An Informatics Framework for Discovery and Ascertainment of Drug-Supplement Interactions

项目摘要 大多数美国成年人（68%）服用膳食补充剂，越来越多的证据表明， 近年来，越来越多的证据支持DS在ADRD中的作用， 预防认知障碍，但证据有限，样本量较小。真实的- 世界数据（RWD），特别是EHR包含患者的详细治疗和反应信息， 可用于检测DS、DSI的使用和效果，更能转化为临床结局 (e.g., MCI到ADRD转换）。据我们所知，没有关于DS使用的调查， 使用EHR数据对MCI和ADRD患者的安全性进行评估。我们目前的家长奖是集中在 开发翻译信息学框架，以发现药物-补充剂相互作用 （DSI）通过链接来自生物医学文献的科学证据。为了响应NOT-AG-20-008， 行政补充申请将在多个方面补充我们的母奖：（1）发展 在EHR中挖掘RWD的新的和先进的数据分析方法，（2）识别DS使用信息 在ADRD患者中，（3）检测DS在现有ADRD患者中的安全性和有效性， EHR数据。在我们的初步工作中，我们研究了识别EHR上DS术语的方法， 开发了自然语言处理（NLP）方法来识别DS的使用状态。我们将继续努力 收集AD患者的DS使用和AE-DS信号的EHR数据集，并开发创新的 信息学方法来提取这些信息。我们的具体目标是：（1）确定DS的使用， 现有EHR数据中MCI和ADRD患者;（2）检测DS安全性信号并探索 DS使用对从MCI到ADRD从现有EHR数据转换的影响。圆满完成 本项目的研究将刺激我们在更大范围内进一步研究DS在ADRD患者中的作用。 规模涉及其他医疗机构的EHR数据。

PROJECT SUMMARY Most U.S. adults (68%) take dietary supplements and there is increasing evidence of drug-supplement interactions (DSIs); In recent years, there has been increasing evidence supporting the role of DSs in ADRD in preventing cognitive impairment but there is limited evidence and the sample sizes have been small. Real- world data (RWD) especially the EHR contain detailed treatment and response information from patients and could be used to detect the usage and effect of DSs, DSIs, which is more translational to clinical outcomes (e.g., MCI to ADRD conversion). To the best of our knowledge, there is no investigation on DSs usage and safety among patients in MCI and ADRD using EHR data. Our current parent award is focusing on the development of a translational informatics framework to enable the discovery of drug-supplement interactions (DSIs) by linking scientific evidence from the biomedical literature. To response to NOT-AG-20-008, this administrative supplement application will complement our parent award in multiple aspects: (1) developing novel and advanced data analytic methods for mining RWD in EHR, (2) identifying DSs usage information among patients with ADRD, and (3) detecting safety and effect of DS among patients with ADRD from existing EHR data. In our preliminary work, we have investigated the methods to identify DSs terms on EHR and developed natural language processing (NLP) methods to identify use status of DSs. We will further our efforts to collect a EHR dataset with DSs usage and AE-DSs signals from AD patients and develop innovative informatics methods to extract such information. Our specific aims are: (1) identifying DSs usage among patients with MCI and ADRD from existing EHR data; and (2) detecting the DSs safety signals and exploring the effect of DSs use on the conversion from MCI to ADRD from existing EHR data. The successful completion of this project will stimulate our further investigation on the role of DS use in patients with ADRD in a larger scale involving EHR data from other healthcare institutions.