a1 Department of Neurology, University of Pittsburgh School of Medicine, Pittsburgh, PA, U.S.A.
a2 Department of Psychiatry, Indiana University School of Medicine, Indianapolis, IN, U.S.A.
a3 Department of Medicine, University of Pittsburgh School of Medicine, Pittsburgh, PA, U.S.A.
a4 Department of Biostatistics, University of Pittsburgh Graduate School of Public Health, Pittsburgh, PA, U.S.A.
a5 Department of Psychiatry, University of Pittsburgh School of Medicine, Pittsburgh, PA, U.S.A.
a6 Department of Medicine, Division of Biostatistics, Indiana University School of Medicine, Indianapolis, IN, U.S.A.
a7 Department of Epidemiology, University of Pittsburgh Graduate School of Public Health, Pittsburgh, PA, U.S.A.
Background: Neuropsychological tests, including tests of language ability, are frequently used to differentiate normal from pathological cognitive aging. However, language can be particularly difficult to assess in a standardized manner in cross-cultural studies and in patients from different educational and cultural backgrounds. This study examined the effects of age, gender, education and race on performance of two language tests: the animal fluency task (AFT) and the Indiana University Token Test (IUTT). We report population-based normative data on these tests from two combined ethnically divergent, cognitively normal, representative population samples of older adults.
Methods: Participants aged ≥65 years from the Monongahela-Youghiogheny Healthy Aging Team (MYHAT) and from the Indianapolis Study of Health and Aging (ISHA) were selected based on (1) a Clinical Dementia Rating (CDR) score of 0; (2) non-missing baseline language test data; and (3) race self-reported as African-American or white. The combined sample (n = 1885) was 28.1% African-American. Multivariate ordinal logistic regression was used to model the effects of demographic characteristics on test scores.
Results: On both language tests, better performance was significantly associated with higher education, younger age, and white race. On the IUTT, better performance was also associated with female gender. We found no significant interactions between age and sex, and between race and education.
Conclusions: Age and education are more potent variables than are race and gender influencing performance on these language tests. Demographically stratified normative tables for these measures can be used to guide test interpretation and aid clinical diagnosis of impaired cognition.
(Received March 11 2009)
(Revised April 09 2009)
(Revised April 27 2009)
(Accepted April 28 2009)
(Online publication July 09 2009)
c1 Correspondence should be addressed to: Beth E. Snitz, PhD, Department of Neurology, University of Pittsburgh, 3471 Fifth Avenue, Suite 802, Pittsburgh, PA 15213, U.S.A. Phone: +1 412–692-4820; Fax: +1 412–692-4031. Email: [email protected].