The objective was to demonstrate the derivation and the preliminary validation of a brief screening score to improve recognition of panic-like anxiety in emergency department (ED) patients with unexplained chest pain.
This cross-sectional study included 507 ED patients with unexplained chest pain randomly assigned to the derivation condition (n = 201) or the validation condition (n = 306). The Anxiety Disorders Interview Schedule for DSM-IV (ADIS-IV) acted as the criterion standard for panic diagnosis. Receiver operating characteristic (ROC) curves were used to determine which of the sociodemographic, medical, and questionnaire response variables best predicted the presence of panic-like anxiety in the derivation condition. The selected predictors were entered in a logistic regression analysis to determine variables to be included in the screening score. The predictive validity of the resulting score was evaluated in both conditions.
The four-item Panic Screening Score (PSS) identified panic-like anxiety with a sensitivity of 63% (95% confidence interval [CI] = 52% to 73%) and a specificity of 84% (95% CI = 76% to 90%) in the derivation phase. The positive predictive value was 74% (95% CI = 62% to 93%) and the negative predictive value was 76% (95% CI = 68% to 89%). In the validation phase, the sensitivity and specificity were 53% (95% CI = 44% to 62%) and 85% (95% CI = 78% to 89%), respectively. The positive predictive value was 72% (95% CI = 62% to 80%) and the negative predictive value was 71% (95% CI = 65% to 77%).
The PSS may be useful to identify patients with unexplained chest pain who could benefit from referral to mental health professionals or their primary care providers for evaluation of panic attack and panic disorder. Prospective studies to improve and validate the PSS, followed by impact analysis, are mandatory before clinical implementation.