Construction and verification of a risk prediction model for ventilator-associated pneumonia in trauma patients
10.3969/j.issn.1008-9691.2024.06.008
- VernacularTitle:创伤患者呼吸机相关性肺炎风险预测模型的建立及验证
- Author:
Zhibing WANG
1
;
Kejing YU
1
;
Qianqian LIU
1
;
Zhongjian LI
1
;
Chunxia ZHANG
1
;
Dongdong HAN
1
Author Information
1. 河北省沧州中西医结合医院急诊医学科,河北 沧州 061000
- Publication Type:Journal Article
- Keywords:
Trauma;
Ventilator-associated pneumonia;
Risk prediction model
- From:
Chinese Journal of Integrated Traditional and Western Medicine in Intensive and Critical Care
2024;31(6):684-689
- CountryChina
- Language:Chinese
-
Abstract:
Objective To construct a risk prediction model for ventilator-associated pneumonia(VAP)in trauma patients and evaluate its efficacy.Methods A single-center retrospective study was conducted,trauma patients admitted to the department of emergency intensive care unit(EICU)of Hebei Cangzhou Hospital of Integrated Traditional Chinese Medicine and Western Medicine from January 1,2018 to January 1,2023 were selected as the study subjects,and the patients were divided into VAP group and non-VAP group.Differences between the two groups in variables including demographic characteristics,clinical data,and clinical scores.To prevent overfitting,differences between the groups were reduced using LASSO regression.Multifactor Logistic regression was used to identify risk factors for VAP in trauma patients and construct a risk prediction model.The model's discrimination was evaluated using the receiver operator characteristic curve(ROC curve)and area under the curve(AUC).The calibration curve was drawn and Hosmer-Lemeshow test were performed to evaluate the calibration degree of the model.Decision curve analysis(DCA)and clinical impact curve(CIC)were used to analyse the model's net benefit at different probability thresholds.Results A total of 888 trauma patients were included,among which 166 cases(18.7%)were diagnosed with VAP.Compared to the non-VAP group,the VAP group showed a significant increase in age,age-adjusted Charlson comorbidity index(aCCI)scores,white blood cell count(WBC),sequential organ failure assessment(SOFA)scores,length of ICU stay,and the proportion of patients with chest trauma,traumatic brain injury,and spinal cord injury.In contrast,hemoglobin(Hb),Glasgow coma scale(GCS)scores,and body mass index(BMI)were significantly lower in the VAP group(all P<0.05).Using LASSO regression,four variables were identified as important predictors for the occurrence of VAP in trauma patients:length of ICU stay,aCCI,WBC,and SOFA score.Multivariate Logistic regression showed that length of ICU stay[odds ratio(OR)and 95%confidence interval(95%CI)was 1.094(1.070-1.117)],aCCI[OR(95%CI)was 1.135(1.065-1.210)],WBC[OR(95%CI)was 1.139(1.104-1.176)],and SOFA score[OR(95%CI)was 1.137(1.080-1.197)]were independent risk factors for the occurrence of VAP in trauma patients(all P<0.05).Based on these influencing factors,a predictive model for VAP occurrence was constructed.ROC curve analysis showed that the AUC for predicting VAP occurrence in trauma patients was 0.876,with a 95%CI was 0.850-0.903,a sensitivity of 86.14%,and a specificity of 75.17%,indicating that the model has a high discriminative ability.Hosmer-Lemeshow test:χ2=7.7,P=0.2,Cox&Snell R2=0.236,Nagelkerke R2=0.387,the calibration curve was very close to the diagonal,and the mean absolute error(MAE)=0.03,indicating the model's predictions were highly consistent with actual clinical observations.The DCA and CIC curves indicate that within the threshold probability of<70%,using this model to identify high-risk groups for VAP in trauma patients and making clinical decisions can provide benefits in clinical practice.Conclusion The risk prediction model of VAP in trauma patients constructed in this study has high discrimination and calibration,which can provide reference for medical personnel to identify high-risk groups of VAP among trauma patients at an early stage and provide targeted intervention measures.
