A semantic verbal fluency test for English- and Spanish-speaking older Mexican-Americans

Size: px

Start display at page:

Download "A semantic verbal fluency test for English- and Spanish-speaking older Mexican-Americans"

Bridget Bennett
5 years ago
Views:

Archives of Clinical Neuropsychology 20 (2005) 199 208 A semantic verbal fluency test for English- and Spanish-speaking older Mexican-Americans Hector M. González a,, Dan Mungas b, Mary N.

1 Archives of Clinical Neuropsychology 20 (2005) A semantic verbal fluency test for English- and Spanish-speaking older Mexican-Americans Hector M. González a,, Dan Mungas b, Mary N. Haan a a University of Michigan, School of Public Health, Department of Epidemiology, 611 Church St., Room 315, Ann Arbor, MI , USA b University of California, Davis, School of Medicine, Department of Neurology, Sacramento, CA, USA Abstract Accepted 11 June 2004 Currently, there are few semantic verbal fluency tests available with appropriate and adequate normative data for older Latinos of Mexican ancestry. The purposes of this study were to examine the effects of language and socioeconomic variables on a semantic verbal fluency test, derive normative data and develop statistical corrections for the effects of significant variables. A large, community-dwelling sample of Latinos of primarily Mexican ancestry over age 60 was asked to name animals with four legs within 1 min in their preferred language of English or Spanish. The total sum of animals named was influenced by age, education, gender, and, to a much lesser extent, language of test administration. Regression coefficients derived from demographic variables were used to adjust raw scores in a general correction formula, and specific percentile cutoff values were identified for these adjusted scores. Normative data tables stratified on demographic variables are also presented National Academy of Neuropsychology. Published by Elsevier Ltd. All rights reserved. Keywords: Hispanic-Americans; Latinos; Mexican-Americans; Neuropsychology; Semantic verbal fluency; Language; Aging 1. Introduction Latinos are now the largest ethnic minority group in the U.S. with about two-thirds being of Mexican ancestry (Census, 2001; Census, 2004; Llorente, Pontón, Taussig, & Satz, 1999). Mexican-Americans in the U.S. are a culturally and socioeconomically diverse group. Within Mexican-Americans, there is considerable variability in not only their adoption of U.S. culture, Corresponding author. Tel.: ; fax: address: hmgonz@umich.edu (H.M. González) /$ see front matter 2004 National Academy of Neuropsychology. Published by Elsevier Ltd. All rights reserved. doi: /j.acn

2 200 H.M. González et al. / Archives of Clinical Neuropsychology 20 (2005) but also in their cultural experiences and practices from their Mexican or U.S. region of origin. Associated with this cultural diversity are the economic and educational opportunities that are less available to some Mexican-Americans (Wade, 1997). Further, for many older Mexican American adults, educational opportunities were fewer than for younger adults. Finally, as is often the case today, the resources available to schools serving low income, ethnic minority communities were often less than those that are available to other communities. In combination, these present challenges to clinicians and investigators involved in the neuropsychological assessment of older adults and in this case older Mexican-Americans. Neuropsychological tests are needed that are minimally affected by cultural, linguistic, and educational biases that could potentially disadvantage population subgroups, and thereby could negatively affect cognitive impairment detection (Ardila, Rosselli, & Puente, 1994; Pontón & Ardila, 1999). Currently, there are few neuropsychological test batteries that are have appropriate and adequate normative data for older Latinos of Mexican ancestry (Mungas, Reed, Marshall, & González, 2000; Pontón, Satz, Herrera, & Ortiz, 1996). Although there are verbal fluency tests available for Spanish-speakers in other countries (Butman, Allegri, Harris, & Drake, 2000; Rosselli et al., 2000) and Latinos of Caribbean ancestry (Jacobs, Sano, Albert, & Schofield, 1997), there are few semantic verbal fluency tests with appropriate normative data for older Mexican-Americans (La Rue, Romero, Ortiz, Liang, & Lindeman, 1999; Pontón et al., 1996). In the assessment of individuals from culturally different backgrounds and lower educational levels, it is reasonable to select tests that first and foremost address the referral or research questions and are likely to provide a measure of functional ability rather than noise associated with test bias. Tests that rely on commonly experienced events or objects (e.g., shopping lists for recall or animal names) would be preferable to those that are culturally bound and highly literacy dependent (e.g., items from the WAIS Information subtest). Semantic or category verbal fluency tests requiring an individual to name objects that are experienced in daily life would be a sensible choice. These tests are designed to assess verbal abilities and executive function. In addition, the task demands experiential or semantic knowledge (e.g., naming animals), regardless of cultural context, that is not absolutely reliant on formal educational experiences of a particular culture. For example, among Hindi-speaking Indians of the rural district of Ballabgarh which has low literacy rates, semantic verbal fluency was shown to be less influenced by educational level than phonemic verbal fluency (Ratcliff et al., 1998). If tests of semantic fluency are less affected by formal education in groups with low levels of education or who speak a given language, then it becomes important to quantify these influences. There are two objectives in this report: (1) to determine the association of formal education and language, on verbal fluency and (2) to provide normative data for a semantic verbal fluency test based from a community-dwelling sample of 1276 of Mexican-Americans ages 60 and over. 2. Methods 2.1. Participants Participants in this study were recruited as part of the Sacramento Area Latino Study on Aging (SALSA), which is a prospective, cohort study of aging and health among Latinos

3 H.M. González et al. / Archives of Clinical Neuropsychology 20 (2005) over age 60 (details of the study design can be found in Haan et al., 2003). Participants were eligible for SALSA if: (1) they self-identified as being of Hispanic or Latino ancestry (parents or grandparents), (2) age 60 or over, and (3) English or Spanish was their primary language. All testing was conducted by bilingual/bicultural technicians who received supervised training from a neuropsychologist (H.M.G). Face-to-face interviews were supervised and observed by the same neuropsychologist in the participants home during testing. For this analysis, participants meeting National Institute of Neurological and Communicative Diseases and Stroke/Alzheimer s Disease and Related Disorders Association criteria for Alzheimer s disease or California Alzheimer s Disease and Treatment Center criteria for ischemic vascular dementia were excluded from the sample. Details of the multi-stage, dementia ascertainment procedures can be found in Haan et al. (2003). The Spanish English Animal Naming Test (SEANT), like other verbal fluency tasks, required the participant to name animals with four legs within 60 s. The SEANT was embedded within the Modified Mini Mental State exam (3MS; Teng & Chui, 1987). The standard 3MS administration has a 30 s animal fluency task as one of its regular items, but limits the number of responses to 10 items for that 30 s interval. For the SEANT, the time was extended to 60 s, and the number of items admissible was unrestricted. This was done to lessen ceiling effects and to provide an additional test of cognitive functioning that would yield normative neuropsychological data. Initial experience with the animal naming item of the 3MS indicated that 41.2% of SALSA participants gave the ceiling number of 10 responses on this item. The SEANT was administered to 1276 SALSA participants in the first follow-up year (FY1) after the baseline year of SALSA (July 1999 to June 2000). Participants self-selected their preferred language (i.e., English or Spanish) for SEANT administration. This was used in data analyses. Education was measured in years of formal school Statistical adjustment for demographic variables and norms Linear regression was used to estimate the effects of the demographic variables on the dependent measure. The dependent measure was the sum of four-legged animals that were given by the respondent in 60 s. The predictor variables were (age and education in years and language group) (English versus Spanish) and gender (M/F). The unstandardized regression coefficients for the demographic variables derived from these analyses were used to adjust raw scale scores for effects of demographic variables. The general formula for this adjustment was: Adjusted Score = Raw Score b Language b Gender (b Age (Age-70)) (b Education (Education-12)) In this equation, b Language is the coefficient corresponding to the language group of the participant, b Gender is the coefficient corresponding to the gender of the participant, and b Age and b Education are the unstandardized coefficients for age and education. The age of 70 and 12 years of education are arbitrary values selected as standard reference points. The resulting adjusted score is uncorrelated with the demographic variables, but maintains a scale of measurement

4 202 H.M. González et al. / Archives of Clinical Neuropsychology 20 (2005) similar to that of the raw scores. This is the same approach previously used to adjust the Folstein Mini Mental State exam (Folstein, Folstein, & McHugh, 1975; Mungas, Marshall, Weldon, Haan, & Reed, 1996) and SEVLT (González, Mungas, Reed, Marshall, & Haan, 2001a,b) for effects of demographic variables. 3. Results In this sample of 1276 older Latinos, the mean age was 70.3 (S.D. = 6.8); mean education was 7.8 (S.D. = 5.4). Women accounted for 57.4% of the sample, 53.6% were Spanish-speaking and 52.5% were U.S.-born. The Spanish-speaking group was older, t(1, 1274) = 3.85, P <.001, and had fewer years of education, t(1, 1274) = 23.23, P <.001, but the groups did not differ in the percentage of females (χ 2 = 1.58, P =.209). Eighty-five percent of the sample was from Mexico and 15% from another Latin American country. In this sample, 53.3% were administered the SEANT in Spanish Statistical adjustment of scores and norms A normal quantile (Q Q) plot of SEANT responses did not indicate a significantly nonnormal distribution. Polynomial regression was used to evaluate non-linear effects of education and age on SEANT performance. Curve estimations of the functions of years of education indicated that there were significant linear, quadratic and cubic effects; however, the alternatives to the linear function offered very small increases in variance explanation. The linear effect alone accounted for 9.9% of SEANT variance, whereas the combined linear, quadratic, and cubic accounted for 10.2%, which indicates that the non-linear terms explained only 0.3% of additional variance. Age did not show a significantly non-linear relationship with the SEANT. The cross-product interaction term for education and language was not statistically significant (P =.234) and was dropped from the model. This indicates that effects of education were similar in English- and Spanish-speakers. Results of the linear regression model are presented in Table 1. Significant main effects were found between the SEANT and age, education, and gender, and language. Education (4.1%) accounted for the largest proportion of variance, followed by age (3.9%), gender (1.2%), and language (0.4%). The derived unstandardized regression coefficients are presented in Table 2. The dichotomous independent variable coefficients for language or gender were either added or subtracted Table 1 Results of the Regression Analyses for the Spanish English Animal Naming Test B S.E. Beta t P Age (years) Education (years) Gender Language

5 H.M. González et al. / Archives of Clinical Neuropsychology 20 (2005) Table 2 Derived Unstandardized Regression Coefficients used to create Adjusted Scores for the Spanish English Animal Naming Test Unstandardized beta coefficients Age (years).09 Education (years).16 Male a.41 Female a.41 English b.24 Spanish b.24 a Coefficient values are added or subtracted from the equation depending on the individual s gender. b Coefficient values are added or subtracted from the equation depending on the language of the individual. from the scores depending on the characteristics of the individual. The coefficients for age and education were multiplied by the difference from the reference values of 70 and 12 in the general correction formula, respectively. In this way, the coefficients were entered into the general correction formula previously described and used to calculate the demographically adjusted scores for each of the seven dependent measures for the entire sample. Along with adjusted sample means and standard deviations, the specific percentile cut-offs are presented in Table 3. Normative data for the SEANT derived from descriptive statistics (numbers per cell, means, and standard deviations) for English- (Table 4) and Spanish-speakers (Table 5) were stratified by 10-year age group, education tertile and gender. Table 3 Specific cutoffs for age, education, gender, and language Adjusted Scores corresponding to different percentile values for Spanish English Animal Naming Test N 1276 M S.D Percentiles

6 204 H.M. González et al. / Archives of Clinical Neuropsychology 20 (2005) Table 4 Normative data for the Spanish English Animal Naming Test for 592 English-speaking Community-dwelling Latinos over age 60 from the Sacramento area Latino study on aging Age groups Education groups Gender N M S.D Male Female Total Male Female Total Male Female Total Total Male Female Total Male Female Total Male Female Total Male Female Total Total Male Female Total Male Female Total Male Female Total Male Female Total Total Male Female Total Total 12+ Male Female Total Male Female Total Male Female Total Total Male Female Total

7 H.M. González et al. / Archives of Clinical Neuropsychology 20 (2005) Table 5 Normative data for the Spanish English Animal Naming Test for 684 Spanish-speaking Community-dwelling Latinos over age 60 from the Sacramento area Latino study on aging Age groups Education groups Gender N M S.D Male Female Total Male Female Total Male Female Total Total Male Female Total Male Female Total Male Female Total Male Female Total Total Male Female Total Male Female Total Male Female Total Male Female Total Total Male Female Total Total 12+ Male Female Total Male Female Total Male Female Total Total Male Female Total

8 206 H.M. González et al. / Archives of Clinical Neuropsychology 20 (2005) Discussion This study provides normative data for a semantic verbal fluency test based upon a very large community-dwelling sample of 1276 older (60 years plus) English- and Spanish-speaking Latinos of predominately Mexican ancestry. The demographics of this large sample closely reflect the Latino population in Sacramento County in terms of age and gender characteristics from the Census 2000 Dress Rehearsal (Census, 1999). Comparisons with data from the 1990 Census indicates that the SALSA sample reflects Latinos over age 60 in California in terms of education, age, and gender (Census, 1995). The representativeness of the SALSA sample would suggest that the SEANT would be appropriate for use with most older Mexican-Americans (Census, 2000). SEANT performance was significantly influenced by the demographic factors studied (i.e., age, education, gender and language). However, in combination, these factors accounted for less than 10% (9.67%) of the variance. Education had the largest affect on SEANT performance, but it is likely that the statistical significance of these factors was largely related to the large sample size in this study. Nevertheless, the findings of this study indicate how the SEANT performance was affected by demographic variables and provide methods for accounting for these effects in interpreting test scores that should be considered when using this exam. It is possible that SEANT performance could be affected by other factors not examined in this study; however, the factors presented above are likely to be of utility to clinicians and researchers. As such, SEANT users that are interested in statistical adjusting test SEANT test performance are encouraged to use the adjustments or normative data provided in the tables above. As in the previous reports, two approaches for deriving the adjusted scores based on significant demographics variables are provided (González, Mungas, & Haan, 2002; González et al., 2001a,b). With the appropriate coefficients for demographics in Table 2, the general formula above can be applied to derive an adjusted score for the acceptable words produced. The derived adjusted score can then be compared to the normative data in Table 3 for deviation of the adjusted score from the means and to determine its percentile rank. Additionally, the SEANT score can be interpreted with a more conventional approach using the normative data for English and Spanish-speakers in Tables 4 and 5, respectively. Although the former approach will yield a more finely calibrated adjusted score, the two approaches will yield similar results. An advantage of the derived adjusted score is that it avoids problems with small sample sizes that are seen some of the cells of the normative tables (Tables 4 and 5), which could potentially result in unstable estimates resulting from the more infrequent demographic variable combinations. The potential disadvantage of the derived score approach involves underlying assumptions for linear regression. The results of this study indicate the linear associations between variables had a better fit than the non-linear associations, which would indicate that this regression assumption was met. The normative data provided for the SEANT in this manuscript should be of utility clinicians and researchers working with older Latino individuals of Mexican ancestry. There is considerable ethnic and cultural variability within populations of Latinos and further research with the SEANT will be needed to demonstrate its applicability to other groups. Tests, like the SEANT, hold promise for assessing the cognitive function of culturally and educationally diverse populations. Providing normative data for assessing the cognitive functioning of

9 H.M. González et al. / Archives of Clinical Neuropsychology 20 (2005) ethnically diverse populations is consistent with the American Psychological Association s General Principles and Ethics Standards and should guide the future direction of neuropsychological test development. References Ardila, A., Rosselli, M., & Puente, A. E. (1994). Neuropsychological evaluation of the Spanish speaker. New York, NY, US: Plenum Press. Butman, J., Allegri, R. F., Harris, P., & Drake, M. (2000). Spanish verbal fluency. Normative data in Argentina. Medicina (B Aires), 60(5 Pt 1), Folstein, M. F., Folstein, S. E., & McHugh, P. R. (1975). Mini-mental state. A practical method for grading the cognitive state of patients for the clinician. Journal of Psychiatric Research, 12(3), González, H. M., Mungas, D., & Haan, M. N. (2002). A verbal learning and memory test for English- and Spanish-speaking older Mexican-American adults. The Clinical Neuropsychologist, 16(4), González, H. M., Mungas, D., Reed, B. R., Marshall, S., & Haan, M. N. (2001). A new Verbal Learning and Memory Test for English- and Spanish-speaking older people. Journal of the International Neuropsychological Society, 7(5), González, H. M., Mungas, D. M., Reed, B. R., Marshall, S., & Haan, M. (2001). A new Verbal Learning and Memory Test for English- and Spanish-speaking older people. Journal of the International Neuropsychological Society, 7, Haan, M. N., Mungas, D. M., Gonzalez, H. M., Ortiz, T. A., Acharya, A., & Jagust, W. J. (2003). Prevalence of dementia in older Latinos: The influence of type 2 diabetes mellitus, stroke and genetic factors. Journal of the American Geriatrics Society, 51(2), Jacobs, D. M., Sano, M., Albert, S., & Schofield, P. (1997). Cross-cultural neuropsychological assessment: A comparison of randomly selected, demographically matched cohorts of English- and Spanish-speaking older adults. Journal of Clinical and Experimental Neuropsychology, 19(3), La Rue, A., Romero, L. J., Ortiz, I. E., Liang, H. C., & Lindeman, R. D. (1999). Neuropsychological performance of Hispanic and non-hispanic older adults: An epidemiologic survey. The Clinical Neuropsychologist, 13(4), Llorente, A. M., Pontón, M. O., Taussig, I. M., & Satz, P. (1999). Patterns of American immigration and their influence on the acquisition of neuropsychological norms for Hispanics. Archives of Clinical Neuropsychology, 14(7), Mungas, D., Marshall, S. C., Weldon, M., Haan, M., & Reed, B. R. (1996). Age and education correction of Mini-Mental State Examination for English and Spanish-speaking elderly. Neurology, 46(3), Mungas, D., Reed, B. R., Marshall, S. C., & Gonzalez, H. M. (2000). Development of psychometrically matched English and Spanish language neuropsychological tests for older persons. Neuropsychology, 14(2), Pontón, M. O., & Ardila, A. (1999). The future of neuropsychology with Hispanic populations in the United States. Archives of Clinical Neuropsychology, 14(7), Pontón, M. O., Satz, P., Herrera, L., & Ortiz, F. (1996). Normative data stratified by age and education for the Neuropsychological Screening Battery for Hispanics (NeSBHIS): Initial report. Journal of the International Neuropsychological Society, 2(2), Ratcliff, G., Ganguli, M., Chandra, V., Sharma, S., Belle, S., Seaberg, E., et al. (1998). Effects of literacy and education on measures of word fluency. Brain and Language, 61(1), Rosselli, M., Ardila, A., Araujo, K., Weekes, V. A., Caracciolo, V., Padilla, M., et al. (2000). Verbal fluency and repetition skills in healthy older Spanish-English bilinguals. Applied Neuropsychology, 7(1), Teng, E. L., & Chui, H. C. (1987). The Modified Mini-Mental State (3MS) examination. Journal of Clinical Psychiatry, 48(8), U.S. Census Bureau. (1995). Education in the United States: 1990 census of population and housing. Washington, DC.

10 208 H.M. González et al. / Archives of Clinical Neuropsychology 20 (2005) U.S. Census Bureau. (2001). Overview of race and Hispanic origin Census 2000 brief. Washington, DC: US Government Printing Office. U.S. Census Bureau. (1999). Census 2000 dress rehearsal 100-percent summary file. Washington, DC: US Government Printing Office. U.S. Census Bureau. (2000). The Hispanic population of the United States. Washington, DC: US Government Printing Office. U.S. Census Bureau. (2004). Current population series (Vol. 15). Washington, DC: US Government Printing Office. Wade, P. (1997) Race and ethnicity in Latin America. London, Chicago, IL: Pluto Press.

Elderly Norms for the Hopkins Verbal Learning Test-Revised*

The Clinical Neuropsychologist -//-$., Vol., No., pp. - Swets & Zeitlinger Elderly Norms for the Hopkins Verbal Learning Test-Revised* Rodney D. Vanderploeg, John A. Schinka, Tatyana Jones, Brent J. Small,