Background/Aim Hospital malnutrition is a prevalent issue affecting 30–50 % of hospitalized patients, leading to prolonged hospital stays, increased complications, and higher healthcare costs. Traditional screening methods often fail to identify malnutrition effectively. Artificial intelligence (AI) has the potential to enhance early detection, improve patient outcomes, and optimize hospital resource utilization. This systematic review aims to evaluate the effectiveness of AI-based interventions in the early identification and management of hospital malnutrition. Methods A systematic search was conducted across multiple databases, including PubMed, the Cochrane Library, Cumulative Index to Nursing and Allied Health Literature, and Excerpta Medica Database, following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. Studies evaluating AI interventions for the detection of malnutrition in hospitalized adult patients were included. Methodological quality and risk of bias were assessed using the Joanna Briggs Institute (JBI) tools. Results Twelve studies met the inclusion criteria, utilizing AI algorithms such as Random Forest, Extreme Gradient Boosting (XGBoost), and Light Gradient Boosting Machine (LightGBM). AI-based models demonstrated superior accuracy compared to traditional screening methods, with area under the curve (AUC) values exceeding 90 % in several studies. These interventions improved malnutrition diagnosis rates, reduced diagnostic delays, and enhanced cost-efficiency by optimizing resource allocation and reducing hospital length of stay. Conclusion AI-driven approaches show strong potential for improving the detection and management of malnutrition, offering greater diagnostic accuracy and operational efficiency. Integrating AI into clinical workflows could enhance patient outcomes and generate cost savings. However, challenges such as data quality, staff training, and ethical considerations must be addressed to ensure effective implementation. Further research is needed to validate AI applications across diverse healthcare settings. Protocol registration This systematic review followed a protocol registered prospectively on Open Science Framework available at: 10.17605/OSF.IO/34KWU.
Artificial intelligence in the management of hospital malnutrition: A systematic review
Cangelosi, Giovanni
;Petrelli, Fabio;
2025-01-01
Abstract
Background/Aim Hospital malnutrition is a prevalent issue affecting 30–50 % of hospitalized patients, leading to prolonged hospital stays, increased complications, and higher healthcare costs. Traditional screening methods often fail to identify malnutrition effectively. Artificial intelligence (AI) has the potential to enhance early detection, improve patient outcomes, and optimize hospital resource utilization. This systematic review aims to evaluate the effectiveness of AI-based interventions in the early identification and management of hospital malnutrition. Methods A systematic search was conducted across multiple databases, including PubMed, the Cochrane Library, Cumulative Index to Nursing and Allied Health Literature, and Excerpta Medica Database, following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. Studies evaluating AI interventions for the detection of malnutrition in hospitalized adult patients were included. Methodological quality and risk of bias were assessed using the Joanna Briggs Institute (JBI) tools. Results Twelve studies met the inclusion criteria, utilizing AI algorithms such as Random Forest, Extreme Gradient Boosting (XGBoost), and Light Gradient Boosting Machine (LightGBM). AI-based models demonstrated superior accuracy compared to traditional screening methods, with area under the curve (AUC) values exceeding 90 % in several studies. These interventions improved malnutrition diagnosis rates, reduced diagnostic delays, and enhanced cost-efficiency by optimizing resource allocation and reducing hospital length of stay. Conclusion AI-driven approaches show strong potential for improving the detection and management of malnutrition, offering greater diagnostic accuracy and operational efficiency. Integrating AI into clinical workflows could enhance patient outcomes and generate cost savings. However, challenges such as data quality, staff training, and ethical considerations must be addressed to ensure effective implementation. Further research is needed to validate AI applications across diverse healthcare settings. Protocol registration This systematic review followed a protocol registered prospectively on Open Science Framework available at: 10.17605/OSF.IO/34KWU.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
