【摘要】 目的 探讨现行Oswestry功能障碍指数(Oswestry disability index, ODI)(v2.1a)在临床应用中的局限性，通过人机协作模式识别问题并提出优化方案，以提升其评估准确性、适用性及患者填写的便利性。方法 结合研究者与生成式人工智能(ChatGPT-4.5与DeepSeek-R1)的多轮交互，系统分析ODI 2.1a存在的问题。基于人机协作提出条目修改、分级调整与漏填处理优化方案，形成更契合现代临床需求的优化版ODI量表。结果 识别出ODI 2.1a在功能障碍分级设置、漏填条目处理、部分条目适应性及表述清晰度等方面存在不足。优化措施包括：将0%得分单列为“正常功能”;明确81%～100%区间的适用人群；以“排尿功能”替代高漏填率的“性生活”条目；细化“提重物”“社交活动”“旅游”等条目的描述。优化版量表在保留原有结构的基础上，提高了评估准确性、填写完整性及临床适应性。结论 在人机协作基础上提出了保持原有框架下的ODI量表优化策略，提升了量表的临床实用性与科学性。虽然人工智能在结构审视与内容优化中发挥了有效辅助作用，但最终仍需临床专业人员的综合判断。更多还原

【Abstract】 Objective To examine the limitations of the current Oswestry Disability Index(ODI v2.1a) in clinical practice and, through human-artificial intelligence(AI) collaboration, propose optimization strategies to improve assessment accuracy, applicability, and patient compliance. Methods This study integrated multiple rounds of interaction between humans(researchers) and generative AI models(ChatGPT-4.5 and DeepSeek-R1) to systematically analyze existing limitations in ODI 2.1a. Using this human-AI collaborative framework, item revisions, scoring adjustments, and improved handling of missing data were proposed to develop a version better suited to contemporary clinical needs. Results It was identified that ODI 2.1a had deficiencies in aspects such as the grading of functional disability, handling of missing items, adaptability of certain entries and clarity of item description. Optimization measures included classifying 0% as normal function, clarifying the clinical meaning of the 81%-100% range, replacing the sexual activity item(frequently left unanswered) with a urinary function item, and refining item descriptions for lifting, social life, and travelling. The optimized version demonstrated greater content accuracy, improved response completeness, and stronger clinical relevance, while preserving continuity with the original structure. Conclusions Through human-AI collaboration, feasible optimization strategies for the ODI are proposed that do not alter its core framework, enhancing its clinical usability and scientific validity. Although AI has played an effective auxiliary role in structural review and content optimization, ultimately, a comprehensive judgment by clinical professionals is still needed.更多还原