Differential validity of psychometric tests in dementia of the Alzheimer type.

Forty-nine patients with a clinical diagnosis of probable dementia of the Alzheimer type underwent an extensive test battery designed to evaluate cognitive deficits according to NINCDS/ADRDA criteria. All patients demonstrated signs of impairment on this test battery. One day later, they were administered a second test battery that consisted of the Mini-Mental State Examination (MMS), the Alzheimer's Disease Assessment Scale (ADAS), the SKT test (SKT), and the Brief Cognitive Rating Scale (BCRS) to assess the construct validity, sensitivity, and possible shortcomings of these tests. A control group of 47 age-matched persons was administered the same test battery to allow a comparison with reference values from other studies. Due to the design of the study, values of controls and patients did not overlap. Intercorrelations in patients were above 0.65 (p < 0.05 after Bonferroni correction) for all four cognitive tests. The ADAS and BCRS appeared to document the whole course of the disease in patients studied. The best differentiation with the SKT test could be obtained in mild to moderate dementia; however, due to the test's construction, a floor effect demonstrated its limitations in the case of severe dementia. Results obtained with the MMSE indicated the contrary: a ceiling effect showed its lack of differentiation in mild dementia. Therefore, a combination of tests should be used in the evaluation of cognitive deficits in the course of dementia of the Alzheimer type.