Memory evaluation in Alzheimer's disease. Caregivers' appraisals and objective testing.

OBJECTIVES

To evaluate if caregivers are reliable informants concerning memory deficits in patients with Alzheimer's disease (AD).

DESIGN

Responses of caregivers of patients with probable AD and responses of healthy control subjects on a standardized memory questionnaire were compared with objective measures of cognition (Consortium to Establish a Registry for Alzheimer's Disease neuropsychological battery) and with clinical estimates of activities of daily living, depression, and psychopathology (Consortium to Establish a Registry for Alzheimer's Disease [CERAD] clinical assessment battery) using the Self-report Memory Questionnaire.

SETTING

A federally funded AD research center.

SUBJECTS

The referred sample included 117 patients with probable AD, their informants, and 41 healthy control subjects age-matched to the patients. Patients and control subjects were between the ages of 58 and 85 years, had between 9 and 19 years of education, and were in good health.

EXCLUSIONS

Patients who did not meet NINCDS-ADRDA criteria of probable AD.

MAIN OUTCOME MEASURE

The optimal number of questionnaire items yielding the best combination of sensitivity and specificity.

RESULTS

An abbreviated version of the scale, renamed the Short-Memory Questionnaire, had excellent specificity and sensitivity for identifying dementia. Positive and negative predictive values were 63.5% and near 100%, respectively. The Short-Memory Questionnaire showed good reliability, internal consistency, and external validity. Caregiver appraisals of memory deficits significantly correlated with objective measures of memory and also with generalized cognitive dysfunction.

CONCLUSIONS

Caregivers of patients with AD are reliable informants of their relatives' deficits. The Short-Memory Questionnaire is an easily administered, informant-based scale that may be useful in clinical settings or epidemiologic studies to screen out persons with memory difficulties.