From Triage to Intensive Care: A Qualitative Study of Nurses' Experiences with AI-Enabled Decision Support.

BACKGROUND: Artificial intelligence (AI)-enabled decision support systems are increasingly used in emergency departments and intensive care units to support triage, prediction of deterioration, sepsis recognition and escalation decisions. Although these tools may enhance patient safety, they also introduce new challenges for nurses who remain professionally accountable for clinical judgement in high-acuity settings. AIM: To explore emergency and critical care nurses' experiences of using AI-enabled decision support systems, with a focus on clinical judgement, escalation decisions, professional responsibility and ethical considerations across the emergency department-to-intensive care unit continuum. STUDY DESIGN: A qualitative study informed by Husserl's descriptive phenomenology was conducted across three emergency departments and three intensive care units in three public hospitals in northern Saudi Arabia. Semi-structured interviews were conducted with 16 nurses who had direct experience using AI-enabled decision support systems. Data were analysed using Braun and Clarke's reflexive thematic analysis. Reporting followed the Standards for Reporting Qualitative Research. FINDINGS: Three themes were identified. 'Judgment under algorithmic pressure' reflected nurses' efforts to interpret AI alerts alongside bedside assessment, clinical uncertainty and concerns about over-reliance. 'Navigating escalation and responsibility' highlighted heightened accountability during patient deterioration and transfer from the emergency department to intensive care. 'Professional identity and ethical framing' captured nurses' concerns about role changes, moral agency and ethical discomfort when AI recommendations conflicted with patient-centred judgement. CONCLUSIONS: Emergency and critical care Nurses experienced AI-enabled decision support as both helpful and burdensome. Rather than reducing responsibility, AI intensified interpretive work, accountability and ethical tension. Human-centred governance, AI literacy and clear accountability frameworks are needed to support safe AI integration in acute care. RELEVANCE TO CLINICAL PRACTICE: Nurses require organisational support, ethical guidance and AI-focused education to critically engage with decision support systems while preserving professional judgement and moral agency.