Title: The Comparison of AI to Traditional Screening Tools for the Early Detection of Sepsis: A Systematic Review
Authors: Brittany Nguyen, BSN, RN and Dr. Bruce Lazar, MBA, DM
Volume: 9
Issue: 8
Pages: 92-102
Publication Date: 2025/08/28
Abstract:
Sepsis is a severe reaction to an infection in the body, which can be life-threatening if not treated early. Therefore, early detection and intervention are crucial for preventing death. This systematic literature review aimed to explore whether the use of artificial intelligence (AI) for early detection of sepsis, compared to traditional manual screening tools, improves early detection and reduces sepsis rates among hospitalized patients. A literature search was conducted using the Cumulative Index to Nursing and Allied Health Literature and Medical Literature Analysis and Retrieval System Online, Ultimate academic databases, following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. Data from twenty relevant articles were analyzed with the research question at the forefront, and then each article was categorized into multiple common themes found throughout the research. Five themes emerged from the thematic analysis including traditional screening tools (10%), artificial intelligence and machine learning for the prediction and detection of sepsis and sepsis-related mortality (40%), comparison of artificial intelligence and machine learning to traditional sepsis screening tools (20%), creation of artificial intelligence and machine learning models to predict and detect sepsis and sepsis-related mortality (35%), and the combination of both artificial intelligence and machine learning and traditional sepsis screening tools (10%). The findings indicate that artificial intelligence and machine learning models have higher accuracy rates for the early prediction of sepsis in comparison to traditional screening tools. The results also demonstrate that while traditional screening tools can help detect sepsis early, combining them with artificial intelligence and machine learning models can further improve the early detection of sepsis, enabling early intervention to reduce sepsis rates. The implications of these findings provide healthcare leaders with an opportunity to determine which artificial intelligence and machine learning model is most applicable to their healthcare facility to improve sepsis rates.