Construction of a risk prediction model for the occurrence of consciousness disorders in acute respiratory failure patients based on DynNom dynamic scoring

Wang Zhen; Sun Haiyan; Wu Fei; Yang Wu; Guo Ge

doi:10.3969/j.issn.1002-1949.2026.03.011

Abstract：Objective To identify the risk factors and develop a DynNom-based dynamic nomogram for predicting consciousness disorders in patients with acute respiratory failure(ARF).Methods This retrospective study included a total of 133 patients with acute respiratory failure hospitalized from June 2022 to February 2025 classified as either having a consciousness disorder or the no consciousness disorder group based on the occurrence of consciousness disorders.Predictive factors were initially screened using LASSO regression,with significant risk factors further identified via multivariable logistic regression.A dynamic nomogram prediction model was then constructed using R software(version 4.2.3)and validated internally.An external validation cohort of 45 ARF patients(admitted March – August 2025)was used for independent assessment.Results The incidence of consciousness disorders was 33.8%(45/133).While baseline demographics(age,gender,comorbidities)were comparable between groups,the consciousness disorder group had significantly higher rates of malnutrition,severe ARF,renal insufficiency,sepsis,and hyponatremia(all P<0.05).Logistic regression analysis confirmed that malnutrition,severe acute respiratory failure,renal insufficiency,sepsis,and hyponatremia were independent risk factors for consciousness disorders in patients with acute respiratory failure(P<0.05).The area under the ROC curve of the nomogram model for predicting consciousness disorders in patients with acute respiratory failure was 0.795(95%CI 0.710-0.880);the predicted values of the calibration curve were basically consistent with the actual values;the decision curve showed that when the threshold probability was 13%-83%,the nomogram had a good benefit value for predicting consciousness disorders in patients with acute respiratory failure.The area under the ROC curve of the validation set was 0.741(95%CI 0.591-0.891),suggesting that the constructed nomogram model had good external predictive efficacy.Conclusions Malnutrition,severe acute respiratory failure,renal insufficiency,sepsis,and hyponatremia are independent risk factors for consciousness disorders in patients with acute respiratory failure.The developed DynNom-based dynamic nomogram provides acceptable predictive accuracy and may serve as a practical tool for individualized risk assessment in clinical settings.