Accuracy of a Brief Neuropsychological Battery for the Diagnosis of Dementia and Mild Cognitive Impairment: An Analysis of the NEDICES Cohort

Abstract

Abstract

Early separation of mild cognitive impairment (MCI) from normal aging and mild cases of dementia remains a challenge, especially in the general population. We aimed to analyze the diagnostic accuracy of a brief neuropsychological battery (BNB) in dementia and MCI cases from the Neurological Disorders in Central Spain (NEDICES) population-based cohort study. We screened 3,891 participants into dementia and non-dementia groups using a two-phase procedure: screening (MMSE-37 and Pfeffer-11) and clinical diagnosis by specialists (DSM-IV criteria). We selected subsequently a subsample of dementia (n = 98), MCI (n = 71), and cognitively healthy (n = 123) participants matched in socio-demographic characteristics. The clinical validity of each test of the BNB was determined by the area under the ROC curve. We determined the best combination of tests to classify individuals into the diagnostic groups by logistic regression analyses. The results indicated that dementia and MCI groups could be best discriminated from the healthy control group on the basis of their scores on the semantic verbal fluency and delayed recall subtests of the BNB. As for discriminating the MCI group from the dementia group, immediate recall tasks (stories and pictures) yielded the highest level of accuracy. Probably the most interesting finding is that the verbal fluency task consistently allowed discrimination among the diagnostic groups. Overall, subtests of the BNB are more accurate in differentiating dementia patients than MCI patients from healthy controls. In this population-based sample, a more fine-grained discrimination that includes MCI patients should follow a systematic subtest-wise analysis and decision.