Causal Estimates of Neighborhood Poverty on Health and Mortality

描述（由申请人提供）：货物和服务的分配以及跨空间的危险和机会的差距正在扩大，突出表明地点与健康之间的联系日益密切。尽管有充分的证据证实，生活在经济上处于不利地位的社区与不良健康结果有关，但对横断面数据的依赖以及对两个主要偏见来源的关注不足，使得因果推论存在问题。居民倾向于根据许多特征将自己划分为不同类型的社区。如果不考虑与结果和邻里环境相关的所有特征，可能会导致对邻里效应的高估。由于回归模型不可能考虑所有相关因素，未观察到异质性的可能性很大，使得邻域效应研究容易受到忽略变量偏差的批评。然而，与此同时，邻域效应研究也同样容易因过度调整而产生偏差。许多在邻里效应模型中受控制的因素，如受教育程度、收入和就业，可能受到过去邻里环境的影响。对这些因素进行调整，消除了社区影响健康的可能关键途径，可能导致对社区影响的过度保守估计。这两种截然相反的偏见来源一直困扰着现有的社区健康研究；因此，目前的研究结果产生了脆弱和模糊的推论。这个拟议的项目将使用新的分析方法和现有观察研究的纵向数据，以解决上述两个主要限制，并恢复邻里贫困对自评健康和死亡率的因果估计。我们将1)使用边际结构模型来适当调整同时是混杂因素和中介因素的协变量，2)进行敏感性分析，以确定邻域效应发现对未观察到的异质性的稳健性。将这种综合方法应用于社区健康研究有可能显著提高我们对地点与健康之间关系的认识，产生深远的政策影响。居住环境与健康之间因果关系的有力发现可以帮助卫生政策制定者判断社区对健康影响的程度和程度，并指导公共卫生政策。社会和结构环境影响生活机会并最终影响健康结果的证据表明，传统上以个人为目标而很少考虑社区情况的卫生政策，在设计和执行最有效和最高效的公共卫生政策时，应考虑住宅环境中存在的潜在制约因素或机会。

DESCRIPTION (provided by applicant): The disparities in the distribution of goods and services, and hazards and opportunities across space are increasing, underscoring the growing connection between place and health. Although ample evidence confirms that living in an economically disadvantaged neighborhood is associated with adverse health outcomes, the reliance on cross-sectional data and inadequate attention to two main sources of bias make causal inferences problematic. Residents tend to sort themselves into different types of neighborhoods based on a multitude of characteristics. Not accounting for all characteristics that are correlated to both the outcome and neighborhood context would likely lead to over-estimations of neighborhood effects. Because regression models cannot possibly account for all relevant factors, the strong possibility of unobserved heterogeneity make neighborhood effect studies open to criticisms of omitted variable bias. Yet, at the same time, neighborhood effect studies are also just as likely to be susceptible to bias due to over- adjustment. Many factors that are controlled for in neighborhood effect models, such as educational attainment, income, and employment, may arguably have been influenced by past neighborhood context. Adjusting for these factors eliminate possible critical pathways through which neighborhoods affect health, likely yielding overly conservative estimates of neighborhood effects. These two sources of bias, working in opposing directions, have plagued extant neighborhood-health research; consequently, results from current research yield tenuous and ambiguous inferences. This proposed project will use novel analytical methods and longitudinal data from an existing observational study to address the two major limitations described above and recover causal estimates of neighborhood poverty on self-rated health and mortality. We will 1) use marginal structural modeling to appropriately adjust for covariates that are simultaneously confounders as well as mediators and 2) conduct a sensitivity analysis to determine the robustness of the neighborhood effect findings to unobserved heterogeneity. Applying this combined methodology to neighborhood-health research has the potential to significantly advance our knowledge of the relationship between place and health, yielding far reaching policy implications. PUBLIC HEALTH RELEVANCE: Relevance Robust findings of a causal connection between residential context and health can help health policymakers judge the extent and magnitude of neighborhood impacts on health and guide public health policies. Evidence that the social and structural environment influences life-chances, and ultimately health outcomes, suggests that health policy, traditionally targeted at the individual level with little regard to neighborhood context, should consider underlying constraints or opportunities present in the residential environment in designing and implementing the most effective and efficient public health policies.