OBJECTIVES: To compare the validity of detecting panic disorder using the Hospital Anxiety and Depression Scale (HADS), the Patient Health Questionnaire (PHQ), a screening question, and physicians' diagnosis, and to test whether modified evaluation algorithms improve the operating characteristics of these questionnaires. Additionally, patient and physician acceptability of the screening questionnaires was investigated. METHODS: The total sample of 499 patients comprised 348 medical outpatients and 151 psychosomatic outpatients. The Structured Clinical Interview for DSM-IV (SCID) was the criterion standard for the presence of a panic disorder. Sensitivity, specificity, predictive values, and overall accuracy were compared for the different measures. The conditional test characteristics were calculated based on the observed prevalence of panic disorder in the medical and psychosomatic subsample, respectively. RESULTS: Panic disorder was diagnosed in 4.3% of the medical outpatients and in 19.2% of the psychosomatic outpatients. The HADS, PHQ, and screening question achieved good operating characteristics. In contrast, physicians detected only 15% of cases with panic disorder. Modified evaluation algorithms of the questionnaires lead to an improvement of test characteristics. Of all measures, the PHQ had the best operating characteristics. The use of screening questionnaires was accepted by 96% of the patients and 97% of the physicians. CONCLUSIONS: The questionnaires as well as the screening question performed well in detecting panic disorder. Thus, the integration of these highly accepted measures into clinical evaluation is suggested. Recommendations for the selection of specific evaluation algorithms are given.
