Validation of the five-item geriatric depression scale in elderly subjects in three different settings

Abstract

Objectives: To test the effectiveness of a five-item version of the Geriatric Depression Scale (GDS) for the screening of depression in community-dwelling older subjects, hospitalized older patients, and nursing home residents.

Design: A cross-sectional study.

Setting: A geriatric acute care ward, a geriatric outpatient clinic, and a nursing home.

Participants: One hundred eighty-one cognitively intact older subjects.

Measurement: All the participants had a comprehensive geriatric assessment including a neuropsychological evaluation by a geriatrician experienced in the management of depression. The five-item GDS was compared with the 15-item version of the GDS using the clinical diagnosis according to the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition criteria as the criterion standard. The sensitivity, the specificity, the overall accuracy, positive and negative predictive values, and positive and negative likelihood ratios were calculated. The agreement between each of two different versions of the GDS and the clinical diagnosis and the test-retest and the interrater reliability of the five-item scale were also evaluated.

Results: In the whole sample, 48.1% of the subjects were depressed. The five-item GDS had a sensitivity of 0.94 (0.91-0.98), a specificity of 0.81 (0.75-0.87), a positive predictive value of 0.81 (0.75-0.87), a negative predictive value of 0.94 (0.90-0.97), a positive likelihood ratio of 4.92 (4.39-5.5), and a negative likelihood ratio of 0.07 (0.06-0.08). The five-item GDS and the 15-item GDS showed a significant agreement with the clinical diagnosis of depression (kappa = 0.74 for both scales). The five-item GDS had good interrater reliability (kappa = 0.88) and test-retest reliability (kappa = 0.84). Similar values were obtained in each setting and in both sexes.

Conclusion: The five-item GDS is as effective as the 15-item GDS for the screening of depression in cognitively intact older subjects.