Performance of DeepSeek-R1 and ChatGPT-5 in the Generation of North American Spine Society Clinical Guidelines for Adult Vertebral Compression Fractures: Comparative Study - Takeaways - MDSpire

Performance of DeepSeek-R1 and ChatGPT-5 in the Generation of North American Spine Society Clinical Guidelines for Adult Vertebral Compression Fractures: Comparative Study

  • By

  • Ruiyuan Chen

  • Yue Pan

  • Minghui Liang

  • Aobo Wang

  • Ziqian Ma

  • Yu Xi

  • Ning Fan

  • Shuo Yuan

  • Peng Du

  • Tianyi Wang

  • Lei Zang

  • July 10, 2026

  • 0 min

Share

  • 1

    Vertebral compression fractures (VCFs) significantly impact public health, affecting approximately 1.5 million American adults annually.

  • 2

    The North American Spine Society (NASS) published updated guidelines for managing osteoporotic and neoplastic VCFs in 2024.

  • 3

    Large language models (LLMs) like DeepSeek-R1 and ChatGPT-5 show potential in providing clinical recommendations across medical domains.

  • 4

    This study evaluated the performance of DeepSeek-R1 and ChatGPT-5 in relation to the 2024 NASS guidelines using a QUEST-aligned framework.

  • 5

    The evaluation aimed to compare the response quality and guideline concordance of the two models based on question type and recommendation grade.

Original Source(s)

Related Content