Park Yujin, Chang Sun Ju, Kim Eunhye
College of Nursing, Seoul National University, Seoul, South Korea.
College of Nursing Research Institute of Nursing Science, Seoul National University, Seoul, South Korea.
Aust Crit Care. 2025 Apr 4;38(4):101225. doi: 10.1016/j.aucc.2025.101225.
The integration of artificial intelligence (AI) into health care has been rapidly advancing, driven by its potential to enhance nursing care quality through improved decision-making and efficiency. Within critical care nursing, where the complexity and urgency of patient data are paramount, AI technologies offer significant advantages, such as enhanced patient monitoring and support in clinical decision-making.
AIM/OBJECTIVE: The aim of this scoping review was to synthesise existing literature on AI applications in critical care nursing and their impact on patient outcomes and nursing practice.
Following Arksey and O'Malley's framework, we conducted a systematic search across seven electronic databases including PubMed, CINAHL, and Embase. Studies were included if they involved AI applications in critical care nursing or reported on AI's impact on patient outcomes and clinical decision-making in critical care settings. A synthesis of the literature was conducted following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews checklist.
Thirty-five studies that addressed this topic were included. The review identified six distinct domains of AI applications: monitoring, nursing intervention, clinical decision support systems, documentation, resource allocation, and predictive analytics. Predictive analytics emerged as the most prevalent application, particularly in forecasting complications such as pressure injuries and sepsis onset. Notably, narrowly focussed AI applications demonstrated superior performance compared to broader applications in clinical decision support systems, particularly in specific tasks like neonatal pain classification. AI-driven documentation systems showed promise in reducing administrative burden and improving accuracy, while resource allocation tools enhanced staffing optimisation and workflow management in intensive care units.
Our findings demonstrate AI's significant potential to enhance critical care nursing practice while highlighting implementation challenges. Future research should focus on developing standardised implementation strategies and clear guidelines for AI integration in nursing workflow while maintaining the balance between technological advancement and human expertise.
Open Science Framework Registries http://osf.io/t2y43/.
在通过改善决策和提高效率来提升护理质量的潜力推动下，人工智能（AI）在医疗保健领域的整合正在迅速推进。在重症护理中，患者数据的复杂性和紧迫性至关重要，人工智能技术具有显著优势，例如加强患者监测以及在临床决策中提供支持。
本范围综述的目的是综合关于人工智能在重症护理中的应用及其对患者结局和护理实践影响的现有文献。
按照阿克西和奥马利的框架，我们在包括PubMed、CINAHL和Embase在内的七个电子数据库中进行了系统检索。如果研究涉及人工智能在重症护理中的应用，或者报告了人工智能对重症护理环境中患者结局和临床决策的影响，则纳入这些研究。按照系统评价和Meta分析扩展版的范围综述清单的首选报告项目进行文献综合。
纳入了35项涉及该主题的研究。该综述确定了人工智能应用的六个不同领域：监测、护理干预、临床决策支持系统、文档记录、资源分配和预测分析。预测分析成为最普遍的应用，特别是在预测诸如压疮和脓毒症发作等并发症方面。值得注意的是，在临床决策支持系统中，与更广泛的应用相比，狭义聚焦的人工智能应用表现更优，尤其是在新生儿疼痛分类等特定任务中。人工智能驱动的文档系统在减轻行政负担和提高准确性方面显示出前景，而资源分配工具则加强了重症监护病房的人员配置优化和工作流程管理。
我们的研究结果表明人工智能在提升重症护理实践方面具有巨大潜力，同时也凸显了实施挑战。未来的研究应专注于制定标准化的实施策略以及人工智能融入护理工作流程的明确指南，同时保持技术进步与人类专业知识之间的平衡。
开放科学框架注册库http://osf.io/t2y43/
