Systematic review of risk prediction models for ventilator-associated pneumonia in mechanically ventilated patients in Intensive Care Unit
10.3760/cma.j.cn115682-20231227-02834
- VernacularTitle:ICU机械通气患者呼吸机相关性肺炎风险预测模型的系统评价
- Author:
Hui WEN
1
;
Qingmei NIE
;
Lili SUN
;
Yueyue BAO
;
Yingying ZHANG
;
Pei LIU
;
Rongrong CAO
Author Information
1. 潍坊市人民医院急诊部，潍坊 261000
- Keywords:
Intensive Care Unit;
Ventilator-associated pneumonia;
Prediction model;
Systematic review
- From:
Chinese Journal of Modern Nursing
2024;30(24):3280-3286
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To systematically search and evaluate risk prediction models for ventilator-associated pneumonia (VAP) of ICU in order to provide references for developing higher-quality VAP risk prediction models.Methods:Relevant literature was retrieved from databases including China Biology Medicine disc, WanFang data, China National Knowledge Infrastructure, Embase, PubMed, CINAHL, Web of Science, and Cochrane Library. The search timeframe was from the establishment of the databases to September 30, 2023, limited to English and Chinese languages. Two researchers independently screened the literature and extracted data, and the PROBAST tool was used to evaluate the risk of bias and applicability of the included studies.Results:A total of 15 studies on VAP risk prediction models were included. The area under the receiver operating characteristic curve for the 15 models ranged from 0.722 to 0.982. The most frequently involved predictors were age, duration of mechanical ventilation, ICU length of stay, and comorbid chronic obstructive pulmonary disease. The overall adaptability was good, but the risk of bias was high. The main sources of bias included insufficient sample size, inappropriate data sources, lack of model performance evaluation, and inadequate attention to missing data.Conclusions:The risk of bias in studies on VAP risk prediction models is high, indicating that the field is still developing. Future research should focus on the effectiveness of different risk assessment methods to construct models with low bias, excellent predictive performance, and suitability for clinical practice in China.
