Health Policy and Management, UAMS College of Public Health, 4301 W. Markham St., Little Rock, AR 72206, USA.
Matern Child Health J. 2009 Nov;13(6):806-13. doi: 10.1007/s10995-008-0311-y. Epub 2008 Feb 16.
The objectives of this study were to explore regional, economic and racial disparities in infant mortality rates between geographic sub-regions within the eight states containing the Delta and to test hypotheses that regional disparities would decrease over time while county poverty level and racial composition would remain significant predictors of infant mortality rates.
The study used secondary data analysis of county level rates, including descriptive statistics, hierarchical multiple regression with interaction effects and linear multiple regression. Models testing the impact of sub-regional geographic differences, percent of poverty, percent of black population and interaction effects were conducted at three time periods, the late 1970s, late 1980s and late 1990s.
In the first time period, regional differences, percent of poverty, percent of black population and the interaction of region and poverty were all predictive of infant mortality (R(2) = 0.31, P < 0.0001). In the subsequent time periods, only percent of poverty and percent of black population were significant predictors (R(2) = 0.20, P < 0.0001 and R(2) = 0.26, P < 0.0001).
During the late 1970s and early 1980s, region, poverty and racial composition of counties all played an important part in predicting life chances for infants born in these eight states. Furthermore, Central Delta infants in counties with poverty levels of 30% or greater were significantly more likely to die than infants in other areas with the same rates of poverty, even after controlling for racial composition. The impact of regional differences was no longer significant at the ends of the subsequent two decades. Both medical and policy changes during these decades may have contributed to the decreased impact of region. However, both poverty and racial composition continue as important factors, accounting for more variance in the late 1990s than a decade before.
本研究旨在探讨密西西比三角洲地区八个州内地理分区之间婴儿死亡率的区域、经济和种族差异，并验证以下假设，即随着时间的推移，区域差异会减少，而县贫困水平和种族构成将仍然是婴儿死亡率的重要预测因素。
本研究使用县一级数据的二次数据分析，包括描述性统计、具有交互效应的层次多重回归和线性多重回归。在三个时期（20 世纪 70 年代末、80 年代末和 90 年代末）进行了检验区域亚区地理差异、贫困百分比、黑人人口百分比以及交互效应影响的模型测试。
在第一个时期，区域差异、贫困百分比、黑人人口百分比以及区域与贫困的交互作用均对婴儿死亡率具有预测作用（R²=0.31，P<0.0001）。在后两个时期，只有贫困百分比和黑人人口百分比是显著的预测因素（R²=0.20，P<0.0001 和 R²=0.26，P<0.0001）。
在 20 世纪 70 年代末和 80 年代初，县的区域、贫困和种族构成在预测这八个州出生的婴儿的生命机会方面都起着重要作用。此外，在控制种族构成的情况下，贫困水平达到 30%或更高的中心三角洲县的婴儿死亡的可能性明显高于其他贫困率相同地区的婴儿。在随后的二十年结束时，区域差异的影响不再显著。这几十年中的医疗和政策变化可能导致区域影响降低。然而，贫困和种族构成仍然是重要因素，在 20 世纪 90 年代末比十年前的解释变量方差更大。
