临床医生与患者对外科人工智能应用认知的范围综述
共归纳出四个主题:对人工智能的熟悉度、信心与信任、使用期望、伦理关注与责任。结论认为成功整合需教育、透明可解释系统及明确治理框架。
要点
- 外科医生和患者认可人工智能的潜在获益,尤其在术前决策方面。
- 人工智能作为决策支持工具时信任度更高,自主性越强舒适度越低。
- 对人工智能概念理解有限,并关注透明度、可靠性、数据隐私和法律责任。
核心结论
- 将人工智能成功整合到外科实践需要为临床医生和患者提供教育、开发透明可解释的系统,并建立明确的治理框架以处理知情同意、隐私和责任问题。
研究设计与样本
- 范围综述,检索MEDLINE、EMBASE、PubMed、CINAHL和Web of Science截至2024年6月发表的报告外科医生或患者对人工智能使用观点的研究,按PRISMA-ScR指南进行定性分析。
主要发现
- 归纳出四个主题:对人工智能的熟悉度、信心与信任、使用期望、伦理关注与责任。
- 外科医生和患者承认人工智能的潜在益处,特别是在术前决策方面。
- 当人工智能作为决策支持工具而非自主系统时信任度更高;随着人工智能系统自主性增加,舒适度下降。
- 报告了对人工智能概念理解有限,以及对透明度、可靠性、数据隐私和立法责任的担忧。
临床怎么用
- 在围术期引入人工智能时,应优先将其定位为决策支持工具而非自主系统,以提升临床医生和患者的接受度。
- 需同步开展针对临床医生和患者的人工智能教育,并选择透明可解释的系统。
- 应建立明确的治理框架,覆盖知情同意、数据隐私和法律责任。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
原文
BACKGROUND: Widespread adoption of artificial intelligence into surgical care heavily depends on clinician and patient attitudes, which remain poorly characterized. This scoping review aimed to synthesize existing literature on clinician and patient perspectives toward the use of artificial intelligence in perioperative settings and to identify reported barriers and facilitators to its integration. METHODS: A scoping review of MEDLINE, EMBASE, PubMed, CINAHL, and Web of Science was conducted for studies published up to June 2024 that reported surgeon or patient perspectives on artificial intelligence use in surgical care. Findings were qualitatively analyzed and synthesized according to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis scoping review guidelines. RESULTS: Two thousand one hundred fifty-eight records were screened, and 16 studies met inclusion criteria. Four themes emerged: (1) familiarity with artificial intelligence, (2) confidence and trust, (3) expectations for use, and (4) ethical concerns and liability. Surgeons and patients acknowledged potential benefits of artificial intelligence, particularly in preoperative decision-making. Trust was greater when artificial intelligence functioned as a decision-support tool rather than an autonomous system. As artificial intelligence systems became more autonomous, comfort levels declined. Limited understanding of artificial intelligence concepts was reported, alongside concerns regarding transparency, reliability, data privacy, and legislative responsibilities. CONCLUSION: Successful integration of artificial intelligence into surgical practice requires education for both clinicians and patients, development of transparent and interpretable systems, and clear governance frameworks addressing consent, privacy, and liability. Incorporating stakeholder perspectives during artificial intelligence development may help facilitate ethical, patient-centered adoption in perioperative care.
引用
- Musa E, Hei Lee JK, Tubin J, et al. Perceptions of clinicians and patients toward artificial intelligence applications in surgery: A scoping review. Surgery. 2026. PMID: 42224967. doi:10.1016/j.surg.2026.110291
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。