Abstract
Youth anxiety disorders are common and heterogeneous. The Pediatric Anxiety Rating Scale (PARS) is a clinician-administered tool designed to assess overall anxiety severity, independent of specific symptoms. The thresholds on PARS for the reliable change index (RCI), the minimal clinically important difference (MCID), treatment response, and full and partial remission remain unclear. Using the Clinical Global Impression (CGI) scales and diagnostic interviews as benchmark measures, several thresholds for the 6-item PARS were estimated using data from 904 youth with anxiety disorders, of which 36 % had an autism diagnosis. Data were drawn from 9 clinical trials conducted in the United States. Threshold accuracy was evaluated in a holdout sample and in an independent Swedish sample (n = 49). The reliable change index (RCI) was 3.39 points. A raw score reduction of ≥ 4 points or a ≥ 20 % reduction best defined the MCID. A raw score reduction of ≥ 8 points or a ≥ 43 % reduction best defined treatment response. Scores of 0-10 after treatment accurately defined full and partial remission, and scores of 0-5 defined full remission. Threshold accuracies ranged from 74-91 % in the holdout and 67-78 % in the Swedish sample. Treatment response accuracy was similar across samples, while MCID accuracy was lower in the Swedish sample. Similar cutoffs emerged in youth with and without autism and across age groups and genders. Threshold accuracies for the 5- and 7-item PARS versions were comparable to the 6-item version. This study establishes response and remission thresholds for PARS that are applicable across age groups, genders, and autism status. We advise against using the RCI as it does not exceed the MCID.