Abstract
To address the growing need for ethnically unbiased cognitive screening, we examined whether the Mini Mental State Exam (MMSE), the abbreviated Fuld Object Memory Evaluation (FOME), or a combination of the two provided optimal detection of dementia in an ethnically diverse group of older adults with no cognitive impairment (normal); cognitive impairment not dementia (CIND); and dementia. Participants included 509 Caucasians, 124 African Americans, and 68 Latinos (>70 years old) from the Aging, Demographics, and Memory Study who completed the MMSE and FOME. Empirically derived decision trees were computed using signal detection software for receiver operator characteristics (ROC). Among the three ethnic groups, ROC analyses revealed that lower scores on both the MMSE and FOME provided better detection of CIND or dementia. Sensitivity and specificity of the MMSE was augmented by the addition of the FOME among Caucasian and African American older adults. The MMSE alone was the best screen in Latino older adults to distinguish any cognitive impairment from normal. When comparing CIND versus dementia, however, the FOME alone was best for detecting dementia among Latinos. The abbreviated FOME is recommended to increase clinical validity and thus minimize ethnic biases when administering the MMSE to Caucasian and African American older adults. The MMSE alone is preferred for older Latinos unless comparing CIND and dementia, in which case the FOME alone would then be recommended. Findings suggest that ethnicity is important in the selection of an appropriate cognitive screen and cut-score to use with older adults.