上市后药品严重罕见不良反应快速循证评价的关键方法研究

Drug safety administration is the key component of national drug policy. The serious and rare adverse drug reaction is a ticklish and serious challenge for the management of post-marketing drugs. The key method for effectively assessing the serious and rare post-marketing drug adverse reactions is to produce research evidence with sufficient statistical power and achievable attribution by rigorously and comprehensively collecting and synthesizing all relevant research data (i.e., evidence-based assessment). This method has been adopted by drug administration departments (including the US FDA). However, the existing routine evidence-based assessment methods cannot meet the demand for rapid and scientific decision-making regarding drug safety administration...Focusing on this key question, this project aims to develop key methods for rapid evidence-based assessment of serious and rare post-marketing drug adverse reactions, so as to establish a high quality rapid assessment mechanism. The details of this project are in four-folds: ① Compare the impact of different models and methods on the time of evidence-based assessment through methodological empirical analysis so as to establish scientific and rapid assessment methods; ② Optimize the method for synthesizing evidence of rare events from clinical trials through statistical model development and statistical simulation; ③ Establish scientific methods of risk of bias assessment, statistical synthesis, quality grading and recommendation for evidence from real world studies through statistical simulation, expert consensus, etc.; ④ Taking the rapid assessment of serious and rare adverse drug reactions of the novel antidiabetic drugs for example, carry out an empirical research to optimize these proposed methods, and then finally establish these key methods through expert consensus. If successful, it could improve the efficiency of serious and rare post-marketing drug adverse reactions assessment.

药品安全是上市后药品监管的关键。药品严重罕见不良反应是上市后药品安全监管的难点和痛点。有效评价上市后药品严重罕见不良反应的关键方法是通过严格全面地收集和合成所有相关研究数据，形成统计效能足够、可实现归因的研究证据（即循证评价）。该方法已被包括美国FDA在内的药监部门采用。但现有常规循证评价方法无法满足药品安全监管对快速、科学决策的需求。本研究拟构建上市后药品严重罕见不良反应快速循证评价的关键方法，解决该重要科学问题。研究内容包括：①通过方法学实证分析，比较和遴选不同模式和手段对循证评价时间的影响；②通过统计模型开发和数值模拟，优化基于临床试验的罕见事件统计合成方法；③通过数值模拟、专家共识等，建立针对真实世界证据的偏倚风险评价、统计合成、质量分级与推荐方法；④以新型降糖药的严重罕见不良反应快速评价为例优化拟建立的方法，最后通过专家共识确立。若获成功，将提高上市药品严重罕见不良反应评价效率。

药品安全是上市后药品监管的关键。药品严重罕见不良反应是上市后药品安全监管的难点和痛点。循证评价已被包括美国FDA在内的药监部门采用，但现有常规循证评价方法无法满足药品安全监管对快速、科学决策的需求。. 本课题针对药品监管的快速评价和科学决策双重需求，建立了上市后药品严重罕见不良反应快速循证评价的关键方法和技术推荐。具体内容包括：（1）系统整理了现有可用于系统评价初筛、全文筛选和数据提取过程的自然语言处理工具，构建了基于人工智能技术的快速文献筛选系统和平台，可用于海量文献的快速筛选；（2）采用统计学模拟比较了连续性校正、Peto法、Mantel-Haenszel法、广义线性混合效应模型（GLMM）和贝叶斯logistic回归模型5种罕见事件统计合成方法的优劣，为合理选择罕见事件统计合成方法提供依据；（3）优化了基于真实世界数据的队列研究、病例对照研究偏倚风险评价工具，形成了基于真实世界证据的统计合成方法选择路径；（4）以SGLT2抑制剂治疗2型糖尿病的酮症酸中毒风险为例开展实证验证研究，最终从研究设计和统计分析两个维度建立了14条上市后药品严重罕见不良反应快速循证评价技术推荐。. 研究结果共发表学术论文8篇，其中SCI论文4篇（含学科顶级杂志3篇）、中文核心期刊论文4篇；申请发明专利2项。协助培养博士生2名。建立的技术推荐被临流顶刊配发评论并呼吁研究者详细报告安全性信息，实证研究结果被美国医师协会期刊《内科学年鉴》作为摘要发布并配发专门评论，还被用于BMJ糖尿病快速指南的制订。形成的关键方法和技术推荐，可为今后基于临床试验和真实世界研究证据快速、科学评价上市药品的严重不良反应提供关键方法支持；解决了现有罕见不良事件证据合成限速和合成模型选择面临方法与技术瓶颈；提升了上市后药品严重不良反应评价的效率。