Psychometric comparison of single-item, short, and comprehensive depression screening measures in Korean young adults

Abstract

BackgroundIntegrating long depression-screening instruments into routine clinical practice and research studies is often impractical, necessitating short-item if not single-item measures with comparable psychometric properties. ObjectiveTo examine whether single-item or short depression-screening measures are comparable to a comprehensive screening measure in reliability (i.e., internal consistency and test–retest reliability) and validity (i.e., convergent, concurrent, and predictive validity) in Korean young adults within a Classical Testing Theory framework. MethodA total of 458 students from six nursing colleges in South Korea completed three depression measures: the 20-item Center for Epidemiologic Studies-Depression screening instrument (CES-D; comprehensive measure); the five-item Profile of Mood States-Brief depression subscale (POMS-B depression subscale; short measure); a single-item Likert measure; and a single-item numeric rating scale. Internal consistency reliability was tested by Cronbach's alpha and item-total correlations; test–retest reliability by intraclass correlation coefficient (ICC); convergent validity by correlation with the CES-D; concurrent validity by the correlation with perceived stress level and sleep quality; and predictive validity by receiver operating characteristic curve to predict the two groups with different depression levels. ResultsThe POMS-B depression subscale was comparable to the comprehensive CES-D scale in internal consistency reliability (alpha=.85); test–retest reliability (ICC=.76); and convergent (r=.81 with CES-D), concurrent (r=.64 with perceived stress level, r=.34 with sleep quality), and predictive validity (area under the curve=.88). The two single-item options were not comparable to the comprehensive CES-D. ConclusionThe short POMS-B depression subscale shows an acceptable balance between practical clinical and research needs and psychometric quality.