Data Harmonization

项目摘要 中风是美国第五大最常见的死亡原因，每年折磨近80万人。大约三 四分之一的中风是第一次事件，强调了初级预防的重要性。设计最优 预防策略需要识别风险因素和估计中风风险。最近的 美国心脏协会（AHA）/美国卒中协会一级预防指南 中风得出结论，“一个理想中风风险评估工具是简单的，广泛适用和接受， 并考虑到多种风险因素的影响是不存在的。”其中最常见的 推荐的预测模型是Fragrance Stroke Profile，开发和更新超过25年 年前已经提出了更新的模型（包括来自 REGARDS研究），但尚未得到彻底验证。因此，中风的一级预防 指导方针要求进行更多的研究“以验证跨年龄、性别和种族/民族群体的风险评估工具” 以及“评估是否有任何最近确定的风险因素增加了预测的准确性， 现有规模”。 我们建议通过汇总和协调收集的现有患者水平数据来解决这些差距， NINDS和NHLBI支持的纵向队列研究的一部分。数据将通过一个 与伯明翰亚拉巴马大学REGARDS研究协调中心合作 并通过向NIH dbGap存储库提交的请求，从Fractionary Offspring，ARIC获得数据 和梅萨同伙与此同时，我们将扩展所开发的先进机器学习技术 作为我们目前资助给杜克的NIH BD 2K奖的一部分。我们将把这些模型应用于协调数据 以促进原发性中风预测工具的开发和验证。这些复杂的分析将 需要先进的计算资源，将利用AHA的精准医学平台（PMP），建立 基于Amazon Web Services。

PROJECT SUMMARY Stroke is the fifth most prevalent cause of death in the U.S. afflicting nearly 800,000 per year. About three quarters of strokes are first events, underscoring the importance of primary prevention. Designing optimal preventive strategies requires identification of risk factors and estimation of the risk of stroke. The most recent American Heart Association (AHA)/American Stroke Association Guidelines for the Primary Prevention of Stroke conclude that “an ideal stroke risk assessment tool that is simple, is widely applicable and accepted, and takes into account the effects of multiple risk factors does not exist.” One of the most commonly recommended predictive models is the Framingham Stroke Profile, developed and updated more than 25 years ago. Newer models have been proposed (including the Self-Reported Stroke Risk Stratification tool from the REGARDS study) but have not been thoroughly validated. Consequently, the Primary Prevention of Stroke guidelines call for more research “to validate risk assessment tools across age, sex, and race/ethnic groups” and “to evaluate whether any of the more recently identified risk factors add to the predictive accuracy of existing scales”. We propose to address these gaps by aggregating and harmonizing existing patient-level data collected as part of longitudinal cohort studies supported by the NINDS and NHLBI. The data will be obtained through a partnership with the coordinating center for the REGARDS Study at the University of Alabama at Birmingham and by a request submitted to the NIH dbGap repository to obtain data from the Framingham Offspring, ARIC and MESA cohorts. At the same time, we will expand the advanced machine learning techniques developed as part of our currently funded NIH BD2K award to Duke. We will apply these models to the harmonized data to facilitate development and validation of prediction tool for primary strokes. These complex analyses will require advanced computational resources that will utilize the AHA's Precision Medicine Platform (PMP), built based on Amazon Web Services.