The intersection of adverse childhood experiences and neighborhood determinants of health: An exploratory spatial analysis.

Abstract

Objective: Adverse childhood experiences (ACEs) are associated with elevated risk for poor physical and psychological health outcomes. Nearly all of the literature on the association between ACEs and poor health focuses on the individual and family level; the potential role of neighborhood environment is overlooked. Understanding the relationship between ACEs and characteristics of the neighborhood environment is a necessary first step in determining if and how place-based, trauma-informed interventions might mitigate the negative effects of ACEs. The purpose of this exploratory study was to describe the neighborhood environment of adults who have experienced $ 4 versus # 3 ACEs. Method: An exploratory secondary analysis of cross-sectional and geospatial data was conducted during 2021. Data sources included 2011/2012 Philadelphia ACE Survey data, a telephone survey of 1,784 randomly sampled Philadelphia adults, linked with geospatial data on 21 neighborhood-level determinants of health. Neighborhood was defined as participant's home census tract at the time of survey. The sample for this secondary analysis included the 1,679 Philadelphia ACE Survey participants for whom home census tract was available. Bivariate logistic regression examined differences between groups (# 3 ACEs vs. $ 4 ACEs). Results: Individuals with $ 4 ACEs lived in neighborhoods with higher neighborhood poverty rates, less socioeconomic resources, worse food access, poorer perceived physical and mental health, more substance overdose deaths, higher crime, and less green space. Conclusions: Findings suggest future work may benefit from considering neighborhood environments when examining and intervening upon the association between ACEs and poor physical and psychological health.

Can adverse childhood experiences predict chronic health conditions? Development of trauma-informed, explainable machine learning models

“…For example, Black, Hispanic, or low-income individuals show the highest prevalence of ACEs ( 2 ). Additionally, social, economic, and environmental inequities are greater in the environments of those who have endured four or more ACEs ( 3 ).…”

“…Multiple pathways connect ACEs to chronic health conditions, including social determinants of health (SDoH) and health behaviors. Individuals with a higher number of ACEs tend to live in areas of greater poverty, fewer economic and health resources, worse food access, less green space, and more community instability ( 3 , 37 , 38 ). ACE survivors are also more likely to engage in harmful behaviors, such as smoking, heavy alcohol consumption, substance use, high-risk sexual behavior, interpersonal violence, excess screen time, and inadequate sleep ( 5 , 27 , 30 , 39–42 ).…”

“…In syndemics, social conditions contribute to disease formation, accumulation, spread, and progression by increasing susceptibility and reducing immune function, particularly among marginalized populations; hence, syndemics are most likely to emerge under conditions of health inequality ( 43 ). A syndemic can be exemplified by the interactions of ACEs, negative social conditions (i.e., SDoH), and risky health behaviors worsening the risk of various chronic health conditions ( 2 , 3 , 44 ). However, an accurate assessment of these complex associations can be methodologically challenging, as the involved risk factors may be highly correlated, interactive, or synergistic.…”

“…We employ classical ML models identified in the literature as robust tools for predicting chronic health conditions: LR, Gaussian NB, SVM, RF, and KNN ( 58 , 60 , 62–64 ) Although neural networks are also promising for this prediction task ( 58 ), they lack interpretability and demand greater computational power and time ( 71 , 72 ). Computational resources are crucial during model deployment, given the higher prevalence of ACEs in disadvantaged communities that can benefit most from the models we developed ( 3 , 37 , 38 , 73–76 ). Accordingly, we focus on classical ML models that can be scalable and adaptable, even in low-resource settings, while empowering end-users with explainable results to aid clinical decision-making.…”

The Children, Caregivers, and Community (C3) study of together growing strong: A protocol for an observational, place-based initiative in Sunset Park, Brooklyn

“…Additionally, the neighborhood built environment that poor and non-white families often live in may reflect socio-political factors that historically disinvest in resources that affect access to air quality, green space, food, social services and housing-key social determinants that affect child development outcomes [15][16][17][18][19]. Analysis of the 2020-2021 National Survey of Children's Health, for example, indicates that most Black children in the U.S. live in neighborhoods that lack amenities such as sidewalks, parks/playgrounds, libraries, and recreational centers [20].…”