Introduction

Anxiety disorders (AD) are among the most common mental health disorders, with a prevalence of approximately 15-31.9%.^1-4 They have high co-morbidity with other psychiatric diseases and a high rate of onset later in life.^5,6 Pediatric AD have been linked to impairing characteristics and difficulties and to lower academic achievement and have a tendency to become chronic.^7,8 These findings highlight the importance of timely diagnosis of pediatric AD to enable prevention of severity, chronicity and harmful effects on other aspects of life through the use of suitable treatments.⁹ Standard instruments are a good option for screening for AD among rural children because non-rating assessment and diagnostic tools are expensive and time consuming and require trained interviewers.¹⁰ Furthermore, the ease of administration of self-report instruments makes them suitable tools for screening, diagnosis and follow-up. The Spence Children's Anxiety Scale (SCAS) is a self-report questionnaire based on criteria from the Diagnostic and Statistical Manual of Mental Disorders, 4th edition (DSM-IV), adapted for children's psychological stages of development.^11,12 It was developed to measure symptoms of generalized anxiety disorder (GAD), obsessive-compulsive disorder (OCD), separation anxiety disorder (SAD), social phobia (SP), physical injury, and panic attack/agoraphobia. The SCAS has been culturally adapted for several cultures and languages and has exhibited good internal consistency for total score and domains.^8,10,13-20 A factor model of SCAS with six correlated factors has been confirmed in some studies, but not in all cultures and populations,^13,19,21,22 since psychopathological symptoms are influenced by cultural factors, genetic differences and educational and socialization conditions.²¹

Since pediatric psychiatric disorders are largely under-recognized and under-treated in developing countries such as Malaysia, screening and treatment of childhood AD, which are common mental health problems, is essential to improve psychological health indicators.²³ There are very few published studies about pediatric AD in urban and rural areas in Malaysia. A trend for mental health problem among Malaysian children to increase was identified from 2006 to 2011. Many researchers believe that this is attributable to recent changes within households and in society, such as family lifestyle; negative and stressful effects of media; and an increased awareness of children's mental health among their parents.²⁴ There is no culturally-adapted Malay instrument specifically for assessment of childhood anxiety symptoms of different AD. Furthermore, there are no studies that have investigated the psychometric properties of the SCAS among rural children. The purpose of this study is to assess anxiety disorder symptoms, as defined by the DSM-IV, in a large community sample of low socioeconomic level rural children and to investigate some of the psychometric properties (factor structure, child-parent correlation, and internal consistency and most frequently experienced symptoms) of the Malay version of the SCAS - Child version (SCAS-C). Additionally, this is one of the first studies using the SCAS in a rural population and the secondary objective is to illustrate differences and similarities in expression of anxiety as measured by the SCAS.

Method

Results

Discussion

Although the core features of AD remain the same in DSM-5 (no age limitation, consideration of minimum duration, exclusion of trauma-related and stressor-related disorders and obsessive-compulsive and related disorders, and separation of agoraphobia and panic attack and some other minor changes),³⁰ there is a growing body of evidence suggesting that patterns of psychopathological symptoms in children are affected by cultural factors which cannot be included in DSM criteria.³¹

There are a number of standard instruments for measuring pediatric AD. In view of its research objectives, the best instrument to match the research objectives of this study had to include the middle childhood age group and had to be DSM-based (the best criteria for diagnosis of AD), time effective, and have both parent and child version to increase the validity of screening.

The present study examined the psychometric properties of the SCAS and the profile of AD in a large sample of Malaysian children in the rural area of Pontian. This study's unique characteristic is that the Malay version of SCAS was used to investigate AD in a sample entirely made up of rural children for the first time. Our results demonstrated that the SCAS-C total score has good internal consistency and the subscales have adequate validity for assessment of anxiety symptoms among Malay children. Cronbach's alpha for the subscales ranged from 0.50 to 0.68.These findings are in the line with previous studies in urban populations,^8,10,14 but internal consistencies were lower for GAD and OCD. The lowest internal consistency score was for physical injury fears, and the highest was for panic/agoraphobia, which is in line with some previous studies^10,17,32 and may be related to the low number of items for fears (each item is related to a specific fear and they are not strongly related to other items) and the high number of items related to symptoms of panic/agoraphobia.¹⁷ There was low internal consistency for GAD and OCD, which could be related to a low degree of variability among GAD and OCD items in our sample's reports and these respondents may have misunderstood or misread the context. There may also be wording and expressions with which the Malay students were unfamiliar, since the cross-culturally adapted version of the Depression Anxiety Stress Scale (DASS) produced similar results in Malaysia.³³ Furthermore, Cronbach's alpha is influenced by scale length (lower numbers of items produce lower coefficients and larger numbers of items produce higher coefficients).³⁴ This effect is significant when the number of items is less than seven.³⁵ Four SCAS subscales have six items and one has five. The reliability of an instrument's scores depends as much on the sample being tested as on the instrument.³⁶ Undoubtedly, the respondents' rural Malay culture might have influenced this study's outcomes and even driven them to give socially desirable answers.³³ In short, it is recommended that items be refined further for the Malay context in order to achieve an SCAS with greater reliability and validity.

The results of CFA showed that a model with five correlated factors had the best fit to the Malay version of SCAS. A five-factor structure is not that surprising, since some other populations also exhibited this structure, such as a sample of South-African children, for example.³¹ These findings are not consistent with the six-factor structure observed in some other countries (e.g., Iran, Bulgaria, and Cyprus).^8,10,13 It is possible that cultural values (social rules, religious worldviews, philosophical points of view, environmental factors, academic and parenting practices) and methodological differences (e.g., age characteristics) could be responsible for these differences in factor structures.⁸ The loading of GAD and SP onto a single construct may indicate that children with SP are more frequently in trouble or plagued by worrisome thoughts.³¹

In agreement with previous studies, a significant moderate correlation was observed between the child and parent versions.^15,37 This merits verification since the mean SCAS-P score was 25.94 while the mean SCAS-C score was 32.84. Furthermore, this significant difference between means was found between all subscales. It is well known that self-reporting children are better informants than their parents responding to parent versions of instruments and that this is related to the internalizing nature of AD which results in the frequency and severity of their problems remaining underestimated by their parents.^17,38

Intercorrelations between domains and total score confirmed that the SCAS is an appropriate instrument for discriminating between different AD on the basis of DSM-IV criteria. In common with reports by Essau et al.⁸ and Mellon & Moutavelis,¹⁸ the lowest intercorrelation between SCAS subscales was between OCD and physical injury fears.

The total SCAS-C score in our research was 32.84, which is in the range of the scores reported in the only five studies that have used SCAS in the middle childhood age group previously, in which scores ranged from 22.3 to 40.02 (Table 6).^{12,16,31,39,40} The results of the present study indicate that the anxiety levels of Malaysian children were relatively high. There are several possible explanations for this finding. Rural children generally have a lower socioeconomic level than urban children, worse living conditions, and inferior education³¹ (all of the other five studies were in urban areas), all of which are predictors of anxiety. Anxiety disorders are a universal phenomenon, but can be more substantial in some cultures than in others. There is extensive evidence showing the influence of cross-cultural variability on the prevalence of AD, even when the same diagnostic instrument is applied.⁴¹

In line with numerous previous studies, boys scored significantly lower for total anxiety, physical injury fears and SAD and lower for other subscales, but without significance.^31,42,43 The higher levels of anxiety in some subscales among older children conflict with the findings of some previous studies. Manifestations of fear and anxiety are at least to some extent culturally determined.⁴⁴ On the other hand, a study by Ahmad et al.²⁴ investigating mental health problems in Malaysia identified a direct relationship between the prevalence of mental health problems and age up to 10-12 years old. When a child grows older, certain academic, peer and family boundaries can be obstacles to sufficient psychological development to reduce illogical anxieties and fears.

The most reported item related to separation anxiety, which is the same as in most Western countries. The most frequently experienced symptoms in Australian and South African children were symptoms of OCD^19,31 and these were the items rated second to fourth by our sample, but they were not ranked within the top ten by German children.²¹ In the Malay educational system, and based on cultural family rules, teachers and parents expect strict discipline from children. In Western countries these expectations have transitioned from heteronomy to autonomy in recent years. Thus, the frequency of OCD symptoms may be related to these exogenous factors.¹⁵ The least frequently endorsed item was in parallel with most of the other countries, e.g., Australia and Japan (Children in the middle childhood age group are accustomed to being in transportation vehicles from birth).

There is no doubt that the characteristics of our rural sample had an influence on some of the differences in this study's results when compared to other studies (mean age, which is among the lowest among samples investigated for SCAS psychometric properties; low parental educational level; low socioeconomic level; crowded families; and some cultural beliefs, such as belief in the existence of ghosts among all age groups).

This study has certain limitations that should be considered when interpreting our findings. First, because of our focus on middle childhood, we only included 9 to 11-year-olds in our study and it is not clear whether our findings can be generalized to other ages covered by the SCAS-C. Second, the participants were recruited from schools and not from clinical settings. Since no diagnostic interview was conducted, the clinical and diagnostic utility of the Malay version of the SCAS has not yet been confirmed. Third, our data sources were self-report and parent-report instruments. Although the best method for diagnosing psychopathology in children is via multiple informants, the internalizing nature of anxiety reduces this accessibility.⁴⁵ Therefore, use of an instrument alone cannot provide a definitive diagnose of the disease. Fourth, this research did not investigate the test-retest reliability of the Malay SCAS in order to examine its stability across time. Finally, this sample may not be representative of all Malaysian children. Future studies in urban areas and also comparative studies between rural and urban areas in Malaysia are recommended.

Conclusion

Although some of the instrument's psychometric properties were different from those observed in some other countries, it nevertheless appears to be useful for assessing childhood anxiety symptoms in this country. The SCAS is inexpensive, easy to administer and score, and time-efficient, so it can be an effective instrument for screening anxiety symptoms to identify at-risk children who may require further evaluation and diagnosis in Malaysia.