Construction and validation of a predictive model for ventilator-associated pneumonia in premature infants undergoing mechanical ventilation in the neonatal intensive care unit
10.3760/cma.j.cn211501-20240222-00393
- VernacularTitle:新生儿重症监护病房机械通气早产儿并发呼吸机相关性肺炎预测模型的构建与验证
- Author:
Yongqin GUO
1
;
Yingying DOU
;
Jianli LI
;
Ruimin CHANG
;
Yanan HAO
Author Information
1. 长治市妇幼保健院新生儿科，长治 046000
- Keywords:
Pneumonia, ventilator-associated;
Mechanical ventilation;
Premature infants;
Risk prediction model
- From:
Chinese Journal of Practical Nursing
2024;40(33):2578-2587
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To construct a prediction model for ventilator-associated pneumonia (VAP) in mechanically ventilated preterm infants in the neonatal intensive care unit (NICU) and to test its clinical effect.Methods:This was a cross-sectional study. A total of 740 preterm infants admitted to the NICU for mechanical ventilation from July 2018 to June 2023 were retrospectively selected as the study subjects, and were divided into the modeling set (518 cases) and the validation set (222 cases) according to the ratio of 7∶3 using the computer-generated random number method. The modeling set was divided into the VAP group (181 cases) and the non-VAP group (337 cases) according to whether VAP occured, and 21 clinical characteristics were analyzed, using single factor difference analysis to screen predictive factors, the independent risk factors of VAP in mechanically ventilated preterm infants were determined by multivariate Logistic regression analysis, and the nomogram model was made by R software. Then, the nomogram model was tested by validating the data of the validation set. The receiver operating characteristic (ROC) curve, Hosmer-Lemeshow goodness-of-fit test, calibration curve and clinical decision curve were used to evaluate the efficacy and practical value of the model.Results:There were 88 males and 93 females in the VAP group, with 156 cases of gestational age<34 weeks and 25 cases of gestational age≥34 weeks. There were 155 males and 182 females in the non-VAP group, with 196 cases of gestational age<34 weeks and 141 cases of gestational age≥34 weeks. Birth weight ( OR=0.114, 95% CI 0.044-0.268, P<0.05) and oral care of breast milk ( OR=0.124, 95% CI 0.0.057-0.249, P<0.05) were protective factors for VAP in mechanically ventilated preterm infants, and Apgar score at 5 min after birth ( OR=2.895, 95% CI 1.318-6.419, P<0.05), serum prealbumin at 72 h of mechanical ventilation ( OR=4.837, 95% CI 2.643-9.063, P<0.05), gastric contents reflux ( OR=6.754, 95% CI 3.156-15.240, P<0.05), and time of mechanical ventilation ( OR=7.784, 95% CI 3.491-18.160, P<0.05) were independent risk factors for VAP in mechanically ventilated preterm infants. The area under the curve (AUC) of the ROC curve of the modeling set was 0.929 (95% CI 0.907-0.950, P<0.01), and the validation set (AUC) was 0.917 (95% CI 0.882-0.952, P<0.01), the model has good discrimination. The C indices of the modeling set and the validation set were 0.93 and 0.92 respectively by sampling 500 times by the Bootstrap method, indicating that the model had good consistency, and the decision curve suggested that the prediction model was far from the extreme curve and the net return value was high, indicating that the nomogram prediction model constructed this time had high prediction value. Conclusions:Birth weight, Apgar score at 5 min after birth, time of mechanical ventilation, oral care of breast milk, serum prealbumin at 72 h of mechanical ventilation, and gastric contents reflux are independent influencing factors for VAP in mechanically ventilated preterm infants. The nomogram prediction model constructed can provide a visual and simple evaluation tool for early identification of high-risk children and reducing the occurrence of VAP.
