Developing i2b2 into a Health Innovation Platform for Clinical Decision Support in the Genomics Era

项目摘要/摘要 精准医学中使用的许多结果预测指标都是基于 在重要方面相似的既往患者的群体。然而，标准 电子健康记录系统(EHR)很少支持实时人口查询和 因此无法将精准医疗作为其综合决策支持的一部分 临床医生。此外，许多精确医学算法所基于的数据并不是 可在标准EMR中使用，因为它来自高级基因组学、成像分析和 属于标准电子病历平台数据域之外的其他数据形态。我们 建议发展生物学与床边一体化的信息学(I2b2)，一口井-- 成熟、开源、集成的大数据分析平台，目前在 140家医院和医疗中心，研究表型/基因型比较并将其纳入 使用称为可替代医疗应用程序的小型互联应用程序和 可重用技术(SMART)。我们将采用合作伙伴医疗基因组学平台， GeneInsight，并使用i2b2将其数据集成到我们的Epic EHR工作流中。然后我们将测试 遗传性心脏病的特定决策支持算法。由此产生的软件将 开源，并允许集成基于基因组学的大数据决策支持算法 广泛地扩展到EHR。我们还将使用这些相同的数据为实验室提供决策支持 根据临床效果对变体进行分类的专业人员。标准的方法 表示这些数据将用于使算法具有可移植性和通用性 适用。创建的决策支持应用程序将不仅针对其 对临床护理的潜在影响，还包括他们对不同 医疗保健环境。

Project Summary/Abstract Many outcome predictors used in precision medicine are based upon experience with populations of previous patients who are similar in important ways. However, standard Electronic Health Record systems (EHRs) seldom support real-time population queries and therefore cannot administer precision medicine as part of their integrated decision support for clinicians. Furthermore, the data upon which many precision medicine algorithms operate is not available in standard EMRs because it comes from advanced genomics, imaging analytics, and other data modalities which fall outside of the data domains of standard EHR platforms. We propose developing Informatics for Integrating Biology and the Bedside (i2b2), a well- established, open source, integrated, big data analytic platform that is currently used at over 140 hospitals and medical centers, to study phenotype/genotype comparisons and incorporate it into the EHR using small, connected applications named Substitutable Medical Applications and Reusable Technologies (SMART). We will take the Partners HealthCare genomics platform, GeneInsight, and integrate its data into our Epic EHR workflow using i2b2. We will then test specific decision support algorithms for inherited cardiac diseases. The resulting software will be open source and allow integration of genomics-based, big data decision support algorithms broadly into EHRs. We will also use these same data to provide decision support to laboratory professionals who classify variants relative to their clinical effects. Standard methods of representing these data will be used to make the algorithms transportable and universally applicable. The Decision Support Apps that are created will be evaluated not only for their potential impact upon clinical care, but also for their durability and adaptability to different healthcare environments.