The role of the social environment in the association between air pollution and cardiovascular disease

摘要 邻里社会经济地位 (NSES) 和种族居住隔离等背景因素 (RRS) 是可能使空气污染流行病学研究产生偏差的潜在混杂因素。很少有研究表明 系统地检查了这些重要背景因素造成的混杂程度。致我们的 目前尚无研究将 RRS 视为空气污染与心血管疾病 (CVD) 的混杂因素 研究。我们建议对 SES 的混杂和协同作用进行系统评估（两者 在社区和个人层面）和 RRS 在一个独特而强大的数据源中，该数据源由八个井组成 慢性疾病队列研究的特征。广泛的协变量数据，一致且长期的随访 参与者、高质量的空气污染暴露和 CVD 结果的标准化收集使其成为 进行这项重要工作的理想数据源。现有项目正在进行中，以协调 这些群体的数据。我们的目标是： 1) 评估 SES 不同指标与 RRS 和长期环境空气污染暴露横截面和随着时间的推移，通过发展最新状态 SES 和 RRS 的艺术测量 2) 进行定量偏差分析以评估混杂的程度 通过 SES 和 RRS 确定长期空气污染与 CVD 死亡率和事件之间的关联，以及 3) 研究长期空气污染以及 SES 和 RRS 对心血管事件和死亡率的联合影响。 我们将根据 1990 年至 2015 年期间的主成分分析创建一个时变 NSES 指数 以及均匀度（相异指数）和隔离度（隔离指数）的测量来评估 RRS。我们将使用 目标 1 的空间回归方法、目标 2 的概率定量偏差分析和生存分析 实现目标 3。该提案对公共卫生产生三重影响。首先，评估混杂的程度 SES在空气污染流行病学研究中的应用将加强国家空气污染的证据 标准。其次，对 RRS 的探索可能会为空气污染建立新的背景混杂因素 流行病学。第三，制定一个严格的、有坚实理论基础的框架来评估 多个空间变化因素将帮助我们更好地了解长期存在的环境健康差异 并可能有助于设计未来的干预措施和政策。

Abstract Contextual level factors like neighborhood socioeconomic status (NSES) and racial residential segregation (RRS) are potential confounders that may bias air pollution epidemiology studies. Few studies have systematically examined the magnitude of confounding by these important contextual factors. To our knowledge no studies have included RRS as a confounder in air pollution-cardiovascular disease (CVD) studies. We propose to conduct a systematic assessment of the confounding and synergist roles of SES (both at the neighborhood and individual levels) and RRS in a unique and robust data source made-up of eight well- characterized chronic disease cohort studies. Extensive covariate data, consistent and lengthy follow-up of participants, high quality air pollution exposures and standardized collection of CVD outcomes makes this an ideal data source within which to conduct this important work. An existing project is underway to harmonize data across these cohorts. Our aims are to: 1) assess the association between different indicators of SES and RRS and long-term ambient air pollution exposures cross-sectionally and over time, by developing state-of-the- art measures of SES and RRS 2) conduct quantitative bias analysis to evaluate the magnitude of confounding by SES and RRS in the association between long-term air pollution and CVD mortality and events and 3) examine the joint effects of long-term air pollution and SES and RRS on cardiovascular events and mortality. We will create a time-varying NSES index from principal components analysis for the time period 1990 - 2015 and measures of evenness (dissimilarity index) and isolation (isolation index) to evaluate RRS. We will use spatial regression approaches for aim 1, probabilistic quantitative bias analysis for aim 2 and survival analysis for aim 3. The public health impact of this proposal is three-fold. First, evaluating the magnitude of confounding by SES in air pollution epidemiology studies will strengthen the evidence used for national air pollution standards. Second, exploration of RRS could establish a new contextual confounder for air pollution epidemiology. And third, developing a rigorous and strongly theoretically grounded framework for assessing multiple spatially varying factors will help us better understand longstanding environmental health disparities and may help design future interventions and policy.