Strand Bjørn Heine, Dalgard Odd Steffen, Tambs Kristian, Rognerud Marit
Norwegian Institute of Public Health, Nydalen, Oslo.
Nord J Psychiatry. 2003;57(2):113-8. doi: 10.1080/08039480310000932.
A great number of questionnaires and instruments have been developed in order to measure psychological distress/mental health problems in populations. The Survey of Level of Living in 1998 conducted by Statistics Norway used both Hopkins Symptom Checklist (SCL-25) and the Short Form 36 (SF-36), including the five-item mental health index (MHI-5). Five-item and 10-item versions of the SCL-25 have also been used in Norwegian surveys. The purpose of this study was to investigate the correlation between the various instruments, and to assess and to compare psychometric characteristics. A random sample of 9735 subjects over 15 years of age drawn from the Norwegian population received a questionnaire about their health containing SCL-25 and SF-36. Response rate was 71.9%. Reliability of the SCLs and MHI-5 were assessed by Cronbach alpha. The scores from full and abbreviated instruments were compared regarding possible instrument-specific effects of gender, age and level of education. The correlations between the instruments were calculated. The capacity of the various instruments to identify cases was assessed in terms of sensitivity, specificity, predictive values, receiver operating characteristics (ROC) and area under the curve (AUC). The reliabilities were high (Cronbach alpha>0.8). All instruments showed a significant difference in the mean scores for men and women. The correlation between the various versions of SCL ranged from 0.91 to 0.97. The correlation between the MHI-5 and the SCLs ranged from -0.76 to -0.78. The prevalence rate was 11.1% for SCL-25 scores above 1.75 and 9.7% for scores below 56 in MHI-5. AUC values indicated good screening accordance between the measures (AUC>0.92). The results suggest that the shorter versions of SCL perform almost as well as the full version. The corresponding cut-off points to the conventional 1.75 for SCL-25 are 1.85 for SCL-10 and 2.0 for SCL-5. MHI-5 correlates highly with the SCL and the AUC indicate that the instruments might replace each other in population surveys, at least when considering depression. An operational advantage of the MHI-5 over the SCL instruments is that it has been widely used not only in surveys of mental health, but also in surveys of general health.
为了测量人群中的心理困扰/心理健康问题，已经开发了大量的问卷和工具。挪威统计局在1998年进行的生活水平调查中使用了霍普金斯症状清单（SCL - 25）和简短形式36（SF - 36），包括五项心理健康指数（MHI - 5）。SCL - 25的五项和十项版本也在挪威的调查中使用过。本研究的目的是调查各种工具之间的相关性，并评估和比较心理测量特征。从挪威人口中随机抽取的9735名15岁以上的受试者收到了一份包含SCL - 25和SF - 36的健康问卷。回复率为71.9%。通过克朗巴哈系数评估SCL和MHI - 5的信度。比较了完整和简化工具的得分，以探讨性别、年龄和教育水平可能对特定工具产生的影响。计算了各工具之间的相关性。根据敏感性、特异性、预测值、受试者工作特征（ROC）和曲线下面积（AUC）评估了各种工具识别病例的能力。信度较高（克朗巴哈系数>0.8）。所有工具在男性和女性的平均得分上均显示出显著差异。SCL不同版本之间的相关性在0.91至0.97之间。MHI - 5与SCL之间的相关性在 - 0.76至 - 0.78之间。SCL - 25得分高于1.75时的患病率为11.1%，MHI - 5得分低于56时的患病率为9.7%。AUC值表明这些测量方法之间具有良好的筛查一致性（AUC>0.92）。结果表明，SCL的简短版本表现几乎与完整版本一样好。与SCL - 25传统的1.75截止点相对应，SCL - 10为1.85，SCL - 5为2.0。MHI - 5与SCL高度相关，AUC表明这些工具在人群调查中可能相互替代，至少在考虑抑郁症时如此。MHI - 5相对于SCL工具的一个操作优势在于，它不仅在心理健康调查中广泛使用，而且在一般健康调查中也广泛使用。
