Neuropsychological Test Development and Normative Data on Hispanics

Transcription

1 Archives of Clinical Neuropsychology, Vol. 14, No. 7, pp , 1999 Copyright 1999 National Academy of Neuropsychology Printed in the USA. All rights reserved /99 $ see front matter PII S (99) Neuropsychological Test Development and Normative Data on Hispanics Gustavo J. Rey, Esther Feldman, and R. Rivas-Vazquez University of Miami School of Medicine Bonnie E. Levin University of Miami and University of Miami School of Medicine Arthur Benton University of Iowa School of Medicine The development of culturally relevant psychological assessment tools and intervention procedures has not been commensurate with the rate of Hispanic population growth in the United States. The development of valid and reliable test measures for the assessment of this population must be based on empirical investigations. In this article, we present normative data on multiple measures from the Benton Laboratory and the Wisconsin Card Sorting Test. Results revealed equivalent findings for our Hispanic subjects and the English-speaking samples utilized in the original normative studies within the United States. We additionally review current trends and specific problems encountered in neuropsychological research with Hispanics, and suggest guidelines and directions for future research National Academy of Neuropsychology. Published by Elsevier Science Ltd Keywords: MAE-S, test development, Hispanics, norms INTRODUCTION The 1990 Census of Population and Housing indicates that there are 22 million Hispanics living in the United States making it the third largest ethnic group in the country (Census of Population, 1990). Census data also indicate that Hispanics constitute one of the fastest growing ethnic minority in the United States. Projections of Hispanic population are estimated to increase to 46 million by the year 2015, and to 88 million by the This rapid growth has created a significant demand for culturally appropriate medical and mental health services. The development of such services, however, has not been commensurate with the population growth. Address correspondence to: Gustavo J. Rey, Division of Neuropsychology, Department of Neurology, University of Miami School of Medicine, 1150 N.W. 14th Street, Suite 715, Miami, FL

2 594 G. J. Rey et al. The need to take cultural factors into account is particularly important in the area of cognitive and neuropsychological examinations, because vocational, legal, and medical decisions are frequently based on the results of these exams. Yet, despite increased interest in the field of test development for the assessment of Hispanics residing in the United States, there is still a dearth of empirical research. Instead, there is a blind reliance on adapted and translated measures without scientific validation. Additionally, as it has been argued by other investigators (Echemendia, Congett, Harris, Diaz, & Puente, 1994), inappropriate clinical neuropsychological services with this minority group are pervasive. Current practices in the assessment of Hispanic patients include the use of literal translations of English instruments and the use of norms derived from samples of English-speaking Americans. This approach incorrectly assumes that the psychometric properties of these translations to different cultural groups is comparable, when, in fact, no empirical data exist to support this claim. Another approach used to test Hispanics is the use of translators or interpreters. Some disadvantages of this procedure include violation of rules for standard administration, interference with the establishment of rapport, and, in many instances, the inability to obtain qualitative information regarding performance that may be crucial to the interpretation of test results. While the ultimate goal of neuropsychological research is the advancement of knowledge regarding brain behavior relations, the contribution of neuropsychological research with Hispanics will depend on the capacity to develop valid and reliable assessment instruments. Over the last several years our laboratories have collected standardization data on a series of neuropsychological measures adapted for Hispanics. This article presents the results of these efforts. This article also provides criterion-related validity data by comparing the results of our normative sample to those utilized in the development of the original instruments with English-speaking populations in the United States. STUDY 1 Methods Subjects. Subjects consisted of the standardization samples of the Multilingual Aphasia Examination-Spanish (MAE-S; Rey & Benton, 1991) and the Multilingual Aphasia Examination (MAE; Benton & Hamsher, 1989). For the Spanish version, subjects consisted of 234 normal adults without a history of neurological disease. All subjects were primarily Spanish-speaking. Subjects were classified according to age (18 60 and years of age), gender, and education ( 4, 5 7, 8 11, 12 15, 16). The original English version was standardized on a sample of 360 subjects ranging in age from 16 to 69 years, whose native language was English and who showed no evidence of hemispheric brain disease. The sample was also divided into age-education-sex cells. Subjects with a history of mental retardation or psychiatric disorder were excluded from both samples. Measures. The original Multilingual Aphasia Examination (Benton & Hamsher, 1989) was developed as a relatively brief battery to evaluate the presence and nature of aphasic disorders following these principles: (a) an instrument capable of providing an accurate assessment of the qualitative and quantitative aspects of language; (b) a selection of tests with clinical and theoretical relevance; (c) an instrument that has sensitivity to mild language deficits; (d) an instrument that is simple to administer and that does not require the use of complex test materials; and (e) a standardized instrument that corrects

3 Test Development and Hispanics 595 for the influence of demographical factors on test performance. These same principles were used in the development of the Spanish version (Rey & Benton, 1991). The measure consists of seven subtests as follows: A. Visual Naming (VN): Ten line drawings requiring 30 naming responses are used to assess the subject s ability to apply semantically correct labels to visually presented stimuli, independently of impaired recognition or articulation. Acceptable responses were developed by consulting dictionaries approved by the Real Academia de la Lengüa Española and by interviewing individuals in the different ethnic subgroups. B. Sentence Repetition (SR): Fourteen sentences of progressively increasing length are read one at a time and the subject is required to repeat each sentence exactly as presented. Items of the English MAE were modified in order to maintain similar sentence length and grammatical complexity. C. Controlled Oral Word Association (COWA): The subject makes associations to three letters of the alphabet by saying all the words that s/he can think of that start with the specified letter during a 60-second interval. Letters of the English MAE were replaced in order to maintain similar frequencies in the respective languages. D. Token Test (TT): Twenty large and small circles and squares in five colors are utilized to assess the subject s ability to follow auditory commands. Items from the English MAE were modified to maintain similar command length and complexity. E. Aural Comprehension of Words and Phrases (AC): Six pages with four line drawings are used to assess the aural comprehension of words and phrases in a multiple choice format. When indicated, the terminology of the English MAE was modified with culturally appropriate stimuli. F. Written Comprehension of Words and Phrases (WC): This is the written version of the AC subtest. The subject is required to write a word or phrase and point to the drawing that best represents it. The terminology of English MAE was modified when indicated. G. Rating of Articulation: A 9-point, subjective rating scale with five descriptors is used to evaluate the subject s speech articulation. Judgment is based on total speech output (conversation plus performance on tests requiring verbal output). Results Following corrections for demographic factors that influenced test results, the researchers compared the median scores and cutoffs for defective performance between the Hispanic and American norms for each subtest (Table 1). These findings indicate that results are quite comparable in both instruments. Small differences in performance can be attributed to sampling variance. The influence of nationality of MAE-S test results was also investigated by comparing the performance of 48 Puerto Rican and 47 Mexican subjects who were matched on age, education, and gender (Table 2). A multivariate analysis of variance for all of the subjects did not yield significant differences. In an attempt to conduct a conservative assessment of the data, we proceeded to perform direct t-test comparisons on these subtests. Bonferroni correction was utilized due to multiple nonorthogonal comparisons, and an alpha level of was established. No significant differences were obtained between subject groups in any of the measures (Table 2).

4 596 G. J. Rey et al. TABLE 1 Performance on the Multilingual Aphasia Examination: Median Scores and Cutoff for Defective Performance of Adjusted Scores Rey and Benton (1991) Scores and Hamsher (1989) Scores Median Cutoff Median Cutoff Visual Naming Md 5% Sentence Repetition Md 5% COWA Md 5% Token Test Md 5% Aural Comprehension Md 5% Reading Comprehension Md 5% Note. Median 5%. COWA Controlled Oral Word Association. STUDY 2 Methods Subjects. Normative standards for Hispanics on a wide range of neuropsychological instruments that include memory, attention, visuoperceptual/constructional skills, and executive functions are currently being developed in our laboratories. Test instructions were adapted or translated from the original versions. Translation instructions and the contents of the tests were reviewed by the ex-chairman and Professor of Spanish from the University of Clemson in Clemson, South Carolina. This study will present preliminary data on some of these measures for the first 75 subjects assessed. The demographic characteristics of this group is presented in Table 3. As this table illustrates, our data is representative of the estimated population distribution of Hispanics in Dade County, Florida. TABLE 2 Performance on the Multilingual Aphasia Examination-Spanish: Comparison of Raw scores of Puerto Rican and Mexican Controls Puerto Rican Mexican M (SD) M (SD) p Age (years) 46.1 (14.6) 41.9 (16.2) Education (years) 8.5 (3.8) 7.29 (3.9) Test Scores VN 38.9 (6.6) 39.7 (7.7) SR 8.0 (1.6) 8.0 (1.9) COWA 28.6 (5.3) 32.5 (5.1) TT 36.7 (0.77) 36.5 (0.74) AC 16.1 (2.00) 15.9 (1.5) RC 16.3 (1.8) 16.7 (1.4) Note. Multivariate analysis of variance procedure for the six subjects, F 1.27; p.28. VN Visual Naming; SR Sentence Repetition; COWA Controlled Oral Word Association; TT Token Test; AC Aural Comprehension of Words and Phrases; RC Reading Comprehension

5 Test Development and Hispanics 597 TABLE 3 Demographic Information of Hispanic Normative Sample (N 75) Age (Years) M (SD) (19.75) Education (Years) M (SD) (3.25) Gender Female 56 Male 19 Hispanics in Dade County (%) Percent of Sample Nationality Cuba Perú Venezuela 1 Puerto Rico Panamá 1 Colombia Honduras Nicaragua Other Measures of orientation, perception, and constructive skills. Some of the instruments that are being standardized with the Hispanic subjects are select measures from Benton s battery (Benton, des. Hamsher, Varney, & Spree, 1983): Temporal Orientation: As the name indicates, requires specification of the date, month, and year; Right Left Orientation: Examines the capacity to make right left discriminations on own body and that of the confronting examiner; Facial Recognition: Evaluates the capacity to identify and discriminate photographs of unfamiliar human faces; Judgment of Line Orientation: Test of angular judgments; Visual Form Discrimination: Brief procedure to assess capacity to make complex visual form discrimination; Tactile Form Perception: Tactile information processing requiring spatial thinking; Finger Localization: Nonverbal procedure of finger recognition following stimulation of individual or two simultaneously touched fingers; Three-Dimensional Block Construction: Assembly of blocks in three dimensional space to match a presented model. Measure of Executive Functioning. Another measure that the researchers are currently standardizing for Hispanics is the Wisconsin Card Sort Test (WCST; Heaton, 1981). This is a well-known instrument of executive functioning that involves complex problem-solving in response to the examiner s feedback, and requires cognitive flexibility and set shifting. Results Table 4 presents a comparison of the published norms (Benton et al., 1983) and our results for median and cutoff scores for each of the measures in Benton s battery. Test re-

6 598 G. J. Rey et al. TABLE 4 Performance on Benton s Neuropsychological Battery: Median Scores and Cutoff for Defective Performance Scores Hispanics Benton et al. (1983) Median Cutoff Median Cutoff Temporal Orientation Right-Left Orientation Facial Recognition Judgment of Line Orientation Visual Form Discrimination Tactile Form Perception-Right Tactile Form Perception-Left Finger Localization Three-Dimensional Block Construction Note. Median 5%. sults reveal nearly identical scores in all of these measures. Differences in some of the subtests are minimal and most likely represent sampling variance. Table 5 presents the results of our subjects on the WCST (Heaton, 1981). Their performance was compared to the corresponding norms published by Heaton in the 1981 manual. The age and education level of both normative groups are comparable (Table 5). As Table 5 indicates, test results for the groups as a whole were largely comparable for all parameters studied. The Hispanic group was further subdivided according to years of education following the same breakdown described in the WCST manual (Table 6). Results again indicate almost identical findings for all the parameters assessed. DISCUSSION Our findings indicate that the instruments investigated can be utilized clinically in the assessment of Hispanic subjects. The test results for the Hispanic groups sampled are largely equivalent to those that have been obtained with English-speaking individuals. TABLE 5 Performance on the Wisconsin Card Sorting Test Hispanics (n 75) Heaton (1981) (n 150) M (SD) M (SD) Age (Years) 33.5 (19.8) 35.9 (15.3) Education (Years) 14.5 (3.3) 13.9 (3.2) Categories completed 5.5 (1.2) 5.4 (1.3) Perseverative errors 10.7 (8.7) 12.6 (10.2) Perseverative responses 11.6 (9.7) 15.6 (11.5) Trials to complete first category 12.9 (8.5) 13.4 (10.6) Failure to maintain set 0.6 (0.8) 0.8 (1.3)

7 Test Development and Hispanics 599 TABLE 6 Performance on the Wisconsin Card Sorting Test by Education Level Hispanics Heaton (1981) years (n 25) 15 years (n 30) years (n 77) 15 years (n 53) M (SD) M (SD) M (SD) M (SD) Categories completed 5.4 (1.0) 5.8 (0.8) 5.2 (1.5) 5.7 (1.0) Perseverative errors 11.6 (9.4) 8.9 (6.6) 16.7 (10.7) 9.2 (7.9) Perseverative responses 12.7 (10.6) 9.6 (7.3) 16.7 (11.8) 11.7 (9.0) Trials to complete first category 12.4 (3.3) 11.7 (2.0) 14.0 (13.8) 11.4 (1.6) Failure to maintain set 0.5 (0.6) 0.6 (1.0) 0.9 (1.5) 0.5 (0.9) The distribution of scores for our Hispanic subjects was nearly identical to published norms on the MAE-S subtests, as well as in all the other measures studied, including Temporal Orientation; Right Left Orientation; Facial Recognition; Judgments of Line Orientation; Visual Form Discrimination; Tactile Form Perception; Finger Localization; Three-Dimensional Block Construction; and the WCST. The equivalence of the MAE-S and the original English norms suggests a valid adaptation of this measure for use with Spanish-speaking individuals. It should be emphasized that, on this measure, the data represented adjusted test scores. That is, scores that had been corrected for age, education, and gender, when these corrections were indicated. The nature and magnitude of these corrections were not equivalent for the MAE and MAE-S, to a large extent the result of lower education levels of the Spanish-speaking sample. Inclusion of these lower education groups was necessary for MAE-S in order to provide normative data for the large number of Hispanics in the United States with low educational attainment. However, other factors probably contributed to the equivalence between both versions. The MAE-S was adapted from the original English MAE according to the guidelines proposed during the 1966 meeting of Research Group on Aphasiology of the World Federation of Neurology (Benton, 1969). As such, the MAE-S is not a literal translation of the original version, but rather an adapted instrument that utilizes, when necessary, different terminology and/or items that are more appropriate to the Spanish language and culture. When these corrections were indicated, particular attention was placed on utilizing alternatives that were accepted by the Real Academia de la Lengüa Española. Additionally, individuals from various Hispanic cultures were interviewed during a pilot study in order to develop items that were equally applicable between subgroups and/or to provide alternate acceptable responses when indicated. When the sample was broken down by country of origin (Puerto Rico and Mexico) and matched on age, gender, and education, there were no significant group differences for any subtest. The clinical validity of this measure has been demonstrated in various studies that have investigated language performance of Hispanic patients with different neurological conditions, including head injury, dementia, and stroke (Rey, Hernandez, Levin, Benton, & LeBlanc, 1993; Rey, Levin, Rodas, Bowen, & Nedd, 1994; Rey et al., 1995). Similar distribution of scores between our subjects and English-speaking North American samples was also obtained for the WCST. To our knowledge, standardization studies with Hispanics have not been previously published for this measure. Given the nature of the conceptual demands of this test, normative data on the specific populations to be examined clinically are indicated. In our study, results were compared to those published in the WCST test manual (Heaton, 1981). Although these investigators have

8 600 G. J. Rey et al. recently published more comprehensive norms divided according to gender, age, and education, the limited size of our sample did not allow for such a breakdown of demographic variables. Our research group continues to gather data prospectively. As sample size increases, a more detailed examination of these characteristics will allow further comparisons between the recently published norms (Heaton, Chelune, Talley, Kay, & Curtiss, 1993) and performance of Hispanic subjects. The need to standardize Benton s measures of nonverbal reasoning and visuo-perceptual and constructive skills may be questioned because subjective impressions may suggest that these instruments are culturally fair. However, while efforts have been made to address cultural factors in psychological testing since the 1940s (Cattell, 1940), the concept of culture-free or fair measures has been challenged, and there is a question as to whether any unbiased measures actually exist (Scarr, 1978). Cultural bias is most likely present, to some extent, on all psychometric instruments, and may have differential negative influence on the performance of minority groups (Arvey, 1972). It has also been documented that ethnic minorities may have less opportunity for educational or vocational placement when evaluations are performed with culture- fair tests instead of more traditional instruments (Sattler, 1988). It is now well-accepted that psychodiagnostic instruments and clinical interventions must take into account cultural factors in the assessment and treatment of minority groups. Specific guidelines for the providers of psychological services to culturally diverse populations are being developed by the American Psychological Association Task Force on the delivery of services to these minority populations (American Psychological Association, 1990). Although our results found similar scores between the Hispanic subjects and the original North American normative samples, these findings are not to be interpreted as indicating that the instruments utilized are culturally unbiased or fair. Rather, they simply suggest that for the specific group sampled, the results are equivalent and that these measures are most likely valid indicators of neuropsychological performance for this cohort of Hispanics. In sum, in this study we have presented preliminary normative data on a series of neuropsychological measures, including the MAE-S, WCST, and various tests from Benton s neuropsychological battery. In all instances, our results were quite comparable with published norms and indicate that they are instruments that can be administered to Hispanic subjects to obtain clinically useful and reliable findings. Additional investigations are necessary to examine and determine the validity of these instruments. Caution is strongly recommended in the application of these norms to Hispanic subjects of different backgrounds. Various Hispanic subgroups were represented in our sample, with a distribution of nationality similar to that reported in census data for Hispanics in Dade County, Florida. It should be noted that this population is not representative of Hispanic subgroups residing in other regions of the United States (see Llorente, Pontón, Taussig, & Satz, 1999, this issue). Additionally, the average education level of the subjects in our study was relatively high and further research is needed to determine the validity of these instruments in the examination of lower education groups. Multicenter studies will be needed in order to obtain adequate comparisons between Hispanic subgroups. This form of collaborative work is necessary to determine the contribution variables, such as education, nationality, and degree of bilingualism and acculturation on neuropsychological test performance. REFERENCES American Psychological Association. (1990). Board of Ethnic Minority Affairs: Guidelines for providers of psychological services to ethnic, linguistic and culturally diverse populations. Task Force on the Delivery of Services to Ethnic Minority Populations. Washington, DC: Author.

9 Test Development and Hispanics 601 Arvey, R. D. (1972). Some comments on culture fair tests. Personnel Psychology, 25, Benton, A. L. (1969). Development of a multilingual aphasia battery: Progress and problems. Journal of Neurological Sciences, 9, Benton, A. L., des. Hamsher, K., Varney, N. R., & Spree, O. (1983). Contributions to neuropsychological assessment. A clinical manual. New York: Oxford University Press. Benton, A. L., & Hamsher, K. (1989). Multilingual Aphasia Examination. Manual of instructions (2nd ed.). Iowa City: AJA Associates. Cattell, R. B. A. (1940). A culture-free test, Part I. Journal of Educational Psychology, 31, Census of Population. (1990). Current population reports. Washington, D.C.: U.S. Department of Commerce, Bureau of the Census. Echemendia, R. J., Congett, S., Harris, J., Diaz, L., & Puente, A. (1994). Neuropsychological assessment and treatment of Spanish speaking individuals: A national survey of neuropsychologists. Paper presented at the 1994 Annual Meeting of the American Psychological Association. Heaton, R. (1981). Wisconsin Card Sorting Test. Manual. Odessa, FL: Psychological Assessment Resources. Heaton, R. K., Chelune G. J., Talley J. L., Kay G. G., Curtiss, G. (1993). Wisconsin Card Sorting Test manual. Revised and expanded., FL: Psychological Assessment Resources. 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, Rey, G. J., & Benton, A. L. (1991). Multilingual Aphasia Examination-Spanish. Manual of Instructions. Iowa City: AJA Associates. Rey, G. J., Hernandez, D., Levin, B. E., Benton, A. L., & LeBlanc, W. (1993). Multilingual Aphasia Examination-Spanish performance post-closed head trauma in Hispanics. Journal of Clinical and Experimental Neuropsychology, 15, 46 (abstract). Rey, G. J., Levin, B. E., Rodas, R., Bowen, B. C., & Nedd, K. J. (1994). A longitudinal examination of cross aphasia. Archives of Neurology, 51, Rey, G. J., Tomer, T., Levin, B. E., Sanchez-Ramos, J., Bowen, B. & Bruce J. H. (1995). Psychiatric symptoms, atypical dementia, and left visual field inattention in corticobasal ganglionic degeneration. Movement Disorders, 10, Sattler, J. (1988). Assessment of children. San Diego: Jerome Sattler Publishing. Scarr, S. (1978). From evolution to Larry P., or what shall we do about IQ tests? Intelligence, 2,