Journal of Child and Adolescent Health

Reach Us +44-1647-403003

Research Article - Journal of Child and Adolescent Health (2018) Volume 2, Issue 2

Describing psychological and behavioural problems in Omani young people: Reliability of the self-reported Strength and Difficulties Questionnaire (SDQ) in Oman.

Al-Mukhaini ZA*, Bekker HL and Cottrell D

Leeds Institute of Health Sciences, School of Medicine, University of Leeds, UK

*Corresponding Author:
Al-Mukhaini ZA
Research Fellow - Leeds Institute of Health Sciences
School of Medicine
University of Leeds, UK
E-mail: [email protected]

Accepted on October 30, 2018

Citation: Mukhaini AZ, Bekker HL, Cottrell D. Describing psychological and behavioural problems in Omani young people: Reliability of the self-reported Strength and Difficulties Questionnaire (SQD) in Oman. J Child Adolesc Health. 2018;2(2):19-23.

Visit for more related articles at Journal of Child and Adolescent Health

Abstract

Aim: The aim of this study is to further assess the reliability of the Strength and Difficulties Questionnaire (SDQ) when completed by school-age children in Oman. The SDQ is an essential tool for screening psychological and behavioural problems in school children but has only been psychometrically tested once in Omani culture without checking the reliability of the impact subscale.

Method: A cross sectional study. Children aged 11-16 from one private and four public schools in two areas of Oman (Muscat; Buraimi) completed the SDQ. Cronbachs alpha and person-product correlations were used to assess internal, and test-re-rest reliability.

Results: 377/422 children completed the SDQ, 138 completed the re-test. The Total Difficulties (T-SDQ) score was normally distributed with 78% scoring normal, 13% borderline and 9% clinical significant scores. Girls’ scores were significantly higher than boys in T-SDQ and in emotional and prosocial subscales. Internal reliability for the T-SDQ was (Cronbach alpha=0.72; Person Product-moment coefficient r=0.71; p<0.001), the impact subscale=0.68 and varied between 0.26 and 0.60 for the other subscales.

Conclusion: The self-reported SDQ using T-SDQ score is a reliable screening tool for psychological and behavioural problems for Omani children. The T-SDQ can be disseminated within private and public-school systems. Having a reliable screening tool for identifying children with psychological and behavioural problems is a significant step to improving mental health pathways of care and outcomes in Oman.

Keywords

SDQ self-reported version, Reliability, Behavioural and psychological problems, Oman, children, Young people.

Introduction

Essential to the early and effective treatment of mental health problems in children is their identification. The Strength and Difficulties Questionnaire (SDQ) is an established measure used to screen for psychological and behavioural problems in children aged between 4 and 17 years old [1,2]. Self (child) and proxy (parent, teacher) completion versions have been used in community (e.g. school) and health care (e.g. hospital) settings. The SDQ consists of 25 items divided into five subscales: Conduct Problems, Hyperactivity Problems, Emotional Symptoms, Peer Problems and Pro-social Behaviour [3-5]. Both the total score excepting the pro-social behaviour items, and sub-scale totals are used to screen children, and assess the impact of treatment. The impact subscale supplement is optionally available for use. It is useful for clinicians and researchers as the respondent would put his/her perception if there is any impact on his/her life at home, school, friend relationship or leisure [1].

The SDQ has been translated into over 80 languages (www.sdq.sdqinfo.com), and psychometric testing indicates it is a reliable screening tool in different cultures [1,6-12]. The self-report version of the SDQ has been translated into Arabic [1] and tested in Yemeni and Palestinian populations [6,13,14]. The SDQ-Arabic has been applied and psychometrically tested once in two large city populations in Oman [15]. In view of shortage in the number of mental health professionals in Oman [16], having a simple and short screening tool will be very useful for the mental health services in the country.

The aim of this study is to assess the reliability of the SDQ with the supplemented impact subscale when completed by school-age children from a port and rural location in Oman, and to make a preliminary assessment of the level of psychological problems in Omani school children. Before the SDQ can be adopted as a general screening tool for psychological problems in Oman, its reliability needs to be assessed. Reliability is the internal consistency of a psychometric scale and its freedom from random error [17,18]. Reliability is defined as the “quantifying precision of a quantitative test” [17]. A reliable scale needs to be reproducible and often assessed by test-re-test as well as the inter-rater reliability [17-19].

Methods

Design and Sample

A cross sectional study employing questionnaire methods in five school across two areas of Oman, Muscat (capital city) and Buraimi (border town and rural). Five schools were selected purposefully to ensure a diverse group of pupils: three Muscat schools - a private school of mixed sex, and two single-sex community schools (one for boys and for girls); two Buraimi schools - single-sex community schools (one for girls and one for boys). Schoolchildren aged 11-16 years were eligible to participate (n=442).

Questionnaire

The Arabic version of the SDQ self-report (child) with Impact Supplement was employed. The subscales scores for Emotional Symptoms (ES), Conduct Problems (CP), Hyperactivity (H), Peer Problems (PP) and ProSocial (PS) are rated 0-10. The total difficulty score (T-SDQ) is calculated by adding up the scores from all items within each sub-scale, excepting those in the pro-social sub-scale, i.e., a range from 0 to 40. Higher scores indicate greater levels of difficulty; 0-15 is indicative of a normal range, 16-19 is indicative of a borderline clinical indicator, and 20+ as clinically significant levels [1]. The impact supplement subscale consists of five questions reflecting the child’s perception that the condition has an impact on her/his life at home or school or leisure or peer-relationship. The impact score provides researchers and clinicians with insight into how the child’s problems interfere with everyday activities and lives. Each item within the SDQ is rated as true (score=2), somewhat true (score=1) or not true (score=0).

Procedure

Ethical approval was granted by the Ministry of Health in Oman. Letters of study participation send to headmasters of all the five schools. and face-to-face meetings between ZA-M and the headmasters; all consented to participate. ZA-M worked with two teachers (ZA and SG) to distribute questionnaires across grade 6-10 classes (age 11-17) in all five schools. Pupils were given a description of the study and asked to complete the SDQ fully at two time-points (a month apart). Participants were informed their results would be treated in confidence by ZA-M and were not associated with their school marks and/or progress. In the private school, questionnaires were distributed by the teachers and the researcher and the researcher was able to address pupil queries; in the public schools, the teachers distributed the questionnaires with guidance from ZA-M on the study purpose and SDQ completion.

Analysis

Data were managed using SPSS. Analyses were carried out to describe the sample and assess the reliability of the SDQ scores using Cronbach’s alpha coefficient for internal consistency [16], i.e., scores greater than 0.70 indicative of relaibility [17]. The test-re-test reliability analyses of the total difficulties SDQ scores was assessed using the Person Product-moment Correlation.

Results

Questionnaires were completed by 442 school children across five different schools in Oman (52 from Buraimi and 390 from Muscat); 201 pupils from the private schools and 241 from the public schools. Children were excluded from the analyses for those not completing the questions necessary to calculate the total difficulties (T-SDQ). There was no difference between those completed the SDQ (n=377) and those with missed data (n=65) in terms of age, sex and schools (Table 1). For the test-retest reliability analysis, a sample of 138 children completed the self-report SDQ again, four weeks later.

Case Types Age in years Sex Number (%) Number of Students defined by school number (%)
Mean ± (SD) 95% CI Male Female Public school Private school
With complete data (n=377) 13.32 ± (1.46) 13.17-13.47 174 (46.2) 203 (53.8) 204 (54.1) 173 (45.9)
With missed data (n=65) 13.41 ± (1.51) 13.03-13.78 22 (33.8) 43 (66.2) 37 (56.9) 28 (43.1)

Table 1. Description of 442 completed questionnaires.

The T-SDQ scores were distributed normally across the sample (Figure 1) and supported by similarities in the mean (11.9 ± 5.3) and median (12.00) average scores, and 5% trimmed mean (11.7) analysis; t (376)=0.673, P=0.50, 95% CI (-.35, 0.72). Most (78%) pupils fell within the normal range for the SDQ; 14% were within the borderline category and 9% clinically significant range (Table 2). The SDQ subscale scores and impact score illustrate all scores were skewed towards the normal range. The mean impact score was 1.26 ± 1.92.

child-adolescent-health-distribution

Figure 1. Histogram illustrating distribution of SDQ responses by participant.

T-SDQ (groups) Male Female Total
Number (%) Number (%) Number (%)
Normal (Scores ≤15) 141 48 152 52 293 78%
Borderline (Scores 16-19) 22 43 29 57 51 13%
Clinically Significant (Scores ≥20) 11 33 22 67 33 9%
Total 174 46 203 54 377 100

Table 2. Distribution of total difficulties score groups in both sexes (n=377).

There was no significant difference in T-SDQ scores between private and public schools (private mean 11.5 ± 5.4; public mean 12.2 ± 5.2; t (170)=-1.58, P=0.116, 95% CI (-1.46,0.16)). Girls scored higher than boys (boys mean 11.2 ± 5.2; girls mean 12.5 ± 5.3; t (375)=-2.37, p<0.05, 95% CI (-2.35, -0.22)). There was no difference between boys and girls in conduct problem, hyperactivity and peer problem subscales’ scores; girls scored higher than boys on the emotional symptoms, pro-social subscales and impact score (Table 3).

Subscales Mean score ± (S.D) Z-value P-value
Both sexes Males Females
Emotional symptoms 3.46 ± 2.22 2.82 ± (2.01) 3.97 ± (2.25) -5.09 <0.001
Peer Problems 2.85 ± 1.70 2.64 ± (1.59) 2.99 ± (1.76) -1.94 NS
Conduct Problems 2.36 ± 1.72 2.38 ± (1.64) 2.34 ± (1.77) -0.791 NS
Hyperactivity 3.35 ± 1.88 3.35 ± (1.88) 3.36 ± (1.89) -0.088 NS
Pro-social 7.91 ± 1.63 7.73 ± (1.59) 8.04 ± (1.65) -2.37 <0.018
Impact Score 1.26 ± 1.92 0.82 ± 1.46 1.62 ± 2.16 -3.86 <0.001

Table 3. Mean scores of all five subscales and impact subscale and Mann-Whitney U test findings of subscales comparing female and male scores (n=377).

Cronbach’s alpha (T-SDQ) was 0.72, indicating good internal consistency; values for the subscales were modest-to-poor: ES (0.60), CP (0.45), Hyperactivity (0.47), PP (0.26), and PS (0.42). For Impact score the Cronbach alpha was 0.68. The test-retest correlation was significantly high (Person Product-moment coefficient r=0.71, p<0.001). Cronbach's alpha (T-SDQ) is like some studies in literature [10,21,22]. Table (4) compares the Cronbach's alpha between present study and literature.

Year & Author Sample size Sample population Country or Ethnicity Internal consistency (Cronbach Alpha)
T-SDQ ES CP Hyperactivity PP PS Impact
Goodman [1] 199 Community UK 0.61 to 0.82 - - - - - -
Goodman, et al. [2] 199 Community and clinic UK Caucasian 0.82 0.8 0.7 0.59 0.7 0.6 -
Koskelainen, et al.  [10] 735 Community Finland 0.71 0.7 0.7 0.57 0.6 0.7 -
Thabet, et al. [13] 322 Community Palestine - 0.5 0.3 0.55 0.2 0.7 -
Emam, et al. [15] 634-815 Community Oman 0.76 0.6 0.5 0.51 0.4 0.5 -
Van Widenfelt, et al.  [20] 1288 Community Netherlands 0.7 0.6 0.5 0.66 0.4 0.6 -
Mansbach-Kleinfeld, et al. [21] 611 Community Isreal 0.72 0.6 0.5 0.71 0.2 0.6 -
Muris, et al. [23] 439 Community Netherlands 0.76 0.6 0.5 0.68 0.4 0.6 -
Giannakopoulos, et al. [26] 1194 Community Greek 0.77 0.73 0.6 0.63 0.6 0.6 -
Du, et al. [27] 690 Community China 0.57 0.6 0.3 0.64 0.3 0.7 -
Ortuno-Sierra, at al. [28] 1474 Community Spain 0.75 0.71 0.6 0.68 0.6 0.6 -
Present study 377 Community Oman 0.72 0.6 0.5 0.47 0.3 0.4 0.68

Table 4. Comparison of Cronbach's alpha of study scores (n= 377) with some other studies in literature.

Compared with children in other countries, Omani children’s total difficulties score is similar to Australian children [23] but higher than children in other cultures [10,22-24]. There were some consistent differences reported between male and female responses on subscales across countries, such as girls scoring high on the Emotional symptoms score [10,24,25-28] and Prosocial score [25,28]. The same findings reported in this study.

Discussion

This study is the first to test the reliability of the supplemented SDQ in a population of Omani schoolchildren from both private (affluent) school and public (general) school. No significant difference in T-SDQ mean score between private and public schoolchildren might indicate that the socio-economic-class has no influence on Omani young people psychological health. Although the T-SDQ has been tested and used in several different English [2-5] and non-English, speaking countries [6-15] there are few examples of it being used in Arabic populations and culture [6,14,16]. The findings indicate the T-SDQ is a reliable screening tool for psychological and behavioural problems in Omani children aged 11-16 years. Having a reliable screening tool for identifying children with psychological and behavioural problems is a first step to improving mental health pathways of care and outcomes. This study illustrated the acceptability of disseminating and utilizing the T-SDQ within private and public-school systems and suggests a model of collaboration between school and health care systems when implemented in practice. The impact subscale showed reasonably good Cronbach alpha coefficient (0.68).

The findings illustrated a normal distribution of responses, and discrimination between normal, borderline and clinical significant responses in accordance with external indicators of mental health prevalence in populations [1]. However, the study did not explore the validity of the items and questionnaire, and it may be Omani pupil’s understanding and/or interpretation of items differ from children in other cultures. This may have implications for the sensitivity and specificity of the test, and implications for service provision. For example, it is unclear if sex differences in responses represent a difference in illness prevalence between girls and boys, or a difference in reporting. It is recognized that girls develop and articulate emotions and social relationships before boys [29], suggesting the T-SDQ may be misclassifying boys as normal or borderline. Future research and audit would be able to address this type of issues and adapt classification algorithms accordingly.

Conclusion

The SDQ looks at strength as well as difficulties and it is more acceptable to respondents and gives richer information. It is short and easy to administer (one side of A4 paper if SDQ alone or two sided A4 paper if the SDQ with impact subscale is used) which makes it a quick, useful screening tool for both clinician and researcher

Acknowledgments

I would like to thank all teachers, students and head teachers for their cooperation without which this study would not be possible. I declare that there is no conflict of interest on this project. I would like to thank government of The Sultanate of Oman that supported me to do this study by sabbatical leave

References