Prediction and Simulation of Clinical Crises in the Intensive Care Unit Using Artificial Intelligence and Machine Learning: scoping review

Background: Intensive care units (ICUs) are environments where patients with critical and life-threatening conditions are treated. Managing these patients has always been a significant challenge for healthcare teams due to the complexity of clinical conditions and the high volume of data. In the current era, technologies based on artificial intelligence (AI) and machine learning (ML) are widely used in medical fields. This study aims to review the role of these technologies in predicting and simulating clinical crises in ICUs. Methods: This scoping review was conducted using a structured search strategy across Google Scholar, PubMed, ScienceDirect, and Scopus. Studies from the past five years, published in English, with full-text availability, and based on reliable clinical data, were included. Non-relevant studies, non-ICU research, and those lacking sufficient data were excluded. Titles and abstracts were screened, and five relevant studies were selected for qualitative analysis. Search Strategy: ("Artificial Intelligence" OR "Machine Learning" OR "AI") AND ("Clinical Crisis" OR "Crisis Prediction" OR "Clinical Prediction" OR "Simulation") AND ("Critical Care" OR "Intensive Care Units" OR "ICU") AND ("Simulation" OR "Modeling") Results: Machine learning models demonstrated high accuracy in predicting ICU patient mortality, early sepsis detection, and acute intraoperative hypotension. Additionally, these models improved mortality prediction in ARDS patients and acute kidney injury by identifying key risk factors, enabling preventive interventions. Conclusion: The use of artificial intelligence (AI) and machine learning (ML) in predicting and simulating clinical crises in intensive care units (ICU) shows significant potential for improving patient outcomes. These models analyze complex data, identify hidden patterns, and facilitate early detection and preventive interventions. However, further studies and validation in diverse clinical settings are necessary to ensure their accuracy and reliability. Keywords: AI, Crisis Prediction, ICU, Simulation

کلیدواژه ها:

AI ، Crisis Prediction ، ICU ، Simulation