Presentation on theme: "Diagnostic Efficiency of Adolescent Self Report: Detecting Conduct Disorder in Community Mental Health Katherine Bobak Kate Bobak, Department of Psychology;"— Presentation transcript:
Diagnostic Efficiency of Adolescent Self Report: Detecting Conduct Disorder in Community Mental Health Katherine Bobak Kate Bobak, Department of Psychology; Kristen Green, Applewood Centers; Maya Brown, Department of Psychology; Oren Meyers, Mandel School of Applied Social Sciences; Eric A. Youngstrom, University of North Carolina at Chapel Hill Case Western Reserve University, University of North Carolina at Chapel Hill Background Conduct Disorder is often underdiagnosed clinically, even though it is associated with poor treatment outcomes and tremendous societal costs. There are concerns about the diagnostic efficiency of various measures (e.g., Youngstrom, Findling, Calabrese, et. al, 2004) and it is not known how well scores on adolescent self-reports correlate with their Conduct Disorder (CD) diagnoses. Promising candidates include the Achenbach Youth Self Report (YSR) or Frick and Hare’s Antisocial Process Screening Device Self Report (APSD) as predictors of Disruptive Behavior Disorder NOS (DBD-NOS), Oppositional Defiant Disorder (ODD), and Conduct Disorder (CD). Purpose Examine the diagnostic dependency of the YSR and the APSD in predicting DBD- NOS, ODD, and CD, using DSM-IV diagnoses for comparison. Participants N=271 72% African American Average age: 13.6 23% White, not Hispanic 52% male 2% Hispanic Antisocial Processes Screening Device (APSD) Predicting DBD-NOS/ODD/CD (AUC=.61, p <.05) YSR Aggressive Behavior Score Predicting DBD-NOS/ODD/CD (AUC=.64, p <.01) Receiver Operating Curves Table 1: AUCs for Personality and Forensic Variables Predicting CD/APD Variable AUC APSD Child Self Report Total Score.61 YSR Aggressive Score.64 Rule Breaking Score.61 Externalizing Score.64 ODD.65 Total Problem Score.58 _____________________________ p <.05 AUCs of.50 indicate chance performance. No AUCs differ from each other, all comparisons p >.10. Discussion ROC analyses demonstrate that no significant differences exist among any of the variables in predicting DBD-NOS, ODD, or CD. Because no variable is significantly better at predicting DBD-NOS/ODD/CD, it is advisable to use the APSD because it has 20 questions to the YSR’s 118 questions. However, if already using the Achenbach YSR clinically, no value added by incorporating APSD. From these results, if the YSR measures are preferred to the APSD measures, one could use only the questions dealing with aggression, and/or externalizing, and/or ODD. Non-target cases included a substantial amount of mood and attention problems, but this is clinically realistic set of circumstances. Future Directions Need to see if APSD or other specialized subscales do a better job of discriminating CD and more serious cases. References Youngstrom, EA., RL. Findling, JR. Calabrese, et. al (2004). Comparing the Diagnostic Accuracy of Six Potential Screening Instruments for Bipolar Disorder in Youths Aged 5 to 17 Years. Journal of the American Academy of Child and Adolescent Psychiatry. Volume 43:7, 847-858. DiagnosisRate (n) Conduct Disorder14% (39) Oppositional Defiant Disorder 25% (67) Disruptive Behavior NOS9% (24) Total “Targets”48% (130) ADHD20% (53) Bipolar Spectrum7% (19) Depression & Dysthymia17% (46) Other8% (23) Total “Nontargets”52% (141) Results Significant AUCs and correlations were found for all variables. No significant differences were found between any of the variables. Abstract Conduct Disorder is often under-diagnosed clinically, even though a misdiagnosis is associated with poor treatment outcomes and tremendous societal costs. There are concerns about the diagnostic efficiency of various measures and it is not known how well scores on adolescent self-reports correlate with their Conduct Disorder diagnoses. I look at the Achenbach Youth Self Report and Frick and Hare ﾕ s Antisocial Process Screening Device Self-Report to look at how well these measures can predict a Conduct Disorder, Oppositional Defiant Disorder, or Disruptive Behavior Disorder-NOS diagnosis. I use DSM-IV (Diagnostic and Statistical Manual of Mental Disorders, 4th edition) diagnoses for comparison. Statistical analyses conducted in the computer program SPSS demonstrated that no significant differences exist among any of the variables in predicting Conduct Disorder, Oppositional Defiant Disorder, or Destructive Behavior Disorder-NOS. The Youth Self Report has 118 questions and the Antisocial Process Screening Device only has 20, so the Antisocial Process Screening Device can yield similar results to that of the Youth Self Report in a fraction of the time. This would shorten the amount of time clients spend taking (and psychologists spend giving) tests. If one is already using the Youth Self Report, however, and does not wish to switch to another measure, one could just look at the questions that deal with aggression, and/or externalizing, and/or Oppositional Defiant Disorder to predict these diagnoses in a shortened amount of time Method Table 1 reports the measures available. Receiver Operating Characteristics (ROC) curves compared the sensitivity and specificity of measures for diagnosing adolescents with DBD-NOS/ODD/CD. Diagnoses made based on Kaufman, Birmaher, Brent, Rao, & Ryan’s Kiddie Schedule for Affective Disorders and Schizophrenia (KSADS) interview by highly trained raters, directly interviewing both youth and caregiver. Questionnaires completed by youth on same day as KSADS. Diagnoses made blind to questionnaire results. Participants were consecutive case series at urban community mental health center, or at academic outpatient clinic.