Bibliometric and visual analysis of emergency department crowding assessment studies——CiteSpace-based knowledge graph construction

Ren Zhen; Yang Yilan; Li Shu; Ma Qingbian

doi:10.3969/j.issn.1002-1949.2025.07.008

Abstract：Objective Based on the bibliometric method,to systematically sort out the research progress,knowledge structure and evolution path in the field of emergency department crowding assessment in order to reveal the development law of the discipline.Methods The Web of Science Core Collection was used as the data source to retrieve 931 documents from January 1,1995 to December 31,2024 related to emergency department crowding assessment.CiteSpace 6.3.1 was used to perform keyword co—occurrence,emergence detection and timeline clustering analysis,combined with author collaboration network,institutional collaboration network and literature co-citation analysis to draw a knowledge map.Results The development of the field was characterized by three phases:①the period of foundation building(1995-2005),focusing on basic indicators and policy response;②the period of technology integration(2006-2015),with data-driven model development and quality standardization and validation management becoming the focus of research.The emergence intensity of"ambulance steering"reached 12.98,and the 12-year emergence period since 2003 indicated that this topic has experienced a complete cycle of"policy response-technology optimization-effect controversy";③Intelligent transition period(2016-2024),"machine learning"(the emergence intensity of 9.31,2021-2024)was first applied to the development of emergency department crowding prediction model,reflecting the explosive growth of technology application.The knowledge graph showed that high-frequency keywords("emergency department""nursing",etc.)and high-mid-frequency keywords("quality""accuracy",etc.)formed a"basics-methods"dual-core structure.Author collaboration was generally characterized by"wide distribution,low publication,and small aggregation",with institutions such as Harvard University and the University of Toronto forming a North American-led collaborative network.Three types of knowledge clusters were identified in the co-citation analysis:highly cited foundational literatures that laid the foundation for research(e.g.,systematic evaluation provided a framework for multidimensional assessment of interventions),highly intermediary-centered bridging literatures that facilitated cross-disciplinary integration(e.g.,internationalized discussion of emergency crowding and the impact of emergency crowding on treatment delay),and emergent literatures that revealed emerging literature trends over time(e.g.,Affleck's"input-throughput-output"model systematically analysed the core mechanisms of resource flow in the emergency department and was still widely used today).Conclusions Research on emergency department crowding assessment has undergone a paradigm shift from standardized indicators to intelligent models,and is now showing a dual-track development of"technology-driven"and"policy-driven".It is recommended that future breakthroughs include:①developing a dynamic assessment system that integrates multimodal data and machine learning,②establishing a policy framework for grading congestion response based on evidence-based medicine.