中国全科医学 ›› 2026, Vol. 29 ›› Issue (22): 3094-3102.DOI: 10.12114/j.issn.1007-9572.2025.0579

• 中国全科医学研究方法学专题系列(一) • 上一篇    下一篇

全科医务人员在人工智能的辅助下提炼高质量研究问题的五个步骤

汪洋1,2,3, 邹川4, 郭梦若5, 何柳华6,7, 陆许航1,7, 桂贤忠7,8, 徐志杰9, 林恺10, 金花1,2,3, 姚弥11, 杨辉12, 于德华1,2,3,*()   

  1. 1.200090 上海市,同济大学附属杨浦医院全科医学科
    2.200090 上海市全科医学与社区卫生发展研究中心
    3.200090 上海市,同济大学医学院全科医学研究中心
    4.611130 四川省成都市,成都中医药大学附属第五人民医院全科医学科
    5.200123 上海市浦东新区迎博社区卫生服务中心
    6.200042 上海市长宁区华阳街道社区卫生服务中心
    7.200331 上海市,同济大学医学院
    8.200072 上海市静安区大宁路街道社区卫生服务中心
    9.310009 浙江省杭州市,浙江大学医学院附属第二医院全科医学科
    10.515041 广东省汕头市,汕头大学医学院第一附属医院
    11.100034 北京市,北京大学第一医院全科医学科
    12.VIC3168澳大利亚墨尔本市,蒙纳士大学医学部公共卫生和预防医学学院全科医学系
  • 收稿日期:2025-12-12 修回日期:2026-06-02 出版日期:2026-08-05 发布日期:2026-07-08
  • 通讯作者: 于德华
  • 汪洋与邹川为共同第一作者

    本文为中文翻译版本,原文"A five-stage, AI-assisted approach for general practitioners to formulate practice-based research questions",已获得授权。翻译与出版遵循COPE和ICMJE关于二次发表的指南。


    作者贡献:

    汪洋提出研究的主要目标,负责方法学的构思与设计,并主导了研究的实施,起草了论文初稿,绘制了图表并编制了指示词,并负责资金获取;邹川为方法学提出了关键的实用转化设计;郭梦若、陆许航、桂贤忠为方法学提供了用于实际检验的实践案例;汪洋、于德华负责项目管理和监督,并提供了研究资源;汪洋、邹川、郭梦若、何柳华、陆许航、桂贤忠、徐志杰、林恺、金花、姚弥、杨辉、于德华进行论文修订;于德华对论文整体负责。

  • 基金资助:
    同济大学附属杨浦医院博士科研启动项目(BS-202404)

A Five-stage Approach for Primary Care Practitioners to Refine High-quality Research Questions with Artificial Intelligence Assistance

WANG Yang1,2,3, ZOU Chuan4, GUO Mengruo5, HE Liuhua6,7, LU Xuhang1,7, GUI Xianzhong7,8, XU Zhijie9, LIN Kai10, JIN Hua1,2,3, YAO Mi11, YANG Hui12, YU Dehua1,2,3,*()   

  1. 1. Department of General Practice, Yangpu Hospital, Tongji University, Shanghai 200090, China
    2. Shanghai General Practice and Community Health Development Research Center, Shanghai 200090, China
    3. General Practice Research Center, Tongji University School of Medicine, Shanghai 200090, China
    4. Department of General Practice, the Fifth People's Hospital of Chengdu University of Traditional Chinese Medicine, Chengdu 611130, China
    5. Yingbo Community Health Service Center, Pudong New Area, Shanghai 200123, China
    6. Huayang Sub-district Community Health Service Center, Changning District, Shanghai 200042, China
    7. Tongji University School of Medicine, Shanghai 200331, China
    8. Daning Road Street Community Health Service Center, Jing'an District, Shanghai 200072, China
    9. Department of General Practice, the Second Affiliated Hospital of Zhejiang University School of Medicine, Hangzhou 310009, China
    10. The First Affiliated Hospital of Shantou University Medical College, Shantou 515041, China
    11. Department of General Practice, Peking University First Hospital, Beijing 100034, China
    12. School of Public Health and Preventive Medicine, Monash University, Melbourne 3168, Australia
  • Received:2025-12-12 Revised:2026-06-02 Published:2026-08-05 Online:2026-07-08
  • Contact: YU Dehua
  • About author:

    WANG Yang and ZOU Chuan are co-first authors

摘要: 提炼高质量研究问题是当前中国全科医务人员在从事相关领域研究时面临的核心挑战之一。他们虽然拥有丰富的实践经验,却普遍缺乏系统的方法学训练和专业指导。为应对此挑战,这项方法学研究开发并初步验证了一套基于人工智能聊天机器人(AI Chatbot)的"人机协作五步法"。该方法创新性地将经典科研方法学理论(如JBI PCC框架、实施科学、矛盾论与实践论等)"封装"为一系列标准化的AI指示词(Prompt),并与全科医务人员共同组成了一套人机协同系统,以完成"实践观察与价值评估""信息提取与要素循证化""文献检索与知识总结""构建研究问题雏形"和"方法选择与可行性评估"5个关键步骤。本研究为科研基础薄弱的中国全科医务人员提供了一个生成以实践导向科学问题的"提炼辅助器",该工具的传播和应用,有望系统性地解决全科医务人员,尤其是基层医务人员开展科研工作时难以实现"从0到1"的启动难题。并为推动中国全科医学科研体系向着务实的"产生本土证据、实践-理论循环"的方向发展,提供了一条切实可行的路径。

关键词: 全科医学, 人工智能, 研究问题, 人机协作, 研究方法学, 提示词工程

Abstract:

Formulating high-quality research questions is one of the most fundamental challenges facing general practice professionals in China. Although they are deeply embedded in frontline practice and accumulate rich experiential knowledge, many still lack systematic methodological training and sustained research support. To address this gap, this methodological study developed and preliminarily validated a five-step human-AI collaborative approach based on AI chatbots. The approach translates classic methodological theories—such as the JBI PCC framework, implementation science, On Contradiction, and On Practice—into a set of standardized prompts, and combines them with stepwise human input to support five core tasks: practice observation and value assessment, information extraction and evidence-based conceptualization, literature searching and knowledge synthesis, construction of a preliminary research question, and method selection with feasibility assessment. This study provides a practical and accessible toolkit for Chinese general practice professionals, especially those with limited research training, to refine practice-based research questions in a structured manner. By lowering the threshold for moving from practical confusion to research initiation, this approach may help frontline practitioners overcome the common "zero-to-one" difficulty in starting research. More broadly, it offers a feasible methodological pathway for strengthening the generation of locally grounded evidence and promoting a practice-theory cycle in China's primary health care system.

Key words: General practice, Artificial intelligence, Research question, Human-AI collaboration, Research methodology, Prompt engineering