Introduction: The purpose of this study was to examine the diagnostic accuracy and clinical utility of the DSM-oriented Child and Adolescent Symptom Inventory (CASI-4R) psychosis scale (parent and teacher versions) to identify clinically significant psychotic symptoms in youth outpatients. We will use vignettes to illustrate how this approach can be used in a clinical setting.
Method: Receiver Operating Characteristic (ROC) analyses tested diagnostic accuracy of the CASI-4R psychotic symptoms scores (parent and teacher forms) to identify clinically significant psychotic symptoms in new outpatients (N=700) aged 6.0 to 12.9 years (M= 9.7, SD=1.8) at 9 child outpatient mental health clinics, who participated in the Longitudinal Assessment of Manic Symptoms (LAMS) Study baseline assessment. Because LAMS undersampled participants with low mania scores by design, present analyses weighted produce unbiased estimates. Psychotic symptoms, operationally defined as a score of 3 or more for hallucinations or 4 or more for delusions based on the Schedule for Affective Disorders and Schizophrenia (K-SADS) psychosis items, occurred in 5% of youth. KSADS diagnoses for those identified with psychotic symptoms above threshold included major depressive disorder, bipolar spectrum disorder, attention deficit/hyperactivity disorder, posttraumatic stress disorder, psychotic disorders, and autism spectrum disorder (ASD). Venkatraman tests compared areas under the curve (AUCs) for the CASI-4R parent and teacher scores at all the operating points instead of just the overall AUC estimates, which gives the test more statistical power to detect significant differences (Venkatraman, 2000; Venkatraman & Begg, 1996).
Results: The AUC for parent and teacher report was .83 and .74, ps<.001). The parent report performed significantly better than teacher report for identifying psychotic symptoms above threshold (p=.03). The optimal psychosis screening cut-off score (maximizing sensitivity and specificity) was 2.75+ [corresponding diagnostic likelihood ratio (DLR) = 4.21] for the parent version and 3.50+ (DLR = 5.51) for the teacher version.
Discussion: The CASI-4R psychosis subscale appears clinically useful for identifying psychotic symptoms in children. Using DLRs with this subscale can help clinicians identify the presence of significant psychotic symptoms above threshold in youths seeking outpatient mental health treatment via a subscale that is brief and easy to administer.
Stephanie Salcedo– Graduate Student, UNC Chapel Hill, Chapel Hill, North Carolina
Sabeen Rizvi– Gargi College, Delhi University
Eric Youngstrom– Professor and ABCT Fellow, University of North Carolina, Chapel Hill, North Carolina
Lindsey Freeman– Student, UNC Chapel Hill, Hillsborough, North Carolina
Kenneth Gadow– Stony Brook University Department of Psychiatry
Mary Fristad– Vice Chair for Research & Academic Affairs, Director of Psychology, Ohio State University, Columbus, Ohio
Boris Birmaher– University of Pittsburgh School of Medicine
Robert Kowatch– Nationwide Children’s Hospital
Sarah Horwitz– New York University Langone Medical Center
Thomas Frazier– Cleveland Clinic
Eugene Arnold– Case Western Reserve University
H Taylor– Case Western Reserve University
Robert Findling– Johns Hopkins University/Kennedy Krieger Institute
Professor and ABCT Fellow
University of North Carolina
Chapel Hill, North Carolina
Vice Chair for Research & Academic Affairs, Director of Psychology
Ohio State University