Exploration of the prediction model for children with severe community-acquired pneumonia admitted to the intensive care unit based on the pediatric early warning score
10.3760/cma.j.issn.1673-4912.2025.08.003
- VernacularTitle:基于儿童早期预警评分构建儿童重症社区获得性肺炎入住重症监护室的预测模型探讨
- Author:
Tianming WANG
1
;
Jiahu HUANG
1
;
Jian LIU
1
;
Zhagen WANG
1
;
Tingjun LI
1
Author Information
1. 上海市儿童医院 上海交通大学医学院附属儿童医院急诊科 200062
- Publication Type:Journal Article
- Keywords:
Severe community-acquired pneumonia in children;
Risk factors;
Pediatric early warning score;
Pediatric intensive care unit;
Clinical Prediction Model
- From:
Chinese Pediatric Emergency Medicine
2025;32(8):573-578
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To analyze the risk factors for children with severe community-acquired pneumonia (CAP) being admitted to the pediatric intensive care unit (PICU),and establish a clinical prediction model,then evaluate the clinical application value of this model.Methods:A retrospective analysis was performed on children diagnosed with severe CAP at the Children's Hospital Affiliated to School of Medicine of Shanghai Jiao Tong University from January to June 2023.The children were divided into the PICU group and the non-PICU group based on whether they were admitted to PICU at admission.The differences in pediatric early warning score（PEWS），clinical characteristics，and laboratory test results between the two groups at their last visit before admission were compared. The independent risk factors for children with severe CAP admitted to PICU were analyzed,and a clinical prediction model was established,which was validated through the receiver operating characteristic (ROC) curve.Results:A total of 274 children were included,including 141 males and 133 females,with a median age of 50 (24,81) months. There were 43 cases in PICU group and 231 cases in non-PICU group.There were no statistically significant differences in gender and age between the two groups of children ( P>0.05). The PEWS score,white blood cell count,neutrophil count,neutrophil/lymphocyte ratio,procalcitonin (PCT),and lactate levels of children in the PICU group were significantly higher than those of children in the non-PICU group.While the duration of fever,peak temperature,and percutaneous arterial oxygen saturation (SpO 2) were significantly lower in the PICU group than those in the non-PICU group. All these differences were statistically significant ( P<0.05).Binary Logistic regression analysis showed that PEWS>4 points（ OR=6.583，95% CI 1.763 - 24.588， P<0.05），PCT>0.42 μg/L（ OR=19.046，95% CI 4.362-83.159， P<0.05），and SpO 2<93%（ OR=21.670，95% CI 3.843-122.184， P<0.05）were independent risk factors for children with severe CAP to be admitted to PICU.A clinical prediction model was constructed based on the above three independent risk factors.The area under ROC curve of the clinical prediction model was 0.941（95% CI 0.913-0.968， P<0.05），the sensitivity was 95.3%，the specificity was 80.5%，the positive predictive value was 83.0%，and the negative predictive value was 94.5%. Conclusion:For children with severe CAP，if they have PEWS > 4,an elevated PCT level，and a decreased SpO 2，it is recommended that they be admitted to PICU for further monitoring and treatment.The clinical prediction model for admission to the PICU for children with severe CAP，constructed by combining PEWS with commonly used clinical information in pediatric emergency，has a relatively high predictive efficacy and can provide a reference for the stratified diagnosis and treatment of children with severe CAP in the future.
