Causal Estimates of Neighborhood Poverty on Health and Mortality

描述（由申请人提供）：货物和服务的分布以及空间上的危险和机会的差异正在增加，强调了地点与健康之间日益增长的联系。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利的健康结果有关，但对横截面数据的依赖和对两个主要偏见来源的关注不足，使得因果推断存在问题。居民们倾向于根据众多的特征将自己分为不同类型的社区。不考虑与结果和邻域背景相关的所有特征可能会导致对邻域效应的高估。由于回归模型不可能解释所有的相关因素，未观察到的异质性的强烈可能性，使邻里效应研究开放的遗漏变量偏倚的批评。然而，与此同时，由于过度调整，邻里效应研究也同样容易受到偏倚的影响。邻里效应模型中控制的许多因素，如教育程度、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，可能会对社区影响产生过于保守的估计。这两种偏见的来源，在相反的方向工作，困扰着现存的邻里健康研究;因此，从目前的研究结果产生脆弱和模糊的推论。该项目将使用新的分析方法和现有观察性研究的纵向数据，以解决上述两个主要限制，并恢复社区贫困对自测健康和死亡率的因果估计。我们将1）使用边际结构模型对同时作为混杂因素和中介因素的协变量进行适当调整，2）进行敏感性分析，以确定邻域效应结果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，有可能大大提高我们对地方与健康之间关系的认识，产生深远的政策影响。公共卫生相关性：住宅环境和健康之间的因果关系的强有力的调查结果可以帮助卫生政策制定者判断邻里对健康的影响程度和幅度，并指导公共卫生政策。证据表明，社会和结构环境影响生活的机会，并最终健康的结果，表明，卫生政策，传统上针对个人层面很少考虑到邻里的情况下，应考虑潜在的限制或机会，目前在住宅环境中设计和实施最有效和最高效的公共卫生政策。

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.