Identifying low dose measurement error corrected effects of multiple pollutants using causal modeling

项目摘要 候选人将通过使用最新的混合空气污染来为父母资助的目标做出贡献 暴露模型和Medicare管理数据，并将它们与新获得的 国家死亡指数(NDI)数据。NDI数据确定了特定的死亡原因，从而使 建立一个解决空气污染物临时影响的补充项目的候选人 在联邦医疗保险中单独和联合使用特定原因的心血管死亡率 人口。具体地说，这项研究将采用准实验设计(Difference-in-in- 差异)，通过分层来控制测量的和许多未测量的混杂因素 对象(控制时间不变或缓慢变化的单个协变量)或 邻域(控制时间不变或缓慢变化的邻域级别协变量)。这 是DID的一个主要优势，因为许多未测量和测量的混杂因素都得到了控制 由设计者设计。我们将在2000-2020年间采用更新的曝光模型， 空间分辨率提高500米。通过使用一种新的混合技术--快速贝叶斯核 机器回归(BKMR)，它加快了传统BKMR的处理速度， 研究可以分析大数据集，如医疗保险。FAST BKMR确定 个别污染物，低浓度下的非线性，以及混合物的整体影响。我们 还将对居住在该地区的一小部分医疗保险人口进行估计 这从未超过目前EPA的标准。这将有助于得出结论，即 目前的法规足以保护心血管健康。在这个支持下 作为补充，我们将结合候选人在心血管流行病学和 混合方法纳入我们目前的医疗保险研究。她为这份副刊建议的工作不仅是 与主要R01的目标保持良好一致，但有潜力为正在进行的 加强环境空气污染监管的努力

Project Summary The Candidate will contribute to the goals of the parent grant by using updated hybrid air pollution exposure models and Medicare administrative data and link them with the newly acquired National Death Index (NDI) data. The NDI data identifies the specific cause of death allowing the Candidate to establish a complementary project that addresses the casual effect of air pollutants with cause-specific cardiovascular mortality both individually and jointly within the Medicare population. Specifically, the study will employ a quasi-experimental design (difference-in- differences) that controls for both measured and many unmeasured confounders by stratifying on either the subject (controlling for time invariant or slowly varying individual covariates) or the neighborhood (controlling for time invariant or slowly varying neighborhood level covariates). This is a major advantage of DID since many unmeasured and measured confounders are controlled for by designed. We will employ updated exposure models for the years 2000-2020 with an increased spatial resolution of 500m. By using a novel mixture technique, fast Bayesian Kernel Machine Regression (BKMR), that accelerates the processing speed of the traditional BKMR, the study can analyze big datasets such as Medicare. Fast BKMR identifies interactions between individual pollutants, nonlinearities at low concentrations, and the overall effect of the mixture. We will also produce estimates on a restricted subset of the Medicare population that lived in areas that never exceeded the current EPA standards. This will help draw conclusions on whether the current regulations are sufficient to protect cardiovascular health. With the support of this supplement, we will incorporate the Candidate’s background in cardiovascular epidemiology and mixture methods into our current Medicare study. Her proposed work for this supplement not only aligns well with the objectives of the primary R01 but has the potential to contribute to the ongoing efforts to tighten ambient air pollution regulations