Wei Hanxue, Spoer Benjamin R, Titus Andrea R, Lampe Taylor M, Gourevitch Marc N, Faber Jacob W, Korzeniewski Steven J, Bauer Samantha J, Thorpe Lorna E
Department of Population Health, New York University Grossman School of Medicine, New York, NY 10016, USA.
Robert F. Wagner School, New York University, New York, NY 10003, USA.
PNAS Nexus. 2024 Aug 9;3(8):pgae301. doi: 10.1093/pnasnexus/pgae301. eCollection 2024 Aug.
Studies have recently begun to explore the potential long-term health impacts of homeownership policies implemented in the New Deal era. We investigated the association between assigned grades of lending risk by the Home Owners' Load Corporation (HOLC) maps from the 1930s and present-day prevalence of three cardiovascular risk factors (diabetes and obesity in 2020, and hypertension in 2019), estimated at the census tract level in the United States. To minimize potential confounding, we adjusted for sociodemographic data from the time period when HOLC maps were made. We calculated propensity scores (predicted probability of receiving a HOLC grade) and created a pseudo-population using inverse probability weighting. We then employed marginal structural models to estimate prevalence differences comparing A vs. B, B vs. C, and C vs. D HOLC grades. Adjusting only for regions, a less desirable HOLC grade was associated with higher estimated prevalence rates of present-day cardiovascular risk factors; however, most differences were no longer significant after applying propensity score methods. The one exception was that the prevalence of diabetes, hypertension, and obesity were all higher in C vs. B graded census tracts, while no differences were observed for C and D and A and B comparisons. These results contribute to a small body of evidence that suggests historical "yellowlining" (as C grade was in color yellow) may have had persistent impacts on neighborhood-level cardiovascular risk factors 80 years later.
近期,研究开始探索新政时期实施的住房自有政策对健康可能产生的长期影响。我们调查了20世纪30年代房主贷款公司(HOLC)地图所划定的贷款风险等级与当今三种心血管危险因素(2020年的糖尿病和肥胖症,以及2019年的高血压)患病率之间的关联,这些数据是在美国普查区层面估算得出的。为尽量减少潜在的混杂因素,我们对制作HOLC地图时的社会人口数据进行了调整。我们计算了倾向得分(获得HOLC等级的预测概率),并使用逆概率加权法创建了一个伪总体。然后,我们采用边际结构模型来估计HOLC等级A与B、B与C、C与D之间患病率的差异。仅对地区进行调整时,HOLC等级较低与当今心血管危险因素的较高估计患病率相关;然而,应用倾向得分方法后,大多数差异不再显著。唯一的例外是,C级普查区的糖尿病、高血压和肥胖症患病率均高于B级,而C与D以及A与B的比较中未观察到差异。这些结果为一小部分证据提供了补充,表明历史上的“黄线划分”(因为C级用黄色表示)可能在80年后对邻里层面的心血管危险因素仍有持续影响。