Vorensky Mark, Peredo Daniel, Ferraro Richard, Paris Emily, Mohammadi Asma, Spano Paul, Rao Smita
Department of Rehabilitation and Movement Sciences, School of Health Professions, Rutgers University, Newark, NJ.
Rusk Rehabilitation, NYU Langone Health, New York, NY.
JOSPT Methods. 2025 Jun;1(2):56-60. doi: 10.2519/josptmethods.2025.0151. Epub 2025 Apr 28.
ChatGPT has been increasingly used in clinical practice, education, and research. In orthopedic research, ChatGPT's accuracy in clinical decision-making has been a major concern, with results ranging from 33% to 80% accuracy. Inaccuracies from ChatGPT can be harmful to clinicians, trainees, or patients when responses appear plausible, are trusted, and acted upon. A critical limitation in orthopedic research is the lack of structured prompt engineering, which significantly impacts ChatGPT's performance. The CRISPE (Capacity/Role, Insight, Statement, Personality, Experiment) framework offers a systematic approach to refining prompts and improving response accuracy. This Viewpoint applies the CRISPE framework to recent orthopedic research and highlights opportunities to optimize prompts in ChatGPT. While research is needed to validate and refine prompt engineering tools in orthopedics, these methods have the potential to enhance the accuracy and reliability of ChatGPT's responses and serve as valuable tools in orthopedic practice, education, and research.
ChatGPT已越来越多地应用于临床实践、教育和研究领域。在骨科研究中，ChatGPT在临床决策方面的准确性一直是主要关注点，其准确率在33%至80%之间。当ChatGPT的回答看似合理、被信任并被付诸行动时，其不准确之处可能会对临床医生、实习生或患者造成伤害。骨科研究中的一个关键限制是缺乏结构化的提示工程，这严重影响了ChatGPT的性能。CRISPE（能力/角色、洞察力、陈述、个性、实验）框架提供了一种系统的方法来优化提示并提高回答的准确性。本观点将CRISPE框架应用于近期的骨科研究，并强调了在ChatGPT中优化提示的机会。虽然需要进行研究以验证和完善骨科中的提示工程工具，但这些方法有可能提高ChatGPT回答的准确性和可靠性，并成为骨科实践、教育和研究中有价值的工具。
