Background: The prevalence of depression in individuals with coronary artery disease (CAD) is high. The Geriatric Depression Scale (GDS) is a broadly used tool to screen for depression in elderly individuals. In Brazil, the psychometric properties of the short version have not been adequately assessed. Objective: To evaluate the psychometric properties of the short version of the GDS in patients with CAD treated at a Cardiology Outpatient Clinic. Methods: The present is a cross-sectional study that assessed 209 elderly individuals (>= 65 years) with CAD using the GDS-15, the Cumulative Illness Rating Scale for Geriatrics (CIRS-G), the Brazilian OARS Multidimensional Function Assessment Questionnaire (BOMFAQ) and The Cambridge Examination for Mental Disorders of the Elderly (CAMDEX). The internal consistency of the GDS-15 was calculated through the KR-20. A factorial analysis of this scale was carried out. The CDS-15 scores were compared with the diagnoses of depression (DSM-IV) for the validity of criteria. At the analysis of concurrent validity the same scores were correlated with those of the CAMDEX, mini-mental state examination (MMSE), Cambridge Cognitive Examination (CAMCOG) and BOMFAQ depression scales. Results: Clinical depression was diagnosed in 35.71% of the sample assessed according to the DSM-IV. For the diagnosis of major depression or dystimia, the cutoff 5/6 presented moderate accuracy (AUROC = 0.84), sensitivity of 79.92% and specificity of 78.29%. The internal consistency was 0.80. At the factorial analysis, three obtained factors explained 52.72% of the total variance that was observed. The GDS-15 scores correlated with those of the CAMDEX depression scale. Conclusion: In general, the GDS-15 presented good reliability and validity (concurrent and of criterion). simple and fast, can be utilized in the screening for depression. (Arq Bras Cardiol 2010;94(5):535-544)
