Establishment and Validation of a Predictive Model for the Risk of Invasive Mechanical Ventilation in Elderly Patients with Sepsis

Abstract

Abstract Background The aim of the research was to discover risk elements and create a useful nomogram for predicting the occurrence of invasive mechanical ventilation (IMV) in elderly patients with sepsis. Methods Sepsis patients who were elderly, aged 65 years or older, were selected from the Medical Information Mart for Intensive Care (MIMIC)-IV database. Demographic and laboratory test information were collected on the first day of Intensive Care Unit (ICU) admission. Participants were 7:3 randomly assigned divisions into training and validation sets. The features of training set were used to determine risk factors for predicting invasive mechanical ventilation. The least absolute shrinkage and selection operator (LASSO) regression was employed to recognize predictors. Subsequently, the training set was utilized to create a nomogram. The validity of the nomogram was evaluated using receiver operating characteristic (ROC) analysis, decision curve analysis (DCA), and calibration curve analysis. Results We enrolled 7,868 patients, with 5,259 (66.8%) receiving invasive mechanical ventilation. In the IMV group, hospital mortality was higher than in the non-IMV group (23% vs. 13%, p < 0.001). Logistic regression analysis identified SpO2, hemoglobin, anion gap, chloride, vasopressor drugs, and ICU length of stay (LOS) as predictors, which were integrated into a nomogram. The AUC of the nomogram was 0.84 in both training set and validation set. The calibration plot demonstrated that the nomogram effectively predicted the requirement for IMV in both datasets. DCA proved the clinical values of the nomogram. Conclusion The nomogram provides a predictive tool for identifying the demand for invasive mechanical ventilation in sepsis patients who aged 65 or older. This model can aid healthcare professionals in identifying high-risk patients earlier and implementing timely interventions to improve their prognosis.