Scholarlink

Objective <p>To systematically evaluate the risk stratification performance of four physiological scoring systems—National Early Warning Score (NEWS), Rapid Acute Physiology Score (RAPS), Glasgow Coma Scale (GCS), and Shock Index (SI)—in prehospital emergency care within resource-limited settings, and to inform evidence-based triage strategies and intervention allocation.</p> Methods <p>This retrospective population-based cohort study analyzed 15,720 prehospital emergency cases from Shaodong Emergency Medical Center (2023–2024), a county-level EMS system operating under resource-constrained conditions in southern China. A novel analytical framework, integrating “score stratification × intervention behavior,” was designed to assess the predictive consistency and clinical utility of the four scoring systems across four essential prehospital interventions: oxygen therapy, ECG monitoring, intravenous (IV) access, and fluid administration. Statistical methods included trend analysis, nonparametric effect size estimation, generalized linear mixed models (GLMMs), and an XGBoost machine learning algorithm.</p> Results <p>NEWS and RAPS demonstrated the strongest dose–response consistency in guiding intervention decisions. Although all effect sizes remained in the weak range (η² &lt; 0.06), NEWS showed the highest trend strength for ECG monitoring (Z = 23.58, <i>P</i> &lt; 0.001) and the largest effect size (η² = 0.0415), with RAPS showing a comparable value (η² = 0.0412). High-risk patients were significantly more likely to receive ECG monitoring than low-risk patients (OR = 2.691, 95% CI: 2.301–3.142). GCS was predictive primarily in patients with severe consciousness impairment (OR = 4.32, <i>P</i> &lt; 0.001), while SI demonstrated minimal discriminatory value (η² &lt; 0.01). In the XGBoost model, disease category (importance = 0.161) and NEWS (0.108) emerged as the most influential predictors of intervention intensity.</p> Conclusion <p>In this real-world, resource-constrained EMS cohort, physiological scores exhibited limited discriminatory capacity. Nevertheless, NEWS and RAPS showed more consistent associations with intervention allocation than GCS or SI. Even modest stratification reduced decision variability, supporting an adjunctive role. Multicenter prospective studies are needed to establish actionable thresholds and validate their impact on intervention timeliness and patient outcomes.</p> Trial registration <p>Not applicable.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Evaluating the real-world utility of four physiological scores for prehospital interventions in resource-limited EMS systems: a population-based study of 15,720 cases

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年1月12日

查看全文
  • Zichao Zhao,
  • Linfei Li,
  • Jinyi Zhang,
  • Xiaoyan Shi,
  • Zheng Jiang,
  • Fengjue Jiang,
  • Xiaofei Yan,
  • Bei Hu

摘要

Objective

To systematically evaluate the risk stratification performance of four physiological scoring systems—National Early Warning Score (NEWS), Rapid Acute Physiology Score (RAPS), Glasgow Coma Scale (GCS), and Shock Index (SI)—in prehospital emergency care within resource-limited settings, and to inform evidence-based triage strategies and intervention allocation.

Methods

This retrospective population-based cohort study analyzed 15,720 prehospital emergency cases from Shaodong Emergency Medical Center (2023–2024), a county-level EMS system operating under resource-constrained conditions in southern China. A novel analytical framework, integrating “score stratification × intervention behavior,” was designed to assess the predictive consistency and clinical utility of the four scoring systems across four essential prehospital interventions: oxygen therapy, ECG monitoring, intravenous (IV) access, and fluid administration. Statistical methods included trend analysis, nonparametric effect size estimation, generalized linear mixed models (GLMMs), and an XGBoost machine learning algorithm.

Results

NEWS and RAPS demonstrated the strongest dose–response consistency in guiding intervention decisions. Although all effect sizes remained in the weak range (η² < 0.06), NEWS showed the highest trend strength for ECG monitoring (Z = 23.58, P < 0.001) and the largest effect size (η² = 0.0415), with RAPS showing a comparable value (η² = 0.0412). High-risk patients were significantly more likely to receive ECG monitoring than low-risk patients (OR = 2.691, 95% CI: 2.301–3.142). GCS was predictive primarily in patients with severe consciousness impairment (OR = 4.32, P < 0.001), while SI demonstrated minimal discriminatory value (η² < 0.01). In the XGBoost model, disease category (importance = 0.161) and NEWS (0.108) emerged as the most influential predictors of intervention intensity.

Conclusion

In this real-world, resource-constrained EMS cohort, physiological scores exhibited limited discriminatory capacity. Nevertheless, NEWS and RAPS showed more consistent associations with intervention allocation than GCS or SI. Even modest stratification reduced decision variability, supporting an adjunctive role. Multicenter prospective studies are needed to establish actionable thresholds and validate their impact on intervention timeliness and patient outcomes.

Trial registration

Not applicable.