切换至 "中华医学电子期刊资源库"

中华临床医师杂志(电子版) ›› 2023, Vol. 17 ›› Issue (12) : 1212 -1218. doi: 10.3877/cma.j.issn.1674-0785.2023.12.002

专题笔谈

人工智能技术在超说明书用药循证中的应用研究
邱凯锋, 王则远, 何志超, 付凯利, 梅童霖, 关英杰, 高飞, 伍俊妍()   
  1. 510123 广州,中山大学孙逸仙纪念医院药学部
    102600 北京,灵犀量子(北京)医疗科技有限公司
  • 收稿日期:2023-08-10 出版日期:2023-12-15
  • 通信作者: 伍俊妍
  • 基金资助:
    2022年广东省医院协会药学科研专项基金(2022YSGL02); 广州市科技计划(登峰医院)基础研究项目(2023A03J0725)

Application of artificial intelligence technology in evidence-based off-label drug use

Kaifeng Qiu, Zeyuan Wang, Zhichao He, Kaili Fu, Tonglin Mei, Yingjie Guan, Fei Gao, Junyan Wu()   

  1. Department of Pharmacy, Sun Yat-sen Memorial Hospital, Sun Yat-Sen University, Guangzhou 510123, China
    Lingxi Quantum (Beijing) Medical Technology Co., Ltd, Beijing 102600, China
  • Received:2023-08-10 Published:2023-12-15
  • Corresponding author: Junyan Wu
引用本文:

邱凯锋, 王则远, 何志超, 付凯利, 梅童霖, 关英杰, 高飞, 伍俊妍. 人工智能技术在超说明书用药循证中的应用研究[J]. 中华临床医师杂志(电子版), 2023, 17(12): 1212-1218.

Kaifeng Qiu, Zeyuan Wang, Zhichao He, Kaili Fu, Tonglin Mei, Yingjie Guan, Fei Gao, Junyan Wu. Application of artificial intelligence technology in evidence-based off-label drug use[J]. Chinese Journal of Clinicians(Electronic Edition), 2023, 17(12): 1212-1218.

超说明书用药在临床中广泛存在。2022年3月1日生效的《中华人民共和国医师法》要求,超说明书用药必须基于循证医学证据才能实施。循证决策需要对相关海量证据进行查找、筛选、评价,但医学证据的指数倍增长使得海量的数据信息成为循证的瓶颈和挑战。人工智能(AI)技术的应用为解决这一问题带来了新的机遇,其可实现自动化证据搜索、分析、汇总和追踪,从而帮助临床药师更快速、更准确、更全面地获取关键信息,并降低人为误差的风险。聚焦超说明书用药循证决策中的痛点,基于BioBERT、T5、UnifiedQA和GPT-2等AI技术,以满足循证流程的语义搜索、文本分类、信息抽取、决策内容生成等功能,笔者最终开发了EviMed系统,实现了全球海量医学证据的实时搜索并自动进行分析和AI辅助决策,具备快速、全面、准确、可自动更新的特点。该系统已广泛应用于全国多家三甲医院,促进了超说明书用药循证决策的高效化和规范化。

Off-label drug use is widely practiced in clinical settings. In 2022, the Law on Doctors of the People's Republic of China mandated that off-label medication must be supported by evidence-based medical evidence before implementation. However, evidence-based decision-making involves searching, screening, and evaluating vast amounts of relevant evidence. The exponential growth of medical evidence has become a bottleneck and challenge for evidence-based decision-making due to the massive amount of data involved. Fortunately, the application of artificial intelligence (AI) technology presents new opportunities to address this issue. AI technology enables automatic evidence search, analysis, summarization, and tracking, assisting clinical pharmacists in obtaining key information quickly, accurately, and comprehensively while minimizing the risk of human error. This paper focuses on addressing the challenges faced during evidence-based decision-making for off-label medication. By utilizing AI technologies such as BioBERT, T5, UnifiedQA, and GPT-2, we designed functionalities such as semantic search, text classification, information extraction, and generation of decision-making content aligned with the evidence-based process. As a result, the EviMed system has been developed to facilitate real-time search, automatic analysis, and AI-assisted decision-making based on a vast array of global medical evidence. The system is fast, comprehensive, accurate, and capable of automatic updates. Widely adopted in numerous tertiary hospitals nationwide, the system has significantly enhanced the efficiency and standardization of evidence-based decision-making for off-label medication.

图1 超说明书用药人工智能(AI)循证决策方法 注:NER Model为命名实体识别;BioBERT为生物学双向语言表征模型;BERT为双向语言表征模型;CRF;条件随机场;T5为生成式语言模型;UnifiedQA为统一问答模型;GPT-2是生成式自然语言处理预训练模型2
表1 决策报告内容清单
表2 人工输出结果与超说明书用药系统输出结果比对
表3 500个超说明书用药问题人工智能(AI)结果与人工结果的比较情况
图2 EviMed超说明书用药循证决策系统(证据列表)
表4 缬沙坦用于治疗儿童高血压临床研究文献检索及纳入情况
表5 缬沙坦治疗儿童高血压相关临床试验
1
Radley DC, Finkelstein SN, Stafford RS. Off-label prescribing among office-based physicians[J]. Arch Intern Med, 2006,166(9):1021-1026.
2
Briesacher BA, Limcangco MR, Simoni-Wastila L, et al. The quality of antipsychotic drug prescribing in nursing homes[J]. Arch Intern Med, 2005,165(11):1280-1285.
3
Tefferi A, Kantarjian H, Rajkumar SV, et al. In support of a patient-driven initiative and petition to lower the high price of cancer drugs[J]. Mayo Clin Proc, 2015, 90(8):996-1000.
4
刘夏. 临床上超说明书用药的分析及对策[J]. 药学实践杂志,2014(6):465-468.
5
中华人民共和国国家卫生健康委员会. 中华人民共和国医师法(修订)[EB/OL].(2021-08-20) [2022-02-24].
6
Andrews J, Guyatt G, Oxman AD, et al. GRADE guidelines: 14. Going from evidence to recommendations: the significance and presentation of recommendations[J]. J Clin Epidemiol, 2013,66(7):719-725.
7
Guyatt GH, Oxman AD, Vist GE, et al. GRADE: an emerging consensus on rating quality of evidence and strength of recommendations[J]. BMJ, 2008,336(7650):924-926.
8
Brouwers MC, Kho ME, Browman GP, et al. AGREE Ⅱ: advancing guideline development, reporting and evaluation in health care[J]. CMAJ, 2010,182(18):E839-E842.
9
Glanville J, Kendrick T, McNally R, et al. Research output on primary care in Australia, Canada, Germany, the Netherlands, the United Kingdom, and the United States: bibliometric analysis[J]. BMJ, 2011,342:d1028.
10
陈耀龙, 罗旭飞, 史乾灵,等. 人工智能如何改变指南的未来[J]. 协和医学杂志,2021,12(1):114-121.
11
Gates A, Johnson C, Hartling L. Technology-assisted title and abstract screening for systematic reviews: a retrospective evaluation of the Abstrackr machine learning tool[J]. Syst Rev, 2018,7(1):45.
12
Simon C, Davidsen K, Hansen C, et al. BioReader: a text mining tool for performing classification of biomedical literature[J]. BMC Bioinformatics, 2019,19(Suppl 13):57.
13
Adams C E, Polzmacher S, Wolff A. Systematic reviews: work that needs to be done and not to be done[J]. J Evid Based Med, 2013, 6(4): 232-235.
14
Gates A, Guitard S, Pillay J, et al. Performance and usability of machine learning for screening in systematic reviews: a comparative evaluation of three tools[J]. Syst Rev, 2019, 8(1): 1-11.
15
Aljanabi M, Ghazi M, Ali A H, et al. ChatGpt: open possibilities[J]. IJCSM, 2023, 4(1): 62-64.
16
Qiu X, Sun T, Xu Y, et al. Pre-trained models for natural language processing: a survey[J]. Sci China Technol Sci., 2020, 63(10): 1872-1897.
17
Vaswani A, Shazeer N, Parmar N, et al. Attention is all you need[J]. Adv Neural Inf Process Syst, 2017, 30.
18
Lee J, Yoon W, Kim S, et al. BioBERT: a pre-trained biomedical language representation model for biomedical text mining[J]. Bioinformatics, 2020,36(4):1234-1240.
19
Palangi H, Deng L, Shen Y, et al. Deep sentence embedding using long short-term memory networks: Analysis and application to information retrieval[J]. IEEE-ACM T AUDIO SPE, 2016, 24(4): 694-707.
20
Xue L, Constant N, Roberts A, et al. mT5: a massively multilingual pre-trained text-to-text transformer[J]. arXiv preprint arXiv:2010.11934, 2020.
21
Dai Z, Wang X, Ni P, et al. Named entity recognition using BERT BiLSTM CRF for Chinese electronic health records[C]//2019 12th international congress on image and signal processing, biomedical engineering and informatics (cisp-bmei). IEEE, 2019: 1-5.
22
Khashabi D, Min S, Khot T, et al. Unifiedqa: crossing format boundaries with a single qa system[J]. arXiv preprint arXiv:2005.00700, 2020.
23
Radford A, Wu J, Child R, et al. Language models are unsupervised multitask learners[J]. OpenAI blog, 2019, 1(8): 9.
24
伍俊妍, 邱凯锋, 曾英彤, 等. 广东省药学会超说明书用药规范管理工作成果写入《 医师法》[J]. 今日药学, 2022,32(6):401-408.
25
Facts and Comparisons®: Drug Referential Resource[N/OL]. Amsterdam: Wolters Kluwer.
26
Nerenberg K, Zarnke K, Leung A, et al. Hypertension Canada's 2018 guidelines for diagnosis, risk assessment, prevention, and treatment of hypertension in adults and children[J]. Can J Cardiol, 2018, 34(5):506-525.
27
中华医学会儿科学分会心血管学组,中国医师协会心血管内科医师分会儿童心血管专业委员会,中华儿科杂志编辑委员会. 儿童心力衰竭诊断和治疗建议(2020年修订版)[J]. 中华儿科杂志,2021,59(2):84-94.
28
Jankauskiene A, Drozdz D, Wasilewska A, et al. Efficacy and safety of valsartan in children aged 1-5 years with hypertension, with or without chronic kidney disease: a randomized, double-blind study followed by open-label phase[J]. Curr Med Res Opin, 2021,37(12):2113-2122.
29
Bapatla K, Jankauskiene A, Drozdz D, et al. Efficacy and safety of valsartan in children aged 1-5 years with hypertension, with or without chronic kidney disease: a 6-week randomised, multicentre, double-blind study followed by open-label phase[J]. EUR HEART J, 2019, 40(s1): ehz745.0946.
30
Schaefer F, Coppo R, Bagga A, et al. Efficacy and safety of valsartan in hypertensive children 6 months to 5 years of age[J]. J Hypertens, 2013,31(5):993-1000.
31
Wells T, Blumer J, Meyers KE, et al. Effectiveness and safety of valsartan in children aged 6 to 16 years with hypertension[J]. J Clin Hypertens (Greenwich), 2011,13(5):357-365.
32
Schaefer F, Litwin M, Zachwieja J, et al. Efficacy and safety of valsartan compared to enalapril in hypertensive children: a 12-week, randomized, double-blind, parallel-group study[J]. J Hypertens, 2011,29(12):2484-2490.
33
Meyers KE, Lieberman K, Solar-Yohay S, et al. The efficacy and safety of valsartan in obese and non-obese pediatric hypertensive patients[J]. J Clin Hypertens (Greenwich), 2011,13(10):758-766.
34
Flynn JT, Meyers KE, Neto JP, et al. Efficacy and safety of the Angiotensin receptor blocker valsartan in children with hypertension aged 1 to 5 years[J]. Hypertension, 2008,52(2):222-228.
[1] 张梅芳, 谭莹, 朱巧珍, 温昕, 袁鹰, 秦越, 郭洪波, 侯伶秀, 黄文兰, 彭桂艳, 李胜利. 早孕期胎儿头臀长正中矢状切面超声图像的人工智能质控研究[J]. 中华医学超声杂志(电子版), 2023, 20(09): 945-950.
[2] 唐玮, 何融泉, 黄素宁. 深度学习在乳腺癌影像诊疗和预后预测中的应用[J]. 中华乳腺病杂志(电子版), 2023, 17(06): 323-328.
[3] 唐蜜, 蔡江晖, 罗尔丹, 郭文玫, 熊丽玲, 林永红, 邢莎莎, 杨霄. 多肽疫苗治疗乳腺癌的临床研究进展[J]. 中华乳腺病杂志(电子版), 2023, 17(05): 309-313.
[4] 范帅华, 郭伟, 郭军. 基于机器学习的决策树算法在血流感染预后预测中应用现状及展望[J]. 中华实验和临床感染病杂志(电子版), 2023, 17(05): 289-293.
[5] 戴雨霖, 张新春. 人工智能在口腔修复诊疗中的应用与进展[J]. 中华口腔医学研究杂志(电子版), 2024, 18(01): 65-69.
[6] 叶长缨, 谢静, 丁桂聪. 乳牙龋病的过渡性治疗研究进展[J]. 中华口腔医学研究杂志(电子版), 2023, 17(05): 365-370.
[7] 李晓阳, 刘柏隆, 周祥福. 大数据及人工智能对女性盆底功能障碍性疾病的诊断及风险预测[J]. 中华腔镜泌尿外科杂志(电子版), 2023, 17(06): 549-552.
[8] 邢晓伟, 刘雨辰, 赵冰, 王明刚. 基于术前腹部CT的卷积神经网络对腹壁切口疝术后复发预测价值[J]. 中华疝和腹壁外科杂志(电子版), 2023, 17(06): 677-681.
[9] 韩冰, 顾劲扬. 深度学习神经网络在肝癌诊疗中的研究及应用前景[J]. 中华肝脏外科手术学电子杂志, 2023, 12(05): 480-485.
[10] 雷漫诗, 邓锶锶, 汪昕蓉, 黄锦彬, 向青, 熊安妮, 孟占鳌. 人工智能辅助压缩感知技术在上腹部T2WI压脂序列中的应用[J]. 中华肝脏外科手术学电子杂志, 2023, 12(05): 551-556.
[11] 陈意志. 老年人肾小球滤过率估算公式的选择及预后意义:基于循证医学的证据解读[J]. 中华肾病研究电子杂志, 2023, 12(06): 360-360.
[12] 刘谨硕, 朱思泉. 前房角检查的临床应用进展[J]. 中华眼科医学杂志(电子版), 2023, 13(05): 316-320.
[13] 王守森, 方翌, 冯添顺, 魏梁锋. 深度学习模型在侵袭性垂体腺瘤诊疗中的应用[J]. 中华脑科疾病与康复杂志(电子版), 2023, 13(06): 382-384.
[14] 林绍涌, 朱先理, 魏梁锋, 王守森. ChatGPT在临床医学实践中的应用进展[J]. 中华临床医师杂志(电子版), 2023, 17(10): 1113-1116.
[15] 胡平, 鄢腾峰, 周海柱, 祝新根. 人工智能在非增强CT图像中颅内出血早期检出和血肿分割的研究进展[J]. 中华脑血管病杂志(电子版), 2023, 17(04): 410-416.
阅读次数
全文


摘要