Large language models in emergency medicine: a bibliometric study

Aims: The aim of this study was to bibliometrically map the research literature on large language model (LLM) applications in emergency medicine (EM), identify publication trends, thematic structures, and citation characteristics, and evaluate global research productivity and collaboration patterns. Methods: A bibliometric analysis was performed using publications retrieved from the Web of Science Core Collection (WoSCC). Bibliographic data were systematically preprocessed through deduplication, harmonization, and standardization. Performance analysis was conducted to assess publication output, citation metrics, journal characteristics, and geographic distribution. Science-mapping methods, including keyword co-occurrence networks, citation-based overlay visualization, and international collaboration analysis, were applied to explore the thematic structure of the field.Results: A total of 156 original research articles published were included. Annual publication output showed peaking in 2025. Nearly half of the articles were published in first-quartile journals. The United States (USA) was the leading contributor, followed by Turkiye and Israel. Keyword co-occurrence analysis identified artificial intelligence, emergency department, LLMs, and ChatGPT as the central thematic core. Citation-based overlay visualization demonstrated that LLM- and decision-support related keywords were associated with higher average citation impact. International collaboration analysis indicated that the USA served as the primary collaboration hub, with increasing cross-national co-authorship.Conclusion: Research on LLMs in EM has increased rapidly since the emergence of generative LLMs, shifting from exploratory studies toward clinically oriented applications. The literature reflects increasing thematic diversity, international collaboration, and emphasis on diagnostic support and clinical decision-making. This bibliometric analysis summarizes the research landscape and highlights trends guiding the responsible integration of LLMs into EM.