Conduct Disorder is one of the most common problems often found in children taken to Puskesmas. Based on SIKM (Mental Health Information System) from 2011 to May 2013, 46.3% of those taken to Puskesmas in Sleman and Yogyakarta were identified with behavioral and emotional disorder. Therefore, screening instrument is needed for early identifying conduct disorder in them. Strengths and Difficulties Questionnaire (SDQ) has been developed to meet the need. The SDQ is a 25-item questionnare about the positive and negative atributes of children and adolescents (4-16 years old). This study aimed to adapt the English version of SDQ-Teacher Reports (TR) into Indonesian version and to examine psychometrical properties and clinical validation of the Indonesian version for screening conduct disorder. The subjects were 161 Elementary students from Sleman and Yogyakarta, boys and girls of 7-13 years old of age. The reliability coeficient using Alpha Cronbach was α=0,773, and the construct validity using Principal Axis Factoring (PAF) showed that SDQ-TR had six structural factors. Clinical validation using Receiver Operating Curve (ROC) revealed that the value of sensitivity were 0.67 with spesificity 0.68 (for the optimum cut-off point ≥5), and using Likelihood Ratio (LR) showed LR (+) was 2.09 and LR (-) was 0.49. Keywords: children, clinical validation, conduct disorder, SDQ-TR

This study aims at examining the agreement level between clinical diagnoses by a senior psychologist in the hospital and diagnoses/screening on Strengths and Difficulties Questionnaire conducted by parents. Using ICD 10 (Indonesian version) as the gold standard, a clinical child psychologist diagnosed 253 male and female elementary school children aged 7-14 years old. Parents of the same children were requested to fill-out the SDQ questionnaire (SDQ-PR).  Psychometric property of SDQ-PR was analyzed using Alpha Cronbach and Principal Axis Factoring Analysis. Screening quality of SDQ-PR was examined using Receiver Operating Characteristic (ROC), Likelihood Ratio (LR+ and LR-), and Chi-square. Reliabilities of SDQ-PR in all subscales were (α = 0.562 to α = 0.684) except subscale of peer-problem (α = 0.174). Different from the original version of SDQ-PR that consists of 5 subscales, this study revealed 8 subscales. LR and ROC analyses for hyperkinetic revealed LR+ = 1.84 and LR– = 0.51, sensitivity = 67.6% and specificity = 63.3% with cut-off score ≥ 6. For behavior disorder, the LR+ = 2.3, LR- = 0.42, sensitivity = 70.96%, and specificity = 69.15% with cut-off score ≥ 4. As for behavior disorder the LR+ = 1.07, LR- = 0.94, sensitivity = 50%, specificity 53.27% with cut-off score ≥ 4. Chi-square score indicated a significant correlation between SDQ-PR and the diagnoses from child psychologist for hyperkinetic and behavior disorder, but not for an emotional problem.