Cross Validation of the Prodromal Questionnaire 16-Item Version in an Adolescent Help-Seeking Population

Abstract

The Prodromal Questionnaire 16-item version (PQ-16) is used as a screener in the early detection of psychosis. We wished to cross-validate it in a help-seeking population of adolescents aged 12-17 who had been referred for assessment and treatment to an outpatient Center for Child and Adolescence Psychiatry (CCAP). To examine the psychometric properties of the PQ-16 for adolescents, we used reliability analysis and receiver-operating-characteristic (ROC) curves. The internal consistency of the 16-item version was good, with Chronbach's α = 0.84. A cutoff of 7 or more items marked as true, showed good sensitivity (.85) and acceptable specificity (.42). Half (51.3%) of the above cutoff sample were classified by the Comprehensive Assessment of At-Risk Mental States (CAARMS) assessment as having Ultra-High Risk (UHR) status or as having reached the psychosis threshold. Choosing a higher cutoff of 8 produces less false positives, but also loses 8% false negatives. Differentiation between boys and girls showed that the use of distress scores improved sensitivity and specificity values for girls but worsened them for boys. When a cutoff score of 7 or more items marked as true is used, the PQ-16 is a feasible instrument with acceptable screening properties for UHR and psychosis in boys and girls aged 12-17 attending a CCAP.