Choi HwaJung, Burgard Sarah, Elo Irma T, Heisler Michele
Department of Internal Medicine, School of Medicine, University of Michigan, United States.
Department of Sociology, University of Michigan, United States; Epidemiology, School of Public Health, University of Michigan, United States; Population Studies Center, University of Michigan, United States.
Soc Sci Med. 2015 Sep;141:82-90. doi: 10.1016/j.socscimed.2015.07.020. Epub 2015 Jul 26.
We assessed the potential contextual effect of income inequality on health by: 1) comparing individuals with similar socioeconomic status (SES) but who reside in counties with different levels of income inequality; and 2) examining whether the potential effect of county-level income inequality on health varies across SES groups. We used the Health and Retirement Study, a nationally representative study of Americans over the age of 50. Using propensity score matching, we selected SES-comparable individuals living in high-income inequality counties and in low-income inequality counties. We examined differences in self-rated overall health outcomes and in other specific physical/mental health outcomes between the two groups using logistic regression (n = 34,994) and imposing different sample restrictions based on residential duration in the area. We then used logistic regression with interactions to assess whether, and if so how, health outcomes differed among participants of different SES groups defined by wealth, income, and education. In bivariate analyses of the unmatched full sample, adults living in high-income inequality counties have worse health outcomes for most health measures. After propensity score matching, adults in high-income inequality counties had worse self-rated health status (AOR = 1.12; 95% CI 1.04-1.19) and were more likely to report diagnosed psychiatric problems (AOR = 1.08; 95% CI 0.99-1.19) than their matched counterparts in low-income inequality counties. These associations were stronger with longer-term residents in the area. Adverse health outcomes associated with living in high-income inequality counties were significant particularly for individuals in the 30(th) or greater percentiles of income/wealth distribution and those without a college education. In summary, after using more precise matching methods to compare individuals with similar characteristics and addressing measurement error by excluding more recently arrived county residents, adults living in high-income inequality counties had worse reported overall physical and mental health than adults living in low-income inequality counties.
1)比较社会经济地位(SES)相似但居住在收入不平等程度不同的县的个体;2)研究县级收入不平等对健康的潜在影响在不同SES群体中是否存在差异。我们使用了健康与退休研究,这是一项针对50岁以上美国人的具有全国代表性的研究。通过倾向得分匹配,我们选择了生活在高收入不平等县和低收入不平等县的SES相当的个体。我们使用逻辑回归(n = 34,994)并根据在该地区的居住时间施加不同的样本限制,来研究两组在自评总体健康结果以及其他特定身体/心理健康结果方面的差异。然后,我们使用带有交互项的逻辑回归来评估不同SES群体(由财富、收入和教育定义)的参与者在健康结果上是否存在差异,若存在差异又是如何表现的。在对未匹配的全样本进行的双变量分析中,生活在高收入不平等县的成年人在大多数健康指标上的健康结果更差。经过倾向得分匹配后,与低收入不平等县的匹配对象相比,高收入不平等县的成年人自评健康状况更差(调整后比值比[AOR] = 1.12;95%置信区间[CI] 1.04 - 1.19),且更有可能报告被诊断出的精神问题(AOR = 1.08;95% CI 0.99 - 1.19)。这些关联在该地区的长期居民中更强。与生活在高收入不平等县相关的不良健康结果在收入/财富分布处于第30百分位或更高百分位的个体以及没有大学学历的个体中尤为显著。总之,在使用更精确的匹配方法比较具有相似特征的个体并通过排除刚到该县不久的居民来解决测量误差后,生活在高收入不平等县的成年人报告的总体身心健康状况比生活在低收入不平等县的成年人更差。