精神分裂症认知功能成套测验中文版临床信度及效度的研究

摘要

目的 对中文版精神分裂症认知功能成套测验(MCCB)的信度及效度进行临床测试.方法 对122例符合美国精神障碍诊断与统计手册第4版精神分裂症诊断标准的住院患者(患者组)进行MCCB测验,4周后重测,同期接受威斯康星卡片分类测验(WCST)、瑞文推理测验(RAVEN)、色词测验(Stroop)及阳性和阴性症状量表(PANSS)测查;并与122名性别、年龄和文化程度与患者组相匹配的社区正常人(对照组)进行比较.结果 (1)MCCB重测相关系数为0.88,P<0.001;(2)评定者间组内相关系数为0.97,P<0.001;(3)MCCB的A、B版本间的复本相关系数为0.64～0.74,P<0.001;A、B版本间的差异无统计学意义(P>0.05);(4)患者组各个分测验得分均低于对照组(P<0.001);逻辑回归分析,用MCCB区分精神分裂症患者与正常人,符合率达到84.8%(P<0.001),敏感性83.6%,特异性86.1%;(5)关联效度:MCCB与WCST、RAVEN和Stroop呈显著性相关(r=0.54～0.55),P<0.001;(6)结构效度:验证性因素分析证明中文版MCCB与英文版7个维度结构模型拟合良好;(7)MCCB平均完成时间为(58±10)min,耐受性和操作性达到中等偏上水平.结论 中文版MCCB的重测信度、评定者间信度、同质性信度、复本信度、关联效度、结构效度和效标效度等指标满足心理测量学要求,MCCB作为精神分裂症患者认知功能疗效评估的新标准,值得进一步修订和完善.%Objective To evaluate the reliability and validity of the Chinese version of Measurement and Treatment Research to Improve Cognition in Schizophrenia (MATRICS) Consensus Cognitive Battery (MCCB). Methods One hundred and twenty-two inpatients with schizophrenia according to the DSM-IV criteria were measured with the MCCB, and re-measured after 4 weeks. The patients were simultaneously measured with the WCST, RAVEN, Stroop and PANSS to test validity. The controls, including 122 normal community people matched to the patient in gender, age and education, were measured with the MCCB only. Results The test-retest reliability of MCCB composite score was 0.88 (P<0.001), and inter-raters reliability ICC of 0.97 (P<0.001). Correlation between A/B alternate forms of MCCB was 0.64-0.74 (P<0.001), with no significant difference between A/B forms (P>0.05). In comparison with normal control, the MCCB scores in schizophrenia were significant lower (P<0.001). Using MCCB scores to classify schizophrenia and normal people by Logistic regression, the correct rate was 84.8%, with sensitivity of 83.6%, and specificity of 86.1%. All the subtests and composite scores of MCCB significantly correlated with the WCST, RAVEN and Strcop scores (P<0.05), which indicated better concurrent validity of MCCB. As to the construct validity, confirmatory factor analysis indicated that seven domains theory model of MCCB was well fired in Chinese version. The tolerability and practicabihty of MCCB were acceptable, and the average performance time was (58±10) minutes. Conclusion The test-retest, inter-raters, consistent and A/B alternate form reliability, concurrent, structure and diagnostic validity of the Chinese version of MCCB meet the requirements of psychometrics, and could be a promising tool for assessing cognitive deficit of schizophrenia patients in China.