Prediction of Intensive Care Unit Length of Stay in the MIMIC-IV Dataset

Abstract

Accurately estimating the length of stay (LOS) of patients admitted to the intensive care unit (ICU) in relation to their health status helps healthcare management allocate appropriate resources and better plan for the future. This paper presents predictive models for the LOS of ICU patients from the MIMIC-IV database based on typical demographic and administrative data, as well as early vital signs and laboratory measurements collected on the first day of ICU stay. The goal of this study was to demonstrate a practical, stepwise approach to predicting patient’s LOS in the ICU using machine learning and early available typical clinical data. The results show that this approach significantly improves the performance of models for predicting actual LOS in a pragmatic framework that includes only data with short stays predetermined by a prior classification.

PSO-XnB: a proposed model for predicting hospital stay of CAD patients

“…For three datasets and various LOS ranges, the highest class scored 96.6% accuracy among other classes. Another work by Hempel et al (2023) performed prediction using the MIMIC-IV dataset; in this work, the LOS was categorized into two classes: short and long stay. The authors used the default parameters of the model in optimization and classifier training using Logistic Regression (LR), RF, SVM, and XGBoost; the result shows that RF scored the highest accuracy of 81%.…”

“…In this subsection, the proposed model PSO-XnB is compared with other ensemble models such as Random Forest (RF) (Hempel et al, 2023), Gradient Boosting (GB) (Naemi et al, 2021), Bagging (Tully et al, 2023), and XGBoost (Hempel et al, 2023), to observe the effectiveness of predicting LOS.…”