文献速递复苏、休克与脓毒症

人工智能心肺复苏导师

An Artificial Intelligence-Enabled Cardiopulmonary Resuscitation Instructor.
PubMed 文献 · JAMA Intern Med · 2026-01-01
一句话结论在模拟院外心脏骤停场景中,专用AI CPR导师对最低可行和最高有效标准的符合率均达100%,911通话中分别为100%和98.9%,高于调度员。

本研究为横断面研究,用模拟应急场景评估6种广泛可用AI模型(ChatGPT、Claude、Gemini、Grok、Llama、Mistral)提供指南一致CPR指令的能力,并据此开发专用AI CPR导师ChatCPR。6种模型对最低可行标准平均符合率89.7%(95%CI 84.8%-93.2%),对最高有效标准平均69.8%(95%CI 65.5%-73.7%)。ChatCPR在相同模拟场景中对最低可行和最高有效标准符合率均为100%(95%CI 89.8%-100%和95%CI 95.4%-100%)。在911通话复测中,ChatCPR对最低可行标准符合率100%(95%CI 88.6%-100%),对最高有效标准98.9%(95%CI 94.0%-99.8%),分别较调度员绝对提高15.5个百分点(95%CI 2.1-27.4,P=.02)和36.1个百分点(95%CI 27.2-44.3,P<.001)。作者认为AI CPR指导对支持旁观者显示出前景,需在多样的一般人群环境中进一步验证。

要点

  • 6种广泛可用AI模型在模拟院外心脏骤停场景中对最低可行CPR标准平均符合率89.7%(95%CI 84.8%-93.2%),对最高有效标准平均69.8%(95%CI 65.5%-73.7%)。
  • 专用AI CPR导师ChatCPR在相同模拟场景中对最低可行和最高有效标准符合率均为100%(95%CI 89.8%-100%和95%CI 95.4%-100%)。
  • 在911通话复测中,ChatCPR对最低可行标准符合率100%(95%CI 88.6%-100%),对最高有效标准98.9%(95%CI 94.0%-99.8%)。
  • ChatCPR较调度员对最低可行标准绝对提高15.5个百分点(95%CI 2.1-27.4,P=.02),对最高有效标准绝对提高36.1个百分点(95%CI 27.2-44.3,P<.001);调度员分别为84.5%(95%CI 73.1%-91.6%)和62.8%(95%CI 55.0%-70.0%)。
  • 作者认为AI CPR指导对支持旁观者显示出前景,但需在多样的一般人群环境中进一步验证以明确其作为可扩展公共卫生干预的作用。

适用人群与场景

  • 模拟院外心脏骤停应急场景中的旁观者CPR指导评估。
  • 911调度员辅助通话中需要CPR的场景。

诊断标准与关键阈值

  • 最低可行标准:包括进行适当深度胸外按压等指令,源自主要协会CPR指南。
  • 最高有效标准:包括确保按压完全回弹等更细致指令,源自主要协会CPR指南。

研究设计与关键结果

  • 研究类型:横断面研究,使用模拟应急场景评估6种广泛可用AI模型(ChatGPT、Claude、Gemini、Grok、Llama、Mistral)的CPR指导能力,并开发专用AI CPR导师ChatCPR,随后用标准重测方法在相同模拟场景和911调度员辅助通话中评估。数据分析时间为2025年4月至12月。
  • 6种AI模型对最低可行标准平均符合率89.7%(95%CI 84.8%-93.2%),范围从Gemini的79.4%(95%CI 63.2%-89.7%)到Grok和Claude的97.1%(95%CI 85.1%-99.5%)。
  • 6种AI模型对最高有效标准平均符合率69.8%(95%CI 65.5%-73.7%),范围从Llama的61.3%(95%CI 50.3%-71.2%)到GPT-4o的75.0%(95%CI 64.5%-83.2%)。
  • ChatCPR在相同模拟场景中对最低可行和最高有效标准符合率均为100%(95%CI 89.8%-100%和95%CI 95.4%-100%)。
  • 在911通话复测中,ChatCPR对最低可行标准符合率100%(95%CI 88.6%-100%),对最高有效标准98.9%(95%CI 94.0%-99.8%)。
  • ChatCPR较调度员对最低可行标准绝对提高15.5个百分点(95%CI 2.1-27.4,P=.02),对最高有效标准绝对提高36.1个百分点(95%CI 27.2-44.3,P<.001);调度员分别为84.5%(95%CI 73.1%-91.6%)和62.8%(95%CI 55.0%-70.0%)。

证据类型、等级与局限

  • 证据类型:横断面研究,早期概念验证研究设计。
  • 局限:作者指出需在多样的一般人群环境中进一步验证。

对床旁实践的改变

  • 作者认为AI-enabled CPR指导对支持旁观者显示出前景,需在多样的一般人群环境中进一步验证以明确AI-based CPR指导作为可扩展公共卫生干预的作用。

提炼说明

  • 本条依据论文摘要整理;摘要未涉及「处置流程与药物剂量」「红旗征与不能漏的情况」「指南或共识出处」,故未列出这些栏目;全文请点「阅读原文」。
英文原文(摘要)· 点击展开

IMPORTANCE: High-quality, timely cardiopulmonary resuscitation (CPR) after out-of-hospital cardiac arrest (OHCA) is vital, but bystanders need help to act. Telecommunicator-CPR by 911 dispatchers promotes bystander CPR but may be limited by inherent variability and delays in human-led systems.

OBJECTIVE: To evaluate the performance of widely available artificial intelligence (AI) models in delivering guideline-concordant CPR instruction and to develop and assess ChatCPR, a purpose-built AI CPR instructor agent in an early proof-of-concept study design.

DESIGN, SETTING, AND PARTICIPANTS: In this cross-sectional study, simulated emergency scenarios were used to evaluate baseline CPR instructing capabilities of widely available AI models (ChatGPT, Claude, Gemini, Grok, Llama, and Mistral). Models were evaluated for their adherence to minimally viable criteria, which included instructions like performing appropriate-depth chest compressions, and maximally effective criteria, which included more nuanced instructions like ensuring that compressions achieved full recoil, all derived from major association CPR guidelines. These insights informed development of the purpose-built AI CPR instructor agent that was subsequently evaluated using standard test-retest methods across the same simulated scenarios and 911 dispatcher-assisted calls where CPR was indicated. Data were analyzed from April through December 2025.

RESULTS: In simulated OHCA scenarios among 6 widely available AI models, the models achieved 89.7% (95% CI, 84.8%-93.2%) of minimally viable criteria, ranging from 79.4% (95% CI, 63.2%-89.7%) for Gemini to 97.1% (95% CI, 85.1%-99.5%) for Grok and Claude. Models achieved a mean of 69.8% (95% CI, 65.5%-73.7%) of maximally effective criteria, ranging from 61.3% (95% CI, 50.3%-71.2%) for Llama to 75.0% (95% CI, 64.5%-83.2%) for GPT-4o. The instructor agent achieved 100% (95% CI, 89.8%-100%) and 100% (95% CI, 95.4%-100%) adherence to minimally viable and maximally effective criteria, respectively, for the same scenarios. In a retest using 911 calls, the agent achieved 100% (95% CI, 88.6%-100%) adherence to minimally viable and 98.9% (95% CI, 94.0%-99.8%) adherence to maximally effective criteria, representing absolute improvements of 15.5 percentage points (95% CI, 2.1-27.4 percentage points; P = .02) and 36.1 percentage points (95% CI, 27.2-44.3 percentage points; P < .001) over dispatchers (84.5%; 95% CI, 73.1%-91.6% and 62.8%; 95% CI, 55.0%-70.0%, respectively).

CONCLUSIONS AND RELEVANCE: These findings suggest that AI-enabled CPR instruction shows promise for supporting bystanders. Further validation in diverse, general population settings is warranted to define the role of AI-based CPR instruction as a scalable public health intervention for OHCAs.

来源:JAMA Intern Med,查看原文。版权归原作者所有。

参考来源

  1. Desai N, Majhail N, Dredze M, et al. An Artificial Intelligence-Enabled Cardiopulmonary Resuscitation Instructor. JAMA Intern Med. 2026. PMID: 42149572. doi:10.1001/jamainternmed.2026.1552

供执业医师抢救与诊疗参考,不替代临床判断;剂量请结合体重、肝肾功能和本院药品规格复核,以指南原文为准。