ABSTRACT
Record review in a sample of 313 geriatric rehabilitation in-patients at an urban medical center examined the diagnostic validity of the short form of the Geriatric Depression Scale (GDS-SF). In a predominantly African American medical patient sample, scores on the GDS-SF were compared to the 30-item form (Geriatric Depression Scale-Long Form, GDS-LF) for both cognitively intact and cognitively impaired persons. Results suggest that the fifteen-item form of the GDS is a valid screening instrument for this population. The two scales were highly correlated (r = .88), and both demonstrated similar rates of sensitivity and specificity for the total and split samples. Results are discussed in terms of screening for mood disorders in the urban population.