Improving the Clinical Research Informatics and Clinical Bioinformatics Infrastructure at the NIH Clinical Center

本摘要按项目的不同方面进行了组织。 改善对数据的访问： = 获取研究数据需要在科学要求和隐私法规之间取得平衡。我们探索了访问存储在电子健康记录中的已故患者数据的监管规则。我们在一份政策文件中提出了目前的法规如何允许使用已故患者数据，这可能是相对未得到充分利用的类别。我们指出了一个问题，即当死亡记录包含已死亡的主要患者以外的个人的私人信息时(参考文献11)。我们首先将我们的发现作为会议提交(参考文献1)，然后作为论文(参考文献2)。本项目基于生物医学翻译研究信息系统(BTRIS)的数据。 数据组织： = 我们探索了组织和维护大型集成数据存储库(IDR)的最佳方法。我们已经制定了需求列表，并评估了现有架构如何满足这些需求(参考文献4) 临床研究企业： = 在NIH IRP中，我们探索了存储在注册中心的临床试验如何与PubMed上发布的数据相联系。基于这些发现，我们探索了2004-2009年的整个临床试验注册，以及与PubMed上所有注册试验和所有文章的链接(参考文献5)。基于这些发现，我们还发表了这种联系缺失的地方。根据ICMJE的政策，结果文章与注册临床试验的正确关联取决于手稿作者正确引用相关临床试验。我们发表了对这一政策遵守情况的评估(参考文献6)。 临床生物信息学： = BTRIS包含完整的外显子组测序数据，可用于生成药物剂量要求。我们探索了如何对整个外显子组数据执行CPIC的剂量指南(参考文献7)。除实施外，我们还参与改进未来剂量指南的努力(参考文献8)。 我们还分析了支持基因组和个性化医学的当前数据库和知识库(参考文献10) 知识表示： = 作为在数据库和本体系统中表示知识的一般努力的一部分，我们参与了为药物警戒产生大规模知识库的努力(参考文献9)。

This summary is organized by the different aspects of the project. Improve access to data: ======================= Access to research data needs to balance scientific requirements with privacy regulations. We explored regulatory rules for accessing data of deceased patients stored in Electronic Health Record. We have proposed in a policy paper how current regulations allow use of deceased patient data and that it may be relatively underused category. We pointed to a problem when a deceased record contains private information on individuals other than the primary patient that is deceased (ref 11). We presented our findings first as a conference submission (ref 1) and later as a paper (ref 2). This project is based on using data from Biomedical Translational Research Information System (BTRIS) (ref 3) Data organization: ======================= We explored the best way to organize and maintain large Integrated Data Repositories (IDRs). We have formulated a list of requirements and evaluated how they are fulfilled by existing architectures (ref 4) Clinical research enterprise: ======================= Within NIH IRP, we explored how clinical trial stored in registries can be linked with published data in PubMed. Based on these findings, we have explored the whole clinical trial registry for years 2004-2009 and linkage to PubMed for all registered trials and all articles in PubMed (ref 5). Based on these findings, we have also published where this linkage is missing. Correct association of a result article with a registered clinical trial is dependent on manuscript writers correctly referencing the relevant clinical trial per policy of ICMJE. We published an evaluation of adherence to this policy (ref 6). Clinical Bioinformatics: ======================= BTRIS contains whole exome sequencing data and this can be used to generate drug dosage requirements. We explored how dosing guidelines from CPIC can be executed against whole exome data (ref 7). In addition to implementation, we participate on efforts to improve future dosing guidelines (ref 8). We have also analyzed the current databases and knowledge bases that support genomic and personalized medicine (ref 10) Knowledge representation: ======================= As part of a general effort to represent knowledge in database and ontology systems, we participate on efforts to produce large scale knowledge bases for pharmacovigilance (ref 9).