Clinical Report: Machine Learning for Forecasting Gram-Negative BSIs in Elderly

Overview

This study developed a machine learning model to identify elderly inpatients at high risk for Gram-negative bloodstream infections (GNB-BSI) following the detection of Gram-negative bacteria in non-blood samples. The XGBoost model showed optimal performance with an AUC of 0.816.

Background

Bloodstream infections (BSI) are a significant public health issue, particularly among the elderly, who are at increased risk due to multiple comorbidities and immune dysfunction. Gram-negative bacteria are the leading cause of BSI, with high incidence and mortality rates. Early identification of at-risk patients is crucial.

Data Highlights

Key Findings

• The study enrolled 9,646 elderly inpatients with Gram-negative bacteria positivity.
• Seven predictive variables were identified: maximum procalcitonin level, max neutrophil percentage, maximum C-reactive protein level, minimum white blood cell count, venous catheter, age, and length of hospital stay.
• XGBoost demonstrated optimal performance.
• External validation confirmed the model's generalizability using the MIMIC-IV database.
• SHAP analysis indicated maximum procalcitonin level and length of hospital stay as top predictors.

Clinical Implications

The machine learning model provides a timely tool for clinicians to identify high-risk elderly patients for GNB-BSI, potentially allowing for earlier interventions. Understanding the predictive variables can guide clinical decision-making and resource allocation.

Conclusion

The developed predictive model aids in the identification of GNB-BSI in elderly patients.

Related Resources & Content

1. Author(s)/Org, Source, Year -- Title
2. Author(s)/Org, Source, Year -- Title
3. Author(s)/Org, Source, Year -- Title
4. Author(s)/Org, Source, Year -- Title
5. Surviving Sepsis Campaign Adult Guidelines | SCCM
6. Fast Antimicrobial Susceptibility Testing for Gram-Negative Bacteremia: The FAST Randomized Clinical Trial | Critical Care Medicine | JAMA
7. Surviving Sepsis Campaign Adult Guidelines | SCCM
8. Fast Antimicrobial Susceptibility Testing for Gram-Negative Bacteremia: The FAST Randomized Clinical Trial | Critical Care Medicine | JAMA | JAMA Network