RESUMEN
OBJECTIVE: This study was conducted to evaluate scorecards for early recognition of high-risk patients of delirium in the emergency department (ED). METHODS: Data from 399 consecutive patients aged 65 years or older between January 1, 2015 and December 31, 2015 were retrospectively analyzed. Delirium was identified by reviewing medical records and was confirmed by a psychiatrist. The study population was divided into a training and validation group. Predisposing factors were evaluated and validated by multivariate logistic regression analysis and a calibration plot, after which a scorecard was constructed using these factors and applying points to double odds to each regression coefficient. RESULTS: Dementia, transfer from a long-term care facility, acute acid-base imbalance, moderate pain, and stroke were independent predisposing factors for delirium in ED, with assigned scores in the scorecard of 3, 2, 2, 2, and 2, respectively. The total score of the scorecard for delirious patients was significantly higher than that for non-delirious patients in both the training and validation groups. The coefficient of determination (R²) of the calibration plot was 0.74 and 0.68 in the training and validation group, respectively. In the receiver operation characteristic curve, the cut-off point of the scorecard for delirium was 2.5 and the sensitivity, specificity, and accuracy were 75.0%, 87.8%, and 86.7% in training group, while they were 76.9%, 85.1%, and 84.2% in the validation group, respectively. CONCLUSION: The scorecard was a useful screening tool for early recognition of patients with a high-risk of developing delirium in the ED.