Effects of accumulated environmental, social and host exposures on early childhood educational outcomes

Abstract

Background: Persistent disparities in academic performance may result from a confluence of adverse exposures accruing disproportionately to specific subpopulations.

Objective: Our overarching objective was to investigate how multiple exposures experienced over time affect early childhood educational outcomes. We were specifically interested in whether there were: racial/ethnic disparities in prevalence of adverse exposures; racial/ethnic disparities in associations observed between adverse exposures and early childhood educational outcomes; and interactions between exposures, suggesting that one exposure augments susceptibility to adverse effects of another exposure.

Methods: We link geocoded North Carolina birth data for non-Hispanic white (NHW) and non-Hispanic black (NHB) children to blood lead surveillance data and 4th grade end-of-grade (EOG) standardized test scores (n = 65,151). We construct a local, spatial index of racial isolation (RI) of NHB at the block group level. We fit race-stratified multi-level models of reading and mathematics EOG scores regressed on birthweight percentile for gestational age, blood lead level, maternal smoking, economic disadvantage, and RI, adjusting for maternal- and child-level covariates and median household income.

Results: There were marked racial/ethnic disparities in prevalence of adverse exposures. Specifically, NHB children were more likely than NHW children to be economically disadvantaged (80% vs. 40%), live in block groups with the highest quintile of RI (46% vs. 5%), have higher blood lead levels (4.6 vs. 3.7 μg/dL), and lower birthweight percentile for gestational age (mean: 39th percentile vs. 51st percentile). NHB children were less likely to have mothers who reported smoking during pregnancy (11% and 22%). We observed associations between key adverse exposures and reading and math EOG scores in 4th grade. Higher birthweight percentile for gestational age was associated with higher EOG scores, while economic disadvantage, maternal smoking, and elevated blood lead levels were associated with lower EOG scores. Associations observed for NHB and NHW children were generally not statistically different from one another, with the exception of neighborhood RI. NHB children residing in block groups in the highest RI quintile had reading and math scores 1.54 (0.74, 2.34) and 1.12 (0.38, 1.87) points lower, respectively, compared to those in the lowest RI quintile; statistically significant decrements in EOG scores associated with RI were not observed for NHW children. We did not find evidence of multiplicative interactions between exposures for NHB or NHW children.

Discussion: Key adverse host, environmental, and social exposures accrue disproportionately to NHB children. Decrements in test scores associated with key adverse exposures were often but not always larger for NHB children, but were not significantly different from those estimated for NHW children. While we did not observe interactive effects, NHB children on average experience more deleterious combined exposures, resulting in larger decrements to test scores compared to NHW children.

Keywords: Childhood lead exposure; Early childhood educational outcomes; Multiple exposures; Racial isolation; Standardized test scores.

Figure 1.

Forces shaping health and developmental outcomes. Image reproduced with permission.

Figure 2.

Host exposure: Proportion of children born small-for-gestational age (<10^th percentile of birthweight percentile for gestational age). The small for gestational age variable, which is derived from the detailed birth records, is shown based on the child’s block group of residence at time of birth.

Figure 3.

Environmental exposures: Proportion of children with mothers who reported smoking during pregnancy (Panel A); and proportion of children with blood lead level ≥5 μg/dL (Panel B). The maternal smoking during pregnancy variable, which is derived from the detailed birth records, is shown based on the child’s block group of residence at time of birth. The blood lead level variable, obtained from the blood lead screening data, is shown based on the child’s residence at time of lead screening. Approximately 10% of children in the final analysis dataset were missing residential address at time of lead screening. These children are not included in the proportion of children with lead levels ≥5 μg/dL.

Figure 4.

Social exposures: proportion of children who are economically disadvantaged (Panel A) and neighborhood racial isolation of NHB (Panel B). The economic disadvantage variable, obtained from the EOG dataset, is shown based on the child’s block group of residence in 4th grade, at time of testing. Approximately 10% of children in the final analysis dataset were missing addresses at time of EOG testing. These children are not included in the proportion calculated for the economic disadvantage map. RI of NHB values are shown for all block groups with non-zero population in the neighborhood environment (i.e., the index block group and adjacent block groups).

Figure 5.

Individual-level economic disadvantage by race and neighborhood (block group) RI quintile

Figure 6.

Individual-level blood lead level by race and neighborhood (block group) RI quintile