The AI Co-Pilot in Nursing: A Systematic Review of Clinical Decision Support Systems at the Bedside

Abstract

Background: Nursing practice is under significant strain due to workforce shortages and increasing patient acuity. Traditional Clinical Decision Support Systems (CDSS) often lead to "alert fatigue" from vague alarms. The introduction of artificial intelligence, especially machine learning, represents a shift towards predictive, individualized CDSS, intended to assist bedside nurses effectively. Aim: This systematic narrative review synthesizes evidence (2010-2024) on the implementation and impact of AI-powered CDSS in nursing, focusing on applications in deterioration prediction, sepsis detection, and harm prevention. Methods: A systematic search of PubMed, CINAHL, Scopus, and IEEE Xplore databases identified peer-reviewed studies evaluating AI-CDS interventions in acute care nursing. Results: AI-CDSS offers improved predictive accuracy compared to traditional tools, enhancing outcomes such as sepsis bundle compliance and reducing adverse events. Its effect on mortality is ambiguous. Successful integration depends on human-centered design; ineffective systems heighten cognitive load, while effective ones bolster situational awareness. Key findings indicate AI can enhance judgment but may also cause automation bias and diagnostic deskilling. Alert fatigue remains a concern but can be alleviated through tiered, intelligent alerting and closed-loop workflows. Conclusion: The AI co-pilot represents a transformative but complex partner. Its value is realized not through algorithmic superiority alone, but through thoughtful design that supports nursing cognition, integrates seamlessly into workflow, and fosters a culture of calibrated trust, ensuring technology augments rather than displaces essential nursing judgment.

Ähnliche Arbeiten