GRAITE-USRA:区域麻醉超声扫查人工智能技术评价报告指南
该国际多学科共识采用EQUATOR Network方法,经文献回顾和两轮改良Delphi(第一轮67名、第二轮63名专家)及25人圆桌会议,制定了GRAITE-USRA报告指南,共40个条目,涵盖超声采集方案和操作者专业能力等现有AI报告指南未涉及的内容。
要点
- GRAITE-USRA含40个报告条目,用于区域麻醉超声扫查AI研究。
- 条目包括超声采集方案与操作者专业能力,为现有AI指南未覆盖。
- 采用改良Delphi共识制定,首轮67名、次轮63名、圆桌25名专家参与。
核心结论
- GRAITE-USRA为区域麻醉超声扫查AI研究提供了最低限度的报告推荐条目。
研究设计与样本
- 遵循EQUATOR Network报告指南制定方法,经文献回顾与专家咨询生成候选条目。
- 采用国际多学科改良Delphi流程,两轮专家咨询后行在线圆桌讨论。
- 第一轮67名、第二轮63名专家参与问卷,25名专家参与圆桌共识会议。
主要发现
- GRAITE-USRA报告指南共包含40个条目,覆盖区域麻醉超声扫查AI研究的关键报告方面。
- 特定条目包括超声采集方案和操作者专业能力,现有AI报告指南未涵盖这些内容。
关键数值与推荐
- 指南共40个条目,为区域麻醉超声扫查AI相关研究提供最低报告推荐集。
临床怎么用
- 采用GRAITE-USRA可促进区域麻醉超声AI研究的一致报告标准与透明度。
- 遵循该指南有助于提高研究可重复性,并支持证据有效整合进临床实践。
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
INTRODUCTION: The application of artificial intelligence to enhance the clinical practice of ultrasound-guided regional anaesthesia is of increasing interest to clinicians, researchers and industry. The lack of standardised reporting for studies in this field hinders the comparability, reproducibility and integration of findings. We aimed to develop a consensus-based reporting guideline for research evaluating artificial intelligence applications for ultrasound scanning in regional anaesthesia. METHODS: We followed methodology recommended by the EQUATOR Network for the development of reporting guidelines. Review of published literature and expert consultation generated a preliminary list of candidate reporting items. An international, multidisciplinary, modified Delphi process was then undertaken, involving experts from clinical practice, academia and industry. Two rounds of expert consultation were conducted, in which participants evaluated each item for inclusion in a final reporting guideline, followed by an online discussion. RESULTS: A total of 67 experts participated in the first Delphi round, 63 in the second round and 25 in the roundtable consensus meeting. The GRAITE-USRA reporting guideline comprises 40 items addressing key aspects of reporting in artificial intelligence research for ultrasound scanning in regional anaesthesia. Specific items include ultrasound acquisition protocols and operator expertise, which are not covered in existing artificial intelligence reporting guidelines. DISCUSSION: The GRAITE-USRA reporting guideline provides a minimum set of recommendations for artificial intelligence-related research for ultrasound scanning in regional anaesthesia. Its adoption will promote consistent reporting standards, enhance transparency, improve study reproducibility and ultimately support the effective integration of evidence into clinical practice. Doctors, scientists and companies are becoming more interested in using artificial intelligence (AI) to help with a type of pain‐blocking treatment called ultrasound‐guided regional anaesthesia. But it is hard to compare studies about this because they are not all written in the same way. Our goal was to create a clear list of rules for how to write these kinds of studies so that everyone can understand and trust the results. We followed expert advice on how to make good rules for research reporting. First, we looked at past research and talked to experts to come up with a list of things that should be included in these studies. Then, we asked a group of experts from different countries and jobs (like doctors, scientists and people from companies) what they thought. They shared their opinions in two rounds of surveys and one online group discussion. In the first round, 67 experts gave feedback. In the second round, 63 experts joined. Then, 25 experts came together for a group meeting. In the end, we created the GRAITE‐USRA guideline. It has 40 important points that researchers should include when writing about AI for ultrasound in regional anaesthesia. Some of these points, like how the ultrasound images are taken and how skilled the person using the machine is, are not covered in other guidelines. The GRAITE‐USRA guideline gives clear and simple rules to help people write better research about using AI in ultrasound for pain‐blocking treatments. If everyone follows these rules, it will be easier to understand, trust and use the results in real medical care.
引用
- Zhang X, Ferry J, Hewson DW, et al. Guidance for reporting artificial intelligence technology evaluations for ultrasound scanning in regional anaesthesia (GRAITE-USRA): an international multidisciplinary consensus reporting framework. Anaesthesia. 2025. PMID: 40963489. doi:10.1111/anae.16733
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。