Data Management and Methods Core (DM&M)

项目摘要 数据管理和方法核心(DM&M)或核心C将集中统一并为所有四个提议的部门提供服务 研究项目以及由Core B通过提供数据管理骨干网资助的试点项目 以及每个拟议的研究计划项目的方法专业知识。Core C将继续与 项目调查员为每个项目建立分析文件，构建和测试独立的和依赖的 在概念上和理论上适合跨项目使用的变量，请为 传播，并针对复杂的评估和行政数据制定专门的统计方法 在项目中使用。具体目标是1)收集项目数据并开发跟踪和 清理纵向数据；2)创建用于项目分析的统一核心措施；3)提供分析和 对所有项目执行其具体目标的统计支持；4)开发新的评估方法 认知功能状态；5)应用和扩展替代方法来测试策略对患者的影响 结果和解决选择偏见问题；以及6)制定措施并在项目网站上传播数据。 核心C是项目重叠的核心，因为所有项目都依赖于相同的纵向核心 联邦医疗保险索赔和受益人评估数据，与联邦医疗保险受益人的登记文件相匹配。 新的方法工作加强并建立在先前工作的基础上，将贝叶斯归因法整合到 跨医疗保险评估的认知功能测量。方法工作还将扩展 由Core PI开发的跨时间匹配技术。所有四个项目都依赖于这些新的 构建认知功能测量方法的改进，用于队列选择和 分层，应用跨时相匹配模型。核心C将通过以下方式整合医疗保险数据 2018年进入我们现有的2000-2015年数据结构，构建至少三到四个新的分析文件程序 为每个项目提供支助，资助试点项目调查员，并继续更新以下方面的信息 我们的网站并将国家汇总数据发布到由合作伙伴维护的国家网站，范围从 联邦基金、美国退休人员协会和NH行业的贸易协会。

Project Summary The Data Management & Methods Core (DM&M), or Core C, will centrally unify and serve all four proposed research projects as well as the pilot projects funded by Core B by providing the data management backbone and methodological expertise for each proposed research program project. Core C will continue to work with project investigators to build analytic files for each project, construct and test independent and dependent variables that are conceptually and theoretically appropriate for use across projects, compile data for dissemination, and develop statistical methods specific to the complex assessment and administrative data utilized in the projects. The specific aims are to 1) assemble project data and develop methods for tracking and cleaning longitudinal data; 2) create uniform, core measures for use in project analyses; 3) provide analytic and statistical support to all projects' execution of their specific aims; 4) develop novel methods of evaluating cognitive functional status; 5) apply and extend alternate approaches to testing the effects of policies on patient outcomes and to address selection bias; and 6) create measures and disseminate data on the project website. Core C is at the heart of the overlap of the projects since all rely upon the same core of longitudinal Medicare claims and beneficiary assessment data, matched to Medicare beneficiaries’ enrollment files. The new methods work enhances and builds upon prior work integrating Bayesian imputation methods to cross-walk cognitive function measures across Medicare assessments. Methods work will also extend the cross-temporal matching techniques developed by the Core PI. All four projects rely on these new method enhancements to construct measures of cognitive function for use in cohort selection and stratification and to apply the cross-temporal matching model. Core C will integrate Medicare data thru 2018 into our existing 2000-2015data structure, construct at least three to four new analysis file programs for each of the projects, support funded pilot project investigators and continue to update the information on our web site and disseminate state aggregates to national websites maintained by partners ranging from the Commonwealth Fund, to AARP and to trade associations in the NH industry.