Predictive power and practical effects of modified early warning score in critical care patients to be transferred into intensive care unit

Abstract

Objective To explore the power of modified early warning score (MEWS) in predicting Class I-Ⅲ patients to be transferred into the intensive care unit (ICU) and to provide an objective assessment tool for medical workers to properly triage patients. Methods Totally 463 patients admitted into the Department of Emergency of a Class Ⅲ Grade A hospital from May 17 to September 27, 2017 by convenient sampling. MEWS was obtained by collecting the patients' data at admission. The outcome indicator was whether the patients were transferred into ICU. The patients were divided into the MEWS≥ 4 group and the MEWS<4 group based on MEWS. The rate of transfer into ICU was compared between the patients. The area under the receiver operating characteristic curves (AUC-ROC) , sensitivity, specificity and the best cutoff point of MEWS and MEWS combined with stratification in predicting whether the patients would be transferred into ICU were calculated. Results The rate of transfer into ICU in the MEWS≥4 group and the MEWS<4 group was 43.4% and 13.5%, respectively (χ2=52.3028, P<0.001) . The AUC-ROC of MEWS in predicting whether the patients would be transferred into ICU was 0.759, whose best cutoff point, sensitivity and specificity was 3.5, 68.91% and 68.90%. The AUC-ROC of MEWS combined with stratification in predicting whether the patients would be transferred into ICU was 0.861. Conclusions The predictive power of MEWS in predicting whether class I-Ⅲ patients would be transferred into ICU is medium, but its predictive power is enhanced when stratification is combined with MEWS. The scoring tool is simple and practical, which provides a reference for properly triaging patients. Key words: Emergency service, hospital; Condition assessment; Stratification and triage; Modified early warning score; Prediction