ISSN 2097-6054（网络） ISSN 1672-9234（印刷） CN 11-5289/R
主管：中国科学技术协会　主办：中华护理学会
出版：中华护理杂志社
收录：中国科学引文数据库（CSCD）来源期刊
　　　中国期刊全文数据库
　　　中国核心期刊（遴选）数据库
　　　Scopus
中华护理教育

中华护理教育 ›› 2025, Vol. 22 ›› Issue (9): 1043-1048.doi: 10.3761/j.issn.1672-9234.2025.09.003

• 专题策划——护理教育数智化发展 • 上一篇    下一篇

基于生成式人工智能的护理信息学智能问答系统设计与应用评价

葛慧(), 张星辰, 胡慧玲, 李佳帅, 吴雪()   

  1. 100191 北京市 北京大学护理学院（葛慧，胡慧玲，李佳帅，吴雪）；999077 香港特别行政区 香港城市大学系统工程系（张星辰）
  • 收稿日期:2025-05-20 出版日期:2025-09-15 发布日期:2025-09-19
  • 通讯作者: 吴雪，博士，副教授，E-mail：wuxue@bjmu.edu.cn
  • 作者简介:葛慧，女，硕士（博士在读），E-mail：gh1105@bjmu.edu.cn
  • 基金资助:
    北京大学医学部教育教学研究课题(2024YB13);国家自然科学基金面上项目(72471006);“未名护理”领军人才科研创新孵化基金项目(LJRC202406)

Design and application evaluation of the nursing informatics intelligent question-answering system based on generative artificial intelligence

GE Hui(), ZHANG Xingchen, HU Huiling, LI Jiashuai, WU Xue()   

  • Received:2025-05-20 Online:2025-09-15 Published:2025-09-19

HTML阅读

PDF下载

赞

摘要：

目的 基于生成式人工智能构建护理信息学智能问答系统并进行应用测试。方法 采用网络爬虫技术获取数据，构建护理信息学知识库，以增强DeepSeek-V3模型输出内容的可靠性。基于fastGPT平台实现系统功能集成，开发护理信息学智能问答系统，并开展实验测试。2025年4月—5月招募30名护理学专业学生，通过问卷调查和半结构化访谈对系统进行可用性和满意度评价。结果 实验测试结果表明，与DeepSeek-V3相比，护理信息学智能问答系统回答准确率提升15.39%，语义相似度提升3.08%。在界面清晰性、系统易用性、回答准确性和可靠性方面，学生持积极态度且意见较为一致，平均满意度得分为（4.33±0.48）分。学生访谈结果显示，该系统有助于提高知识获取效率，可以借助其更好地了解护理信息学研究动态和学术交流信息，但在知识库内容和系统功能方面仍有改进空间。结论 护理信息学智能问答系统具有良好的有效性和可用性，未来可在护理教育实践中进一步投入使用，推动护理教育的数字化和智能化转型，提升教学质量。

关键词: 生成式人工智能, 护理教育, 护理信息学, 智能问答系统

Abstract:

Objective To construct a nursing informatics intelligent question-answering system based on generative artificial intelligence and conduct preliminary study for applying and testing. Methods Data was collected using web crawling technology to build a nursing informatics knowledge base，enhancing the reliability of the content output by the DeepSeek-V3 model. The system functions were integrated using the fastGPT platform to develop the nursing informatics intelligent question-answering system，and experimental testing was conducted. From April to May 2025，30 nursing students were recruited to evaluate the system’s usability and satisfaction through questionnaires and semi-structured interviews. Results Experimental test results showed that compared to Deep-Seek-V3，the nursing informatics intelligent question-answering system improved answer accuracy by 15.39% and semantic similarity by 3.08%. Students expressed positive attitudes and consistent opinions regarding interface clarity，system usability，answer accuracy，and reliability，with an average satisfaction score of（4.33±0.48） point. Besides，interview results indicated that the system could improve knowledge acquisition efficiency and facilitates better understanding of nursing informatics research trends and academic exchange information. However，there was still room for improvement in terms of the content of the knowledge base and system functionality. Conclusion The nursing informatics intelligent question-answering system demonstrates good effectiveness and usability. In the future，it can be further implemented in nursing education to promote the digital and intelligent transformation of nursing education and enhance teaching quality.

Key words: Generative artificial intelligence, Nursing education, Nursing informatics, Intelligent question-answering system