Predictive Factors of Length of Stay in Intensive Care Unit after Coronary Artery Bypass Graft Surgery based on...

Area of Science:

• Cardiovascular Surgery
• Medical Informatics
• Health Services Research

Background:

• Coronary artery bypass grafting (CABG) surgery is associated with prolonged intensive care unit (ICU) length of stay (LOS).
• Accurate prediction of ICU LOS is essential for resource allocation and patient management.
• Machine learning offers potential for identifying complex patterns influencing patient outcomes.

Purpose of the Study:

• To predict factors influencing ICU LOS after CABG surgery.
• To apply machine learning models for enhanced predictive accuracy.
• To identify key demographic and clinical variables impacting prolonged ICU stays.

Main Methods:

• Literature review and expert confirmation of factors affecting ICU LOS post-CABG.
• Retrospective analysis of 605 patient records from Farshchian Specialized Heart Hospital.
• Training and testing of four machine learning models to predict ICU LOS.

Main Results:

• Key predictors of ICU LOS included intubation duration, body mass index (BMI), age, surgery duration, and packed red blood cell transfusions.
• The Random Forest model demonstrated superior performance in predicting effective factors.
• Performance metrics for the Random Forest model: MSE = 1.64, MAE = 0.93, R² = 0.28.

Conclusions:

• Machine learning models effectively highlight the importance of demographic and clinical variables in predicting ICU LOS.
• Understanding these predictors enables healthcare professionals to identify high-risk patients for extended ICU stays.
• This predictive capability can optimize patient care and resource management following CABG surgery.