The Relationship Between Adverse Childhood Experiences and Child Well-Being: Comparing Cumulative Risk and Latent Class Approaches Using National Data

Introduction: Research on Adverse Childhood Experiences (ACEs) has demonstrated strong associations between ACEs and various health outcomes, including asthma, depression, and overall well-being. Measures to screen for ACEs are now employed in many clinical settings and used in state-level public health surveillance. Effective risk assessment is central to prevention efforts seeking to improve child well-being and address health disparities. Studies using ACEs create a cumulative risk index to demonstrate that a greater number of ACEs is associated with greater risk for negative health outcomes.  However, effective tailoring of prevention strategies requires an understanding of the relationship between the different adverse experiences and related health outcomes. Person-centered methods, such as latent-class analysis (LCA), are used to identify subgroups of similar individuals within a heterogeneous population. However, there are no known studies that attempt to identify latent classes of ACEs. The purpose of this study is to compare the utility of cumulative risk and LCA approaches applied to ACEs research.

Methods: This study used data from the 2011/2012 National Survey of Children’s Health (NSCH) containing nationally representative data from 95,677 caregivers of children ages 0-17. The predictor variables of interest included nine ACES: financial hardship, parent divorce, death of parent, parent in jail, domestic violence in the home, neighborhood violence, parent with mental health diagnosis, drug use in household, and child experienced discrimination. Dependent variables included well-being outcomes associated with ACEs in prior research: overall health status, depression, asthma, developmental delay, and flourishing. Two sets of generalized linear models incorporating NSCH’s complex survey weights were used to predict well-being outcomes. The first set of models used a cumulative risk index (a count of ACEs). Second, LCA was used to identify the number of latent classes in the data and group membership was used as a predictor of well-being outcomes.

Results: The cumulative risk index significantly predicted all child well-being outcomes with a dose-response relationship. The LCA identified a 5-class solution (low risk, primary poverty, high violence exposure/discrimination, high parental risk, and high cumulative risk). However, the associations between latent classes and well-being outcomes were not consistent. Notably, the high cumulative risk class did not always have the largest effect on risk for poor outcomes. These findings suggest that different groups of variables, not just a higher total number, may be a better predictor of certain child outcomes. 

Conclusions and Implications: Results from this study indicate that understanding the unique impact of clusters of ACEs may serve as better predictors for child well-being outcomes when compared to use of a cumulative risk index. Research that seeks to understand how childhood experiences impact the healthy development of youth will likely require a more nuanced understanding of the complex interplay of overlapping risk factors. Further, to implement effective strategies to interrupt the chain from ACEs to negative outcomes, we must have an understanding of how the different combinations of ACEs relate to the outcomes. This study provides practical implications for utilizing ACEs screenings in clinical risk assessment settings.