Objective: The objective was to provide a systematic review of the screening accuracy of both versions of the Geriatric Depressions Scale (GDS-30, GDS-15).
Method: An electronic search was performed by using Medline, Embase, Cinahl, Psyndex and the Cochrane library. The selection and examination of papers were performed by two reviewers independently.
Results: Among the 42 papers included, important methodological aspects such as sampling methods or blinding of research workers often were not reported. For both GDS versions, similar validity indices were found (GDS-30: sensitivity 0.753, specificity 0.770; GDS-15: sensitivity 0.805, specificity 0.750). Using comparative studies based on the identical samples, both GDS versions showed significantly better validity indices than the 'Yale-1-question' screen, but were similar to the 'Center for Epidemiological Studies Depression scale' (CES-D).
Conclusion: The GDS does not show a better criterion validity than the CES-D, but methodological limitations of primary studies hamper the generalizability of pooled analyses.