Comparing machine learning and regression models for mortality prediction based on the Hungarian Myocardial Infarction Registry

Abstract The objective of the current study is to compare the relative performance of decision tree, neural network, and logistic regression for predicting 30-day and 1-year mortality in a real-word, unfiltered dataset ( n = 47 , 391 ) of patients hospitalized with acute myocardial infarction. Area under the ROC curve (AUC) was used for evaluating performance of a learning algorithm. For 30-day mortality, we achieved an average of 0.788 for decision tree models, 0.837 for neural net models and 0.836 for regression models on training set (on validation sets: 0.774, 0.835 and 0.834, respectively). For 1-year mortality, the averages were 0.754 for decision tree models, 0.8194 for neural net models and 0.8191 for regression models (on validation sets: 0.743, 0.8179 and 0.8176, respectively). Differences were non-significant between neural network and regression, but both significantly outperformed decision trees. The machine learning methods investigated in the present study could not outperform traditional regression modelling for mortality prediction in myocardial infarction.

[1]  R. Dickerson,et al.  Evaluation of an artificial neural network to predict urea nitrogen appearance for critically ill multiple-trauma patients. , 2005, JPEN. Journal of parenteral and enteral nutrition.

[2]  Mattias Ohlsson,et al.  Risk factor identification and mortality prediction in cardiac surgery using artificial neural networks. , 2006, The Journal of thoracic and cardiovascular surgery.

[3]  Tamas Ferenci,et al.  An Overview of Myocardial Infarction Registries and Results from the Hungarian Myocardial Infarction Registry , 2017, SoMeT.

[4]  Jingjing Lu,et al.  Comparing naive Bayes, decision trees, and SVM with AUC and accuracy , 2003, Third IEEE International Conference on Data Mining.

[5]  Andrew P. Bradley,et al.  The use of the area under the ROC curve in the evaluation of machine learning algorithms , 1997, Pattern Recognit..

[6]  D. Rubin.  Multiple imputation for nonresponse in surveys , 1989 .

[7]  Stef van Buuren,et al.  MICE: Multivariate Imputation by Chained Equations in R , 2011 .

[8]  A. Jánosi,et al.  [Incidence of myocardial infarction in Hungary. Population study in five districts of Budapest and Szabolcs-Szatmár-Bereg county]. , 2013, Orvosi hetilap.

[9]  Ewout W Steyerberg,et al.  Regression trees for predicting mortality in patients with cardiovascular disease: What improvement is achieved by using ensemble-based methods? , 2012, Biometrical journal. Biometrische Zeitschrift.

[10]  Hamido Fujita,et al.  Computer Aided detection for fibrillations and flutters using deep convolutional neural network , 2019, Inf. Sci..

[11]  Peter C. Austin,et al.  Predicting mortality among patients hospitalized for heart failure: derivation and validation of a clinical model. , 2003, JAMA.

[12]  Jing Zhao,et al.  Medical sample classifier design using fuzzy cerebellar model neural networks , 2016 .

[13]  Joseph Brian Adams,et al.  Logistic and neural network models for predicting a hospital admission , 2005 .

[14]  R Core Team,et al.  R: A language and environment for statistical computing. , 2014 .

[15]  Peter C Austin,et al.  A comparison of regression trees, logistic regression, generalized additive models, and multivariate adaptive regression splines for predicting AMI mortality , 2007, Statistics in medicine.

[16]  Babak Mohammadzadeh Asl,et al.  Automated diagnosis of coronary artery disease (CAD) patients using optimized SVM , 2017, Comput. Methods Programs Biomed..

[17]  A. Jánosi,et al.  Does Gender Have Prognostic Value Among Patients with Myocardial Infarction? Analysis of the Data from the Hungarian Myocardial Infarction Registry. , 2018, Journal of women's health.

[18]  Shigeo Yamamura,et al.  Clinical application of artificial neural network (ANN) modeling to predict pharmacokinetic parameters of severely ill patients. , 2003, Advanced drug delivery reviews.

[19]  U. Rajendra Acharya,et al.  Deep convolutional neural network for the automated diagnosis of congestive heart failure using ECG signals , 2018, Applied Intelligence.

[20]  J V Tu,et al.  Development and validation of the Ontario acute myocardial infarction mortality prediction rules. , 2001, Journal of the American College of Cardiology.

[21]  R. Orr,et al.  Use of a Probabilistic Neural Network to Estimate the Risk of Mortality after Cardiac Surgery , 1997, Medical decision making : an international journal of the Society for Medical Decision Making.

[22]  K. Alexander,et al.  Risk adjustment for in-hospital mortality of contemporary patients with acute myocardial infarction: the acute coronary treatment and intervention outcomes network (ACTION) registry-get with the guidelines (GWTG) acute myocardial infarction mortality model and risk score. , 2011, American heart journal.

[23]  William N. Venables,et al.  Modern Applied Statistics with S , 2010 .

[24]  G. Clermont,et al.  Predicting hospital mortality for patients in the intensive care unit: A comparison of artificial neural networks with logistic regression models , 2001, Critical care medicine.

[25]  Helmut Schulte,et al.  Prediction of risk of coronary events in middle-aged men in the Prospective Cardiovascular Münster Study (PROCAM) using neural networks. , 2002, International journal of epidemiology.

[26]  B. Seifert,et al.  Gender differences in the decrease of in-hospital mortality in patients with acute myocardial infarction during the last 20 years in Switzerland , 2017, Open Heart.

[27]  M. Goldacre,et al.  Determinants of the decline in mortality from acute myocardial infarction in England between 2002 and 2010: linked national database study , 2012, BMJ : British Medical Journal.