Abstract:Background: Extensively drug-resistant Acinetobacter baumannii (XDRAB) has become a significant pathogen in hospital environments, particularly in intensive care units (ICUs). XDRAB's resistance to conventional antimicrobial treatments and ability to survive on various surfaces pose a substantial threat to patient health, often resulting in severe infections such as ventilator-associated pneumonia (VAP) and bloodstream infections (BSI). Methods: We retrospectively analyzed clinical data from 559 patients with XDRAB infections admitted to Jinhua Central Hospital between January 2021 and December 2023. Patients were randomly divided into a training set (391 cases) and a testing set (168 cases). Variables were selected using Lasso regression and logistic regression analysis, and a predictive model was constructed and validated internally and externally. Model performance and clinical utility were evaluated using the Hosmer-Lemeshow test, C-index, ROC curve, decision curve analysis (DCA), and clinical impact curve (CIC). Results: Lasso regression analysis was used to screen 35 variables, selecting features through 10-fold cross-validation. We chose lambda.1se=0.03450 (log(lambda.1se)=-3.367), including 10 non-zero coefficient features. These features were then included in a multivariate logistic regression analysis, identifying 8 independent risk factors for XDRAB infection: ICU stay of 1-7 days (OR=3.970, 95%CI=1.586-9.937), ICU stay >7 days (OR=12.316, 95%CI=5.661-26.793), hypoproteinemia (OR=3.249, 95%CI=1.679-6.291), glucocorticoid use (OR=2.371, 95%CI=1.231-4.564), urinary catheterization (OR=2.148, 95%CI=1.120-4.120), mechanical ventilation (OR=2.737, 95%CI=1.367-5.482), diabetes mellitus (OR=2.435, 95%CI=1.050-5.646), carbapenem use (OR=6.649, 95%CI=2.321-19.048), and β-lactamase inhibitor use (OR=4.146, 95%CI=2.145-8.014). These 8 factors were used to construct a predictive model visualized through a nomogram. The model validation showed a C-index of 0.932 for the training set and 0.929 for the testing set, with a Hosmer-Lemeshow test p-value of 0.47, indicating good calibration. Furthermore, the DCA curve demonstrated good clinical decision-making performance, and the CIC curve confirmed the model's reliable clinical impact. Conclusion: Regression analysis identified ICU stay duration, hypoproteinemia, glucocorticoid use, urinary catheterization, mechanical ventilation, diabetes mellitus, carbapenem use, and β-lactamase inhibitor use as independent risk factors for XDRAB infection. The corresponding predictive model demonstrated high accuracy and stability.

Derivation and validation of a predictive mortality model of in-hospital patients with Acinetobacter baumannii nosocomial infection or colonization

Machine Learning Algorithms to Predict the Mortality of Carbapenem Resistant Klebsiella Pneumoniae Bacteremia

Risk factors for mortality in Acinetobacter baumannii bloodstream infections and development of a predictive mortality model

Predictive modeling of mortality in carbapenem-resistant Acinetobacter baumannii bloodstream infections using machine learning

Risk factors and predictive model for nosocomial infections by extensively drug-resistant Acinetobacter baumannii

Clinical and molecular predictors of mortality in patients with carbapenem-resistant Acinetobacter baumannii bacteremia: A retrospective cohort study

Microbiological characteristics and risk factors on prognosis associated with Acinetobacter baumannii bacteremia in general hospital: A single-center retrospective study

Clinical prediction models for multidrug-resistant organism colonisation or infection in critically ill patients: a systematic review protocol

A Machine Learning Predictive Model of Bloodstream Infection in Hospitalized Patients

Derivation Of A Mortality Prediction Model In Critical Care Patients With Cirrhosis And Sepsis

Development and validation of a novel prediction model for Carbapenem-resistant organism infection in a large-scale hospitalized patients

Data-Driven Machine Learning Approaches for Predicting In-Hospital Sepsis Mortality

Monitoring Immune Dysfunction in Critically Ill Patients with Carbapenem-Resistant Acinetobacter baumannii Sepsis Treated with Regimens Including Cefiderocol: A Pilot Study to Identify Accessible Biomarkers to Stratify Patients' Prognosis

Mortality attributable to different Klebsiella susceptibility patterns and to the coverage of empirical antibiotic therapy: a cohort study on patients admitted to the ICU with infection

The prediction of in-hospital mortality in elderly patients with sepsis-associated acute kidney injury utilizing machine learning models

Prediction of Mortality in Inpatients of Covid-19 using Statistical and Artificial Intelligence Approaches: A Case Study in Sakarya

Predicting Mortality in Sepsis-Associated Acute Respiratory Distress Syndrome: A Machine Learning Approach Using the MIMIC-III Database

A Model for Predicting Nosocomial Carbapenem-Resistant Klebsiella Pneumoniae Infection.

Prognostic Factors That Affect Mortality Patients with Acinetobacter baumannii Bloodstream Infection

Risk factors of mortality in patients with bloodstream infections due to Acinetobacter baumannii

Bloodstream Infections Caused by Carbapenem-Resistant Pathogens in Intensive Care Units: Risk Factors Analysis and Proposal of a Prognostic Score