目的 了解当前我国医院大模型技术应用潜在价值和面临挑战,为技术进一步落地应用提出建议。方法 本研究采用焦点小组访谈法进行资料收集,邀请到24位受访者参与访谈,并选取主题分析法对访谈资料进行归纳梳理。结果 对四次访谈转录资料展开逐行编码后,最终定义了关于医院大模型应用的19个编码,8个子主题,3个主题。结论 访谈结果显示,当前大模型技术已经在医疗机构内的多种场景中展现出了降本增效与诊疗提质的作用,并对医疗资源配置、诊疗模式产生了影响。但在技术落地应用过程中,面临着硬件与算力带来的高成本限制,以及模型产出内容质量亟待提高的挑战。在后续技术发展过程中,还需解决如何让医生与患者群体接受技术变革、如何使用新技术产生更高经济效益的问题。
Abstract
Objective To understand the potential value and challenges faced by the current application of hospital large language model technology in China, and to provide suggestions for further implementation and application of the technology. Methods This study used focus group interviews to collect data, invited 24 interviewees to participate in the interviews, and selected thematic analysis to summarize and organize the interview data. Results After encoding the transcribed data from four interviews line by line, 19 codes, 8 sub themes, and 3 themes were finally defined for the application of the hospital large language model. Conclusion The interview results show that the current large language model technology has demonstrated its role in reducing costs, increasing efficiency, and improving diagnosis and treatment quality in various scenarios within medical institutions, and has had an impact on the allocation of medical resources and diagnosis and treatment models. However, in the process of technology implementation and application, there are high cost limitations brought by hardware and computing power, as well as challenges in improving the quality of model output content. In the subsequent process of technological development, it is necessary to address how to make doctors and patient groups accept technological changes, and how to use new technologies to generate higher economic benefits.
关键词
大模型技术 /
医疗机构 /
大模型技术应用
Key words
large model technology /
medical institutions /
application of large model technology
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] 倪明选, 张黔, 谭浩宇, 等. 智慧医疗—从物联网到云计算[J]. 中国科学: 信息科学, 2013, 43(4): 515-528.
[2] 刘漫, 沈鹏熠, 张茹梦. 人工智能技术在医疗护理中的应用研究[J]. 中国农村卫生事业管理, 2025, 45(03): 177-182.
[3] 国家卫生健康委办公厅, 国家中医药局综合司, 国家疾控局综合司. 国家卫生健康委员会办公厅关于印发卫生健康行业人工智能应用场景参考指引的通知[EB/OL].(2024-11-06)[2026-01-28]. http://www.nhc.gov.cn/guihuaxxs/gongwen12/202411/647062ee76764323b29a1f0124b64400.shtml.
[4] 北京市人民政府办公厅. 北京市人民政府办公厅关于印发《北京市加快医药健康协同创新行动计划 (2024—2026年) 》的通知[EB/OL].(2024-05-15)[2026-01-28]. https://www.beijing.gov.cn/zhengce/zhengcefagui/202405/t20240523_3692425.html.
[5] 上海市人民政府办公厅. 上海市人民政府办公厅关于印发《上海市发展医学人工智能工作方案 (2025—2027年) 》的通知[EB/OL].(2024-11-23)[2026-04-28]. https://www.shanghai.gov.cn/nw12344/20241224/0f60c01551784720899dbb911e9d5f08.html.
[6] 肖革新, 陈善吉, 王博远, 等. 医疗大模型的应用现状与展望[J]. 中国数字医学, 2025, 20(2): 39-45.
[7] 孙磊, 汪安安, 宋一敏, 等. 大模型在临床医学领域的应用、挑战和展望[J]. 解放军医学院学报, 2025, 46(1): 50-60.
[8] SHEN T, LI Y, CAO Y, et al.Rapid deployment of large language model DeepSeek in Chinese hospitals demands a regulatory response[J]. Nat Med, 2025, 31(10): 3233-3238.
[9] 郑琰莉, 韩福海, 李舒玉, 等. 人工智能大模型在医疗领域的应用现状与前景展望[J]. 医学信息学杂志, 2024, 45(06): 42-49.
[10] 田雪晴, 李泉江, 游茂, 等. 我国医疗机构大模型建设现状调查与分析[J]. 中国卫生信息管理杂志, 2025, 22(1): 38-44.
[11] 胡浩. 焦点小组访谈理论及其应用[J]. 现代商业, 2010, (26): 282.
[12] BRAUN V, CLARKE V. A critical review of the reporting of reflexive thematic analysis in Health Promotion International[J]. Health Promot Int, 2024, 39(3): daae049.
[13] PARASURAMAN A.Technology readiness index (TRI): a multiple-item scale to measure readiness to embrace new technologies[J]. J Serv Res, 2000, 2(4): 307-320.
[14] 梁志刚, 黄跃, 刘思凯, 等. 大模型技术在病历内涵质控中的应用研究[J]. 中国数字医学, 2025, 20(8): 20-28.
[15] 沈兆福. 基于多模态大模型与AR融合技术的智能导诊服务模式研究[J]. 信息系统工程, 2025, (6): 101-104.
[16] GOH E, GALLO RJ, STRONG E, et al.GPT-4 assistance for improvement of physician performance on patient care tasks: a randomized controlled trial[J]. Nat Med, 2025, 31(4): 1233-1238.
[17] FELDMAN MJ, HOFFER EP, CONLEY JJ, et al.Dedicated AI expert system vs generative AI with large language model for clinical diagnoses[J]. JAMA Netw Open, 2025, 8(5): e2512994.
[18] ZADA T, TAM N, BARNARD F, et al.Medical misinformation in AI-assisted self-diagnosis: development of a method for analyzing large language models[J]. JMIR Form Res, 2025, 9: e66207.
[19] 颜见智, 何雨鑫, 骆子烨, 等. 生成式大语言模型在医疗领域的潜在典型应用与面临的挑战[J]. 医学信息学杂志, 2023, 44(09): 23-31.
[20] 肖仰华, 徐一丹. 大规模生成式语言模型在医疗领域的应用: 机遇与挑战[J]. 医学信息学杂志, 2023, 44(09): 1-11.
基金
四大慢病重大专项“科技创新政策与科技项目管理综合应用研究”(2023ZD0509600); 北京市科技计划“AI+健康”协同创新培育项目“人工智能医疗技术应用价值评估体系及示范验证研究”(Z241100007724003)