The performances of shorter versions of the Geriatric Depression scale (GDS) are examined. A cutoff of 4/5 gives the best sensitivity (80%) and specificity (77%) for the 15-item version (GDS15). A cutoff of 3/4 gives the best sensitivity (75%) and specificity (77%) for the 10-item version (GDS10). A cutoff of 0/1 gives the best sensitivity (85%) and specificity (60%) for the 4-item version (GDS4). Similar cutoff values, specificities and sensitivities were demonstrated in the dementia subsample for GDS15 and GDS10. The GDS4 had poor sensitivity and specificity in the dementia subsample. The GDS15 and the GDS10 may be utilized in screening for depression among continuing care geriatric inpatients, but would require prospective evaluation among acutely ill geriatric inpatients.