National Cohort Studies of Alzheimer's Disease, Related Dementias and Air Pollution

项目摘要 背景资料。家长基金(R01 AG066793)的具体目标是进行全国流行病学调查 医疗保险和医疗补助的研究声称评估长期暴露在空气污染中的影响 阿尔茨海默病(AD)和相关痴呆(ADRD)住院和疾病进展(目标1)， 应用机器学习方法来确定个人水平、环境和社会因素的共同出现 这导致脆弱性增加(目标2)，并制定统计方法以理清空气的影响 暴露于其他混杂因素的污染，并纠正潜在的结果错误分类(目标 3)。母公司R01依赖于广泛的流行病学数据，从环境暴露到 索赔数据受气象和社会经济因素的影响。虽然我们为 父R01，数据尚未存储到公共数据存储库，我们也没有公开 可用。 总体目标。通过这一补充，我们的目标是实现高维数据的有效传播和重用 大容量数据，包括来自父R01的数据产品。这些目标将通过形成一个 与哈佛数据中心建立新的协作和合作伙伴关系，扩展其当前的存储和共享容量 地理空间公共卫生数据。我们的具体目标是：实现高层次元数据的自动抽取 维度数据集格式NetCDF和HDF5(目标1)，实现与Jupyter Binder的集成 允许从Dataverse(目标2)探索和查看复杂的高维数据，增强重用， R01研究的社区参与度和再现性，以及使用 综合合作医疗索赔数据(目标3)。 冲击力。Dataverse软件平台中的每个新功能，如AIMS 1和2中提出的功能，都是 通常传播到所有77个Dataverse安装，从而增强了它们在全球的影响。DataVerse还 有一个由开源贡献者和数字图书馆组成的充满活力的社区，并组织了年度社区 在哈佛大学持续几天的会议。我们将利用这个平台来推动本副刊的工作 让社区参与进来，并以同样的目标参加其他研讨会和会议。我们将密切合作 通过与Make Data Count的现有数据中心协作，跟踪这些发展的影响 项目，提供使用指标标准化(如浏览量、下载量和引用量 数据)，并使我们能够自我修改和改进我们的数据发布。这个副刊也会有一个直接的 对母公司R01的影响，提高可重用性、社区参与度和可再生性，这将反过来导致 得出更有力的流行病学结论(目标3)。

Project Summary Background. The specific aims of the parent grant (R01 AG066793) are to conduct national epidemiological studies of Medicare and Medicaid claims to estimate the effects of long-term exposures to air pollution on Alzheimer’s disease (AD) and related dementias (ADRD) hospitalization and disease progression (Aim 1), to apply machine learning methods to identify co-occurrence of individual-level, environmental, and societal factors that lead to increase vulnerability (Aim 2), and to develop statistical methods to disentangle the effects of air pollution exposure from other confounding factors and to correct for potential outcome misclassification (Aim 3). The parent R01 relies on a wide range of epidemiological data ranging from environmental exposures, to claims data, to meteorological and socioeconomic factors. While we curated massive amounts of data for the parent R01, the data has not been deposited to a public data repository and we have not made it publicly available. Overall Goals. With this supplement our goal is to enable effective dissemination and reuse for high-dimensional high-volume data, including the data products from the parent R01. These goals will be achieved by forming a new collaboration and partnership with Harvard dataverse, expanding their current capacity to store and share geospatial public health data. Our specific aims are to: implement automatic metadata extraction for high dimensional dataset formats NetCDF and HDF5 (Aim 1), implement an integration with Jupyter Binder that allows exploration and viewing of complex high-dimensional data from Dataverse (Aim 2), enhance reuse, community engagement and reproducibility of R01 research with the demonstration of data analysis using synthetic CMS claims data (Aim 3). Impact. Each new feature in the Dataverse software platform, such as the ones proposed in Aims 1 and 2, is typically propagated to all 77 Dataverse installations, which enhances their impact worldwide. Dataverse also has a vibrant community of open-source contributors and digital libraries, and organizes annual community meetings at Harvard that last for several days. We will use this platform to promote the work in this supplement and engage the community and attend other workshops and conferences with the same goal. We will closely follow the impact of these developments through an existing Dataverse collaboration with the Make Data Count project, which provides usage metrics standardization (such as the number of views, downloads, and citations of data), and enables us to self-modify and improve our data releases. This supplement will also have a direct impact on the parent R01, enhancing reuse, community engagement and reproducibility which will in turn lead to more robust epidemiological conclusions (Aim 3).