美国FDA不良事件监测系统底层架构特征及风险治理研究

doi:10.19803/j.1672-8629.20260388

中国药物警戒 ›› 2026, Vol. 23 ›› Issue (9): 961-967.
DOI: 10.19803/j.1672-8629.20260388

• 数智化药物警戒与创新生物制品上市后监测专栏 • 上一篇下一篇

美国FDA不良事件监测系统底层架构特征及风险治理研究

金凯莉¹, 陈曦¹, 熊佳敏¹, 叶小飞^2#, 郭晓晶^2,*

¹海军军医大学基础医学院,上海 200433;
²海军军医大学卫生勤务学系军队卫生统计学教研室,上海 200433

收稿日期:2026-05-18 出版日期:2026-09-15 发布日期:2026-09-15
通讯作者: ^*郭晓晶,女,博士,副教授,药物流行病学与药物警戒研究。E-mail: guoxiaojing1003@163.com。^#为共同通信作者。
作者简介:金凯莉,女,在读本科,药物流行病学与药物警戒研究。
基金资助:
国家自然科学基金资助项目（82073671、81703296）; “癌症、心脑血管、呼吸和代谢性疾病防治研究”国家科技重大专项（2025ZD0548100）; 海军军医大学校级面上孵化基金（2024MS004）

Underlying architecture characteristics of the U.S. FDA AEMS and risk management

Jin Kaili¹, Chen Xi¹, Xiong Jiamin¹, Ye Xiaofei^2#, Guo Xiaojing^2,*

¹College of Basic Medical Sciences, Naval Medical University, Shanghai 200433, China;
²Department of Army Medical Statistics, Faculty of Medical Services, Naval Medical University, Shanghai 200433, China

Received:2026-05-18 Online:2026-09-15 Published:2026-09-15

摘要/Abstract

摘要： 目的探讨美国食品药品监督管理局（FDA）新一代不良事件监测系统（AEMS）底层架构,分析其在数据治理与风险管控方面的经验,为我国“人工智能+药品监管”格局构建及提升上市后风险感知能力提供参考。方法系统梳理FDA从分散独立数据库向AEMS全品类统一平台的演进,评价其在跨模态数据集成、本体映射及人工智能算法应用方面的优劣势。结果 AEMS借助云原生架构消除信息孤岛,依托应用程序接口（API）实现数据动态更新,利用自然语言处理（NLP）模型优化预处理流程,提升信号监测效能、缩短预警周期;但其高透明度公开策略易引发透明化悖论。结论建议我国推进“两品一械”统一警戒数据库建设,强化跨学科协作与人机协同分级审核,引导我国企业构建前瞻性主动风险管理体系。

关键词: 美国食品药品监督管理局, 美国不良事件监测系统, 药物警戒, 风险, 底层架构

Abstract: Objective To explore the underlying architecture of the U.S. FDA's new unified Adverse Event Management System (AEMS) and to analyze its approaches to data governance and risk management so as to offer insights for China's AI+drug regulation framework and improve post-marketing risk perception capabilities. Methods We traced the FDA's evolution from a collection of fragmented, standalone databases to the integrated AEMS platform and identified both the strengths and weaknesses of its design in terms of cross-modal data integration, ontology mapping, and the use of AI algorithms. Results Our analysis found that the cloud-native AEMS eliminated information silos, supported dynamic data updates through APIs, and used natural language processing (NLP) models to streamline data preprocessing. These capabilities could improve signal detection and shorten early-warning cycles. However, the system's high level of transparency could give rise to a transparency paradox, which might potentially trigger unnecessary public alarm resulting from exposing industry to the dual pressures of compliance burdens and information overload. Conclusion It is recommended that China develop a unified pharma-covigilance database covering drugs, cosmetics, and medical devices while strengthening interdisciplinary collaboration and human-machine collaborative tiered review to encourage domestic enterprises to build forward-looking, proactive risk management systems.

Key words: FDA, AEMS, Pharmacovigilance, Risk, Under-lying Architecture

中图分类号:

R951
R994.11

金凯莉, 陈曦, 熊佳敏, 叶小飞, 郭晓晶. 美国FDA不良事件监测系统底层架构特征及风险治理研究[J]. 中国药物警戒, 2026, 23(9): 961-967.

Jin Kaili, Chen Xi, Xiong Jiamin, Ye Xiaofei, Guo Xiaojing. Underlying architecture characteristics of the U.S. FDA AEMS and risk management[J]. Chinese Journal of Pharmacovigilance, 2026, 23(9): 961-967.

导出引用管理器 EndNote|Ris|BibTeX

链接本文: https://zgywjj.magtechjournal.com/CN/10.19803/j.1672-8629.20260388

https://zgywjj.magtechjournal.com/CN/Y2026/V23/I9/961

参考文献

[1] Davis R, Dang O, De S, et al.Characterizing the FDA Adverse Event Reporting System (FAERS) as a network to improve pattern discovery[J]. Drug Saf, 2026, 49(2): 239-251.
[2] FDA. FDA Adverse Event Monitoring System (AEMS) public dashboard [EB/OL]. (2026-03-11)[2026-04-01]. https://www.fda.gov/drugs/fda-adverse-event-monitoring-system-aems/fda-adverse-event-monitoring-system-aems-public-dashboard.
[3] Jamieson K H, Johnson K B, Cappola A R.Misinformation and the Vaccine Adverse Event Reporting System[J]. JAMA, 2024, 331(12): 1005-1006.
[4] Lakhan S E.Postmarket safety surveillance of FDA-cleared prescription digital therapeutics using the manufacturer and user facility device experience (MAUDE) database: a pharmacovigilance study[J]. Cureus, 2025, 17(6): e85343.
[5] Wang S, Cheng Y, Wang X, et al. Mining and analysis of amphotericin B adverse reaction signals: a real-world study based on the FAERS database[J/OL]. Expert opinion on drug safety, 2025-03-12. https://pubmed.ncbi.nlm.nih.gov/40070343/.
[6] He L, Li J, Cheng X, et al.Association between GLP-1 RAs and DPP-4 inhibitors with biliary disorders: pharmacovigilance analysis[J]. Frontiers in Pharmacology, 2025, 16: 1509561.
[7] Ding C, Ma L, Liang Y, et al.Gastrointestinal adverse events associated with Lenvatinib versus Lenvatinib plus Pembrolizumab: a pharmacovigilance study in FDA adverse event reporting system[J]. Scientific Reports, 2025, 15(1): 15047.
[8] Bosma J S, Dercksen K, Builtjes L, et al.The DRAGON benchmark for clinical NLP[J]. NPJ Digital Medicine, 2025, 8(1): 289.
[9] Gucciardo D, Kimbrell M, Seeley A. Adverse events centralized: what the launch of FDA's new adverse event monitoring system means for companies[EB/OL]. (2026-03-17)[2026-06-17]. https://insider.thefdagroup.com/p/fda-adverse-event-monitoring-system-aems.
[10] FDA. Artificial intelligence for drug development. FDA center for drug evaluation and research[EB/OL]. (2026-01-14)[2026-04-07]. https://www.fda.gov/about-fda/center-drug-evaluation-and-research-cder/artificial-intelligence-drug-development.
[11] EMA. International council on harmonisation of technical requirements for registration of pharmaceuticals for human use (ICH)[EB/OL]. (2026-04-26)[2026-04-09]. https://www.ema.europa.eu/en/partners-networks/international-activities/multilateral-coalitions-initiatives/international-council-harmonisation-technical-requirements-registration-pharmaceuticals-human-use-ich.
[12] FDA. FDA data modernization action plan (DMAP)[EB/OL]. (2021-03-03)[2026-04-10]. https://www.fda.gov/about-fda/reports/data-modernization-action-plan.
[13] Wang X, Xu X, Tong W, et al.InferBERT: a transformer-based causal inference framework for enhancing pharmacovigilance[J]. Front Artif Intell, 2021, 4: 659622.
[14] FDA. FDA launches new adverse event look-up tool[EB/OL]. (2026-03-11)[2026-04-11]. https://www.fda.gov/news-events/press-announcements/fda-launches-new-adverse-event-look-tool.
[15] FDA. Postmarketing safety reporting for combination products[EB/OL]. (2021-07-23)[2020-04-12]. https://www.fda.gov/combination-products/guidance-regulatory-information/postmarketing-safety-reporting-combination-products.
[16] Harpaz R, Dumouchel W, Shah N H, et al.Novel data-mining methodologies for adverse drug event discovery and analysis[J]. Clin Pharmacol Ther, 2012, 91(6): 1010-1021.
[17] Bate A, Evans S J.Quantitative signal detection using spontaneous ADR reporting[J]. Pharmacoepidemiol Drug Saf, 2009, 18(6): 427-436.
[18] Zhou Z, Hultgren K E.Complementing the US food and drug administration adverse event reporting system with adverse drug reaction reporting from social media: comparative analysis[J]. JMIR Public Health Surveill, 2020, 6(3): e19266.
[19] Piening S, Haaijer-Ruskamp F M, De Graeff P A, et al. Healthcare professionals'self-reported experiences and preferences related to direct healthcare professional communications: a survey conducted in the Netherlands[J]. Drug Saf, 2012, 35(11): 1061-1072.
[20] He Y, Sarntivijai S, Lin Y, et al.OAE: the ontology of adverse events[J]. J Biomed Semantics, 2014, 5: 29.
[21] Vinayak Rai K B K M. Cloud Computing in Healthcare Industries: Opportunities and Challenges[M]. Singapore: Springer, 2022.
[22] Atiku S, Okoh A, Olakotan O.The contingent payoff of transparency: the mediating roles of citizen participation and enforcement in health service delivery improvement[J]. J Eval Clin Pract, 2026, 32(2): e70401.
[23] Lakhan S E.Early evaluation of the food and drug administration (FDA) adverse event monitoring system (AEMS): an analysis of over 32 million pharmacovigilance reports[J]. Cureus, 2026, 18(4): e106259.
[24] Zhang X, Sumathipala M, Zitnik M.Population-scale identification of differential adverse events before and during a pandemic[J]. Nat Comput Sci, 2021, 1(10): 666-677.
[25] Bate A, Stegmann J U.Artificial intelligence and pharmacovigilance: What is happening, what could happen and what should happen?[J]. Health Policy and Technology, 2023, 12(2): 100743.
[26] Goddard K, Roudsari A, Wyatt J C.Automation bias: a systematic review of frequency, effect mediators, and mitigators[J]. J Am Med Inform Assoc, 2012, 19(1): 121-127.
[27] FDA. FDA's technology modernization action plan[EB/OL]. (2019-09-17)[2026-04-14]. https://www.fda.gov/about-fda/reports/fdas-technology-modernization-action-plan.
[28] Monnot A D, Fung E S, Compoginis G S, et al.An evaluation of the FDA adverse event reporting system and the potential for reporting bias[J]. J Cosmet Dermatol, 2021, 20(6): 1849-1854.
[29] Bouvy J C, De Bruin M L, Koopmanschap M A. Epidemiology of adverse drug reactions in Europe: a review of recent observational studies[J]. Drug Saf, 2015, 38(5): 437-453.
[30] CDR, NMPA. New AI-empowered standards for pharmacovigilance released,helping advance intelligent drug safety monitoring[EB/OL]. (2026-01-08)[2026-06-13]. https://www.cdr-adr.org.cn/.

美国FDA不良事件监测系统底层架构特征及风险治理研究

Underlying architecture characteristics of the U.S. FDA AEMS and risk management

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

[1]	冯思琪, 倪雨晨, 胡冯译, 刘玥, 叶小飞, 郭晓晶. 细胞治疗产品上市后长期安全性监测的国际经验与中国优化路径[J]. 中国药物警戒, 2026, 23(9): 975-981.
[2]	李静, 廖畅, 左甜甜, 林永强, 黄哲. 高丽红参、西洋参及甘草中重金属与有害元素概率风险评估[J]. 中国药物警戒, 2026, 23(9): 982-988.
[3]	宋雅娜, 赵燕, 刘巍, 邓刚. 医疗器械警戒制度体系文件的核心框架与实践要求[J]. 中国药物警戒, 2026, 23(9): 1023-1027.
[4]	向群, 傅大莉. 药源性认知障碍的致病药物及疾病风险评估研究进展[J]. 中国药物警戒, 2026, 23(9): 1075-1080.
[5]	叶倩, 寇小旋, 牛茜怡, 宋捷, 李晓瑜, 文海若. 伊布莫仑体外致突变性风险研究[J]. 中国药物警戒, 2026, 23(8): 841-847.
[6]	萨日娜, 史彦蕾, 马媛媛, 马燕萍, 潘新波, 梁曦, 葛斌, 禄金晶, 张冰. 全身麻醉药的药物警戒实践探析[J]. 中国药物警戒, 2026, 23(8): 898-903.
[7]	孙倩, 俞颖. 艾托格列净片联合聚乙二醇洛塞那肽注射液致血糖正常的糖尿病酮症酸中毒1例分析[J]. 中国药物警戒, 2026, 23(8): 941-943.
[8]	李聪慧, 高振宇, 张京梅. 药品检查缺陷关键词风险评估研究[J]. 中国药物警戒, 2026, 23(6): 667-671.
[9]	王春婷, 张瑞芳, 邓彬, 邹箴蕾, 王琨. 右旋糖酐铁注射液不良反应探析[J]. 中国药物警戒, 2026, 23(6): 672-676.
[10]	许卓欣, 张冰, 张晓朦, 孟敏, 萨日娜. 基于风险管理闭环模型构建的医疗机构药物警戒实践策略思考[J]. 中国药物警戒, 2026, 23(4): 438-442.
[11]	费毅琴, 肖凌, 吕盼, 汪波, 聂晶. 蕲艾中重金属元素暴露水平及综合风险评估[J]. 中国药物警戒, 2026, 23(3): 280-284.
[12]	龙欢, 王凯鑫, 杨悦琦, 安明惠, 梁雨薇, 刘红冉, 高旭, 张丽, 丁佳欣, 宫建. 辐射防治药物警戒研究[J]. 中国药物警戒, 2026, 23(3): 355-360.
[13]	王春婷, 张瑞芳, 陈亚飞. 盐酸雷尼替丁注射制剂安全性风险分析[J]. 中国药物警戒, 2026, 23(2): 181-184.
[14]	张晓朦, 曲聪聪, 季宇鑫, 林志健, 张冰. 基于中医五行与现代机制整合分析的雷公藤器官毒性特征与防治研究[J]. 中国药物警戒, 2026, 23(1): 7-11.
[15]	季宇鑫, 林彦, 冯鑫煜, 王雨, 张晓朦, 张冰. 基于计算预测与斑马鱼观察雷公藤甲素治疗系统性红斑狼疮的效-毒研究[J]. 中国药物警戒, 2026, 23(1): 12-18.