Differences in Rates of Dementia and Quality of Education among Diverse Older Adults

Article excerpt

Exploration of health and educational disparities among racial/ethnic groups also yields clues about the workings of cognition in later life.

This article addresses differences in rates of dementia and levels of education across ethnic/racial groups. The discussion is based on a series of studies we conducted on the accuracy of neuropsychological measures of memory and other cognitive functions in order to refine the accuracy of diagnosis of dementia among ethnic/ racial minorities and address differences in rates of dementia across groups. The following case example from a memory disorders clinic in New York City illustrates the role that language and cultural differences can play in patient assessment

Mrs. A's situation is frequently encountered when working with a diverse research cohort or patient population. In her case, the accuracy of the initial assessment was very likely affected by factors that complicate interpretation of cognitive test scores, such as the language of assessment, low levels of formal schooling and literacy, the possible contribution of depressed mood to her cognitive symptoms, and cultural expectations for cognitive aging. Furthermore, interpretation of the informant's report could have been affected by discrepancies between the patient's view of her memory function and her daughter's report.

Our research program was developed in response to clinical difficulties like this one and to several studies that report higher incidence and prevalence of Alzheimer's disease and dementia among African Americans and Latinos/Hispanics as compared to Caucasians (reviewed in Manly and Mayeux, 2004). Although the difference in rates of dementia across racial/ethnic groups has been discussed as a health inequity or "disparity," it is our opinion that use of this terminology in this instance may be premature, or at least should be clarified. Not all differences in rates of disease reflect inequity (Carter-Pokras and Baquet, 2002), and because we do not yet know the cause of different rates of dementia across groups, it is unclear if the differences are unavoidable or reflect unequal access to resources that would promote healthy cognitive aging. In many studies, rates of cognitive impairment and dementia remain higher among older ethnic minorities than among whites even after adjusting for years of education, income, or occupation. However, these findings do not help determine whether ethnic differences in rates are considered disparities because the measurement of socioeconomic variables across ethnic groups is not commensurate (Kaufman, Cooper, and McGee, 1997).

The correlation between years of school and neuropsychological test performance is well known. It is assumed that the experiences and skills acquired during school improve performance on cognitive tests because of increased familiarity with test-taking, practice, and development of problem-solving skills. It is well accepted that test performance should be adjusted for years of school in order to improve specificity of measures used to diagnose cognitive decline associated with dementia. However, some studies have demonstrated that people with fewer years of school are actually at higher risk for cognitive decline and dementia (Stern, 2002).

Our work has shown that in the United States, quality of education is extremely variable among people with the same quantity of schooling. For example, African Americans educated in the South before the 1960s attended segregated schools that were underfunded, understaffed, and had a shorter school year as compared to integrated schools in the North or white Southern schools (Anderson, 1988). Among Spanishspeaking immigrants in the Washington Heights neighborhood of New York City, cognitive test performance and literacy levels among people who attended school in the Dominican Republic does not correspond to those of people with the same quantity of education from Puerto Rico, Cuba, or the United States (Manly et al, 1999). …