Kessler Ronald C, Avenevoli Shelli, Green Jennifer, Gruber Michael J, Guyer Margaret, He Yulei, Jin Robert, Kaufman Joan, Sampson Nancy A, Zaslavsky Alan M, Merikangas Kathleen R
Drs. Kessler, Green, He, and Zaslavsky, Mr. Gruber, and Ms. Sampson are with the Department of Health Care Policy, Harvard Medical School; Dr. Avenevoli is with the Division of Developmental Translational Research, National Institute of Mental Health; Dr. Guyer is with the Massachusetts Mental Health Center; Dr. Kaufman is with the Department of Psychiatry, Yale Medical School; and Dr. Merikangas is with the Genetic Epidemiology Branch, Intramural Research Program, National Institute of Mental Health.
Drs. Kessler, Green, He, and Zaslavsky, Mr. Gruber, and Ms. Sampson are with the Department of Health Care Policy, Harvard Medical School; Dr. Avenevoli is with the Division of Developmental Translational Research, National Institute of Mental Health; Dr. Guyer is with the Massachusetts Mental Health Center; Dr. Kaufman is with the Department of Psychiatry, Yale Medical School; and Dr. Merikangas is with the Genetic Epidemiology Branch, Intramural Research Program, National Institute of Mental Health.
J Am Acad Child Adolesc Psychiatry. 2009 Apr;48(4):386-399. doi: 10.1097/CHI.0b013e31819a1cbc.
To report results of the clinical reappraisal study of lifetime DSM-IV diagnoses based on the fully structured lay-administered World Health Organization Composite International Diagnostic Interview (CIDI) Version 3.0 in the U.S. National Comorbidity Survey Replication Adolescent Supplement (NCS-A).
Blinded clinical reappraisal interviews with a probability subsample of 347 NCS-A respondents were administered using the Schedule for Affective Disorders and Schizophrenia for School-Age Children (K-SADS) as the gold standard. The DSM-IV/CIDI cases were oversampled, and the clinical reappraisal sample was weighted to adjust for this oversampling.
Good aggregate consistency was found between CIDI and K-SADS prevalence estimates, although CIDI estimates were meaningfully higher than K-SADS estimates for specific phobia (51.2%) and oppositional defiant disorder (38.7%). Estimated prevalence of any disorder, in comparison, was only slightly higher in the CIDI than K-SADS (8.3%). Strong individual-level CIDI versus K-SADS concordance was found for most diagnoses. Area under the receiver operating characteristic curve, a measure of classification accuracy not influenced by prevalence, was 0.88 for any anxiety disorder, 0.89 for any mood disorder, 0.84 for any disruptive behavior disorder, 0.94 for any substance disorder, and 0.87 for any disorder. Although area under the receiver operating characteristic curve was unacceptably low for alcohol dependence and bipolar I and II disorders, these problems were resolved by aggregation with alcohol abuse and bipolar I disorder, respectively. Logistic regression analysis documented that consideration of CIDI symptom-level data significantly improved prediction of some K-SADS diagnoses.
These results document that the diagnoses made in the NCS-A based on the CIDI have generally good concordance with blinded clinical diagnoses.
报告基于完全结构化的由外行人实施的世界卫生组织综合国际诊断访谈（CIDI）3.0版，对美国全国共病调查复制青少年补充调查（NCS-A）中终生DSM-IV诊断进行临床重新评估研究的结果。
采用学龄儿童情感障碍和精神分裂症量表（K-SADS）作为金标准，对347名NCS-A受访者的概率子样本进行盲法临床重新评估访谈。DSM-IV/CIDI病例进行了过度抽样，对临床重新评估样本进行加权以调整这种过度抽样。
CIDI和K-SADS患病率估计值之间总体一致性良好，尽管CIDI对特定恐惧症（51.2%）和对立违抗障碍（38.7%）的估计值明显高于K-SADS估计值。相比之下，CIDI中任何障碍的估计患病率仅略高于K-SADS（8.3%）。对于大多数诊断，发现CIDI与K-SADS在个体水平上有很强的一致性。接受者操作特征曲线下面积是一种不受患病率影响的分类准确性度量，任何焦虑障碍为0.88，任何情绪障碍为0.89，任何破坏性行为障碍为0.84，任何物质障碍为0.94，任何障碍为0.87。尽管酒精依赖以及双相I型和II型障碍的接受者操作特征曲线下面积低得不可接受，但这些问题分别通过与酒精滥用和双相I型障碍合并得到解决。逻辑回归分析表明，考虑CIDI症状水平数据显著改善了对一些K-SADS诊断的预测。
这些结果表明，NCS-A中基于CIDI做出的诊断与盲法临床诊断总体上具有良好的一致性。
