Page 36 - Age of onset of disruptive behavior of residentially treated adolescents -Sjoukje de Boer
P. 36

To examine whether information from YHC files could be used as a predictor for orthopsychiatric treatment later in life, a ROC curve analysis was made (see Figure 1). The remarks of both teachers and YHC professionals were allocated to the following values: 1 “No remarks”, 2 “Neutral remarks”, 3 “Positive remarks”, 4 “Remarks indicating disruptive behavior”. Belonging to the inpatient group was the state variable. The area’s under the curve (AUC), indicating sensitivity (the probability that a child with disruptive behavior at the age of five will belong to the orthopsychiatric sample) and specificity (the probability that a child without disruptive behavior at the age of five will belong to the control group) was significant.
For the teachers’ remarks the AUC was .79 (p=.000, 95% CI = .67-.91), for the YHC professionals’ remarks the AUC was .73 (p=.001, 95% CI = .59-.86). The sensitivity of the teachers remarks was .42 and the specificity .70 when “remarks about disruptive behavior” was taken as a criterion. When “any remark of the teacher” was taken as the criterion, sensitivity was .79 and specificity .95. The sensitivity of the YHC professionals remarks was .50 and the specificity .92 when “remarks about disruptive behavior” was taken as a criterion. When “any remark of the doctor or nurse” was taken as the criterion, sensitivity was 1.00 and specificity .12.
Discussion
The theoretical framework of present study concerned the distinction between individuals with life-course-persistent and adolescence-limited antisocial behavior. In this light, the main goal of this paper was to examine whether an adolescent inpatient group differed from a matched, non-treatment control group on information about early disruptive behavior registered in youth health care (YHC) files. These files contained information that had been gathered at fixed times from birth to age nineteen, and were therefore considered usable to determine early-onset of disruptive behavior reliably. Because much information of a later age was missing in the files, the study focused on information of health scans at the second grade, when the infant was about five years of age.
The first finding was that of the inpatient group over half of the files could not be used, because either the files were untraceable or the information about the examination in the second grade was missing. Of the control group, the percentages of
34





























































































   34   35   36   37   38