Objective: To determine whether the perceptions of patients' cognitive deficits by either the patient or an informant could predict who would develop Alzheimer disease (AD) in a group of 120 memory-impaired patients without dementia.
Methods: At entry into the study, patients were assessed by several measures that included neuropsychological tests and the Cambridge Mental Disorders of the Elderly Examination Interview Schedule. The latter schedule asks patients and their informants about their perceptions of cognitive deficits in the patients. After 2 years, patients underwent a diagnostic workup for AD: 29 had developed probable AD, whereas the other 91 did not develop dementia. We used logistic regression analyses to examine the predictive accuracy of patients' and informants' perceptions of deficits at entry into the study.
Results: Informants' perceptions, not patients' perceptions, contributed significantly to the prediction of AD. The best prediction of AD was obtained by the regression model that included both informants' perceptions and 2 neuropsychological tests.
Conclusions: These results demonstrate the clinical use of including informant perceptions about patients' cognitive deficits in the diagnostic assessment of AD. They also indicate that patients' perceptions of their own deficits are not predictive of AD, but are related to depressive affect.