EAGER: Collaborative Research: Computational Public Drug Surveillance

药物不良反应（ADR）（不期望的或过度的药物反应）与显著的发病率和死亡率有关，占所有入院人数的5%。药物-药物相互作用（DDI）是一种涉及两种或两种以上药物的不良反应。报告显示，1999年至2003年，美国食品和药物管理局（FDA）撤回的药物中有50%与严重的ddi有关。当药物被FDA批准时，药物的不良反应概况很少是完整的。因此，在一般人群（在种族、性别、年龄、生活方式上存在显著差异）使用某种药物后，通常会发现以前未发现的ddi。更复杂的是，某些患者群体，如精神病患者，经常同时使用多种药物治疗。多种药物之间的潜在相互作用既没有得到很好的理解，也没有完全表征。自愿报告是FDA用来监督新药的基本机制，它存在报告少报、报告延迟、报告质量参差不齐，甚至缺乏罕见的ddi报告的问题。在此背景下，本合作项目旨在探索一种新的计算方法来解决药物-药物相互作用监测问题的可行性。它寻求开发新的方法，通过从在线资源和数字社交媒体收集的数据来预测药物之间分子水平的相互作用。该项目旨在验证这样一种假设，即这种在线数据与传统药物相关数据库的数据相结合，可以比使用现有方法更快地可靠地预测潜在的ddi。通过对未来报告的预测进行验证来评估该方法的有效性。如果取得成功，该项目将为药物相互作用监测带来有效、主动的计算方法，使联邦、地方和公共卫生机构、制药公司、临床从业人员、患者和广大公众受益。早期发现不良的ddi可改善患者护理，并显著减少医疗费用和涉及ddi的诉讼。该项目为具有计算科学和健康科学专业知识的研究人员之间的合作提供了更多的机会。它还为西弗吉尼亚大学和弗吉尼亚大学的学生提供基于研究的培训机会。研究结果将自由地传播给更广泛的学术和研究界。

Adverse drug reactions (ADR) (undesired or excessive responses drugs) have been linked with significant morbidity and mortality, and account for as much as 5% of all admissions. A drug-drug interaction (DDI) is a type of ADR involving two or more drugs. Reports suggest that 50% percent of the drugs withdrawn in the U.S. by the Food and Drug Administration (FDA) from 1999 to 2003 were linked with significant DDIs. The ADR profile of a given drug is rarely complete at the time the drug is approved by FDA. Hence, after a drug has been in use by the general population (with significant diversity in race, gender, age, lifestyle), often previously unidentified DDIs are discovered. To complicate matters, certain populations of patients, e.g., psychiatric patients, are often concurrently treated with multiple medications. The potential interactions between multiple drugs are neither well understood nor completely characterized. Voluntary reporting, the basic mechanism used by the FDA to monitor new drugs, suffers from underreporting, delayed reporting, uneven quality of reports, and even lack of reports of rare DDIs.Against this background, this collaborative project aims to explore the feasibility of a novel computational approach to the problem of drug-drug interaction surveillance. It seeks to develop new methods for predicting molecular level interactions between drugs from data gleaned from online sources and digital social media. The project aims to test the hypothesis that such online data, in combination with with data from traditional drug related databases can be used to reliably predict potential DDIs much sooner than possible using current methods. The effectiveness of the approach is assessed through verification of predictions against future reports. If successful, the project could lead to effective, proactive computational approaches to drug interaction surveillance, with benefits to federal, local and public health agencies, drug companies, clinical practitioners, the patients, and the public at large. Early detection of adverse DDIs could lead to improved patient care, and significant reduction in healthcare costs and lawsuits involving DDIs. The project offers enhanced opportunities for collaboration among investigators with expertise in computational and health sciences. It also offers research-based training opportunities to students at West Virgina University and the University of Virginia. Results of the research will be freely disseminated to the broader academic and research community.