Diagnostic Accuracy of Self and Parent Rating Scales in the Prediction of Psychiatric Diagnoses in Children and Adolescents

Abstract

Although parent and self rating scales are often used in clinical and research settings, their accuracy in the prediction of psychiatric disorders according to ICD-10 or DSM-IV often remains unclear. In the present thesis the diagnostic accuracy of three parental rating scales in the prediction of Attention-Deficit-Hyperactivity Disorder (ADHD) and Oppositional Defiant Disorder (ODD) and two self rating scales in the prediction of adolescent depression were tested in three separate studies. The first study found the recently introduced DSM-oriented attention problem scale of the Child Behavior Checklist (CBCL) more adequate than the previous empirical defined attention problem scale in the identification of ADHD. This was in particular true for subjects in a clinical sample referred for various psychiatric disorders. A cut-off score of 5 was recommended for clinical practice. In a second study a sample of adolescents with clinical depression was compared to a sample of unreferred community controls. The Youth Self Report (YSR) and the Center of Epidemiological Studies- Depression Scale (CES-D) showed excellent ability in the discrimination of these two samples. A range of acceptable cut-off scores between 5 and 9 on the YSR affective problem scale and between 12 and 31 on the CES-D scale served best in the prediction of clinical depressive episodes in adolescents. In a third study the Conners’ Parent Ratings Scale Revised (CPRS-R) and the parent version of the Strength and Difficulties Questionnaire (PSDQ) were tested in the prediction of ODD in a large transnational sample of ADHD referred children and adolescents. Furthermore, the construct validity of three previously described dimensions of ODD was examined and finally the accuracy of the CPRS-R and the PSDQ were tested in the prediction of these separate ODD dimensions. The CPRS-R oppositional scale and the PSDQ conduct problem scale showed adequate diagnostic accuracy. Furthermore, the construct validity of three ODD dimensions labeled ODD-irritable, ODD-headstrong and ODD-hurtful was confirmed. Furthermore, our results convincingly show that a three factor structure of ODD is more appropriate than a single general factor of ODD. The CPRS-R emotional lability scale was able to predict ODD-irritable significantly. Overall, these three studies confirmed the diagnostic accuracy of clinical rating 2 scales in the prediction of psychiatric disorders in youth. Furthermore, these results are of clinical importance as newer and diagnosis-oriented rating scales showed better diagnostic accuracy and can be recommended for the initial psychiatric assessment of children and adolescents. However, despite the good validity of rating scales, further information on age of onset, continuity, impairment, specificity of symptoms and information about other psychiatric disorders should be included in order to arrive at the final diagnosis.