Preprint
Large Language Models

Use, knowledge and perception of large language models in clinical practice: a cross-sectional mixed-methods survey among clinicians in Switzerland

Simon Bruno Egli(Division of Internal Medicine, University Hospital Basel, Basel, Switzerland), Armon Arpagaus(Division of Internal Medicine, University Hospital Basel, Basel, Switzerland), Simon Adrian Amacher(Medical Communication and Psychosomatic Medicine, University Hospital Basel, Basel, Switzerland), Sabina Hunziker(Medical Communication and Psychosomatic Medicine, University Hospital Basel, Basel, Switzerland), Stefano Bassetti(Division of Internal Medicine, University Hospital Basel, Basel, Switzerland)
September 1, 2025BMJ Health & Care Informatics3 citations
DOI

3

Citations

1

Influential Citations

BMJ Health & Care Informatics

Venue

2025

Year

Abstract

Objectives Large language model (LLM)-based tools offer potential for clinical practice but raise concerns regarding output accuracy, patient safety and data security. We aimed to assess Swiss clinicians’ use, knowledge and perception of LLMs and identify associated factors. Methods An anonymous online survey was distributed via 34 medical societies in Switzerland. The primary outcome was frequent use of LLMs (at least weekly use). The secondary outcome was higher knowledge regarding LLMs (score above the median in an 11-item test). Qualitative analysis explored clinicians’ perceptions of LLM-related opportunities and risks. Results Among 685 participants (response rate 29.0%), 225 (32.8%) reported frequent use of LLMs, 25 (3.6%) reported having used a specific medical LLM and 42 (6%) reported the availability of workplace LLM guidelines. The median knowledge test score was 6 points (IQR 4–8 points). Multivariable analysis showed that younger age, male sex and research activity were significantly associated with frequent use and higher knowledge. Qualitative analysis identified administrative support, analytical assistance and access to information as key opportunities. The main risks identified were declining clinical skills, poor output quality and legal or ethical concerns. Discussion The study highlights a notable adoption of LLMs among Swiss clinicians, particularly among younger, male and research-active individuals. However, the limited availability of workplace guidelines raises concerns about safe and effective use. Conclusion The gap between widespread LLM use and the scarcity of workplace guidelines underscores the need for accessible educational resources and clinical guidelines to mitigate potential risks and promote informed use.

Analysis

Why This Paper Matters

This paper provides a timely snapshot of how clinicians in Switzerland are adopting large language models (LLMs) in their daily practice. With the rapid proliferation of tools like ChatGPT, understanding real-world usage patterns is critical for healthcare institutions, policymakers, and AI developers. The finding that nearly a third of surveyed clinicians use LLMs at least weekly—while only 6% have workplace guidelines—reveals a significant safety and governance gap. This is particularly concerning given the potential for inaccurate outputs and data privacy risks in clinical settings.

The study also highlights demographic and professional disparities in adoption. Younger, male, and research-active clinicians are more likely to use LLMs and score higher on knowledge tests. This suggests that early adopters may be self-selected, potentially leaving other groups behind. The lack of formal training and guidelines could exacerbate these disparities, leading to uneven quality of care. This paper is a call to action for healthcare organizations to develop structured educational programs and clear policies.

Technical Contributions

  • Mixed-methods design: Combines quantitative survey data with qualitative thematic analysis, providing both statistical associations and rich contextual insights.
  • Knowledge assessment: Uses an 11-item test to objectively measure clinicians' understanding of LLMs, going beyond self-reported familiarity.
  • Multivariable analysis: Adjusts for confounders to identify independent predictors of frequent use and higher knowledge.
  • Qualitative framework: Categorizes perceived opportunities and risks, offering a structured view of clinician attitudes.

Results

  • Frequent use: 225/685 (32.8%) reported using LLMs at least weekly.
  • Specific medical LLMs: Only 25 (3.6%) had used a dedicated medical LLM, indicating reliance on general-purpose tools.
  • Workplace guidelines: Only 42 (6%) had access to institutional guidelines.
  • Knowledge score: Median 6 out of 11 (IQR 4–8), suggesting moderate understanding.
  • Associated factors: Younger age, male sex, and research activity were significantly associated with both frequent use and higher knowledge.
  • Qualitative themes: Opportunities included administrative support, analytical assistance, and information access; risks included declining clinical skills, poor output quality, and legal/ethical concerns.

Significance

This study is among the first to systematically assess LLM adoption in a national clinician population. It provides a baseline for future research and policy development. The stark gap between usage and guidelines highlights an urgent need for healthcare systems to catch up with technological adoption. The findings also inform AI developers about the specific needs and concerns of clinical users, potentially guiding the design of more trustworthy and clinically integrated LLM tools. As LLMs become more embedded in healthcare, studies like this are essential for ensuring that adoption is safe, equitable, and effective.

large-language-modelsclinical-practicesurveyhealthcareadoptionknowledgeguidelines
Large Language ModelsAi Safety

Related Papers