Why the study?
Do artificial neural networks (ANN) models improve the prediction of mortality and morbidity after cardiac surgery compared to logistic regression and Parsonnet score?
Population
Patients undergoing cardiac surgery
Comparison
Artificial neural networks (ANN) models vs Logistic regression model and Parsonnet score
Design
Cohort
Follow-up
in-hospital
Key result
Artificial neural networks predicted in-hospital mortality (AUC 0.873) and major morbidity (AUC 0.852) after cardiac surgery better than logistic regression and Parsonnet scores.
Authors
Loading...
May aid risk stratification after cardiac surgery; leaves open need for prospective validation before clinical use.
Observational
Do artificial neural networks (ANN) models improve the prediction of mortality and morbidity after cardiac surgery compared to logistic regression and Parsonnet score?
Artificial neural network models provide superior discrimination for predicting in-hospital mortality and major morbidity after cardiac surgery compared to traditional logistic regression and Parsonnet scores.
Peng et al. (2008) conducted an observational in Cardiac surgery. Artificial neural networks (ANN) models vs. Logistic regression model and Parsonnet score was evaluated on In-hospital mortality and major morbidity. Artificial neural networks predicted in-hospital mortality (AUC 0.873) and major morbidity (AUC 0.852) after cardiac surgery better than logistic regression and Parsonnet scores.