Scholarlink

Background <p>Artificial intelligence (AI), including large language models (LLMs), is increasingly transforming medicine and surgery by improving precision, efficiency, and communication. Despite significant potential, adoption is limited by technical, ethical, and regulatory challenges, requiring clinician oversight and foundational knowledge. This study assesses AI familiarity, use, and perceptions among obesity care professionals, including metabolic and bariatric surgeons and integrated health (IH) professionals.</p> Methods <p>The IFSO AI Task Force conducted a confidential online survey of metabolic bariatric surgeons and IH professionals to assess familiarity, use, benefits, concerns, barriers, and future trust regarding LLMs. The 46-item questionnaire was validated by experts, piloted, and disseminated via professional channels. Responses were analyzed descriptively, with open-ended items categorized for context, while incomplete responses were included without imputation.</p> Results <p>A total of 243 global metabolic bariatric surgeons and IH completed the survey. 76% had used LLMs, primarily ChatGPT, for research, presentations, patient education, and documentation, with daily or weekly use common. Respondents cited efficiency, time-saving, improved patient education, and decision support as benefits. Concerns included data privacy, bias, hallucinations, and workflow integration. Few had formal AI training. Overall, clinicians viewed LLMs as supportive assistants requiring oversight, transparency, and institutional guidance.</p> Conclusion <p>Clinicians use LLMs cautiously in metabolic and bariatric practice, mainly for documentation, education, and information synthesis. They recognize risks, limitations, and the need for human oversight, transparency, and institutional governance. Current use reflects exploration rather than full endorsement, highlighting the importance of evaluation, training, and structured guidance as AI evolves in medicine.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

IFSO Survey: Use of Large Language Models (LLMs) by Metabolic Bariatric Surgeons and Integrated Health Professionals

  • 期刊论文

  • 发表日期：

    2026年3月16日

查看全文
  • Mohammad Kermansaravi,
  • Athanasios Pantelis,
  • Bart Torensma,
  • Shahab Shahabi,
  • Manuela Mazzarella,
  • Stefanie D’Arco,
  • Yung Lee,
  • Allan Okrainec,
  • Silvia Leite,
  • Thomas H Shin,
  • Ricardo V Cohen

摘要

Background

Artificial intelligence (AI), including large language models (LLMs), is increasingly transforming medicine and surgery by improving precision, efficiency, and communication. Despite significant potential, adoption is limited by technical, ethical, and regulatory challenges, requiring clinician oversight and foundational knowledge. This study assesses AI familiarity, use, and perceptions among obesity care professionals, including metabolic and bariatric surgeons and integrated health (IH) professionals.

Methods

The IFSO AI Task Force conducted a confidential online survey of metabolic bariatric surgeons and IH professionals to assess familiarity, use, benefits, concerns, barriers, and future trust regarding LLMs. The 46-item questionnaire was validated by experts, piloted, and disseminated via professional channels. Responses were analyzed descriptively, with open-ended items categorized for context, while incomplete responses were included without imputation.

Results

A total of 243 global metabolic bariatric surgeons and IH completed the survey. 76% had used LLMs, primarily ChatGPT, for research, presentations, patient education, and documentation, with daily or weekly use common. Respondents cited efficiency, time-saving, improved patient education, and decision support as benefits. Concerns included data privacy, bias, hallucinations, and workflow integration. Few had formal AI training. Overall, clinicians viewed LLMs as supportive assistants requiring oversight, transparency, and institutional guidance.

Conclusion

Clinicians use LLMs cautiously in metabolic and bariatric practice, mainly for documentation, education, and information synthesis. They recognize risks, limitations, and the need for human oversight, transparency, and institutional governance. Current use reflects exploration rather than full endorsement, highlighting the importance of evaluation, training, and structured guidance as AI evolves in medicine.