Predicting Stroke After Heart Surgery Using Machine Learning
Author Information
Author(s): Chen Shiqi, Wang Kan, Wang Chen, Fan Zhengfeng, Yan Lizhao, Wang Yixuan, Liu Fayuan, Shi JiaWei, Guo QianNan, Dong NianGuo
Primary Institution: Union Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, Hubei, China
Hypothesis
Can machine learning models effectively predict the risk of postoperative stroke in patients undergoing coronary artery bypass grafting (CABG) surgery?
Conclusion
The study identified key preoperative and intraoperative factors associated with a significant risk of postoperative stroke, enabling proactive measures for high-risk patients.
Supporting Evidence
- Mechanical ventilation time and preoperative creatinine value were identified as significant predictors of postoperative stroke risk.
- Random forest model showed the best performance with an AUC of 0.9008.
- Diabetes and preoperative renal insufficiency were also linked to increased stroke risk.
Takeaway
Doctors can use a special computer program to help figure out which heart surgery patients might have a stroke after their operation, based on certain health clues.
Methodology
The study analyzed data from 1,200 CABG patients using machine learning models, including LASSO for feature selection and random forest for prediction.
Potential Biases
Potential bias due to the single-center design and incomplete data collection.
Limitations
The study was conducted at a single center, which may introduce regional bias, and the sample size may lead to overfitting.
Participant Demographics
The average age of participants was 60.26 years, with 75.8% being men.
Statistical Information
P-Value
0.05
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website