抑郁症诊断筛查量表的编制及其效度检验

Development and validity of a diagnostic screening instrument for depression

目的为了提高国内躯体疾病病人抑郁症的识别率,制定一个适合在中国使用的抑郁症诊断筛查量表并评估其效度。方法在《DSM-IV轴I障碍用临床定式检查病人版》(SCID)的“重性抑郁发作”基础上多次修订编制“抑郁症诊断筛查量表”。预试验由受过培训的精神科护士对二级综合医院的99例住院病人进行了评估。结果显示,评定员间针对有无重性抑郁发作的一致性极好(ICC=1.00);其中85例完成了盲法重测,重测一致性良好(ICC=0.91)。由受过培训的精神科护士采用筛查量表对分层、随机整群抽样的北京35家一级、8家二级和7家三级综合医院年龄≥15岁、病情允许且交流无明显困难的2877例顺序就诊的门诊病人和调查期间在院的2925例住院病人进行调查,每个病例的筛查时间为15～20min。对所有筛选阳性和随机选择的10%筛选阴性的病人共565例由不知道筛选结果的精神科医师进行SCID检查。结果护士使用的抑郁症筛查量表诊断为重性抑郁发作与精神科医师使用的金标准SCID诊断为重性抑郁障碍的准确度为96.8%,Kappa值0.87,敏感度98.6%,特异度96.6%,假阳性率3.4%,假阴性率1.4%。结论抑郁症筛查量表的敏感性和特异性高,信效度好,适合非精神科医师使用,可以在抑郁症流调和临床中使用以代替复杂的SCID检查。

Objective To develop a screening instrument to improve the identification of depression in persons with physical illnesses in China and to assess its validity. Methods We developed and revised a culturally sensitive screening instrument for depression based on DSM-IV diagnostic criteria of major depressive disorder. During the pilot test for the screening instrument psychiatric nurses who administered the Scale to 99 secondary-level general hospital inpatients had excellent inter-rater reliability （ ICC = 1.00） and when blindly re-administered to 85 of the patients the test- retest reliability was also excellent （ ICC = 0. 91 ）. In the main study a stratified random sample of 35 primary, 8 secondary and 7 ter- tiary general hospitals in Beijing was identified and trained psychiatric nurses administered the screening instrument to 2, 877 sequentially identified outpatients and 2,925 inpatients hospitalized at the time of the survey who were 15 years of age or older, well enough to complete the interview and who did not have serious difficulties communicating. The interview took 15～20 minutes to complete. All subjects screened positive and a random selection of 10% of those screened negative （565 subjects in total） were administered the ＇Structured Clinical Interview for DSM-IV Axis I Disorders＇ （SCID） by a trained at- tending-level psychiatrist who did not know the results of the screening interview. Results The nurses＇ screening result of ＇Major Depressive Episode＇ accurately predicted a ＇gold-standard＇ psychiatrist＇s SCID diagnosis of ＇Major Depressive Disor- der＇ in 96. 8% of cases （ Kappa = 0. 87） ; the sensitivity was 98.6%, specificity 96.6%, false positive rate 3.4%, and false negative rate 1.4%. Conclusions The screening instrument we developed has good reliability, validity, sensitivity and specificity and can be effectively used by non-psychiatrists, so it is suitable for use as an alternative to the more complex SCID examination in epidemiological or clinical studies of depression.