K-CPT in a Brazilian sample: description of performance and comparison with North American norms

Abstracts

INTRODUÇÃO: O presente estudo investigou o desempenho de uma amostra de crianças brasileiras de 4 a 5 anos de idade no Conners' Kiddie CPT (K-CPT), uma tarefa computadorizada utilizada na avaliação da atenção e do controle inibitório em crianças. MÉTODO: Os escores de crianças brasileiras, que passaram por uma triagem para excluir déficit de atenção, foram comparados com as normas norte-americanas. RESULTADOS: Os efeitos de gênero e idade na nossa amostra foram similares aos descritos na literatura. Entretanto, a falta de triagem para déficit de atenção pode ter distorcido os dados normativos norte-americanos, pois meninos e meninas brasileiros tiveram melhores escores em quase todas as medidas. CONCLUSÕES: Os procedimentos de triagem aqui empregados, juntamente com a obtenção dos efeitos de idade e gênero previamente descritos, confirmam a adequação dos escores obtidos como referência local de desempenho, que podem ser úteis na avaliação de crianças pré-escolares no Brasil até que normas nacionais sejam estabelecidas.

Teste de desempenho contínuo; crianças pré-escolares; atenção; estudo comparativo

ORIGINAL ARTICLE

K-CPT in a Brazilian sample: description of performance and comparison with the North American norms*

Mônica Carolina Miranda;^I Elaine Girão Sinnes;^II Sabine Pompeia;^I Orlando Francisco Amodeo Bueno^III

^IPhD, Researcher.

^IISpecialist, Psychologist.

^IIIAssociate professor, Universidade Federal de São Paulo (UNIFESP), São Paulo, SP, Brazil.

Correspondence

ABSTRACT

INTRODUCTION: This study investigated the performance of a sample of Brazilian children aged 4-5 years on the Conners' Kiddie CPT (K-CPT), a computer-based task used to assess attention and inhibitory control in children.

METHODS: Scores of Brazilian children pre-screened for attention disorders were obtained using the K-CPT, and data were compared with North American norms.

RESULTS: Age and gender effects on the Brazilian sample were similar to those previously described. However, the lack of screening for attention deficits might have distorted normative scores of the K-CPT in the USA because Brazilian boys and girls had better scores than the North American children in almost all measures.

CONCLUSION: The screening procedures used in Brazil, in addition to the previously described age and gender effects on sustained attention, indicate that the results described in the present study constitute appropriate local performance scores and, as such, are useful in the evaluation of pre-school children in Brazil until national norms are established.

Keywords: Continuous performance test, preschool children, attention, cross-cultural comparative study.

Introduction

Continuous performance tests (CPTs) have become an increasingly popular objective measure of aspects of attention¹ that are often used as an aid in the diagnosis of attention disorders such as attention deficit and hyperactivity disorder (ADHD). At present, the most popular commercial version of a CPT probably is the Conners' Continuous Performance Test (CPT II²), a computerized visual paradigm used for the evaluation of sustained attention. The Conners' CPT-II can be administered to children who are 6 years of age and older and is a widely used paradigm in the investigation of attention deficit disorders in school-age children.^3-5

Recent studies have showed that ADHD symptoms are quite common in children under the age of 6.⁶ It appears that one-third of all ADHD behaviors listed in the DSM-IV-based questionnaire are present in at least 40% of all children.⁶ However, the development of diagnostic instruments directed toward neuropsychological assessment of preschoolers seems to lag significantly behind those available for school-age children.^7,8 An exception is the Conners' Kiddie CPT⁹ (K-CPT), which was recently developed for the evaluation of children aged 4 to 5. It provides a cost-effective method for assessing attention disorders in this age group and can also be used for evaluation of treatment effectiveness in this population.⁹

Compared to the earlier CPTs, the K-CPT is shorter in duration (7.5 minutes instead of 14 minutes) and it uses pictures instead of letters as stimuli. The interstimulus interval (ISI) is either 1.5 or 3 seconds, and stimuli are displayed for 500 milliseconds.

The normative data in the USA, which are described in the manual of the test,⁹ are based on a sample of 314 nonclinical, 4 to 5 year-old North American children. The authors analyzed reaction times, omission and commission errors and found that scores improved with age, as observed for older children,¹⁰ and 3 to 6 year-olds using a different paradigm (C-CPT⁸). Gender effects were not reported. However, in older children (aged 6-18),¹⁰ girls present lower commission errors, higher signal detection abilities, and are more cautious (β ) when responding to the test, but make more omission errors, and have longer and less consistent reaction times then boys.^5,10 In the only study that assessed gender differences in performance in younger children (3 to 6 year-olds), no gender effects were found using C-CPT.⁸

Few cross-cultural studies have been carried out comparing norms from different countries using the Conners' CPT, and none have been published that used the K-CPT. However, local norms are widely recognized as necessary since cultural factors can influence cognitive performance.¹¹ In our study with a sample of Brazilian school-age children using the Conners' CPT II,⁵ age and gender effects were similar to the ones found in other CPT studies. Nevertheless, when comparing the Brazilian and North American samples, the results showed that the Brazilian children had better scores in most measures. Miranda et al.⁵ concluded that these conflicting results were due to differences in sample selection, the North American results having been determined without pre-screening for possible attention deficits or developmental disabilities. Hence, the North American norms might have been established including children with possible attention problems, which may have decreased overall performance scores. This is particularly worrisome, since norms in this task are used for diagnostic purpose.^4,12 Taken together, these findings suggest that local data must be obtained in different cultures to be used as local reference scores. It is also important to establish norms to aid in the diagnosis of attention disorders that are not contaminated with data from subjects who may have the disorder the instrument is used to assess.

Thus, the objective of the present study was to determine performance in a sample of 4 to 5 year-old Brazilian children for the K-CPT. Analyses of age and gender effects were carried out, and the data were compared with those of the standardized North American norms.⁹

Methods

The children selected as subjects for the present study were students of public and private schools in city of São Paulo and reflected the socioeconomic class distribution for this city provided by the Brazilian Association of Market Research Institutes (ABIPEME).

The children initially selected were odd numbers on the schools' attendance list and were enrolled in appropriate grades for their ages, totalizing 312 children aged 4-5. The second step was to administration of the abbreviated Conners scale, adapted to the Brazilian population,¹³ a widely used screening tool for the selection process of research on patients with ADHD.¹³ The scale is rated by the children's teacher and evaluates behavioral problems such as hyperactivity and inattention. The 10 children that presented scores above the cutoff point for age and gender following Brazilian norms of the Conners' scale¹³ were excluded.

After that, the parents were invited to sign informed consent forms authorizing the children's participation. Parents of 82 children did not reply or did not authorize the child's participation. Seventeen children were excluded for presenting a history that could be related to development disorders (meningitis, head trauma, convulsion, and learning disabilities). After the aforementioned exclusions, a total of 104 children took the K-CPT test. Of these, 10 administrations were invalidated (failure to execute task or data not recorded by program), and three children were excluded for being outliers. The final sample comprised 91 children.

Procedures

All procedures of the present study were approved by the local Research Ethics Committee - UNIFESP. The Conners´ K-CPT was carried out exactly as described in the manual.⁹ Children were assessed individually at the school they attended, in rooms with appropriate lighting and low noise levels. The test was presented in a laptop computer. The total time for the administration was around 7.5 minutes after initial training that ensured the child understood the instructions.

During the task, pictures are displayed on the screen and the child is instructed to press the spacebar every time a picture appears except the picture of a ball. Pictures are displayed for 500 milliseconds. There are five blocks, with two sub-blocks each of 20 trials, one with 1.5-second ISI and another with 3-second ISI.

The following measures generated by the test program were analyzed.

1. Errors: measures of lack of response accuracy, divided into two categories:

a) Omission: number of targets the subject fails to respond to, i.e., when the response is not given (omitted) for pictures that are not the ball. High rates normally indicate non-orientated and slow responses.

b) Commission: number of times the subject responds to a non-target stimulus, i.e., when there are responses to the ball picture. Long reaction time (RT) combined with a great number of omission and commission errors indicate inattention, whereas fast RT combined with a lot of commission errors, but with few omission errors, reflect impulsivity.

2. Hit RT: speed and consistency of reaction faster than 100 milliseconds (see perseverations below). The following subtypes of measures are provided:

a) Hit RT: mean RT for all target responses in all time blocks.

b) Hit RT Standard Error (Hit RT SE): consistency of RT expressed as SE for responses to target. High scores indicate highly variable reactions frequently linked to inattention.

c) Variability of SE: variability presented by subjects to their own general SE.

3. Signal Detections Theory - derived measures:

a) Detectability (d'): a measure of the subjects' discrimination level between targets and non-targets. Higher d values indicate better discrimination between target and foil stimuli.

b) Response Style (β): indicates the subjects' response criterion. Cautious subjects who do not respond often present higher values for this index. Subjects who respond more freely and are less concerned about failing have lower rates.

c) Perseverations: due to physiological limitations, responses are impossible in less than 100 milliseconds after the stimuli are presented, thus reaction times faster than 100 milliseconds are classified as perseverations. A high rate of perseverations is a result of anticipatory responses (indicating impulsivity), random responding (indicating severe impairment), or very slow responses to the previous stimulus (indicating inattention).

4. By Blocks Results: the K-CPT is presented in five blocks, which allows the assessment of changes over time and vigilance, and also the consistency of responses as the test progresses. Two measures are provided:

a) Hit RT Block Change: changes in mean RT over time blocks. Positive values indicate a slowing RT as the blocks are presented, whereas negative values indicate that the RT is faster as the test progresses.

b) Hit SE Block Change: changes in RT SEs as blocks are presented. Positive values indicate less consistency in reaction times, suggesting a possible loss of vigilance, whereas negative values indicate higher response consistency as the test progresses.

5. By ISI Results: enable the assessment of subjects' ability to adjust to changing ISI. There are two measures:

a) Hit RT ISI Change: changes in mean RT over the ISI sub-blocks. Positive values indicate a slowing RT as the ISI increases, whereas negative values indicate a faster RT.

b) Hit SE ISI Change: changes in mean SE over the ISI sub-blocks. Positive values indicate less consistency in the RT during longer ISI, and negative values indicate more consistency during longer ISI.

Statistical analysis

The statistic programs used were SPSS version 11.0 and SAS's proc. GLM version 8.01. The significance level was set at p < 0.05.

The performance of the sample of children from the present study, according to age (in years) and gender, was determined using the scores of each variable made available by the program package, except for the t-score and percentiles because these are transformed values in relation to the normative data obtained in the USA. To evaluate age and gender effects, two-way ANOVAs were applied for each of the individual variables. For the analysis of proportion of children in each gender per age groups (4 and 5 years), the chi-square test was used. For the analysis of the mean age of each gender in both age groups, the Student t test for independent samples was carried out. The comparison between data obtained from the Brazilian authors and data provided by the authors of the North American standardization was carried out by age group and gender, using the Z test for a mean with known variance.¹⁴ Comparison between the Brazilian and North American samples in relation to the mean age in each age group, number of male and female children, and socioeconomic status was not undertaken due to the fact that such data from the North American norms were not provided by the authors.

Results

Brazilian sample

Table 1 shows the distribution of the sample in the present study in relation to the age group and gender; Table 2 shows the statistical results; and Table 3 displays the performance scores: the number of male and female subjects was similar in all age groups (p = 0.75), and mean age of boys and girls in both age groups was equivalent (p = 0.29). There was no interaction between gender and age effects. Age effects revealed that the 4-year-olds had worse scores on the percentage of omissions, Hit RT, Hit RT SE, variability, perseverations, Hit RT ISI Change, and Hit SE ISI Change measures. Gender effects showed that girls had a lower percentage of commission errors and higher d' and indexes.

Table 2 - Click to enlarge

Table 3 - Click to enlarge

Comparison of the Brazilian and North American samples (Tables 4 and 5)

Table 4 - Click to enlarge

Table 5 - Click to enlarge

Both boys and girls aged 4 from the Brazilian sample presented better scores in percentage of omission errors, Hit RT, Hit RT SE, and perseverations. The Hit SE Block Change measure was significantly better only for Brazilian boys aged 4, whereas Brazilian girls had better performance on Hit RT ISI Change. The remaining measures for the 4 year-olds in both samples did not differ significantly. Better scores were also observed for the Brazilian 5 year-old boys and girls for percentage of omissions, Hit RT, Hit RT SE, and Hit RT Block Change. The Brazilian boys aged 5 also had better scores for the percentage of commissions and Hit SE ISI Change, and the Brazilian 5 year-old girls also had better Hit SE Block Change rates.

Discussion

In the present study, age and gender effects on performance were observed. These factors are highly relevant in determining performance on several CPT measures in older children.^15,16 It must be kept in mind, however, that there are very few studies investigating these effects, and that the CPT versions administered in each one of them differ significantly, rendering comparisons between them difficult.¹⁷

In terms of age, we found improved performance in the older children, which is in agreement with prior studies.^2,8,18 Specifically, age was important in determining score improvement on variability, perseverations, Hit RT ISI Change and Hit SE ISI Change, as observed in school-age children.^5,10

Brazilian girls made less commission errors, had better ability to distinguish signals (d' index) and were more careful while responding (β index) than Brazilian boys, confirming previous findings in older children.^5,8 Also, gender effects on d,^5,15 β and commission errors^2,10,15,16 are consistent with previous findings in older children when different CPT versions were used.

As regards to gender differences in sustained attention of children aged 6 and younger, Conners,² who provided the North American norms for the K-CPT, did not analyze such effects. Kerns & Rondeau,⁸ who did so, did not observe gender differences in performance in contrast with our findings, possibly because they used a different test version (C-CPT test), and their subjects had a wider age range (3-6 year-olds), which could have masked the effects observed here.

In the comparison between the Brazilian and North American samples, we found differences in several measures as observed previously for older children.⁵ Brazilian boys and girls had better scores than the North American children for RT (Hit RT, Hit RT SE) and percentage of omission errors. In the remaining measures, the differences depended, additionally, on age and gender, and again showed better scores for the Brazilian sample.

The finding that the Brazilian children had better scores can be interpreted in two ways considering that sociocultural experiences are supposed to influence cognitive functioning in an unknown manner,¹⁹ as well as gender, age, and schooling:¹¹ either this test is sensitive to such yet unspecific cultural factors or differences arose from differences in methodology.⁵ Concerning the first hypothesis, it could be argued that the age and gender effects obtained in the present study were somewhat equivalent to those previously reported in different locations using other CPT versions in older children, corroborating the widespread idea that performance on CPT measures is relatively constant in different cultures.¹¹ As regards to the second hypothesis, because the present and the North American samples were matched for age and schooling, results could only be accounted for by differences in socioeconomic class, different proportion of girls and boys in each age group, and/or attention deficit screening. The author of the K-CPT manual⁹ did not provide subjects' socioeconomic status, so it could be that his sample included only children who were underprivileged in comparison to the Brazilian sample, and thus had worse scores. This would be surprising because, in the present study, subjects were selected to reflect the socioeconomic status distribution of the city of São Paulo, located in a developing country in relation to the USA and in which a much greater percentage of people are underprivileged in comparison to most cities in North America. In terms of gender distribution, despite overall equivalent proportion of boys and girls in the North American sample, information was not provided on the equivalence of this proportion in each age group. We do not believe that possible differences in this distribution is a plausible explanation for our results, since a preponderance of boys or girls would have led to gender effects different from those previously reported.

We believe that our findings of better overall scores in comparison to the North American norms are more easily explained by differences in the screening procedure.⁵ Firstly, in the present study, the percentage of subjects excluded for possible behavioral and/or developmental disorders is compatible with that described in the literature, despite variations between studies.²⁰ Since children with such disorders have worse sustained attention performance,^7,18 if their scores had been included in the overall means, performance would have likely decreased and may have matched data obtained in the USA. Conners⁹ in fact acknowledges that their subjects were not pre-screened "so a proportion of the general population data will include cases with attention problems" (p. 52). The main implication of this is that normative CPT studies must include screening for attention disorders if they are to be used to assess children's likelihood of having attention deficits. Otherwise, a great number of false negatives may be obtained.

Because "the pattern of performance may be entirely different in clinical samples"¹⁸ (p. 16), lack of screening makes it difficult to determine the effect of cultural background on sustained attention during development, as well as that of age and gender. Also, lack of screening may distort scores on tests such as the K-CPT, thus possibly leading to problems in diagnosing ADHD and other attention disorders in this age group.

Some limitations of our study must be pointed out, however. First, there was a considerable loss of subjects from the original sample due mostly to lack of parental consent, so that our final sample size was greatly reduced. In addition, we evaluated subjects from only one city. Hence, future studies are necessary to obtain Brazilian norms for the K-CPT that include larger samples, children from various regions of Brazil, as well as clinical groups, so that adequate psychometric parameters for our population are determined.

Despite these limitations, our study is consistent with previous data on age and gender effects and clearly shows that the North American norms are not adequate for local use. Our data may be thus used as a guide that enables the determination of the deviation some 4 to 5 year-old children may present from expected scores for their age and gender, enabling the identification of possible attention disorders in this country.

Conclusions

In sum, the screening procedures used here, in addition to the finding of previously described age and gender effects on sustained attention, indicate that the results described in the present study constitute appropriate local performance scores and, as such, are useful in the evaluation of preschool children in Brazil until local norms are established.

Acknowledgments

The authors would like to thank and express their appreciation to Camila Cruz Rodrigues for her assistance in collecting data; the Board of Schools EE Pandiá Calógeras, Externato N. Sra. Menina, EE Heitor Carusi, EE Padre Manoel de Paiva, EE Homero dos Santos Fortes, for permission and collaboration in data collection; and the volunteer children and their parents.

References

• 1. Riccio CA, Reynolds CR. Continuous performance tests are sensitive to ADHD in adults but lack specificity. A review and critique for differential diagnosis. Ann N Y Acad Sci. 2001;931:113-39.
• 2. Conners CK. Conners' continuous performance test for Windows [computer program]. Toronto: Multi-Health; 2002.
• 3. Epstein JN, Erkanli A, Conners CK, Klaric J, Costello JE, Angold A. Relations between Continuous Performance Test performance measures and ADHD behaviors. J Abnorm Child Psychol. 2003;31(5):543-54.
• 4. McGee RA, Clark SE, Symons DK. Does the Conner's Continuous Performance Test aid in ADHD diagnosis? J Abnorm Child Psychol. 2000;28(5):415-24.
• 5. Miranda MC, Sinnes EG, Pompeia S, Bueno OFA. A comparative study of performance in the Conner's Continuous Performance Test between Brazilian and North American children. J Atten Disord. 2008;11(5):588-98.
• 6. Smidts DP, Oosterlaan J. How common are symptoms of ADHD in typically developing preschoolers? A study on prevalence rates and prenatal/demographic risk factors. Cortex; 2007;43(6):710-7.
• 7. Mahone EM, Pillion JP, Hoffman J, Hiemenz JR, Denckla MB. Construct validity of the auditory continuous performance test for preschoolers. Dev Neuropsychol. 2005;27(1):11-33.
• 8. Kerns KA, Rondeau LA. Development of a continuous performance test for preschool children. J Atten Disord. 1998;2(4):229-38.
• 9. Conners CK. Conners´ Kiddie continuous performance test for Windows [computer program]. Toronto: Multi-Health; 2000.
• 10. Conners CK, Epstein JN, Angold A, Klaric J. Continuous performance test performance in a normative epidemiological sample. J Abnorm Child Psychol. 2003;31(5):555-62.
• 11. Levav M, Mirsky AF, French LM, Bartko JJ. Multinational neuropsychological testing: performance of children and adults. J Clin Exp Neuropsychol. 1998;20(5):658-72.
• 12. Losier BJ, McGrath J, Klein RM. Error patterns on continuous performance test in non-medicated and medicated samples of children with and without ADHD: a meta-analytic review. J Child Psychol Psychiatry. 1996;37(8):971-87.
• 13. Brito GN. The Conners Abbreviated Teacher Rating Scale: development of norms in Brazil. J Abnorm Child Psychol. 1987;15(4):511-8.
• 14. Bussab WO, Morettin PA. Estatística básica. 4Ş ed. São Paulo: Atual; 1987. Pp. 245-6.
• 15. Lin CC, Hsiao CK, Chen WJ. Developmental of sustained attention assessed using the continuous performance test among children 6-15 years of age. J Abnorm Child Psychol. 1999;27(5):403-12.
• 16. Greenberg LM, Waldman ID. Developmental normative data on the test of variables of attention (T.O.V.A.). J Child Psychol Psychiatry. 1993;34(6):1019-30.
• 17. Denney CB, Rapport MD, Chung K. Interactions of task and subject variables among continuous performance tests. J Child Psychol Psychiatry. 2005;46(4):420-35.
• 18. Hagelthorn KM, Hiemenz JR, Pillion JP, Mahone EM. Age and task parameters in continuous performance tests for preschoolers. Percept Mot Skills. 2003;96(3 Pt 1):975-89.
• 19. Ratner C. Outline of a coherent, comprehensive concept of culture. Cross-Cult Psychol Bull. 2000;34(1-2):5-11.
• 20. Rohde LA, Halpern R. Transtorno de déficit de atenção/hiperatividade: atualização. J Pediatr (Rio J). 2004;80(2 supl):S61-70.

*

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) (02/09395-3) e Associação Fundo de Incentivo à Psicofarmacologia (AFIP).

Publication Dates

• Publication in this collection
  24 Aug 2009
• Date of issue
  2009

History

• Accepted
  11 Dec 2008
• Received
  04 Sept 2008