Adapting ACEs Assessment Toward Explaining Adolescent Mental Health Outcomes in Population Data

Background/Purpose:  In addition to physical health disparities, mental health disorders have been consistently identified as risk outcomes for adverse childhood experiences (ACEs) (McLaughlin et al., 2012). Extending epidemiologic research, we address a gap in testing childhood adversity effects in relation to risk and protective factors that may compound or buffer these effects (Nurius et al., 2012 ). Consistent with social work approaches, we incorporate social disadvantage (income, racial identity) as co-occurring risk factors, as well as inclusion of socioemotional contexts (family, school), and health behaviors (sleep) linked to the neurobiology of stress, and to the value of testing for both cumulative and unique explanatory effects.

Methods:  Data derive from the Washington State 2010 Healthy Youth Survey (HYS) applying a cluster sampling design to yield a population sample 76,758 students from 8th, 10th, and 12th grades; 51.74% females, with racial/ethnic composition representative of the state. Childhood adversity was assessed through low income and an interpersonal maltreatment index (summing 4 subscales: physical abuse, peer bullied, dating aggression, unsafe at school), with protective factors assessed through sleep, family support and school engagement. Using theory-guided hierarchical regressions, we sequentially regressed each of three measure of mental health (depression, suicidality, and psychological well-being), controlling for demographics, onto the two explanatory variable sets: childhood adversity and protective factors. Component analysis comparing the maltreatment indexed effects to those of its four constituent subscales was also examined through a parallel set of regressions (linear and logistic regressions).

Results:  For each of the mental health outcomes, the full regression models achieved significance, and each of the predictor blocks added significantly to the R² change (and likelihood ratio test, X²) in each model. As hypothesized, both maltreatment and socioeconomic factors significantly explained poorer mental health. Although racial minority students reported significantly higher levels of interpersonal maltreatment than white students, race was not consistently patterned in the regression models, although females were more adversely affected than males Each of the three protective factors contributed to better mental health and attenuated the effects of maltreatment and low income. Component analysis revealed that each maltreatment subscale sustained independent significance, but that the cumulative index was far more robust in effect size.

Conclusions/Implications:  The findings demonstrate the value of using routinely collected surveillance and system data to assess not only the prevalence of youth adversities such as maltreatment, but to embed this analysis within risk and protective factor frameworks that allow testing of unique and cumulative factor effects. Measured as a cumulative index, youth maltreatment was nearly exclusively the single strongest predictor—a finding not yet reported within these surveillance data. Three of the 4 indicators reflected peer maltreatment forms, accentuating the importance of ACEs-related assessment to tap peer interactions within childhood and adolescence. We elaborate the theoretical basis for interpreting findings within ecobiodevelopment models attentive to stress embodiment (Shonkoff et al, 2012), and the resilience potential of protective factors. Implications for preventive and remedial interventions among family, school, and community levels are discussed.