Machine learning-based in-hospital mortality prediction for transcatheter mitral valve repair in The United States
Suarez DFH, Kim G, Villablanca P, Wiley JM, and Roche-Lima A. Machine learning-based in-hospital mortality prediction for transcatheter mitral valve repair in The United States. Catheterization and Cardiovascular Interventions 2020; 95:S68.
Catheterization and cardiovascular interventions
Background: Transcatheter mitral valve repair (TMVr) utilization has increased significantly in the United States over the last years. Yet, a risk prediction tool for adverse events has not been developed. We aimed to generate a machine learning-based algorithm to predict inhospital mortality after TMVr.
Methods: Patients who underwent TMVr between 2012 and 2015 were identified using the national inpatient sample (NIS) database. The study population was randomly divided into a training set (n=636) and a testing set (n=213). Prediction models for in-hospital mortality were obtained using five supervised machine learning classifiers.
Results: A total of 849 TMVr were analyzed in our study. The overall in-hospital mortality was 3.1%. A naive Bayes (NB) model had the best discrimination for fifteen variables with an Area Under the Curve of 0.83 (95% CI, 0.80-0.87) compared to 0.77 for logistic regression (95% CI, 0.58-0.95), 0.73 for artificial neural network (95% CI, 0.55-0.91) and 0.67 for both random forest and support vector machine (95% CI, 0.47-0.87). However, both random forest and logistic regression models obtained for ten variables were as good as the best NB model with an AUC=0.82 (95% CI, 0.79-0.86, p=0.34). History of coronary artery disease, renal failure and smoking were the three most significant predictors of in-hospital mortality.
Conclusions: We developed a robust machine learning-derived model to predict in-hospital mortality in patients undergoing TMVr. While the best model was obtained by NB, conventional logistic regression generated an alternative model with a comparable performance.