article · IEEE Access
Intensive care units admit critically ill patients needing constant observation, making early mortality prediction vital for timely life-saving interventions. Existing severity scores and machine-learning approaches face ongoing challenges in precision and clinical usability. A patient-specific stacking ensemble model addresses this by predicting patient mortality more accurately and intuitively. Feature selection and data partitioning into six distinct modalities were guided directly by an intensive care domain expert. Five diverse classifiers, including linear discriminant analysis, decision trees, multilayer perceptrons, k-nearest neighbours, and logistic regression, were assigned to individual modalities before their outputs were fused into an optimized ensemble classifier. Testing on records from 10,664 patients within the MIMIC III benchmark dataset examined time series data from the first 6, 12, and 24 hours of admission. The model achieved 94.4% accuracy, a 93.7% F1 score, and an area under the ROC curve of 93.3%, outperforming earlier methods.
Intensive care teams need rapid, dependable methods to identify patients facing the highest risk of death. By accurately predicting outcomes using data gathered within the first 6 to 24 hours of admission, medical staff can make better-informed, timely decisions about urgent clinical interventions and critical care resource allocation, potentially saving lives.
The model could form the basis of clinical decision-support software used by hospital intensive care staff to flag deteriorating patients early. As the approach has only been evaluated on the historical MIMIC III benchmark dataset, it currently represents applied research tested on retrospective data, requiring further validation in live hospital environments before commercial use.
AI-generated from the published abstract. Always read the original work before citing.
The intensive care unit (ICU) admits the most seriously ill patients requiring extensive monitoring. Early ICU mortality prediction is crucial for identifying patients who are at great risk of dying and for providing suitable interventions to save their lives. Accordingly, early prediction of patients at high mortality risk will enable their provision of appropriate and timely medical services. Although various severity scores and machine-learning models have recently been developed for early mortality prediction, such prediction remains challenging. This paper proposes a novel stacking ensemble approach to predict the mortality of ICU patients. Our approach is more accurate and medically intuitive compared to the literature work. Data were prepared and feature selection was processed under the supervision of the ICU domain expert. The data were split into six modalities based on the expert's decisions. For the prediction process, a separate classifier was selected for each modality based on the performance of the classifiers. We utilized the most popular and diverse classifiers in the literature, including linear discriminant analysis, decision tree (DT), multilayer perceptron, k-nearest neighbor, and logistic regression (LR). Then, a stacking ensemble classifier was constructed and optimized based on the fusion of these five classifier decisions. The framework was evaluated using 10,664 patients from the medical information mart for intensive care (MIMIC III) benchmark dataset. To predict patient mortality, extensive experiments were conducted using the patients' time series data of different lengths. For each patient, the first 6, 12, and 24 hours of the first stay were tested. The results indicate that our model outperformed the state-of-the-art approaches in terms of accuracy (94.4%), F1 score (93.7%), precision (96.4%), recall (91.1%), and area under the receiver operator characteristic (ROC) curve (93.3%). These results demonstrate the ability and efficiency of our approach to predict ICU mortality.
This page summarises published work. The authoritative version sits with the publisher.
DOI: 10.1109/access.2020.3010556
Is something wrong with this record? Report it or request removal.
Discussion
Have you built on this work, tried to replicate it, or seen it applied in practice? Share what you know. Verified researchers and MARATTO™ domain experts can open a discussion, and any member can reply. Contributions are reviewed before they appear.
No discussion yet. Open the first thread.
New to MARATTO™? Create a free account.
