Predictive performance of an extended NEWS model for acute deterioration in general ward inpatients: a retrospective observational study

Vollständiger Abstract

Worum geht es in dieser Arbeit?

Background: This study evaluated whether an extended model combining the National Early Warning Score (NEWS) with blood lactate and serum total CO2 improves prediction of acute deterioration in adult general ward inpatients.Methods: The study analyzed electronic medical records from 437 patients admitted to a tertiary general hospital in Ulsan Metropolitan City from January 1 to December 31, 2024. Data were extracted from electronic medical records using a literature-informed data collection form. Analyses included descriptive statistics, the chi-square test, the t-test, logistic regression, and receiver operating characteristic analysis with area under the receiver operating characteristic curve (AUROC) estimation, performed using SPSS version 24.0.Results: In the extended model, each 1-point increase in NEWS and each 1 mmol/L increase in blood lactate were associated with higher odds of unplanned intensive care unit (ICU) admission (odds ratio [OR] = 1.23 and 1.22, respectively), whereas each 1 mmol/L increase in serum total CO2 was associated with lower odds (OR = 0.94). For 7-day mortality, NEWS and blood lactate were associated with higher odds of death (OR = 1.17 and 1.16, respectively), whereas serum total CO2 was associated with lower odds (OR = 0.91). The AUROC increased from .69 to .74 (p = .009) for unplanned ICU admission and from .68 to .77 (p = .003) for 7-day mortality.Conclusion: Adding blood lactate and serum total CO2 to NEWS improved discrimination for unplanned ICU admission and 7-day mortality. The extended model may support clinical decision-making and timely rapid response team intervention in general ward settings.