From general to specific: Tailoring large language models for real‐world medical communications
2025

Tailoring AI for Medical Communication

Commentary Evidence: moderate

Author Information

Author(s): Sun Xinti, Tang Wenjun, Huang Zigeng, Long Erping, Wan Peixing

Primary Institution: State Key Laboratory of Respiratory Health and Multimorbidity, Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences and Peking Union Medical College

Hypothesis

Can large language models be effectively tailored for real-world medical communications?

Conclusion

The site-specific prompt engineering chatbot (SSPEC) significantly improves healthcare worker efficiency and reduces patient-provider conflicts.

Supporting Evidence

  • The SSPEC framework enhances the model's logical reasoning and adaptability to medical settings.
  • A knowledge-aligned alert system was developed to mitigate hallucinations in AI responses.
  • 62% of doctors and 43.8% of nurses in China reported being overburdened, highlighting the need for improved communication tools.

Takeaway

This study shows how a special AI chatbot can help doctors and nurses communicate better with patients, making healthcare smoother.

Methodology

The study developed a site-specific prompt engineering chatbot (SSPEC) using a fine-tuned GPT model and a structured prompt template.

Potential Biases

Over-reliance on AI could weaken essential human connections in patient-provider interactions.

Limitations

Challenges include maintaining human connection in healthcare and ensuring AI acceptance across different age groups.

Participant Demographics

Healthcare providers in China, including doctors and nurses.

Digital Object Identifier (DOI)

10.1002/ctm2.70157

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication