Objective  To explore the application status and development trend of machine learning in the field of pharmacovigilance worldwide, and to provide reference for the research on the application of machine learning in the field of pharmacovigilance.

Methods Relevant literature was searched in the Web of Science with the key words of "machine learning" and "pharmacovigilance" from the inception to March 1, 2023. R language and other software were used to quantitatively analyze the literature data in this field. The clustering, co-occurrence and emergence visual analysis were carried out on the characteristics of annual published papers, institutions, countries, keywords and other aspects.

Results  A total of 904 literature were included. The number of literature published showed a fluctuating upward trend since 1994. There was cross-regional, cross-regional and cross-agency cooperation among the cooperative network institutions. The top 5 countries in the number of publications were the United States, China, Japan, South Korea and India, China and the United States had relatively close cooperation in this field. Signal detection, social media and electronic health records were high-frequency keywords in this field. Clustering and association rule analysis showed that this field focused on three aspects signal recognition, unstructured text mining and analysis, and processing and analysis of electronic medical information. At present, machine learning has made significant progress in signal recognition, social media information mining, and unstructured text processing of electronic medical information, which broaden the data sources of pharmacovigilance, improve the real-time monitoring ability of adverse drug reactions, bringing innovation impetus to the field of pharmacovigilance.

Conclusion  The rapid development of big data and artificial intelligence technologies has led to an increasing integration of machine learning into the field of pharmacovigilance, which promotes technical exchanges and cooperation and cross-disciplinary integration. It is necessary to optimize each machine learning algorithm to improve its accuracy and stability in pharmacovigilance, strengthen the protection measures of data privacy and security to ensure the safety of patient information. Integrating expertise in the fields of science, medicine, and data statistics with a view to promoting technological progress in the field of pharmacovigilance.

1.杨悦. 我国药物警戒制度实施与ICH药物警戒指导原则转化适用[J]. 中国药物警戒, 2020, 17(2): 65-71. [Yang Y. Implementation of pharmacovigilance system in drug administration law and application of ICH E2 guideline[J]. Chinese Journal of Pharmacovigilance, 2020, 17(2): 65-71.] DOI: 10.19803/j.1672-8629.2020.02.01.

2.Hauben M, Hartford CG. Artificial intelligence in pharmacovigilance: scoping points to consider[J]. Clin Ther, 2021, 43(2): 372-379. DOI: 10.1016/j.clinthera. 2020.12.014.

3.Kompa B, Hakim JB, Palepu A, et al. Artificial intelligence based on machine learning in pharmacovigilance: a scoping review[J]. Drug Saf, 2022, 45(5): 477-491. DOI: 10.1007/s40264-022-01176-1.

4.Bate A, Lindquist M, Edwards IR, et al. A Bayesian neural network method for adverse drug reaction signal generation[J]. Eur J Clin Pharmacol, 1998, 54(4): 315-321. DOI: 10.1007/s002280050466.

5.Nikfarjam A, Sarker A, O'Connor K, et al. Pharmacovigilance from social media: mining adverse drug reaction mentions using sequence labeling with word embedding cluster features[J]. J Am Med Inform Assoc, 2015, 22(3): 671-681. DOI: 10.1093/jamia/ocu041.

6.刘娟娟, 陈奕, 张佳玲, 等. 21世纪全球视域下陇药红芪科学协作与热点前沿知识图谱构建及可视化分析[J]. 中草药, 2023, 54(12): 3932-3948. [Liu  JJ, Chen  Y, Zhang JL, et al. Construction and visual analysis of scientific collaboration and hot frontier knowledge graph of Longyao Hedysari Radix from global perspective in 21st century[J]. Chinese Traditional and Herbal Drugs, 2023, 54(12): 3932-3948.] DOI: 10.7501/j.issn.0253-2670. 2023.12.020.

7.Leal MM, Sanz MM, Ferrando JRC, et al. A comparative analysis of the pharmacovigilance systems of brazil, spain, the european union and the united states based on the information provided by their regulatory agency websites[J]. Daru, 2019, 27(1): 379-387. DOI: 10.1007/s40199-019-00249-4.

8.杨婷, 张晓朦, 张冰, 等. 国内外痛风性关节炎的研究现状—基于CiteSpace软件的可视化分析[J]. 中国实验方剂学杂志, 2020, 26(20): 169-177. [Yang T, Zhang  XM, Zhang B, et al. Research status of gouty arthritis at home and abroad-visual analysis based on citespace software[J]. Chinese Journal of Experimental Traditional Medical Formulae, 2020, 26(20): 169-177.] DOI: 10.13422/j.cnki.syfjx.20202018.

9.Caster O, Aoki Y, Gattepaille LM, et al. Disproportionality analysis for pharmacovigilance signal detection in small databases or subsets: recommendations for limiting false-positive associations[J]. Drug Saf, 2020, 43(5): 479-487. DOI: 10.1007/s40264-020-00911-w.

10.Cocos A, Fiks AG, Masino AJ. Deep learning for pharmacovigilance: recurrent neural network architectures for labeling adverse drug reactions in Twitter posts[J]. J Am Med Inform Assoc, 2017, 24(4): 813-821. DOI: 10.1093/jamia/ocw180.

11.Bian J, Topaloglu U, Yu F. Towards large-scale twitter mining for drug-related adverse events[J]. SHB12 (2012), 2012, 2012: 25-32. DOI: 10.1145/2389707.2389713.

12.Masino AJ, Forsyth D, Fiks AG. Detecting adverse drug reactions on twitter with convolutional neural networks and word embedding features[J]. J Healthc Inform Res, 2018, 2(1-2): 25-43. DOI: 10.1007/s41666-018-0018-9.

13.Chen X. Online health communities influence people's health behaviors in the context of COVID-19[J]. PLoS One, 2023, 18(4): e0282368. DOI: 10.1371/journal.pone. 0282368.

14.Munkhdalai T, Liu F, Yu H. Clinical relation extraction toward drug safety surveillance using electronic health record narratives: classical learning versus deep learning[J]. JMIR Public Health Surveill, 2018, 4(2): e29. DOI: 10.2196/publichealth.9361.

15.Zhao Y, Yu Y, Wang H, et al. Machine learning in causal inference: application in pharmacovigilance[J]. Drug Saf, 2022, 45(5): 459-476. DOI: 10.1007/s40264-022-01155-6.

16.Zhao J, Henriksson A, Bostrm H. Cascading adverse drug event detection in electronic health records[C]. IEEE International Conference on Data Science & Advanced Analytics, 2015. DOI: 10.1109/DSAA.2015.7344869.

17.Henriksson A. Ensembles of semantic spaces: on combining models of distributional semantics with applications in healthcare[EB/OL]. (2015) [2023-07-11]. https://xueshu.baidu.com/usercenter/paper/show?paperid=582cbcb8f6e8a12212253cae16827812&site=xueshu_se&hitarticle=1

18.Daluwatte C, Schotland P, Strauss DG, et al. Predicting potential adverse events using safety data from marketed drugs[J]. BMC Bioinformatics, 2020, 21(1): 163. DOI: 10.1186/s12859-020-3509-7.

19.Schotland P, Racz R, Jackson DB, et al. Target adverse event profiles for predictive safety in the postmarket setting[J]. Clin Pharmacol Ther, 2021, 109(5): 1232-1243. DOI: 10.1002/cpt.2074.

20.Edrees H, Song W, Syrowatka A, et al. Intelligent telehealth in pharmacovigilance: a future perspective[J]. Drug Saf, 2022, 45(5): 449-458. DOI: 10.1007/s40264-022-01172-5.

21.周虎子威,张云静,于玥琳,等. 机器学习方法在预测麻精药品不合理使用风险中的应用现状和思考[J]. 药物流行病学杂志, 2023, 32(4): 446-457. [Zhou HZW, Zhang YJ, Yu YL, et al. Application of machine learning methods in predicting the risk of irrational use of narcotic and psychotropic drugs: current status and considerations[J]. Chinese Journal of Pharmacoepidemiology, 2023, 32(4): 446-457.] DOI: 10.19960/j.issn.1005-0698.202304010.

22.Vilar S, Friedman C, Hripcsak G. Detection of drug-drug interactions through data mining studies using clinical sources, scientific literature and social media[J]. Brief Bioinform, 2018, 19(5): 863-877. DOI: 10.1093/bib/bbx010.

23.Meldau EL, Bista S, Rofors E, et al. Automated drug coding using artificial intelligence: an evaluation of WHO drug koda on adverse event reports[J]. Drug Saf, 2022, 45(5): 549-561. DOI: 10.1007/s40264-022-01162-7.

24.Lian AT, Du J, Tang L. Using a machine learning approach to monitor COVID-19 vaccine adverse events (VAE) from Twitter data[J]. Vaccines (Basel), 2022, 10(1): 103. DOI: 10.3390/vaccines10010103.

25.李海龙,赵厚宇,周一帆,等. 基于电子医疗数据库的药品不良反应信号挖掘方法概述[J]. 药物流行病学杂志, 2018, 27(8): 541-549. [Li HL, Zhao HY, Zhou  YF, et al. Data mining methods for adverse drug reaction signals detection in healthcare databases: a literature review[J]. Chinese Jourmal of Pharmacoepidemiology, 2018, 27(8): 541-549.] DOI: 10.19960/j.cnki.issn1005-0698.2018.08.012.

26.Xu Q, Ahmadi E, Amini A, et al. Identifying and mitigating potential biases in predicting drug approvals[J]. Drug Saf, 2022, 45(5): 521-533. DOI: 10.1007/s40264-022-01160-9.

27.Scaboro S, Portelli B, Chersoni E, et al. Increasing adverse drug events extraction robustness on social media: case study on negation and speculation[J]. Exp Biol Med (Maywood), 2022, 247(22): 2003-2014. DOI: 10.1177/ 15353702221128577.

28.Bate A, Luo Y. Artificial intelligence and machine learning for safe medicines[J]. Drug Saf, 2022, 45(5): 403-405. DOI: 10.1007/s40264-022-01177-0.

29.Rocca E, Copeland S, Ralph Edwards I. Pharmacovigilance as scientific discovery: an argument for trans-disciplinarity[J]. Drug Saf, 2019, 42(10): 1115-1124. DOI: 10.1007/s40264-019-00826-1.