Predicting Length of Stay of Coronary Artery Bypass Grafting Patients Using Machine Learning
Austin J Triana1, Rushikesh Vyas2, Ashish S Shah3
1Vanderbilt University School of Medicine, Nashville, Tennessee.
Insights
Machine learning identified key predictors for postsurgery length of stay (LOS) after coronary artery bypass grafting (CABG). Duration intubated, creatinine, age, and transfusions significantly impact LOS, enabling simpler predictive models.
Area of Science:
- Cardiovascular Surgery
- Health Informatics
- Machine Learning in Healthcare
Background:
- Healthcare providers need to identify high-impact patient information.
- Predicting postsurgery length of stay (LOS) is crucial for resource management.
Purpose of the Study:
- To identify the most significant variables for predicting postsurgery LOS in coronary artery bypass grafting (CABG) patients.
- To compare machine learning models with traditional regression for LOS prediction.
Main Methods:
- Analysis of 2121 isolated CABG patients' data from the Society of Thoracic Surgeons (STS) Registry.
- Application of random forest and artificial neural networks (ANNs) to identify high-impact variables.
- Comparison of model performance against multiple linear regression with out-of-sample validation.
Main Results:
- Four key predictors of LOS identified: duration intubated, preoperative creatinine, age, and intraoperative transfusions.
- An ANN model using the top 10 variables achieved the best performance (MAE = 1.685 d, R² = 0.232).
- The ANN model demonstrated consistent performance in out-of-sample validation (MAE = 1.612 d, R² = 0.150).
Conclusions:
- Machine learning effectively identified novel and known predictors of postsurgery LOS in CABG patients.
- A small set of variables holds significant predictive power for LOS.
- A simplified linear regression model based on these findings could be applicable after further validation.
Background:
There is a growing need to identify which bits of information are most valuable for healthcare providers. The aim of this study was to search for the highest impact variables in predicting postsurgery length of stay (LOS) for patients who undergo coronary artery bypass grafting (CABG).
Materials And Methods:
Using a single institution's Society of Thoracic Surgeons (STS) Registry data, 2121 patients with elective or urgent, isolated CABG were analyzed across 116 variables. Two machine learning techniques of random forest and artificial neural networks (ANNs) were used to search for the highest impact variables in predicting LOS, and results were compared against multiple linear regression. Out-of-sample validation of the models was performed on 105 patients.
Results:
Of the 10 highest impact variables identified in predicting LOS, four of the most impactful variables were duration intubated, last preoperative creatinine, age, and number of intraoperative packed red blood cell transfusions. The best performing model was an ANN using the ten highest impact variables (testing sample mean absolute error (MAE) = 1.685 d, R2 = 0.232), which performed consistently in the out-of-sample validation (MAE = 1.612 d, R2 = 0.150).
Conclusion:
Using machine learning, this study identified several novel predictors of postsurgery LOS and reinforced certain known risk factors. Out of the entire STS database, only a few variables carry most of the predictive value for LOS in this population. With this knowledge, a simpler linear regression model has been shared and could be used elsewhere after further validation.
More Related Videos
07:25Predicting Amputation using Local Circulating Mononuclear Progenitor Cells in Angioplasty-treated Patients with Critical Limb Ischemia
Published on: September 22, 2020
09:12Surgical Swine Model of Chronic Cardiac Ischemia Treated by Off-Pump Coronary Artery Bypass Graft Surgery
Published on: March 27, 2018