目的回顾2014-2024年临床决策支持系统在脑卒中患者中应用的研究热点及前沿，以期为医务人员推动脑卒中防治工作提供参考。方法应用CiteSpace可视化分析软件对Web of Science核心合集数据库的文献进行分析，涵盖年发文量、国家、机构、共被引期刊和关键词聚类及突现等。结果共1681篇文献纳入分析，发文量最多的国家为美国（633篇，占37.66%），其次是中国（247篇，占14.69%）；哈佛大学医学院是发文数量最多的机构（55篇，占3.27%）；Stroke是文献共被引最多的期刊。研究热点包含临床决策支持系统辅助脑卒中风险评估与早期预警、急性期脑卒中治疗与院内管理、康复与长期预后，未来研究将聚焦基于机器学习的动态预测模型、多模态数据融合的快速诊断以及基于人工智能的自适应决策系统。结论临床决策支持系统在脑卒中防治中的研究呈现“风险预警-精准治疗-康复管理”的全周期技术辅助，未来可通过人工智能技术优化脑卒中防治，为临床实践提供数据驱动的决策支持。

Objective To review the research hotspots and frontiers in the application of clinical decision support systems (CDSS) for stroke patients from 2014 to 2024,aiming to provide references for healthcare professionals in promoting stroke prevention and treatment.Methods CiteSpace visualization analysis software was used to analyze literature from the Web of Science Core Collection database,covering annual publication volume,countries,institutions,co-cited journals,keyword clustering,and burst detection.Results A total of 1,681 articles were included in the analysis.The United States contributed the most publications (633 articles,37.66%),followed by China (247 articles,14.69%).Harvard Medical School was the most prolific institution (55 articles,3.27%),and “Stroke” was the most frequently co-cited journal.Research hotspots included CDSS-assisted stroke risk assessment and early warning,acute stroke treatment and in-hospital management,and rehabilitation and long-term prognosis.Future research will focus on machine learning-based dynamic prediction models,multimodal data fusion for rapid diagnosis,and AI-based adaptive decision systems.Conclusions Research on CDSS in stroke prevention and treatment demonstrates a full-cycle technical assistance model of “risk warning-precision treatment-rehabilitation management.” Future efforts should strengthen the optimization of AI technologies in stroke care,so as to provide data-driven decision support for clinical practice.

R473.74

国家自然科学基金项目（72304196）