Introduction: The Anxiety Symptom Scale (ASS) is a ten-item psychometric test, which is frequently used for screening purposes in Danish general practice. The ASS includes nine items on symptoms of anxiety disorders and one item on functional impairment caused by these symptoms. Previous research indicates that reducing the number of response categories may benefit the test and that it may provide a unidimensional measure of global anxiety.
Methods: We made a modified version of the ASS with fewer response categories per item. We compared the structure of the two versions in a cohort of respondents recruited from three different settings: general practitioner during a consultation (setting 1), mental health associations (setting 2), and waiting room in a single general practice (setting 3).
Results: Both versions were fitted to a bifactor model; the nine anxiety items loaded on a general factor, and the individual items addressing specific anxiety disorders loaded on their specific factor. The original version provided a better fit than the modified version. The model suggests that the ASS measures primarily a single underlying trait. Functional impairment was well predicted by the general factor.
Conclusion: The first nine items of the ASS can be considered a unidimensional measure of global anxiety, showing a substantial correlation with functional impairment. This makes the instrument useful for general practitioners in making initial treatment decisions. The performance of the instrument did not improve by reducing the number of response categories or removing one of the panic anxiety attack items.
Keywords: Anxiety; general practice; psychometrics; validation.