Using machine learning to predict bleeding after cardiac surgery
Victor Hui1,2, Edward Litton3,4, Cyrus Edibam3
1Department of Anaesthesia and Pain Medicine, Royal Melbourne Hospital, Melbourne, VIC, Australia.
Insights
Machine learning models accurately predict bleeding after cardiac surgery using diverse patient data. This approach enhances the prediction of perioperative bleeding events, improving patient outcomes.
Area of Science:
- Medical Informatics
- Cardiovascular Surgery
- Machine Learning in Healthcare
Background:
- Post-cardiac surgery bleeding is a significant complication.
- Accurate prediction of bleeding is crucial for patient management.
- Existing prediction methods may not fully utilize comprehensive patient data.
Purpose of the Study:
- To develop and evaluate machine learning models for predicting perioperative bleeding after cardiac surgery.
- To integrate data from multiple sources including surgical, perfusion, ICU, and laboratory records.
- To compare the performance of different machine learning algorithms in bleeding prediction.
Main Methods:
- Utilized data from 2000 cardiac surgery patients (February 2015 - March 2022).
- Trained machine learning models to predict bleeding using Papworth and Dyke et al. definitions.
- Assessed model performance using metrics like AUROC and AUPRC.
Main Results:
- The Ensemble Voting Classifier demonstrated the best performance.
- Achieved AUPRC of 0.310 (AUROC 0.738) for Papworth definition.
- Achieved AUPRC of 0.452 (AUROC 0.797) for Dyke definition of bleeding.
Conclusions:
- Machine learning effectively predicts post-cardiac surgery bleeding.
- Routinely collected data from various sources can be integrated for prediction.
- This predictive capability can aid in clinical decision-making and patient care.
Objectives:
The primary objective was to predict bleeding after cardiac surgery with machine learning using the data from the Australia New Zealand Society of Cardiac and Thoracic Surgeons Cardiac Surgery Database, cardiopulmonary bypass perfusion database, intensive care unit database and laboratory results.
Methods:
We obtained surgical, perfusion, intensive care unit and laboratory data from a single Australian tertiary cardiac surgical hospital from February 2015 to March 2022 and included 2000 patients undergoing cardiac surgery. We trained our models to predict either the Papworth definition or Dyke et al.'s universal definition of perioperative bleeding. Our primary outcome was the performance of our machine learning algorithms using sensitivity, specificity, positive and negative predictive values, accuracy, area under receiver operating characteristics curve (AUROC) and area under precision-recall curve (AUPRC).
Results:
Of the 2000 patients undergoing cardiac surgery, 13.3% (226/2000) had bleeding using the Papworth definition and 17.2% (343/2000) had moderate to massive bleeding using Dyke et al.'s definition. The best-performing model based on AUPRC was the Ensemble Voting Classifier model for both Papworth (AUPRC 0.310, AUROC 0.738) and Dyke definitions of bleeding (AUPRC 0.452, AUROC 0.797).
Conclusions:
Machine learning can incorporate routinely collected data from various datasets to predict bleeding after cardiac surgery.
More Related Videos
Related Concept Videos
01:28Cardiomyopathy VII: Pre and Post Operative Nursing Management
01:22Aneurysm IV: Nursing Management