Objective:

To explore primary care physicians' perceptions and experiences with large language models (LLMs) in the United States and the Netherlands, focusing on clinical usability, communication, teamwork, and integration into routine workflows.

Approach:

• Research Design: Qualitative research design using interviews to gather in-depth insights from primary care physicians.
• Participant Recruitment: Participants were recruited through purposive and snowball sampling, ensuring relevant experience with LLM tools.
• Ethical Considerations: The study was approved by an institutional review board, with informed consent obtained from participants.

Key Findings:

• PCPs face increasing complexity in care, administrative burdens, and workforce shortages.
• LLMs like ChatGPT show potential for improving process efficiency and clinical decision-making.
• Concerns exist regarding the accuracy, privacy, and potential overreliance on LLMs in clinical practice.

Interpretation:

Understanding PCPs' experiences with LLMs is essential for their effective integration into primary care.

Limitations:

• The study may not capture all perspectives due to the purposive sampling method.
• Findings are based on qualitative data, which may not be generalizable.

Conclusion:

The study provides insights that can inform the integration of LLMs in primary care, addressing both potential benefits and challenges.