ,respectively. The re-test reliability for MADRS at 2,4,6 and 8 week was 0. 737 ,0. 651,0. 543 and 0. 524 (P＜0. 01 ) ,respectively.MADRS had higher ES than HAMD when taken as clinical endpoint outcome measurement (0.41 vs 0.40,0.87 vs 0. 72,1.14 vs 0. 88,1.20 vs 0. 96 for 2nd,4th,6th and 8th week, respectively). Conclusion MADRS has good reliability and validity for patients with MDD. It is more sensitive to assess drug effect than HAMD.