Primary care screen for early dementia.

OBJECTIVES To determine whether the Alzheimer's Disease Screen for Primary Care (ADS-PC) is more sensitive to early dementia than the Mini-Mental State Examination (MMSE) and whether it has as high a misclassification rate in minority patients and patients with limited education. DESIGN Cross-sectional validation study. SETTING Urban geriatric primary care practice. PARTICIPANTS Three hundred sixteen African-American and Caucasian patients, including 55 patients with early dementia (Clinical Demential Rating of 0.5). MEASUREMENTS The ADS-PC is a two-stage strategy for identifying early dementia that consists of a brief high-sensitivity dementia screen, applied to all patients aged 65 and older, and a second stage to identify memory impairment, applied to patients who fail the first stage. Differences in the sensitivities or specificities of the ADS-PC and the MMSE were evaluated using the McNemar test. RESULTS Receiver operating characteristic curves were used to examine differences in the operating characteristics of the ADS-PC across a range of cutscores. When the specificities of both tests were equated (0.90), the sensitivities were significantly different (ADS-PC, 0.75; MMSE, 0.56). The MMSE failed to identify five times as many cases of dementia as the ADS-PC. When the sensitivities were equated, the specificities differed significantly (ADS-PC, 0.95; MMSE, 0.73). The MMSE misclassified five times as many noncases as the ADS-PC. This pattern of significantly higher sensitivity and specificity for the ADS-PC than for the MMSE was repeated in the results according to race and for individuals with high school or more education but not in individuals with less education. CONCLUSION The ADS-PC outperformed the MMSE in identifying early dementia in a racially and educationally diverse primary care cohort.