How accurate are vital signs in predicting clinical outcomes in critically ill emergency department patients.

OBJECTIVES: We aimed to evaluate the predictive value of pulse rate (PR), systolic blood pressure (SBP), diastolic blood pressure, respiratory rate (RR), oxygen saturation (SaO2), and the Glasgow Coma Scale (GCS) for cardiac arrest and death in critically ill patients. METHODS: In total, 1025 patients had vital signs recorded at triage at our Emergency Department and were followed up for three clinical outcomes: cardiac arrest in 72 h, admission to ICU, and death within 30 days. Vital signs were used in univariate and multivariate analyses for outcomes. Age was added in multivariate analysis. RESULTS: PR, SBP, RR, SaO2, and GCS were significantly associated with cardiac arrest within 72 h, whereas PR, SBP, RR, SaO2, and GCS were associated with death within 30 days. Only PR and GCS were associated with ICU admission. In the multivariate analysis, age, PR (>100) [odds ratio (OR) 1.65; 95% confidence interval (CI) 1.00-2.71], SBP (>140; OR 0.41; 95% CI: 0.21-0.79), RR (>20; OR 2.90; 95% CI: 1.67-5.03), and GCS (<15; OR 5.71; 95% CI: 3.40-9.57) were significantly associated with death. Vital signs with age have low sensitivity (cardiac arrest 11.54%, death 22.73%, ICU 12.50%) and high specificity (cardiac arrest 99.28%, death 97.22%, ICU 93.80%). Age and GCS were found to be independent predictors of all three outcomes. CONCLUSION: Not all vital signs are useful in the prediction of clinical outcomes. Vital signs had high specificity but very low sensitivity as predictors of clinical outcomes. Clinicians should always remember to treat patients and not numbers.

An intelligent scoring system and its application to cardiac arrest prediction.

Traditional risk score prediction is based on vital signs and clinical assessment. In this paper, we present an intelligent scoring system for the prediction of cardiac arrest within 72 h. The patient population is represented by a set of feature vectors, from which risk scores are derived based on geometric distance calculation and support vector machine. Each feature vector is a combination of heart rate variability (HRV) parameters and vital signs. Performance evaluation is conducted on the leave-one-out cross-validation framework, and receiver operating characteristic, sensitivity, specificity, positive predictive value, and negative predictive value are reported. Experimental results reveal that the proposed scoring system not only achieves satisfactory performance on determining the risk of cardiac arrest within 72 h but also has the ability to generate continuous risk scores rather than a simple binary decision by a traditional classifier. Furthermore, the proposed scoring system works well for both balanced and imbalanced datasets, and the combination of HRV parameters and vital signs shows superiority in prediction to using HRV parameters only or vital signs only.