Validation study of the three-objects-three-places test: a screening test for Alzheimer's disease.

The aim of the present study was to validate a short, ecological test of episodic memory for the screening of Alzheimer's disease (AD). The validation was performed by computing intrarater reliability, homogeneity, internal coherence, convergent, discriminant and known group validities in the performance of normal subjects (N = 65), mild cognitive impairment (MCI) patients (N = 114), and AD (N = 44) and non-AD demented (N = 39) patients. Intrarater reliability was 0.88, homogeneity ranged from 0.81 to 0.97, and internal coherence was 0.87. With respect to convergent and discriminant validities, the test loaded strongly on memory factor (value = 0.64) and weakly on other nonmemory factors. The known group validity showed a specificity between 87% and 91% and a sensitivity between 92% and 100% in correctly identifying AD in age classes ranging from 50 to 65 and 66 to 80 years. The test is a valid instrument for the screening of AD.