Current Attitudes and Use Patterns of Large Language Models in Clinical Practice and Medical Education Among Trauma Providers

Introduction: Large language models (LLMs) are increasingly used in clinical practice and medical education, yet uncertainty remains regarding appropriate use, oversight, and potential effects on clinical reasoning and patient safety. This study examined current LLM use patterns, attitudes, perceived risks, and priorities for responsible integration among individuals engaged in trauma care, clinical guidance development, and medical education. Methods: Participants attending the third Design for Implementation: The Future of Trauma Clinical Guidance and Research Conference (DFI) completed surveys examining LLM use, attitudes, perceived inhibitors to personal use, and clinical applications. During facilitated sessions, participants used open-ended note-taking guides to document their perspectives regarding LLM use in clinical practice, education, and guidance development. Survey data were analyzed descriptively, and qualitative responses were analyzed using rapid qualitative content analysis. Results: 77 participants completed the pre-conference survey; 62 clinicians provided data regarding clinical LLM use and attitudes. Among clinicians, 15% reported daily LLM use, whereas 26% reported never using LLMs. Respondents expressed strongest agreement with the need for regular quality auditing of LLM-supported clinical decisions. Among 52 post-conference survey respondents, 96% somewhat or strongly agreed that clinical training curricula should include education on LLM use. Barriers to LLM use included time required to verify accuracy, concern about patient harm, and insufficient training in appropriate prompting. Qualitative data came from 23 note-taking documents administered during the conference. Both positive and negative perspectives arose regarding AI use, centering on topics including efficiency gains, deskilling and never-skilling, governance and oversight, accuracy and hallucination, patient safety, and education. Discussion: DFI participants recognized efficiency and informational benefits of LLMs while identifying important concerns related to accuracy, patient safety, skill development, and governance. Findings support a need for institutional guidance, education in responsible LLM use, and evaluation of the incorporation of these technologies into clinical practice and health professions education.