SHS Web of Conferences (Jan 2024)

Predictive Analysis of Patient Risk of Death in ICU: A Bibliometric Analysis

  • Kuan Li Chung,
  • Chin Lin Yen,
  • De Li Jin,
  • Cheng Cheng Yu,
  • Tuao Zhang,
  • Zixian Yang,
  • Roy Debopriyo

DOI
https://doi.org/10.1051/shsconf/202419401005
Journal volume & issue
Vol. 194
p. 01005

Abstract

Read online

This bibliometric analysis explores the synergy of artificial intelligence (AI), particularly machine learning, and biomedical signal processing in predicting patient mortality risk within the intensive care unit (ICU). Utilizing a comprehensive literature review, the study assesses the research landscape by applying these techniques to ICU data. Examining diverse data sources like vital signs and electronic health records, the analysis identifies trends and gaps in existing work, emphasizing AI’s potential for resource allocation and preventative care to enhance ICU outcomes. Structured within a bibliometric framework, the review encompasses methodological approaches, results, and discussions, while addressing clinical and ethical perspectives on mortality prediction. Challenges related to data, model performance, and fairness are evaluated through a bibliometric lens. The research questions underscore the importance of understanding past literature trends in predictive analysis for ICU patients. The review methodologically explores recent studies employing word representation models, impact assessments, and risk prediction of vital signs. Global research trends in AI for critical care are identified based on bibliographic data between 2013 and 2022. Noteworthy contributions, such as a sepsis dataset, are highlighted within the bibliometric analysis. In conclusion, this bibliometric analysis positions itself at the intersection of AI and critical care, emphasizing the importance of bibliographic data in understanding past trends, methodologies, and impactful contributions. It sets the stage for future directions in the evolving landscape of ICU predictive analytics within a bibliometric framework.