OBJECTIVE: The aim of this study was to evaluate the validity and reliability of the Geriatric Depression Scale (GDS) in the versions of 30 and 15 items for the identification of suspicion of depression in the elderly in the infirmary of general practice of the Hospital Universitário Lauro Wanderley, in João Pessoa, Paraíba. METHODS: An observational and transversal study, with the application of GDS-30 and GDS-15 by a trained interviewer, was held at the same day the mental exam of the patients by a psychiatrist, according to the criteria of ICD-10. The concordance of the two versions of the GDS was evaluated by the test-retest method and inter-observers equivalence. RESULTS: The concordance between the application of the GDS-30 and the psychiatric exam was significant and moderated (kappa = 0.48, p = 0.04), while the GDS-15 didn't present statistics concordance (p = 0.62). The GDS-30 presented sensitivity of 83% and specificity of 57%, but the GDS-15 presented sensitivity of 50% and specificity of 62%. Scores of GDS-30 presented significant statistically correlation in the test-retest method and inter-observers equivalence, but the version of 15 items weren't reliable. CONCLUSION: The conclusion was that the GDS-30 is more sensitive and reliable than the GDS-15, however its sensitivity was inferior to the numbers related in other studies, while his specificity was not high. Therefore, the GDS-30 is useful to detect as negatives the patients who really do not have depression, however further studies are necessary to adjust the items of this scale to our clientele, in order to enhance its specificity and positive predictive value.
Old age health; depression; GDS
