Identifying Individuals With Generalised Anxiety Disorder: A Receiver Operator Characteristic Analysis of Theoretically Relevant Measures

Abstract

This article appraises the ability of several measures, assessing symptomatology theoretically relevant to generalised anxiety disorder (GAD), to accomplish successful classification of individuals with DSM-IV GAD and individuals with mood or other anxiety disorders. Participants were 197 individuals (average age = 36.43; 67.5% female) receiving treatment at a clinic, who completed pretreatment self-report measures. Receiver operator characteristic analyses and logistic regression analyses were employed to determine the classificatory abilities, including sensitivity and specificity, of several GAD relevant measures. The Penn State Worry Questionnaire (PSWQ) accomplished significant classification in both comparison group dyads: GAD-Anxiety and GAD-Mood. Whereas the general distress of depression subscale of the Mood and Anxiety Symptoms Questionnaire (MASQ) accomplished significant classification in the GAD-Anxiety group dyad, both the general distress of anxiety and anxious arousal subscales of the MASQ did so in the GAD-Mood group dyad. Only the PSWQ significantly predicted the presence of GAD, while controlling for other symptom variables. Though evidence supports the uniqueness of some symptoms of GAD, including worry, often considered pathognomonic to GAD, the modest areas under the curve, sensitivity, and specificity of the measures afford tenuous support for the utility of these measures. The lack of highly sensitive and specific symptomatology comport with critiques of GAD as a diagnostically delimited category.