[Comparison of national early warning score, rapid emergency medicine score and acute physiology and chronic health evaluation II score for predicting outcome among emergency severe patients].

ABSTRACT

OBJECTIVE: To analyze the comparation of national early warning score (NEWS), rapid emergency medicine score (REMS) and acute physiology and chronic health evaluation II (APACHE II) score in predicting prognosis of critically ill patients in emergency department (ED). METHODS: A retrospective study was conducted. Critically ill patients, aged > 16 years, hospitalized > 24 hours, and admitted to the ED of Nanhua Hospital Affiliated to South China University from January 2016 to June 2017 were enrolled. NEWS, REMS and APACHE II score were calculated based on the worst value of each index within 24 hours after emergency admission. The primary endpoint was 28-day mortality. The relationship between the three scoring systems and the prognosis of patients was analyzed. The predictive value of three scoring systems for the prognosis of critically ill patients in ED was analyzed by receiver operating characteristic curve (ROC). RESULTS: A total of 119 emergency severe patients were enrolled in the study, and the 28-day mortality was 21.0%. The scores of NEWS, REMS and APACHE II in the death group were significantly higher than those in the survival group (NEWS score 9.40±3.19 vs. 5.72±2.35, REMS score 12.64±4.46 vs. 7.97±3.28, APACHE II score 26.64±6.92 vs. 16.19±5.48, all P < 0.01). With the increase of NEWS, REMS and APACHE II score, the 28-day mortality of patients gradually increased [28-day mortality of NEWS < 5, 5-6, ≥ 7 was 3.03% (1/34), 13.33% (4/34), 64.25% (20/51); 28-day mortality of REMS < 12, 12-16, ≥ 17 was 10.99% (10/91), 50.00% (11/22), 66.67% (4/6); 28-day mortality of APACHE II < 15, 15-24, ≥ 25 was 2.33% (1/43), 15.09% (8/59), 69.57% (16/23), respectively, all P < 0.01]. The ROC curve analysis showed that the areas under the ROC curve (AUC) of NEWS, REMS and APACHE II score for predicting the prognosis of emergency critically ill patients were 0.830 [95% confidence interval (95%CI) = 0.737-0.923], 0.782 (95%CI = 0.671-0.892) and 0.878 (95%CI = 0.800-0.956), respectively (all P = 0.000), and the accuracy of prediction was 57.4%, 48.6%, 65.4%, respectively. CONCLUSIONS: The scores of NEWS, REMS and APACHE II were useful in predicting prognosis of critically ill patients, with the highest accuracy of APACHE II forecast, followed by NEWS, and the lowest of REMS. After comprehensive consideration of cost-effectiveness, NEWS is more reliable in ED.