Screening for Conduct Disorder in an Adolescent Male Sample from Colombia

Abstract

This study assessed the validity, reliability, and utility of a screening measure for detecting the signs or symptoms of Conduct Disorder in male adolescents from schools in Medellín, Colombia. A first study examined the differences between 70 male offender adolescents (aged 12-16 years) attending alternative education institutions, and 68 (sex, age, and socioeconomic status (SES)-matched) adolescent controls attending regular schools. A CD checklist (CD-CL) was constructed to be used by mothers (CD-CL-M) and for self-report by adolescents (CD-CL-SR). The validity of the screener for CD diagnosis was supported by significant differences between groups (ANOVA, p<.001). The CD-CL-SR had better sensitivity/specificity with a diagnostic cut-off point of 5 (sensitivity=95.3% and specificity=90.5%) than the CD-CL-M. A second study used the CD-CLSR with a random sample of 190 male adolescents (aged 12 to 16 years) from schools of low, middle and high SES. Reliability as assessed by Cronbach's alpha was 0.86. An epidemiological cut point of 5 classified 35.5% of the adolescents assessed as probable CD cases. A psychometric cut point at T score>59 (85th percentile) estimated 16.8% of the sample as probable CD cases. A psychometric cut point at T score>64 (90th percentile) revealed that 10.5% of the sample would be severe CD cases. Some significant differences (p<0.001) in proportions of CD adolescents were found between age and SES groups. It was concluded that such a high frequency of conduct problems in adolescence argues for the need for preventive programs in Colombian schools.