Comparison of application value of two risk prediction models for prediction of intolerance risk in critically ill patients with enteral nutrition
Objective:To assess the predictive accuracy and practical utility of established risk prediction models for enteral nutrition intolerance in critically ill patients. Methods:A meta-analysis was conducted to identify existing risk prediction models for enteral nutrition intolerance in critically ill patients. Eligible patients admitted to the Department of Critical Care Medicine and various ICUs of General Hospital of Eastern Theater Command from March 2023 to August 2023, meeting natriuresis criteria, were included in the study. The discrimination and calibration of the two models were assessed using the area under the receiver operating characteristic curve (AUROC) and the Hosmer-Lemeshow goodness-of-fit test (H-L test). Results:Two models were analyzed, encompassing a total of 395 patients, among whom 161 experienced intolerances, resulting in an incidence rate of 40.8％. Model 1 demonstrated an AUROC of 0.838 (95％CI:0.798 ～ 0.873), while model 2 yielded an AUROC of 0.744 (95％CI:0.698 ～ 0.786). The Delong method was utilized to compare the AUROC values of the two models, revealing a statistically significant difference (P=0.0043). Notably, the model 1 exhibited superior performance compered to model 2. The H-L test for model 1 indicated fair calibration (X2=61.116, P<0.001), whereas model 2 demonstrated better calibration (X2=3.659, P=0.887). Conclusion:Model 1 exhibits superior discriminatory ability compared tomodel 2, while the calibration of model 2 surpasses that of model 1. Model 1 is well-suited for dynamic prediction, accommodating changes in patient condition over time. Conversely, Model 2 is appropriated for initial prediction following enteral nutrition initiation. Healthcare professionals can integrate bothmodels based on the specific clinical conditions to enhance predictive accutacy. Additionally, they can undertake high-quality research to develop a novel risk prediction model.
Critical illnessEnteral nutritionICUPrediction modeRisk of occurrence
万方数据
维普
