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

项目摘要 背景父母补助金（R01 AG066793）的具体目的是进行国家流行病学调查， 研究医疗保险和医疗补助索赔，以估计长期暴露于空气污染对 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识别个人层面、环境和社会因素的共同出现 （目标2），并制定统计方法， 污染暴露与其他混杂因素，并纠正潜在的结果错误分类（目的 3）。母体R01依赖于广泛的流行病学数据，从环境暴露到 索赔数据，气象和社会经济因素。当我们整理了大量的数据， 父R01，数据尚未存放到公共数据存储库，我们也没有公开 available. 总体目标。有了这个补充，我们的目标是实现有效的传播和重用， 大容量数据，包括来自父R01的数据产品。这些目标将通过建立一个 与哈佛dataverse建立新的合作伙伴关系，扩大他们目前的存储和共享能力 地理空间公共卫生数据。我们的具体目标是：实现自动元数据提取高 维数据集格式NetCDF和HDF5（Aim 1），实现了一个集成与Binder， 允许探索和查看Dataverse中的复杂高维数据（目标2），增强重用， R01研究的社区参与和再现性，以及使用 合成CMS声明数据（目标3）。 冲击Dataverse软件平台中的每个新功能，例如目标1和2中提出的功能， 通常传播到所有77个Dataverse安装，这增强了它们在全球的影响力。Dataverse还 拥有一个充满活力的开源贡献者和数字图书馆社区，并组织年度社区 在哈佛的会议持续了几天。我们将利用这个平台来推动这一补充中的工作 并参与社区，参加其他具有相同目标的研讨会和会议。我们会密切 通过现有的Dataverse与Make Data Count的合作，跟踪这些发展的影响 一个项目，它提供了使用度量标准（如查看、下载和引用的数量 的数据），并使我们能够自我修改和改进我们的数据发布。该补充也将有直接的 对父R01的影响，提高重用，社区参与和再现性，这反过来又会导致 更强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).