指南与共识医学人工智能与信息学医学教育与科研方法

现有医学人工智能报告指南方法学严谨性不足且对生成式人工智能覆盖有限:一项范围综述

Lack of methodological rigor and limited coverage of generative artificial intelligence in existing artificial intelligence reporting guidelines: a scoping review.
指南 / 共识 · J Clin Epidemiol · 2025-07-18

该范围综述系统梳理了截至2024年12月31日检索到的68部医学人工智能报告指南,评估其制定方法、适用范围及对生成式人工智能的覆盖情况。结果显示,48.5%的指南聚焦通用人工智能,仅7.4%涉及生成式人工智能或大语言模型,39.7%描述了制定过程,42.6%有多学科专家参与,33.8%遵循EQUATOR建议。作者认为现有指南方法学严谨性欠佳、内容冗余且对生成式人工智能覆盖不足,建议未来指南优先规范制定流程、加强多学科协作并拓展对新兴人工智能技术的关注。

要点

  • 共纳入68部医学人工智能报告指南,48.5%聚焦通用人工智能,仅7.4%涉及生成式人工智能或大语言模型。
  • 39.7%的指南描述了制定过程,42.6%有多学科专家参与,33.8%遵循EQUATOR网络建议。
  • 医学影像领域指南重叠明显,占20.6%;生成式人工智能专项指南占14.7%但覆盖不全面且方法学透明度不足。

核心结论

  • 现有医学人工智能报告指南方法学严谨性欠佳、存在冗余,且对生成式人工智能应用的覆盖不足。
  • 作者认为未来及更新的指南应优先规范制定流程、加强多学科协作,并拓展对生成式人工智能等新兴人工智能技术的关注。

研究设计与样本

  • 本研究为遵循PRISMA-ScR的范围综述。
  • 检索MEDLINE(通过PubMed)、EQUATOR网络、中国知网、FAIRsharing和谷歌学术五个信息源,检索时间自建库至2024年12月31日。
  • 由两名评价者独立筛选文献并使用预设Excel模板提取数据,分歧由第三名评价者解决。
  • 提取数据包括指南特征(如制定方法、目标受众、人工智能领域)、对EQUATOR网络建议的遵循情况及共识方法学。
  • 共纳入68部人工智能报告指南。

主要发现

  • 48.5%的指南聚焦通用人工智能,仅7.4%涉及生成式人工智能或大语言模型。
  • 39.7%的指南描述了制定过程,42.6%有多学科专家参与,33.8%遵循EQUATOR建议。
  • 指南间存在显著重叠,尤其在医学影像领域,占指南的20.6%。
  • 生成式人工智能专项指南占14.7%,但覆盖不全面且方法学透明度不足。

局限性

  • 原文未报告局限性。

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

OBJECTIVES: This study aimed to systematically map the development methods, scope, and limitations of existing artificial intelligence (AI) reporting guidelines in medicine and to explore their applicability to generative AI (GAI) tools, such as large language models (LLMs). STUDY DESIGN AND SETTING: We reported a scoping review adhering to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews. Five information sources were searched, including MEDLINE (via PubMed), Enhancing the QUAlity and Transparency Of health Research (EQUATOR) Network, China National Knowledge Infrastructure, FAIRsharing, and Google Scholar, from inception to December 31, 2024. Two reviewers independently screened records and extracted data using a predefined Excel template. Data included guideline characteristics (eg, development methods, target audience, AI domain), adherence to EQUATOR Network recommendations, and consensus methodologies. Discrepancies were resolved by a third reviewer. RESULTS: Sixty-eight AI reporting guidelines were included; 48.5% focused on general AI, whereas only 7.4% addressed GAI/LLMs. Methodological rigor was limited; 39.7% described development processes, 42.6% involved multidisciplinary experts, and 33.8% followed EQUATOR recommendations. Significant overlap existed, particularly in medical imaging (20.6% of guidelines). GAI-specific guidelines (14.7%) lacked comprehensive coverage and methodological transparency. CONCLUSION: Existing AI reporting guidelines in medicine have suboptimal methodological rigor, redundancy, and insufficient coverage of GAI applications. Future and updated guidelines should prioritize standardized development processes, multidisciplinary collaboration, and expanded focus on emerging AI technologies like LLMs.

来源:J Clin Epidemiol,查看原文。版权归原作者所有。

引用

  1. Luo X, Wang B, Shi Q, et al. Lack of methodological rigor and limited coverage of generative artificial intelligence in existing artificial intelligence reporting guidelines: a scoping review. J Clin Epidemiol. 2025. PMID: 40684889. doi:10.1016/j.jclinepi.2025.111903

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。