The Rise of Large Language Models in the Medical Field: A Bibliometric Analysis

Large language models, through pre-training on massive textual data, have become powerful tools for processing medical texts, extracting key information, and insights. Objective: This study, through literature data mining, analyzes and evaluates the key contributions and development scale of large language models in the medical field. Methods: Literature search was conducted using the Web of Science core database, covering publications up to December 2023. Bibliometrics and content analysis were utilized to explore the current status, hotspots, and trends of research. Results: A total of 583 papers were included in the analysis, published across 328 journals, involving 1275 institutions from 70 countries, and 3369 authors. The field saw rapid development within 2023, with a significant increase in publications, accounting for 71.7% (418/583) of all papers. BERT and ChatGPT are the most used large language models in the medical field. Their primary applications are in medical education, medical literature processing, and electronic health records. Conclusion: Large language models are rapidly evolving in the medical field, especially in the United States, China, and the United Kingdom. Collaboration among different author groups should continue to be strengthened. Their application in disease will be a major research hotspot and trend in the future, but challenges such as privacy and ethical issues that accompany their development must be cautiously addressed.