Department of Psychological Medicine, Division of Academic Psychiatry, Institute of Psychiatry, Psychology and Neuroscience (IoPPN), King's College London, London SE5 8AF, UK.
CIBER Epidemiología y Salud Pública (CIBERESP), 28029 Madrid, Spain.
Int J Environ Res Public Health. 2020 Oct 29;17(21):7955. doi: 10.3390/ijerph17217955.
Though self-reported items (SRD, self-reported depression) are commonly used in health surveys and cohort studies, their metric properties as a depression indicator remain unclear. The aims were to evaluate the measurement properties of SRD using the Patient Health Questionnaire-8 (PHQ-8) as reference and to identify factors related to the agreement between both indicators.
Data from the European Health Interview Survey in Spain in 2014/2015 (n = 22,065) were analyzed. Two indicators of depression were considered: SRD based on two items yes/no (positive: both yes), and the PHQ-8 (positive ≥ 10). Socioeconomic factors and use of health services were considered as independent variables. The prevalence of depression, sensitivity, specificity, global agreement, and positive and negative predictive values (PPV and NPV) of SRDs were evaluated using the PHQ-8 as a reference. Logistic regression models were fitted to determine factors associated with the agreement between indicators.
The prevalence of depression was lower when assessed with PHQ-8 (5.9%) than with SRD (7.7%). SRD sensitivity and PPV were moderate-low (52.9% and 40.4%, respectively) whereas global agreement, specificity, and NPV were high (92.7%, 95.1%, and 97.0%, respectively). Positive agreement was associated with marital status, country of birth, employment status, and social class. Negative agreement was related to all independent variables except country of birth.
SRD items tend to overestimate the current prevalence of depression. While its use in health surveys and cohorts may be appropriate as a quick assessment of possible depression, due to their low sensitivity, its use in clinical contexts is questionable.
尽管自我报告项目（SRD，自我报告的抑郁）在健康调查和队列研究中经常被使用，但作为抑郁指标的度量特性仍不清楚。本研究旨在使用患者健康问卷-8（PHQ-8）作为参考来评估 SRD 的测量特性，并确定与两个指标之间一致性相关的因素。
对 2014/2015 年西班牙欧洲健康访谈调查的数据（n=22065）进行了分析。考虑了两种抑郁指标：基于两个是/否项目的 SRD（阳性：均为是）和 PHQ-8（阳性≥10）。将社会经济因素和卫生服务利用情况作为自变量。使用 PHQ-8 作为参考，评估 SRD 的抑郁患病率、敏感性、特异性、总体一致性以及阳性和阴性预测值（PPV 和 NPV）。拟合逻辑回归模型以确定与指标一致性相关的因素。
使用 PHQ-8 评估的抑郁患病率（5.9%）低于 SRD（7.7%）。SRD 的敏感性和 PPV 为中低水平（分别为 52.9%和 40.4%），而总体一致性、特异性和 NPV 较高（分别为 92.7%、95.1%和 97.0%）。阳性一致性与婚姻状况、出生国、就业状况和社会阶层有关。阴性一致性与除出生国外的所有自变量有关。
SRD 项目可能会高估当前抑郁的患病率。虽然它可能适用于健康调查和队列研究作为可能抑郁的快速评估，但由于其敏感性较低，在临床环境中的使用值得怀疑。
