一家AI X光工具将急诊室17名病人中的误诊病例减少了1:1,特别是帮助中级医生。 An AI X-ray tool cut misdiagnoses by 1 in 17 patients in ERs, especially helping mid-level doctors.

悉尼的一项研究发现,Annalise.ai是AI胸前X射线工具,在急诊部门评估的17名病人中,将误诊人数减少1人,在有大约3年经验的医生中,这一改进幅度最大。 A Sydney study found an AI chest X-ray tool, Annalise.ai, reduced misdiagnoses by one for every 17 patients assessed in emergency departments, with the biggest improvement among doctors with about three years of experience. 在五家医院进行测试,每家有200名医生对18个病例进行评估,AI帮助提高了高压环境的准确性,帮助诊断肺炎和创伤等严重疾病。 Tested across five hospitals with 200 doctors evaluating 18 cases each, the AI helped improve accuracy in high-pressure settings, aiding diagnosis of serious conditions like pneumonia and trauma. 该工具已在一些新南威尔士州医院使用,在现实条件下进行了测试,包括假结果,以评估临床医生的反应。 The tool, already used in several NSW hospitals, was tested under real-world conditions, including false results, to assess clinician responses. 虽然一项隐私调查发现,未经明确同意使用I-MED扫描数据没有违反规定,但仍有令人关切的问题。 While a privacy probe found no breach in using I-MED scan data without explicit consent, concerns remain. 这项研究结果发表在"紧急医学杂志"上,强调了人工智能在作为决策支持工具时,有助于提高诊断准确度的潜力. The findings, published in the Emergency Medicine Journal, highlight AI’s potential to enhance diagnostic accuracy when used as a decision-support tool.