The validity of informant report for grading the severity of Alzheimer's dementia.

The validity of informant-based techniques has been established for the detection of dementia cases by non-pathological individuals, but is still controversial for the assessment of the severity of dementia. This study aimed at ascertaining whether informant-based evaluation (the so-called informant report) of the cognitive and behavioral impairment of a patient is valid for grading the severity of dementia, and consistent with objective assessment of the patient's cognitive and behavioral functioning. We enrolled 96 community-dwelling outpatients and 56 controls assessed at the Geriatric Evaluation Unit of the University of Modena, Italy. All patients scored lower than 27 on the MMSE, and met DSM-IV inclusion criteria for Alzheimer's dementia. Patients and controls were administered the CAMDEX interview, containing a section which collects participant (patient or control) and informant evaluations on dementia-related cognitive and behavioral deficits. The informant report resulted effective at MANOVA for grading the severity of dementia in 4 of its 5 measures (namely, memory, everyday activities, general mental functioning and depressed mood), and was correlated with the scores of several scales of the CAMDEX cognitive section (i.e., CAMCOG). Instead, the participant's (patient or control) report showed a lower capacity for grading dementia, and was poorly correlated with the psychometric outcomes of cognitive functioning. On the whole, the results corroborated the validity of the informant report in the diagnostic work-up for grading dementia, given its sensitivity to the severity of dementia, and its consistency with cognitive psychometric outcomes.