Machine Learning In Predicting Mortality Rates In Intensive Care Units: An Integrative Review Of The Literature

Authors

DOI:

https://doi.org/10.22279/navus.v16.2091

Abstract

Background: The Intensive Care Unit is a complex environment designed to care for critically ill patients who require constant monitoring and intensive interventions. The computerization of medical records has brought significant advances in information management, enabling large-scale data analysis. In this scenario, machine learning (ML) has emerged as a revolutionary tool for mortality prediction.  Methodology: An integrative review was conducted, covering studies published between 2020 and 2024 in the following databases: Web of Science, LILACS, SciELO, and PubMed. A total of 2,627 studies were analyzed, of which 24 were included. The research focused on the application of ML for mortality prediction. Results: In total, the studies used data from 525,081 patients from databases such as MIMIC-III, MIMIC-IV, and eICU. The most prominent algorithms were XGBoost, LightGBM, Random Forest and neural networks (LSTM, GRU), with superior performance in AUROC, sensitivity and specificity. Key variables included age, lactate, creatinine, SOFA, SAPS II/III and inflammatory biomarkers such as IL-6 and LDH. Final considerations: ML models have demonstrated the validated ability to transform the management of critically ill patients by providing personalized analyses and more accurate predictions.

Downloads

Download data is not yet available.

Author Biographies

Enathanael Ribeiro Soares, Universidade Federal do Cariri (UFCA)

Mestrando em Ciências da Saúde.

Maria Angélica Farias Grangeiro, Universidade Federal do Cariri (UFCA)

Mestranda em Ciências da Saúde.

Ana Heloísa dos Santos, Universidade Federal do Cariri (UFCA)

Mestranda em Ciências da Saúde.

Joel Freires de Alencar Arrais, Universidade do Estado do Rio Grande do Norte (UERN)

Mestrando em Ciências da Saúde.

Glêbia Alexa Cardoso, Universidade Federal da Paraíba (UFPB)

Doutora em Ciências da Saúde.

Estelita Lima Cândido, Faculdade de Medicina do ABC (FMABC)

Pós-Doutor em Ciências da Saúde.

Published

2025-10-02

Issue

Section

Articles