Atkinson Connor John, Seth Ishith, Seifman Marc Adam, Rozen Warren Matthew, Cuomo Roberto
Department of Plastic and Reconstructive Surgery, Frankston Hospital, Peninsula Health, Frankston, VIC, Australia.
Department of Surgery, Central Clinical School, Monash University, Alfred Hospital, Prahran, VIC, Australia.
J Hand Surg Glob Online. 2024 Apr 30;6(4):524-528. doi: 10.1016/j.jhsg.2024.03.014. eCollection 2024 Jul.
The integration of artificial intelligence and machine learning technologies into the medical field has brought about remarkable advancements, particularly in the domain of clinical decision support systems. However, it is uncertain how they will perform as clinical decision-makers.
This prospective cohort study evaluates the potential of incorporating ChatGPT-4 plus into the management of subcapital fifth metacarpal fractures. The treatment recommendations provided by ChatGPT-4 plus were compared with those of the two control groups-the attending clinic plastic surgeon and an independent expert panel. The primary outcome measures, operative or conservative, were compared between the groups. Intraclass correlation of 0.61 infers moderate reliability in the consistency of recommended management plans across all groups.
Key predictors for opting for operative management, regardless of the decision-maker, included clinical signs of scissoring, extension deficit, and radiographic evidence of intra-articular extension.
These findings support the potential for artificial intelligence applications in enhancing diagnostic and treatment decisions.
TYPE OF STUDY/LEVEL OF EVIDENCE: Therapeutic IV.
将人工智能和机器学习技术整合到医学领域带来了显著进展，尤其是在临床决策支持系统领域。然而，它们作为临床决策者的表现尚不确定。
这项前瞻性队列研究评估了将ChatGPT-4 plus纳入第五掌骨颈骨折管理的潜力。将ChatGPT-4 plus提供的治疗建议与两个对照组——主治门诊整形外科医生和一个独立专家小组的建议进行比较。比较了各组之间的主要结局指标，即手术或保守治疗。组内相关系数为0.61，表明所有组推荐的管理计划一致性具有中等可靠性。
无论决策者是谁，选择手术治疗的关键预测因素包括交叉畸形、伸展受限的临床体征以及关节内伸展的影像学证据。
这些发现支持了人工智能应用在增强诊断和治疗决策方面的潜力。
研究类型/证据水平：治疗性IV级。
