Family- and neighborhood-level environmental associations with physical health conditions in 9- and 10-year-olds

Abstract

Objective: To determine how environmental factors are associated with physical health conditions in 9- to 10-year-old participants in the Adolescent Brain Cognitive Development (ABCD) Study, and how they are moderated by family-level socioeconomic status (SES).

Method: We performed cross-sectional analyses of 8,429 youth participants in the ABCD Study, in which nine physical health conditions (having underweight or overweight/obesity, not participating in sports activities, short sleep duration, high sleep disturbances, lack of vigorous and strengthening-related physical activity, miscellaneous medical problems, and traumatic brain injury) were regressed on three environmental factors [neighborhood disadvantage (area deprivation index [ADI]), risk of lead exposure, and concentrations of particulate matter 2.5 (PM_2.5)] and their interaction with family-level SES (i.e., parent-reported annual household income). Environmental data were geocoded to participants' primary residential addresses at 9- to 10-year-olds.

Results: Risk of lead exposure and ADI were positively associated with the odds of having overweight/obesity, not participating in sports activity, and short sleep durations. ADI was also positively associated with high sleep disturbances. PM_2.5 was positively associated with the odds of having overweight/obesity and reduced vigorous physical activity. Family-level SES moderated relationships between ADI and both underweight and overweight/obesity, with high SES being associated with more pronounced changes given increased ADI.

Conclusions: Policymakers and public health officials must implement policies and remediation strategies to ensure children are free from exposure to neurotoxicant and environmental factors. Physical health conditions may be less of a product of an individual's choices and more related to environmental influences. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

Figure 1.. Environmental Predictability of Current Physical-Health Conditions.

(A) Adolescent Brain Cognitive Development (ABCD) Study sites’ states are shaded in blue; red markers refer to the approximate location of each site. (B) Boxes represent the strength of the partial correlation coefficient (r_p) from linear models between the two corresponding environmental variables. Analyses included youths’ age and sex at birth, study site, and caregiver-reported annual family-level socioeconomic status (SES; i.e., household income) and maximum education level across caregivers as covariates. Analyses also included interactions between the environmental factor and family-level SES. (C) Each box represents the strength of the association (r_p) between the presence of a physical-health condition (ordinate) and the corresponding environmental factor (abscissa) [risk of lead exposure, area deprivation index (ADI), and particulate matter 2.5 (PM_2.5)]. Associations were derived from generalized linear models (binomial distribution, logit link) predicting the presence of each physical-health condition (0 = not present, 1 = present) in youth participants by each environmental factor. As in (B), analyses included age, sex at birth, study site, the caregiver-reported annual household income and maximum education level across caregivers, and the environmental factor-by- family-level-SES interaction. The r_p value is shown for each association. Single asterisks (*) indicate the associations that passed false-discovery rate correction (q < .05).

Figure 2.. Socioeconomic Moderation of Environmental Associations.

(A) The interaction between area deprivation index (ADI) and family-level socioeconomic status (SES) on probability of youths’ having underweight was driven by youth in mid-SES families showing significant increases with ADI and youth in high-SES families showing significant decreases with ADI. (B) The ADI × SES interaction on probability of youths’ having overweight/obesity was driven by participants of high-SES families showing greater increases with ADI than youth in mid-SES families, and youth in mid-SES families showing greater such increases than youth in low-SES families. Data points represent the observed means. Error bars represent ± 1 between-subjects standard error of the observed means. The solid lines represent means of the fitted values of the model. Analyses included youths’ age and sex at birth, study site, and the caregiver-reported annual household income and maximum education level across caregivers. ADI was converted to deciles relative to the national scale for graphing. Total participants represented for each data point are shown in Supplementary Table 36. The ordinates refer to the probabilities (P) of youth having underweight (A) or overweight/obesity (B).