Jim Crow & health disparities: exploring age-period-cohort effects

描述(由申请人提供)：我们建议的跨学科探索性研究是针对PAR-10-137准备的，重点关注一个重要但很少被探索的问题：如何在概念和方法上最好地整合年龄-时期-队列(APC)分析的使用，以促进对健康不平等程度趋势的社会决定因素的理解。借鉴社会流行病学和政治社会学的框架和方法，以及APC模型的新进展，我们的创新的、以理论为动机的案例涉及我们所称的“吉姆·克罗的死亡地理”，这是一个缺乏实证工作的主题。我们项目的双重贡献将是：(1)概念性的，为使用APC模型分析政策变化如何影响人口健康和健康不平等提供新的见解和指导；(2)实质性的，因为缺乏关于1964年美国民权法案废除合法种族歧视对健康影响的研究。研究结果将是全原因死亡率和癌症死亡率(65年(早产))和婴儿死亡率(1年)；这两组差异涉及美国黑人和白人人口以及吉姆·克罗和非吉姆·克罗各州美国黑人人口的比较，分析与时空尺度(生命过程、历史代际、空间影响)和水平(州、县、 个人)。因此，我们的具体目标是：(1)目标1：对由1960-2006年美国县级黑人和白人死亡率和分母数据组成的数据库进行探索性和图形APC分析，该数据库与年度县级家庭收入中位数数据和新链接到州Jim Crow状态的数据(即，在1964年美国民权法案之前没有使种族歧视合法化的数据)相关联；步骤包括创建详细的A、P和C类别，计算研究结果的年率和差异，以及执行APC表格和图形分析；目的2：探索使用不同的APC建模方法来分析废除Jim Crow对特定死亡率结果的影响：(A)理论上，与病因学和统计学假设有关；(B)在经验上，使用：(I)广泛使用的传统解决方案，通过假设3个因素中只有2个影响结果，来解决由于A、P和C的线性相关性而引起的“识别问题”；(Ii)一种新的和有争议的统计学方法，称为“内蕴估计器”；以及(Iii)一种基于广义线性混合模型的方法，该方法可以扩展到解决群体差异和空间效应；和(3)目标3：传播调查结果，并利用结果为NIH R01拨款的准备工作提供信息，以进行政策变化对卫生不平等程度的影响的APC分析。因此，拟议的项目将直接针对PAR-10-137以及健康人(HP)2010和HP 2020的主要目标，因为它将对APC方法进行理论上有动机的实证测试，这些方法与了解政策、社会背景、水平、生活经历和种族歧视对健康不平等程度的影响有关，实质上与Jim Crow有关，更广泛地与其他政策有关。

DESCRIPTION (provided by applicant): Our proposed interdisciplinary exploratory study, prepared in response to PAR-10-137, focuses on a significant yet little explored problem: how best, conceptually and methodologically, to integrate use of age-period-cohort (APC) analysis to advance understanding of societal determinants of trends in the magnitude of health inequities. Drawing on social epidemiology and political sociology frameworks and methods, and also new advances in APC models, our innovative theoretically-motivated case example concerns what we term the "Jim Crow geography of mortality," a topic on which scant empirical work exists. The two-fold contributions of our project will be: (1) conceptual, providing new insights and guidance regarding use of APC models to analyze how policy changes can influence population health and health inequities; and (2) substantive, given the paucity of research on the health impact of the abolition of legal racial discrimination by the 1964 US Civil Rights Act. Study outcomes will be all-cause and cancer mortality (total and <65 y (premature)) and infant death (<1 y); the two sets of disparities pertain to comparisons of the US black vs white population, and the US black population in Jim Crow vs non-Jim Crow states, analyzed in relation to spatiotemporal scale (lifecourse, historical generation, spatial effects) and level (state, county, individual). Our Specific Aims accordingly are: (1) Aim 1: Conduct exploratory and graphical APC analysis of a database comprised of 1960-2006 county-level US black and white mortality and denominator data linked to annual county median family income data and newly linked to data on state Jim Crow status (i.e., did vs did not have legalized racial discrimination prior to te 1964 US Civil Rights Act); steps include creating detailed A, P, and C categories, calculating the study outcome annual rates and disparities, and performing APC tabulations and graphical analyses; Aim 2: Explore use of diverse APC modeling approaches to analyze the impact of the abolition of Jim Crow on the specified mortality outcomes: (a) theoretically, in relation to etiologic and statistical assumptions, and (b) empirically, using: (i) a widely-used conventional solution to address the "identification problem" due to linear dependence of A, P, and C, by assuming only 2 of the 3 affect the outcome; (ii) a new - and debated - statistical approach, called the "intrinsic estimator"; and (iii) an approach based on generalized linear mixed models which can be extended to address group disparities and spatial effects; and (3) Aim 3: Disseminate findings and use results to inform preparation of an NIH R01 grant to conduct APC analyses of the impact of policy change on the magnitude of health inequities. The proposed project thus will directly address key objectives of PAR-10-137, and also Healthy People (HP) 2010 and HP 2020, since it will conduct theoretically-motivated empirical tests of APC approaches relevant to understanding the impact of policy, social context, levels, lifecourse, and racial discrimination on the magnitude of health inequities, substantively in relation to Jim Crow and more broadly for other policies as well.