利用国际非专利名称词干加强药物警戒信号检测
本文提出利用WHO自1950年代建立的国际非专利名称(INN)系统中的词干来增强药物警戒信号检测。通过分析历史案例和当前药物警戒实践,作者讨论了基于词干的分类如何有助于理解每个词干的不良反应特征,作为早期识别偏离预期类效应的药物不良反应的基准。作者提出了将基于词干的分析整合到现有药物警戒数据库的潜在框架,并辅以机器学习等人工智能方法。
要点
- INN词干标记药物间的药理关系,可用于建立类效应基准。
- 基于词干的分类有助于早期识别新上市药物或老药新用的非预期不良反应信号。
- 提出的框架可将词干分析整合入药物警戒数据库并辅以机器学习。
核心结论
- 基于INN词干的上市后监测可提高信号检测效率,通过更早识别非预期不良反应促进患者安全。
研究设计与样本
- 本文为方法学论述,通过分析历史案例和当前药物警戒实践提出利用INN词干增强信号检测的框架。
主要发现
- 基于词干的分类可促进理解每个词干的不良反应特征,作为早期识别偏离预期类效应的药物不良反应的基准。
临床怎么用
- 监管机构和医疗专业人员可在上市后监测中采用基于词干的分析,辅以机器学习等人工智能方法。
局限性
- 存在词干变异性和报告偏倚等局限性。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
原文
'Stems', which mark pharmacological relationships between substances, form the backbone of the International Nonproprietary Name (INN) system, developed by the WHO in the 1950s. In this paper, we propose using the INN stems to enhance pharmacovigilance signal detection. After analysis of historical cases and current pharmacovigilance practices, we discuss how stem-based classification could facilitate understanding of the adverse-effects profile of each stem, to be used as a benchmark for early identification of adverse drug reactions that deviate from expected class effects, in other words signals associated with newly marketed medicines or different uses of well-known medicines. We propose a potential framework for integrating stem-based analysis into existing pharmacovigilance databases, supplemented by artificial intelligence approaches, such as machine learning. While acknowledging limitations, such as stem variability and reporting bias, we suggest that this approach offers potential advantages for regulatory authorities and healthcare professionals in post-marketing surveillance. Implementation of stem-based post-marketing surveillance could enhance signal-detection efficiency and contribute to improved patient safety through earlier identification of unexpected adverse effects and adverse reactions.
引用
- Balocco R, Aronson JK, Malan SF, et al. Strengthening Signal Detection in Pharmacovigilance by Using International Nonproprietary Name (INN) Stems. Drug Saf. 2025. PMID: 41138033. doi:10.1007/s40264-025-01620-y
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。