Prediction of Mortality in Surgical Intensive Care Unit Patients Using Machine Learning Algorithms

Abstract

Objective: Predicting prognosis of in-hospital patients is critical. However, it is challenging to accurately predict the life and death of certain patients at certain period. To determine whether machine learning algorithms could predict in-hospital death of critically ill patients with considerable accuracy and identify factors contributing to the prediction power. Materials and Methods: Using medical data of 1,384 patients admitted to the Surgical Intensive Care Unit (SICU) of our institution, we investigated whether machine learning algorithms could predict in-hospital death using demographic, laboratory, and other disease-related variables, and compared predictions using three different algorithmic methods. The outcome measurement was the incidence of unexpected postoperative mortality which was defined as mortality without pre-existing not-for-resuscitation order that occurred within 30 days of the surgery or within the same hospital stay as the surgery. Results: Machine learning algorithms trained with 43 variables successfully classified dead and live patients with very high accuracy. Most notably, the decision tree showed the higher classification results (Area Under the Receiver Operating Curve, AUC = 0.96) than the neural network classifier (AUC = 0.80). Further analysis provided the insight that serum albumin concentration, total prenatal nutritional intake, and peak dose of dopamine drug played an important role in predicting the mortality of SICU patients. Conclusion: Our results suggest that machine learning algorithms, especially the decision tree method, can provide information on structured and explainable decision flow and accurately predict hospital mortality in SICU hospitalized patients.

Keywords: anesthesia and intensive care; informatics; intensive care; machine learning; surgery.

Figure 1

Flow diagram of participant selection.

Figure 2

Neural network model for prediction of life and death in hospital. Neural network model to transform 43 Surgical Intensive Care Unit (SICU) medical data for prediction of life and death. Tansig = hyperbolic tangent sigmoid function; Purelin = linear transfer function.

Figure 3

Parameter optimization for neural networks. (A) Area Under Curve (AUC) values from different number of neurons (single layer, 10, 20, 50, 100, 200, 300 neurons) (10-fold cross validation results). Maximum AUC at 200 neurons or more. (B) AUC values from different number of layers with 200 neurons (10-fold cross validation results). Maximum AUC at single layer (Error bars represent standard errors).

Figure 4

Receiver operating characteristics curve of machine learning algorithms. ROC curve of Decision tree (AUC = 0.75), neural net (AUC = 0.80), and Bayes (AUC = 0.73) classification algorithms. The results are based on 10-fold cross validations.

Figure 5

Optimized decision tree for the classification of life/death of patients in surgical intensive care unit.

Figure 6

Contribution of 43 variables in predicting life or death of ICU patients (32, 33).