Cross-Cultural Aspect of Behavior Assessment System for Children-2, Parent Rating Scale-Child: Standardization in Korean Children

Accurate, objective and replicable assessment of psychopathology not only has advantages in clinical practice but also is essential to successful research.1 Understanding, synthesizing and interpreting results from research across diverse populations is only possible if there are comparable studies on similarly characterized research participants who were examined with assessment instruments that can be used reliably and validly in multiple cultural and ethnic groups. The scarcity of such instruments has limited cross-national studies, in particular, the contribution of psychiatry research on Asian children and adolescents to the international research community.

The Behavior Assessment System for Children (BASC) is a comprehensive behavioral rating and self-report scale for the evaluation of not only negative (clinical) but also positive (adaptive) dimensions. It has well-established psychometric properties.2 The BASC-2, a revision of the original BASC, has 100–185 questions, depending on the version (parent, teacher, child), for evaluating psychopathology, personality and adaptive functioning in children and young adults, from 2–25 years old.3 The BASC-2 includes a Teacher Rating Scale, Parent Rating Scale, Self-Report of Personality, Structured Developmental History, and a Student Observation System. The Parent Rating Scale offers a broad assessment of the child's behavioral problems and adaptive functions in both the community and home settings. It consists of three forms: preschool (age 2–5), child (age 6–11), and adolescent (age 12–21). The Parent Rating Scale-Child (PRS-C) includes 9 clinical scales (hyperactivity, aggression, conduct problems, anxiety, depression, somatization, atypicality, withdrawal, and attention problems), 5 adaptive scales (adaptability, social skills, leadership, activities of daily living and functional communication) and 4 composite scales (internalizing problems, externalizing problems, adaptive skills and behavioral symptoms index).3 The reliability and validity of the BASC-2 have been established in various samples. For example, the BASC-2 PRS-C has item consistency in general norm samples with Cronbach's alpha ranging from 0.75 to 0.88, as well as in clinical samples: 0.81 to 0.88.3

The BASC-2 also has demonstrated utility for assessing children with attention deficit/hyperactivity disorder (ADHD),4, 5 in part because the BASC-2 not only detects ADHD symptoms but also other common, comorbid conditions. Indeed, this serves as an illustration that instruments with integrated, multiple behavior ratings, such as the BASC-2, facilitate a more comprehensive appreciation of co-occurring symptoms of developmental psychopathology in relatively common conditions such as ADHD or autism spectrum disorder (ASD).

Parent questionnaires, such as the BASC-2 PRS-C, are often used to assess psychopathology in children. However, parental identification of their offspring's behavioral problems, as well as parental perception of the need for care, can differ depending on cultural background,6 suggesting that it is important to validate the psychometric properties of assessment tools across diverse ethnic and cultural populations when studying developmental psychopathology. This facilitates interpretation and integration of research findings from a growing international developmental psychopathology research community.

Cultural differences in children's behavioral problems and adaptations have been examined for some parental questionnaires, including the Child Behavior Checklist (CBCL) and BASC-2. Crijnen, et al.7 examined variations in CBCL scores across 12 cultures, and reported cross-cultural differences in overall mean scores on the CBCL syndrome scales. Another international comparison study revealed that there were small to medium differences in mean CBCL scale scores among 24 different cultural groups.8 With regard to the BASC-2, a study comparing Colombian and US school children reported cross-cultural variability in anxiety and somatization.9 Another study, comparing Korean, Korean American, and Caucasian American children, reported differences in social, emotional, behavioral, and school adjustment functioning among these groups.10 Such efforts are limited, especially beyond Spanish and English speaking populations, because the BASC-2 PRS-C currently has established reliability and validity only in these two languages.3 There remains a significant need to demonstrate the applicability and generalizability of the BASC-2 in multiple languages and cultures.

The need for common measurement tools across diverse populations led to this study to complete: 1) translation/backt-ranslation and cultural adaptations of the BASC-2 PRS-C for use with Korean children, including an examination of the psychometric properties of the Korean version of BASC-2 PRS-C (K-BASC-2 PRS-C); and 2) cross-cultural comparisons of the psychometric properties and scores between US and Korean children in both general and clinical populations from various diagnostic groups.

Our study indicates that the K-BASC-2 PRS-C is an excellent instrument with psychometric properties that are suitable for comprehensive screening and assessment of behavioral problems and adaptive functions in Korean children. The corresponding subscale mean T scores in the Korean clinical population were significantly higher than those in the general population. T scores on the adaptive scales for the clinical population were near or below 40 and significantly lower than those in general population, indicating that the K-BASC-2 PRS-C validly distinguishes the functional levels between clinical and non-clinical populations in Korean children.

In both the general and clinical population, the internal consistency and reliability of clinical and composite scales were comparable to the US population. The Cronbach's α for the clinical scales ranged from 0.69 to 0.83 in the general population which is almost above the conventionally acceptable level.15 Most of the alpha values in the clinical sample for our study are comparable to those reported in US children (0.77 to 0.96), with the exception of a few subscales. This is likely due to the small number of subjects in the clinical subsamples as small sample size has a tendency to deflate alpha values.16

There are some subtle, but interesting, differences noted in the mean scores of the BASC-2 PRS-C between Korean and US populations. Raw scores for most of the subscales in the same percentile ranges were lower for Korean children than for US children. The most prominent of these were behavioral problems with a mean hyperactivity subscale T score of 58, in the ADHD group which was just below the “at risk” range. Previous studies reported similar findings in Korean children with ADHD using the ADHD Rating Scale (ARS):17, 18 the ARS raw scores in a general Korean population sample were lower than those reported for American children.” Two plausible explanations for these findings seem most likely: 1) different thresholds for reporting the same externalizing behavioral problems between US and Korea parents; or, 2) differences in prevalence of ADHD. Previous studies of developmental psychopathology in Korean children indicated that the prevalence of disruptive behavioral problems, including ADHD, is comparable to those in Western countries,20, 21 making the second explanation seem unlikely. Therefore, observed discrepancies in the hyperactivity scores between Korean and US children may reflect cultural differences in parental reporting of their children's behaviors.

Cultural differences in reporting children's behavior problems have been also identified in other emotional/behavioral problems. In a multi-cultural comparison study of the CBCL in preschool children, mean Total Problem Scores for Korean children ranked 4th from the bottom among 24 societies and lowest for DSM-IV Oppositional Defiant Disorder.8 Other studies have documented that Korean parents report fewer problems in their offspring on the CBCL internalizing and externalizing behavior scales than do parents from the US22 or Australia.23 It has been hypothesized that Korean parents are reluctant to report behavior problems of their children due to negative societal stigma related to having behavior and emotional problems.22

Additionally, attention problems scores were high in all clinical groups in our study. Our finding is also consistent with previous studies that the attention problems score was elevated not only in children with ADHD but also for children with depression24 and/or ASD.25

It is plausible that patterns of parental underreporting of psychopathological signs and symptoms may be associated with lower rates of service-seeking behaviors for mental health problems in Korean children. Indeed, only 10% of Korean children with ADHD receive treatment,26 which is much lower than the 59% of US children.27 This difference warrants further investigation.

Our study demonstrates that most of the adaptive scales were positively correlated with IQ scores in ASD. In ASD children, previous research indicates that IQ predicts adaptive behavior, particularly in the communication sub-domain.28, 29 In our study, the positive correlation between IQ scores and adaptive function with the strongest correlation on the functional communication subscale was also driven by ASD. Functional communication and IQ appear to be two strong predictors of outcomes in ASD.30 Therefore, these findings lend further support to the importance of functional communication as part of the therapeutic process in ASD.

On the other hand, the atypicality and withdrawal subscales were negatively correlated with performance and/or full scale IQ in the ASD samples. These findings are somewhat different from a previous study indicating that verbal IQ was correlated with autism spectrum symptomatology.31 Additionally, a negative correlation between attention problems and IQ was observed in our ASD sample which is also not consistent with a previous study that reported no correlations between ADHD symptoms and IQ in children with ASD.32 Such discrepancies may have stemmed from differences in the study populations; the use of community-ascertained, rather than a clinical sample, likely led to a difference in the severity of the ASD cases and/or the presence and severity of comorbidities.

Strengths of our study include the large number of subjects from a general population sample, inclusion of a unique community-ascertained sample of children with ASD, and comparison of psychometric properties with existing, similar measures that are already validated for use with Korean children. The limitations include: 1) The general population sample was drawn from only one area in Korea. However, one quarter of the entire Korean population lives in metropolitan Seoul from where our study population was drawn. While the generalizability of our findings to the children in rural areas of Korea will be tested in a future study, our current findings should be generalizable to those in urban and suburban areas. 2) Individuals with ASD in our study were from an epidemiologically-ascertained, community sample. They have clinical characteristics that are distinct, especially in comparison to ASD samples ascertained mainly from clinics. For the epidemiologically ascertained sample, the mean IQ was 96.0±25.0 and ASD symptom severity, as measured by the Social Responsiveness Scale total T score, was 69.7±16.6. While these are important differences that may limit the comparability of our results to findings from studies conducted solely in clinically-ascertained children with ASD, they may still more accurately reflect a broader ASD phenotype. 3) The diagnoses in the clinical sample were limited mostly to ADHD and ASD, along with a small number of emotional and disruptive behavioral disorders. Inclusion of other diverse diagnoses is required in a future study. 4) Inter-rater and test-retest reliability should be examined in a future study.

Taken together, our findings indicate the importance of standardization and cultural adaptation for instruments that measure developmental psychopathology. This is especially pertinent when the instruments are introduced to and for use in populations outside of those in which the instruments were originally developed. Direct comparisons of the results or integration of research findings across different populations using instruments for which there is not a complete, rigorous standardization and cultural adaptation may lead to potentially misleading or erroneous results. On the other hand, standardization and cultural adaptation may create remarkable opportunities to not only increase sample sizes but also provide insights into cultural variations in development and psychopathology. In addition, cross national and cross cultural use of instruments increases the opportunity to establish large samples necessary for genetic, gene-environment and other studies.

Our study indicates that the K-BASC-2 PRS-C is an effective and useful instrument with psychometric properties that allow to measure general developmental psychopathology. Further study in independent, general population samples from various areas and large clinical samples with diverse diagnostic groups from different cultural and ethnic backgrounds is needed to corroborate our findings.

Supplementary Table 1

Correlation between Subscale Scores of K-BASC-2 PRS-C and K-CBCL in a Clinical Sample (n=42)