P3‐253: Comparison of neuropsychological test scores among English and Spanish speakers in a multi‐regional sample of cognitively normal subjects
Year: 2012
Type: article
Abstract: Few studies have examined whether neuropsychological test scores vary by primary language of patients, and no known studies have examined the impact of primary language and test language simultaneously. Data on 338 Spanish-speaking and 8,374 English-speaking, cognitively normal subjects were obtained from the 33 Centers contributing to the National Alzheimer's Coordinating Center's Uniform Data Set (UDS). Each of the neuropsychological tests in the UDS were treated as an outcome variable in 11 separate multivariable models, controlling for age (years), sex, race (White, Black, other), and education (0-12 years, 13-16 years, 17+ years). A language-test variable was constructed to differentiate English speakers who tested in English (Eng-Eng), Spanish speakers who tested in English (Sp-Eng), and Spanish speakers who tested in Spanish (Sp-Sp). Multiple linear regression was conducted using Generalized Estimating Equations (GEE) to account for clustering by Center. Possible interaction between the language-test variable and education was evaluated. To account for multiple comparisons, a Bonferroni correction was applied and resulted in the use of an alpha level <0.005 to test for statistical significance. Sp-Sp test scores were significantly worse than Eng-Eng test scores for 10 tests: MMSE (-0.78 points), Digit Span Forward (-2.33 points), Digit Span Backward (-1.82 points), Category Fluency-Animals (-2.17 points), Category Fluency-Vegetables (-2.64 points), Trail Making Part A (+18.9 points), Trail Making Part B (+58.8 points), WAIS-R Digit Symbol (-15.1 points), Logical Memory Delayed (-1.95 points), and Boston Naming Test (-2.94 points). Sp-Eng test scores were also significantly worse than Eng-Eng scores for all tests except MMSE, Trail Making Part A, and Logical Memory Delayed. For many of the tests, the difference between Sp-Sp and Eng-Eng scores was greater than the difference between Sp-Eng and Eng-Eng scores. Significant interactions were observed between language-test and education for the Trail Making tests and WAIS-R Digit Symbol. Results from this study suggest that primary language should be adjusted for when neuropsychological tests, such as the MMSE, are analyzed as the outcome of interest. Future studies are needed to assess whether test score differences by language can be explained by differences in test translation, educational quality, or culture.
Source: Alzheimer s & Dementia
Authors: Lilah Besser, Sarah Monsell, Walter Kukull
Language: English
Subfield: Psychiatry and Mental health
Field: Medicine
Domain: Health Sciences
Open Access status: bronze
OpenAlex ID: W2027755088