OBJECTIVE: To evaluate the discriminative and diagnostic values of neuropsychological tests for identifying schizophrenia patients. METHODS: A cross-sectional study with 36 male schizophrenia outpatients and 72 healthy matched volunteers was carried out. Participants underwent the following neuropsychological tests: Wisconsin Card Sorting test, Verbal Fluency, Stroop test, Mini Mental State Examination, and Spatial Recognition Span. Sensitivity and specificity estimated the diagnostic value of tests with cutoffs obtained using Receiver Operating Characteristic curves. The latent class model (diagnosis of schizophrenia) was used as gold standard. RESULTS: Although patients presented lower scores in most tests, the highest canonical function for the discriminant analysis was 0.57 (Verbal Fluency M). The best sensitivity and specificity were obtained in the Verbal Fluency M test (75 and 65, respectively). CONCLUSIONS: The neuropsychological tests showed moderate diagnostic value for the identification of schizophrenia patients. These findings suggested that the cognitive impairment measured by these tests might not be homogeneous among schizophrenia patients.
Schizophrenia; Cognition disorders; Psychological tests; Diagnostic value
OBJETIVO: Avaliar os valores discriminativos e diagnósticos de testes neuropsicológicos na identificação de pacientes com esquizofrenia. MÉTODOS: Estudo transversal com 36 pacientes ambulatoriais esquizofrênicos masculinos e 72 voluntários saudáveis pareados. Os testes neuropsicológicos usados foram o Wisconsin Card Sorting Test, Fluência Verbal, teste de Stroop, Mini Exame do Estado Mental e Span de reconhecimento espacial. Valores diagnósticos dos testes foram estimados por sensibilidade e especificidade, com os pontos de corte obtidos através da curva ROC (Receiver Operating Characteristic). O modelo de classe latente (diagnóstico de esquizofrenia) foi aplicado como padrão ouro. RESULTADOS: Embora os pacientes tenham apresentado escores inferiores na maioria dos testes, a função canônica mais alta pela análise discriminante foi 0,57 (na fluência verbal com a letra M). Os melhores valores de sensibilidade e especificidade foram obtidos pelo teste de fluência verbal com a letra M (75 e 65, respectivamente). CONCLUSÕES: Os testes neuropsicológicos demonstraram valores diagnósticos moderados na identificação de pacientes com esquizofrenia. Os achados sugerem que o prejuízo cognitivo, medido por esses testes, pode não ser homogêneo entre os pacientes esquizofrênicos.
Esquizofrenia; Transtornos cognitivos; Testes psicológicos; Valor diagnóstico
Schizophrenia in males of cognitive performance: discriminative and diagnostic values
Esquizofrenia em homens pelo desempenho cognitivo: valor discriminativo e diagnóstico
Analuiza Camozzato and Márcia L F Chaves
Faculdade de Medicina da Universidade Federal do Rio Grande do Sul. Porto Alegre, RS, Brasil
Schizophrenia, diagnosis. Cognition disorders. Psychological tests. Diagnostic value.
To evaluate the discriminative and diagnostic values of neuropsychological tests for identifying schizophrenia patients.
A cross-sectional study with 36 male schizophrenia outpatients and 72 healthy matched volunteers was carried out. Participants underwent the following neuropsychological tests: Wisconsin Card Sorting test, Verbal Fluency, Stroop test, Mini Mental State Examination, and Spatial Recognition Span. Sensitivity and specificity estimated the diagnostic value of tests with cutoffs obtained using Receiver Operating Characteristic curves. The latent class model (diagnosis of schizophrenia) was used as gold standard.
Although patients presented lower scores in most tests, the highest canonical function for the discriminant analysis was 0.57 (Verbal Fluency M). The best sensitivity and specificity were obtained in the Verbal Fluency M test (75 and 65, respectively).
The neuropsychological tests showed moderate diagnostic value for the identification of schizophrenia patients. These findings suggested that the cognitive impairment measured by these tests might not be homogeneous among schizophrenia patients.
Esquizofrenia, diagnóstico. Transtornos cognitivos. Testes psicológicos. Valor diagnóstico.
Avaliar os valores discriminativos e diagnósticos de testes neuropsicológicos na identificação de pacientes com esquizofrenia.
Estudo transversal com 36 pacientes ambulatoriais esquizofrênicos masculinos e 72 voluntários saudáveis pareados. Os testes neuropsicológicos usados foram o Wisconsin Card Sorting Test, Fluência Verbal, teste de Stroop, Mini Exame do Estado Mental e Span de reconhecimento espacial. Valores diagnósticos dos testes foram estimados por sensibilidade e especificidade, com os pontos de corte obtidos através da curva ROC (Receiver Operating Characteristic). O modelo de classe latente (diagnóstico de esquizofrenia) foi aplicado como padrão ouro.
Embora os pacientes tenham apresentado escores inferiores na maioria dos testes, a função canônica mais alta pela análise discriminante foi 0,57 (na fluência verbal com a letra M). Os melhores valores de sensibilidade e especificidade foram obtidos pelo teste de fluência verbal com a letra M (75 e 65, respectivamente).
Os testes neuropsicológicos demonstraram valores diagnósticos moderados na identificação de pacientes com esquizofrenia. Os achados sugerem que o prejuízo cognitivo, medido por esses testes, pode não ser homogêneo entre os pacientes esquizofrênicos.
Significant cognitive impairment across multiple ability domains has been considered a core characteristic of schizophrenia and is not caused by chronic illness, treatment, or institutionalization.1,10-13,15 Whether cognitive deficits in schizophrenia are general (impairment across multiple ability domains) or specific (reduced intellectual functions, such as attention, memory and executive functions) is a question still under debate.4,13
From a clinical point of view, if the cognitive impairment were a prominent characteristic of schizophrenia, a reliable neuropsychological evaluation would be an important instrument for the assessment of schizophrenia patients, as in the case of Alzheimer's disease. The neuropsychological examination is today considered an important part of the diagnostic procedure.2 Recently, a neuropsychological battery was suggested as a cognitive screening test for schizophrenia and a useful prognostic indicator.8
Most studies on cognitive performance in schizophrenia have shown poorer average performances of patients compared to healthy volunteers. The studies applied statistical comparisons based on group variability, distribution and central tendency parameters, however, no clinical value has been carried out. The clinical application of this type of result remains poorly developed. How can any specific result of such test be interpreted? Which value identifies schizophrenia patients (i.e., cutoffs)? If a disrupted cognitive process may be seen as part of the disease symptoms, the application of specific neuropsychological tests should detect these changes and demonstrate efficiency in diagnosing the condition (i.e., the poor cognitive performance of schizophrenia). Further information on the prevalence and diagnostic values of the schizophrenia disrupted cognitive process is necessary to elucidate the practical use and the theoretical directions of this concept.
The main goal of this study was to evaluate the performance in neuropsychological tests of schizophrenia patients and healthy matched volunteers for discriminative and diagnostic values. There were selected neuropsychological tests commonly used for the assessment of cognitive functions of the frontal lobes, and for more diffuse functions to determine their clinical applicability in assessing schizophrenia patients.
A cross-sectional study was designed for evaluating the ability of neuropsychological tests to identify cognitive pattern in schizophrenia. The diagnosis of schizophrenia itself (latent class for gold standard)7 was applied in the diagnostic analysis.
Sample definition and subject selection
The group of schizophrenia patients comprised of 36 male outpatients, recruited from the Schizophrenia Outpatient Clinic of a large university hospital in Southern Brazil. Healthy volunteers were 72 male hospital employees, matched for age and educational level. All patients and volunteers signed an informed consent after the nature of all procedures and ethical guarantees were fully explained. Sample size was supported by the rate of the general population's cognitive deficit (5%), an estimated relative risk of 5 for schizophrenia, and a ratio unexposed: exposed of 2:1. Size estimation for cross-sectional studies was processed using the Epi-Info 6.4 software, with alpha and beta errors of 5% and 20%, respectively.
The corroboration of patients' diagnosis and ruling out psychiatric disorders among volunteers were performed using the semi-structured interviews of DSM IV by a certificate psychiatrist. Presence of other psychiatric disorder, neurological or medical conditions, use of psychoactive drugs and substance abuse, which affect cognition, were excluded. Patients kept their regular use of anti-psychotics (Table 1).
Demographic data presented no significant differences between groups. Mean ± SD age of schizophrenia patients was 33.14±6.53 years, and educational attainment was 9.83±2.88 years. Mean ± SD age of healthy volunteers was 32.75±6.95 years, and educational attainment was 10.18±3.61 years. Mean duration ± SD (range) of schizophrenia in years was 11.83±1.07 (2-31) (Table 2).
Each participant underwent neuropsychological tests administered by research assistants who were not aware of the clinical diagnosis of subjects at the time of evaluation (schizophrenia/healthy). The cognitive tests were as follows: Mini Mental State Examination MMSE;6 Spatial Recognition Span;6 Verbal Fluency;5 Stroop test A and B abbreviated version;12 and Wisconsin Card Sorting Test computerized version (WCST).
The last three tests were selected for the inclusion of functions frequently attributed to the frontal lobes and widely used in the study of cognition in schizophrenia. The Mini Mental is a brief and more comprehensive examination measuring several cognitive areas (temporal and spatial orientation, memory acquisition and recall, attention and calculation, and language). The Mini Mental cutoffs were worldwide studied to screen cognitive deficit in several conditions. The Spatial Recognition Span measures attention and visual memory, and consists in the subject indicating the position of the last white circle consecutively placed on a black board out of the view of the examinee. In the abbreviated Stroop test, the subject is first asked to read, as quickly as possible, ten rows with the name of five colors printed with black ink on a white card (Stroop A). Afterwards, a second card containing ten rows of five names of colors printed with unmatched colors are presented and the subject is asked to name, as quickly as possible, the ink color of in print (Stroop B). Time (in seconds) to finish the reading and the number of incorrect ink color naming are the scores of Stroop test (time and error scores, respectively). Verbal fluency consists of free recall, as many as possible, of words beginning with letter "S" and "M" during a 1-minute period, and the score is the number of non-repeated words.
The computerized version of the Wisconsin Card Sorting test is the same test as the manual one. Cards in different shapes, colors and quantity are displayed one at a time in the middle of screen. Individuals have to identify the rule related to each new card, associating to one of the four fixed cards below. The main measures of this test are correct responses, perseverative errors, perseverative responses, responses of conceptual level, non-perseverative errors, number of categories reached, trials to complete the first category, total errors, learning-to-learn score, and failure to maintain set.
All variables were submitted to a pegboard test (normal probability plot) before analyzing them using discriminant and Student's t Test for independent samples. Discriminative power and diagnostic values of cognitive tests were carried out using discriminant analysis and cutoff estimation from receiver operating characteristic (ROC) curves for sensitivity and specificity values. Statistical analyses were carried out using the Statistical Package for the Social Sciences (SPSS PC+) in a personal computer IBM-compatible.
The neuropsychological tests' scores of schizophrenia patients and healthy participants are displayed in Table 3. Patients presented significant lower scores of the Mini Mental Status, spatial recognition span, verbal fluency with letter S and M, the WCST correct answers, WCST responses of conceptual level, WCST categories. Duration, in seconds, to complete the Stroop test part A and B was longer among schizophrenia patients than healthy volunteers. Total error, perseverative error, perseverative response, first category, and percentage of perseverative answer of the WCST were higher among schizophrenia patients. Both groups showed similar performance only for the scores of the Stroop test A and B, and in the WCST failure to maintain set and learning score.
Correct classification for all neuropsychological tests was double checked using discriminant analysis. The canonical functions of the discriminant analysis for verbal fluency (M and S), Stroop test part B (duration), MMSE, Spatial Recognition Span, and the WCST response of conceptual level were higher (Table 4). The percent of grouped cases correctly classified was 83.5% for all tests in the analysis.
The diagnostic value of tests (for the cognitive pattern of schizophrenia) was measured by sensitivity and specificity, and cutoffs were obtained from the ROC curve procedure. Selected tests for this analysis were those presenting higher discriminant values. None of the tests showed adequate values for the identification of cognitive dysfunction for schizophrenia. The best result was that from the Verbal Fluency with letter M (Figure) that showed sensitivity and specificity values of 75 and 65 for the cutoff 11 (Table 5). False negatives were 25% and false positives 35%.
The study aimed the assessment of the clinical applicability (diagnostic value) of neuropsychological tests for the identification of schizophrenia patients because cognitive impairment has been recognized as a core characteristic of schizophrenia. Many studies in the last decades emphasized the cognitive deficit in schizophrenia, specially related to attention, memory and executive tasks,9-13,15 although classification and discriminative power of the neuropsychological evaluation was rarely assessed.4,10,14 These studies raised the hypothesis of cognitive impairment as a frequent clinical characteristic of disease and the cognitive evaluation has been considered a powerful tool for studying the brain and biological mechanisms in schizophrenia. It was intended to establish a correlation between measures of central tendency and variance of cognitive tests and frequency of patients whose test scores were below a cutoff point and differentiated them from the healthy participants. This approach was used to define the value of these instruments for undoubtedly differentiating schizophrenia patients from the age- and education-matched healthy subjects. The comparison of the neuropsychological tests scores between the groups showed lower performance of the schizophrenia group in the most tests. Although statistically significant, similar to the literature, a parallel clinical applicability of these tests was not achieved. The sensitivity of 75 and specificity of 65 for the Verbal Fluency "M" test were just intermediate for a diagnostic tool. The other tests showed poor clinical values (e.g., sensitivity 74 and specificity 56 for the Stroop test, 72 and 56 for the Mini Mental). The scores distribution analysis is based on the groups "central" parameters, which are strongly influenced by extreme values; however, they did not express either discriminative or clinical values.
The absence of a gold standard for cognitive deficit in schizophrenia required the application of the diagnosis itself as the gold standard (latent class).7 The low sensitivity and specificity of various cutoffs and tests assessed in this study did not support the hypothesis of cognitive deficit as a core feature in schizophrenia. If this were a common condition, the tests would be able to identify the deficit or at least the different performance of schizophrenia patients and healthy participants. The cognitive impairment of schizophrenia patients may be of such a sort that is not assessable by these tests. Other neuropsychological tests for the detection of more generalized or specific impairment (e.g., verbal memory) could yield diagnostic efficiency. In the clinical practice today, neuropsychological examinations are often included in the diagnostic procedure, and their results also have an impact on the treatment planning.2 However, this statement is true if the cognitive deficit was a frequent characteristic of the illness.
Identifying the nature, definition, and frequency of symptoms is crucial for the understanding and conceptualization of disease. Schizophrenia remains a disorder of great clinical heterogeneity with symptoms that vary within and between subjects. Some groups of schizophrenia patients may present higher frequency of cognitive deficit.1,3,9,12 The study sample consisted of outpatients with probably less severe illnesses; however, most studies evaluated hospitalized, probably more severe, cases. Further studies for the evaluation of the discriminative power and clinical applicability of neuropsychological tests in sub-groups of illness, as well as older patients, patients presenting a different time course of disease or in the presence of more severe negative symptoms should be carried out.
Márcia L. F. Chaves
Serviço de Neurologia do Hospital de Clínicas de Porto Alegre
Rua Ramiro Barcelos, 2350 sala 2040
90035-003 Porto Alegre, RS, Brazil
Submitted on 31/1/2002. Reviewed on 19/6/2002. Approved on 15/7/2002.
- 1. Aleman A, Hijman R, de Haan EH, Kahn RS. Memory impairment in schizophrenia: a meta-analysis. Am J Psychiatry 1999;156:1358-66.
- 2. American Psychiatric Association. Practice guideline for the treatment of patients with schizophrenia. Am J Psychiatry 1997;154(4 Suppl):1-63.
- 3. Baxter RD, Liddle PF. Neuropsychological deficits associated with schizophrenic syndromes. Schizophr Res 1998;30:239-49.
- 4. Blanchard JJ, Neale JM. Neuropsychological signature of schizophrenia: generalized or differential deficit? Am J Psychiatry 1994;151:40-8.
- 5. Borkowsky JG, Benton AL, Spreen N. Word fluency and brain damage. Neuropsychol 1967;5:135-40.
- 6. Chaves ML, Izquierdo I. Differential diagnosis between dementia and depression: a study of efficiency increment. Acta Neurol Scand 1992;85:378-82.
- 7. Faraone SV, Tsuang MT. Measuring diagnostic accuracy in the absence of a "Gold Standard". Am J Psychiatry 1994;151:650-7.
- 8. Gold JM, Queern C, Iannone VN, Buchanan RW. Repeatable battery for the assessment of neuropsychological status as screening test in schizophrenia, I: sensitivity, reliability and validity. Am J Psychiatry 1999;156:1944-50.
- 9. Gold S, Arndt S, Nopoulos P, O'Leary DS, Andreasen NC. Longitudinal study of cognitive function in first-episode and recent-onset schizophrenia. Am J Psychiatry 1999;156:1342-8.
- 10. Heaton R, Paulsen JS, McAdams LA, Kuck J, Zisook S, Braff D et al. Neuropsychological deficits in schizophrenics: relationship to age, chronicity, and dementia. Arch Gen Psychiatry 1994;51:469-76.
- 11. Hoff AL, Sakuma M, Wieneke M, Horon R, Kushner M, DeLisi LE. Longitudinal neuropsychological follow-up study of patients with first-episode schizophrenia. Am J Psychiatry 1999;156:1336-41.
- 12. Liddle P, Morris D. Schizophrenic syndromes and frontal lobe performance. Br J Psychiatry 1991;158:340-5.
- 13. Mohamed S, Paulsen JS, O'Leary D, Arndt S, Andreasen S. Generalized cognitive deficits in schizophrenia: a study of first-episode patients. Arch Gen Psychiatry 1999;56:749-54.
- 14. Palmer BW, Heaton RK, Paulsen JS, Kuck J, Braff D, Harris MJ et al. Is it possible to be schizophrenic yet neuropsychologically normal? Neuropsychol 1997;11:437-46.
- 15. Weinberger D, Gold J, Goldberg T. Prefrontal function and schizophrenic symptoms. Neuropsych Neuropsychol Behav Neurol 1992;5:253-61.
Publication in this collection
16 Dec 2002
Date of issue
15 July 2002
31 Jan 2002
19 June 2002