人工智能在全科临床诊疗与医患沟通及全科住培教学中的应用前景

金昊晗; 翟文清; 金沈樱; 姜隽; 马亚军

doi:10.16766/j.cnki.issn.1674-4152.004229

人工智能在全科临床诊疗与医患沟通及全科住培教学中的应用前景

doi: 10.16766/j.cnki.issn.1674-4152.004229

1.
南京市江宁区禄口街道社区卫生服务中心全科，江苏南京 210000
2.
南京市江宁区汤山街道上峰社区卫生服务中心全科，江苏南京 210000
3.
南京市江宁医院全科医学科，江苏南京 210000
4.
江苏卫生健康职业学院临床医学院，江苏南京 210000
5.
南京市江宁医院科教处

基金项目:

江苏省卫生健康委科研项目 Z2021092

详细信息

通讯作者:
金沈樱，E-mail：936996955@qq.com

中图分类号: R192 R499
计量
- 文章访问数: 1
- HTML全文浏览量: 1
- PDF下载量: 0
- 被引次数: 0
出版历程
- 收稿日期: 2025-03-16

Prospective applications of artificial intelligence in general practice: clinical decision-making, patient communication, and residency education

1.
Department of General Practice, Lukou Community Health Service Center, Jiangning District, Nanjing, Jiangsu 210000, China

摘要

摘要: 目的本研究旨在评估人工智能(AI)在全科医学领域的应用前景，探索AI作为应对全科患者常见问题的实用工具和临床教学工具的潜力。方法选取20个全科常见问题，由AI[ChatGPT(GPT)和DeepSeek(DS)]及全科医师(一年级住培生Y1，三年级住培生Y3，带教老师)回答。从准确性、全面性、详细性、人文关怀及连续性管理5个维度评分，采用单因素方差分析和Tukey HSD检验进行统计分析。结果在准确性、全面性、详细性维度，GPT与DS的评分显著高于Y1、Y3及带教老师。在人文关怀方面，AI评分GPT[(3.16±0.42)分]、DS[(3.23±0.09)分]高于Y1[(2.33±0.45)分]，与Y3[(3.76±0.27)分]相近，但显著低于带教老师[(4.26±0.27)分]。连续性管理方面的评分与人文关怀结果类似。结论 AI在回答标准化全科问题时具备准确性高、信息全面翔实的优势，展现巨大的应用潜力，可用作基础工具及住培生知识辅助。但在人文关怀和连续性管理等需经验判断的领域表现弱于资深医生。未来需针对性提升AI在复杂情境中的表现，其在教学中应定位为知识巩固工具，医患沟通与人文关怀实践仍需带教指导。
- 全科医学 /
- 人工智能 /
- 全科住培 /
- 医学教育 /
- 基层医疗
Abstract: Objective To evaluate the application prospects of artificial intelligence (AI) in general practice and explore its potential as both a practical tool for addressing common patient inquiries and a clinical teaching resource. Methods Twenty standardized clinical queries frequently encountered in general practice were systematically selected. Responses were independently generated by AI systems [ChatGPT (GPT) and DeepSeek (DS)] and human participants comprising first-year residents (Y1), third-year residents (Y3), and attending physicians specialized in general practice. Responses were evaluated across five domains using a 5-point Likert scale: accuracy, comprehensiveness, detail, patient-centered communication, and continuity of care management. Statistical analyses included one-way ANOVA and Tukey's HSD post hoc tests. Results GPT and DS demonstrated significantly superior performance to all physician groups in accuracy, comprehensiveness, and detail. In patient-centered communication, AI performance (GPT: 3.16±0.42; DS: 3.23±0.09) exceeded that of Y1 residents (2.33±0.45), was comparable to Y3 residents (3.76±0.27), but remained significantly inferior to attending physicians (4.26±0.27). Performance in continuity of care management followed an analogous pattern. Conclusion AI exhibits significant potential for implementation in general practice, particularly for delivering accurate, comprehensive, and intelligible responses to standardized clinical queries, highlighting its significant application potential as a foundational tool and a knowledge aid for resident trainees. However, limitations persist in domains requiring experiential judgment(e.g., patient-centered communication, continuity of care), where AI underperforms relative to senior clinicians. Future development should focus on enhancing AI's contextual adaptability in complex scenarios. Within residency training frameworks, AI tools may effectively serve as knowledge consolidation aids, yet cultivating advanced communication skills and empathetic patient engagement necessitates direct clinical mentoring and iterative practice under physician supervision.
- General practice /
- Artificial intelligence /
- General practice residency /
- Medical education /
- Primary care
利益冲突 无

HTML全文

图 1 研究方法流程

Figure 1. Research workflow

下载: 全尺寸图片幻灯片

表 1 全科临床常见问题

Table 1. Common Clinical Queries in General Practice

类别	问题
概念类	1.糖尿病患者：我血糖好好的，怎么就开始高了？
	2.类风湿关节炎患者：我的这个关节痛为什么老是反反复复发作好不了？
	3.感染甲流，或身边有甲流患者：现在甲流这么多我还能不能出门，有什么好办法避免？
	4.对于患有慢性疾病，如冠心病，高血压的患者：随着年龄增长，我的这些慢性病会不会有什么并发症，会不会危及生命？
	5.患者对于早期胃癌：有哪些不明显的症状提示我可能是胃癌，或者早期胃癌？
	6.高血压患者有时候不能准确识别血压升高时的身体症状：头晕是不是就是表示我血压高了？
症状类	7.哮喘患者：马上天凉了，我怕哮喘发作，要不要提前吃药预防？
	8.慢性肾病患者：有的时候小便多，有的时候少，会不会加重了？
诊断类	9.怀疑自己得了甲状腺疾病：同事体检一查甲状腺有问题，我也掉头发没有劲，我是不是甲减，要做什么检查？
	10.对于疑似肿瘤的患者：胸部CT说我肺上有个5 mm×5 mm的结节，我是不是得了癌了？
	11.我在多个医院检查结果不同，会不会是误诊了？
治疗类	12.腰椎间盘突出症患者：这个病要不要做手术？
	13.肺结核患者：这个病能不能治好？大概要治多久？
	14.放化疗患者：听说化疗对身体有副作用，能不能化疗？
药物类	15.高血压患者：如果血压控制得较好，后面能不能减少药量或者停药？
	16.血糖控制不佳需用胰岛素的患者：我不想打胰岛素，行不行？
康复与预防类	17.心肌梗死患者康复期：我什么时候可以下床活动？
	18.有家族癌症病史的人：我家里人有胃癌，我几年查一次，要注意什么吗？
人文关怀类	19.我家小孩最近咳嗽，吃了你们开的药也不好，怎么回事？
	20.医生跟我说我有多囊卵巢综合征，我还能生小孩吗？

下载: 导出CSV

表 2 各组各评价标准得分比较(x±s)

Table 2. Comparative analysis of evaluation metrics across groups(x±s)

组别	例数	准确性	全面性	详细性	人文关怀	连续性管理
ChatGPT	20	4.16±0.22	4.16±0.22	4.50±0.22	3.16±0.42	3.36±0.27
DeepSeek	20	4.00±0.00	4.33±0.43	4.46±0.00	3.23±0.09	3.16±0.23
一年级住培医生	20	2.66±0.42^ab	2.16±0.42^ab	2.16±0.42^ab	2.33±0.45^ab	2.33±0.45^ab
三年级住培医生	20	3.00±0.00^abc	3.16±0.22^abc	3.83±0.22^abc	3.76±0.27^c	3.58±0.13^c
带教老师	20	3.66±0.23^abcd	3.86±0.22^abcd	4.00±0.22^abcd	4.26±0.27^abcd	4.34±0.32^abcd
F值		39.475	30.568	52.669	37.529	30.118
P值		< 0.001	< 0.001	< 0.001	0.001	0.001
注：与ChatGPT比较，^aP < 0.05；与DeepSeek比较，^bP < 0.05；与一年级住培医生比较，^cP < 0.05；三年级住培医生比较，^dP < 0.05。

下载: 导出CSV

表 3 专家对问题评判的Kendall协调系数

Table 3. Kendall's coefficient of concordance (W) for expert evaluations

问题编号	Kendall' s W	P值
1	0.438	0.174
2	0.063	0.779
3	0.063	0.779
4	0.188	0.472
5	0.250	0.368
6	0.813	0.039
7	0.063	0.779
8	0.813	0.039
9	0.063	0.779
10	0.250	0.368
11	0.063	0.779
12	1.000	0.018
13	0.438	0.174
14	1.000	0.018
15	0.063	0.779
16	0.438	0.174
17	0.563	0.105
18	0.063	0.779
19	0.188	0.472
20	0.063	0.779

下载: 导出CSV

参考文献(17)

[1]	DAVE T, ATHALURI S A, SINGH S. ChatGPT in medicine: an overview of its applications, advantages, limitations, future prospects, and ethical considerations[J]. Front Artif Intell, 2023, 6: 1169595. DOI: 10.3389/frai.2023.1169595.
[2]	TAN S, XIN X, WU D. ChatGPT in medicine: prospects and challenges[J]. Int J Surg, 2024, 110(6): 3701-3706. doi: 10.1097/JS9.0000000000001312
[3]	国家卫生健康委员会. "十四五"卫生健康人才发展规划[J]. 中国实用乡村医生杂志, 2022, 29(9): 12-18. The National Health Commission. The 14th Five-Year Plan for the Development of Health and Wellness Talents[J]. Chinese Journal of Practical Rural Doctors, 2022, 29(9): 12-18.
[4]	AYERS J W, POLIAK A, DREDZE M, et al. Comparing physician and artificial intelligence chatbot responses to patient questions posted to a public social media forum[J]. JAMA Intern Med, 2023, 183(6): 589-596. doi: 10.1001/jamainternmed.2023.1838
[5]	GRUNEBAUM A, CHERVENAK J, POLLET S L, et al. The exciting potential for ChatGPT in obstetrics and gynecology[J]. Am J Obstet Gynecol, 2023, 228(6): 696-705. doi: 10.1016/j.ajog.2023.03.009
[6]	LEE T C, STALLER K, BOTOMAN V, et al. ChatGPT answers common patient questions about colonoscopy[J]. Gastroenterology, 2023, 165(2): 509-511.e7. DOI: 10.1053/j.gastro.2023.05.013.
[7]	YEO Y H, SAMAAN J S, NG W H, et al. Assessing the performance of ChatGPT in answering questions regarding cirrhosis and hepatocellular carcinoma[J]. Clin Mol Hepatol, 2023, 29(3): 721-732. doi: 10.3350/cmh.2023.0089
[8]	袁波, 代华, 伍佳, 等. 人工智能在全科医学领域的应用[J]. 中华全科医学, 2021, 19(9): 1433-1436. doi: 10.16766/j.cnki.issn.1674-4152.002079 YUAN B, DAI H, WU J, et al. Application of artificial intelligence in general practice[J]. Chinese Journal of General Practice, 2021, 19(9): 1433-1436. doi: 10.16766/j.cnki.issn.1674-4152.002079
[9]	ALKADHIMI A, AL-MOGHRABI D, FLEMING P S. The nature and accuracy of Instagram posts concerning marketed orthodontic products[J]. Angle Orthod, 2022, 92(2): 247-254. doi: 10.2319/070421-530.1
[10]	GASPARELLO G G, MOTA-JUNIOR S L, HARTMANN G C, et al. Orthodontics social media: perceptions of science-and non-science-based posts among orthodontists, dentists, students and laypeople[J]. PLoS One, 2023, 18(9): e0286927. DOI: 10.1371/journal.pone.0286927.
[11]	吴健, 谭莉, 肖斯柔, 等. 人工智能助力双导师在全科住院医师规范化培训中的应用探讨[J]. 中国现代医生, 2024, 62(36): 88-90. WU J, TAN L, XIAO S R, et al. Application of AI-assisted dual mentors in standardized residency training for general practice[J]. Chin Mod Doct, 2024, 62(36): 88-90.
[12]	汪苗, 潘庆. 我国老年人焦虑状况城乡差异及影响因素分析[J]. 中国全科医学, 2021, 24(31): 3963-3970. WANG M, PAN Q. Urban-rural differences and influencing factors of anxiety in elderly Chinese[J]. Chinese Journal of General Practice, 2021, 24(31): 3963-3970.
[13]	闫温馨, 胡健, 曾华堂, 等. 人工智能大语言模型在基层医疗卫生服务中的应用与挑战[J]. 中国全科医学, 2025, 28(1): 1-6. YAN W X, HU J, ZENG H T, et al. Applications and challenges of large language models in primary healthcare[J]. Chinese Journal of General Practice, 2025, 28(1): 1-6.
[14]	胡佳敏, 邱艳, 任菁菁. AI在基层医疗慢性病管理中的应用研究进展[J]. 中华全科医学, 2024, 22(3): 481-485. doi: 10.16766/j.cnki.issn.1674-4152.003431 HU J M, QIU Y, REN J J. Research progress on AI applications in chronic disease management in primary care[J]. Chinese Journal of General Practice, 2024, 22(3): 481-485. doi: 10.16766/j.cnki.issn.1674-4152.003431
[15]	RASMUSSEN M L R, LARSEN A C, SUBHI Y, et al. Artificial intelligence-based ChatGPT chatbot responses for patient and parent questions on vernal keratoconjunctivitis[J]. Graefes Arch Clin Exp Ophthalmol, 2023, 261(10): 3041-3043. doi: 10.1007/s00417-023-06078-1
[16]	张俊祥, 李传富, 吕维富. 人工智能在医学教育、科研和临床实践中的应用前景与挑战[J]. 中华全科医学, 2024, 22(7): 1085-1089. doi: 10.16766/j.cnki.issn.1674-4152.003572 ZHANG J X, LI C F, LYU W F. Application prospects and challenges of artificial intelligence in medical education, scientific research and clinical practice[J]. Chinese Journal of General Practice, 2024, 22(7): 1085-1089. doi: 10.16766/j.cnki.issn.1674-4152.003572
[17]	黄宇璐, 王珊, 李万昌, 等. "教育共同体"模式下全科专业住院医师规范化培训基地同质化管理的探讨[J]. 广西医学, 2022, 44(13): 1562-1564. HUANG Y L, WANG S, LI W C, et al. Homogeneous management of standardized residency training bases for general practitioners under the "Educational Community" model[J]. Guangxi Medical Journal, 2022, 44(13): 1562-1564.