A Scoping Review of Machine Learning Applications in Nursing Practice: Clinical Decision Support, Risk Prediction, and Workflow Optimization

Abstract

Machine learning (ML) is rapidly transforming nursing practice by enabling advancements in clinical decision support, risk prediction, and workflow optimization. This scoping review synthesizes evidence from empirical studies, reviews, and implementation reports published between 2018 and 2025, identified through Scopus and ScienceDirect. The findings indicate that supervised learning algorithms, deep learning, and natural language processing are widely utilized for risk assessment, early detection of patient deterioration, and enhancement of administrative efficiency. Natural language processing (NLP) also supports automation of nursing documentation and improved data quality. Despite favorable performance metrics, including AUROC values above 0.85 in many applications, most studies are limited by single-institution data, insufficient external validation, and heterogeneous reporting standards. Major barriers include ethical and legal concerns, data quality issues, algorithmic bias, infrastructural limitations, and limited nurse involvement in model development. Enhancing AI literacy and fostering nurse engagement in system design are highlighted as critical for successful clinical integration. Future research priorities include multicenter validation, development of explainable AI, adoption of standardized reporting guidelines, and interdisciplinary collaboration to address ethical, technical, and regulatory challenges. Overall, this scoping review demonstrates that machine learning offers substantial potential to improve patient outcomes and nursing operations, but responsible adoption requires rigorous validation, transparent governance, and active participation of nursing professionals throughout the technology lifecycle

Ähnliche Arbeiten