Abstract
BACKGROUND: Efficient neuropsychological tests are needed to measure cognitive impairment in moderate to severe dementia. OBJECTIVE: To examine construct validity of the Severe Impairment Battery Short Version (SIB-S) in nursing home patients with moderate to severe dementia, and to examine potential floor effects for the SIB-S. METHODS: Cross-sectional comparison of cognitive measures, dementia severity and functional dependency. RESULTS: A total of 290 patients were included 264 of whom had complete SIB-S protocols. Internal consistency of the SIB-S was very high (Cronbach alpha = 0.97). Principal-component analysis produced three factors, the first of which explained more than 50% of common score variance. Semantic memory items loaded highly on the first factor. Total SIB-S scores were associated with cognitive impairment (SIB-S--Mini Mental State Examination (MMSE) rho = 0.91, p<0.001), and with functional dependency (SIB-S--ADL scale rho = -0.61, p<0.001). SIB-S total scores differentiated between dementia stages as measured with the Global Deterioration Scale (F = 164.6; df 3,260, p<0.001). Comparisons of SIB-S total score variance across patients with moderate to severe dementia and patients with below or above average Mini Mental State Exam scores indicate an absence of large floor effects. CONCLUSION: In this first study examining an independently administered SIB-S, the scale proved to be a homogeneous and valid measure of cognitive impairment. The SIB short version can be used to assess moderately to severely demented patients, who may find it difficult to complete traditional, lengthier neuropsychological tests.
