中国医学科学院基础医学研究所近日宣布,其团队研发的一款具备共情能力的导诊大模型——“特定场景提示增强对话机器人”,已通过临床试验,并获得了循证医学支持的证据。这款模型不仅在专业性和共情能力上展现出明显优势,还能够降低医患沟通中的重复沟通和冲突比例。
该模型构建于不同医学场景中全方位采集的真实导诊对话基础上,提取对话信息形成知识库作为训练数据,能够解决各种特定问题。研究团队成立了“真实医学对话语料联盟”,建立了标准化流程,收集整理了超过120万例真实医患对话。在前瞻性随机对照试验中,导诊大模型在事实性、安全性、共情能力上均优于人类导诊,在实际应用中,该模型可降低11.2%的重复沟通和5.4%的医患冲突比例。
该研究成果已经在医学期刊《自然医学》上发表,标志着中国在人工智能在医疗领域的应用上取得了重要进展。未来,这款导诊大模型有望被用于患者健康教育、术前谈话等更多复杂场景,为医患沟通提供更为高效和人性化的服务。
英语如下:
News Title: “China Develops Empathy-Guided Consultation Model: Reducing Conflicts Between Doctors and Patients Before Surgery”
Keywords: Consultation Guide, Empathy, Clinical
News Content: The Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences, recently announced that a large-model dialogue robot enhanced by situational prompts with empathy capabilities, developed by its team, has passed clinical trials and received evidence-based medical support. This model not only demonstrates significant advantages in professionalism and empathy but also reduces the proportion of repetitive communication and conflicts between doctors and patients.
The model is built on the basis of comprehensive real-world consultation dialogue collected in various medical scenarios, extracting dialogue information to form a knowledge base as training data, capable of solving a variety of specific issues. The research team established the “Real Medical Dialogue Corpus Alliance” and established standardized processes, collecting and organizing over 1.2 million real doctor-patient dialogues. In a prospective randomized controlled trial, the consultation large-model excelled over human consultants in factual accuracy, safety, and empathy, and in practical applications, the model can reduce the proportion of repetitive communication by 11.2% and conflicts between doctors and patients by 5.4%.
The research findings have been published in the medical journal “Nature Medicine,” marking an important progress in China’s application of artificial intelligence in the medical field. In the future, this consultation large-model is expected to be utilized in more complex scenarios such as patient health education and pre-operative talks, providing more efficient and humane services for doctor-patient communication.
【来源】https://www.ithome.com/0/788/635.htm
Views: 3