July 28, 2022Frontiers in Cardiovascular MedicineOpen Access

Machine learning algorithms to predict major bleeding after isolated coronary artery bypass grafting

Q: What is the clinical evidence from this study?

Study design: Cohort. Population: Postoperative major bleeding (n=1045). Intervention: Machine learning algorithms (Conditional Inference Random Forest) vs. Logistic regression model. Primary outcome: Prediction of major bleeding (UDPB classes 3-4) measured by Area Under the Curve (AUC) (AUC 0.831, 95% CI 0.732-0.930, p=0.027).

View Full Paper

Ask AI

Bookmark

Key result

The conditional inference random forest machine learning model predicted major bleeding after isolated CABG with an AUC of 0.831, significantly outperforming the reference logistic regression model.

Why the study?

Postoperative major bleeding is common in cardiac surgery and linked to poor outcomes, prompting evaluation of machine learning methods to predict it.

Do machine learning algorithms improve the prediction of postoperative major bleeding in patients undergoing isolated CABG compared to conventional risk scores?

Population

1,045 patients who underwent isolated CABG

Comparison

Machine learning algorithms vs reference logistic regression model, TRUST, and WILL-BLEED scores

Design

Prediction model development and validation study

Authors

YGYuchen GaoShenyang Aerospace University XLXiaojie LiuXi'an Technological University LWLijuan WangGuizhou Electric Power Design and Research Institute

Discussion

Member takes

Implication

ML models may aid post-CABG bleeding risk stratification; hypothesis-generating and requires prospective validation before clinical use.

Study Design

Type

Cohort (n=1,045)

Multicenter

Structured PICO

Do machine learning algorithms improve the prediction of postoperative major bleeding in patients undergoing isolated CABG compared to conventional risk scores?

Population

1,045 adult patients undergoing elective isolated coronary artery bypass grafting at a single center in China, evaluated for predictors of postoperative major bleeding.

Exposure

Machine learning (ML) prediction models (including conditional inference random forest, stochastic gradient boosting, random forest, support vector machines, XGBoost, etc.)

Comparator

Logistic regression model and conventional risk scores (TRUST and WILL-BLEED)

Outcome

Postoperative major bleeding defined according to the universal definition of perioperative bleeding (UDPB) classes 3 and 4safety

Main Result

Effect estimate: AUC 0.831 (95% CI 0.732-0.930)

Absolute Event Rate: 0.831% vs 0.702%

p-value: p=0.027

Machine learning models, particularly the conditional inference random forest, provide significantly better discrimination for predicting postoperative major bleeding after isolated CABG compared to conventional clinical risk scores.

Limitations

Retrospective data from a single center at a tertiary hospital limits generalizability
Inability to obtain certain degrees of risk, such as relative risk
Unable to compute other risk scores due to lack of appropriate data
The most important variables reported in the ML model are not modifiable
Retrospective data from a single center, limiting generalizability
Inability to obtain certain degrees of risk (e.g., relative risk) from ML algorithms
Inability to compute other risk scores due to lack of appropriate data
Most important variables in the ML model are not modifiable

Cite This Study

Gao et al. (2022) conducted a cohort in Postoperative major bleeding (n=1,045). Machine learning algorithms (Conditional Inference Random Forest) vs. Logistic regression model was evaluated on Prediction of major bleeding (UDPB classes 3-4) measured by Area Under the Curve (AUC) (AUC 0.831, 95% CI 0.732-0.930, p=0.027). The conditional inference random forest machine learning model predicted major bleeding after isolated CABG with an AUC of 0.831, significantly outperforming the reference logistic regression model.

synapsesocial.com/papers/6a8ae81d039371f2a389331d https://doi.org/10.3389/fcvm.2022.881881

View Full Paper

Ask AI

Bookmark

July 28, 2022Frontiers in Cardiovascular MedicineOpen Access

Machine learning algorithms to predict major bleeding after isolated coronary artery bypass grafting

View Full Paper

Ask AI

Bookmark

Key result

The conditional inference random forest machine learning model predicted major bleeding after isolated CABG with an AUC of 0.831, significantly outperforming the reference logistic regression model.

Why the study?

Postoperative major bleeding is common in cardiac surgery and linked to poor outcomes, prompting evaluation of machine learning methods to predict it.

Do machine learning algorithms improve the prediction of postoperative major bleeding in patients undergoing isolated CABG compared to conventional risk scores?

Population

1,045 patients who underwent isolated CABG

Comparison

Machine learning algorithms vs reference logistic regression model, TRUST, and WILL-BLEED scores

Design

Prediction model development and validation study

Authors

YGYuchen GaoShenyang Aerospace University XLXiaojie LiuXi'an Technological University LWLijuan WangGuizhou Electric Power Design and Research Institute

Discussion

Member takes

Implication

ML models may aid post-CABG bleeding risk stratification; hypothesis-generating and requires prospective validation before clinical use.

Study Design

Type

Cohort (n=1,045)

Multicenter

Structured PICO

Do machine learning algorithms improve the prediction of postoperative major bleeding in patients undergoing isolated CABG compared to conventional risk scores?

Population

1,045 adult patients undergoing elective isolated coronary artery bypass grafting at a single center in China, evaluated for predictors of postoperative major bleeding.

Exposure

Machine learning (ML) prediction models (including conditional inference random forest, stochastic gradient boosting, random forest, support vector machines, XGBoost, etc.)

Comparator

Logistic regression model and conventional risk scores (TRUST and WILL-BLEED)

Outcome

Postoperative major bleeding defined according to the universal definition of perioperative bleeding (UDPB) classes 3 and 4safety

Main Result

Effect estimate: AUC 0.831 (95% CI 0.732-0.930)

Absolute Event Rate: 0.831% vs 0.702%

p-value: p=0.027

Limitations

Retrospective data from a single center at a tertiary hospital limits generalizability
Inability to obtain certain degrees of risk, such as relative risk
Unable to compute other risk scores due to lack of appropriate data
The most important variables reported in the ML model are not modifiable
Retrospective data from a single center, limiting generalizability
Inability to obtain certain degrees of risk (e.g., relative risk) from ML algorithms
Inability to compute other risk scores due to lack of appropriate data
Most important variables in the ML model are not modifiable

Cite This Study

synapsesocial.com/papers/6a8ae81d039371f2a389331d https://doi.org/10.3389/fcvm.2022.881881

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

View Full Paper

Ask AI

Bookmark