Predicting Critical Care Patient Health using Deep Learning Methodologies

The Intensive Care Unit is a fast-paced environment where the most critically ill patients are treated. The patients are monitored by a team of nurses and doctors who are responsible for the patient’s care. Such an environment requires a high level of expertise and experience to ensure the patient’s health is monitored and maintained. Understanding a patient’s health trajectory is crucial to ensuring that the patient receives the correct treatment. Furthermore, this information is crucial in deciding to discharge the patient from critical care. While few projects have been done in this area, the majority of them have focused on directly predicting the risk of mortality or the length of stay in the ICU. This project aims to develop a system that can directly forecast the next day a patient’s health in the Intensive Care Unit through the use of different deep learning techniques. The study will explore the effectiveness of different models on different health metrics. Due to the sensitivity of patient health data, very few studies have been undertaken in this area. As such, this field of study is at the frontier of research in using deep learning in critical care. The study successfully predicted the patient’s health trajectory using a deep learning model. The experiments found that despite a highly limited dataset, deep learning models show great potential in predicting a patient’s health in the ICU.Faculty of Science and Engineerin