Abstract
Adolescents with a history of child maltreatment experience increased risk for psychopathology that sets them on a trajectory towards a range of difficulties in adulthood. Various factors influence caregivers' decisions to seek mental health services (MHS) that could improve developmental outcomes. The present study applied a machine learning algorithm, elastic net, to a sample of 878 adolescent-caregiver dyads from the Longitudinal Studies of Child Abuse and Neglect. Analyses simultaneously examined a large number of factors to determine their ability to discriminate between caregivers who perceived a need for MHS and those who did not, as well as caregivers who utilized MHS and those who did not. Results highlight family demographics, chronic parental stressors, youth psychopathology, and exposure to recent adversities as good classifiers of caregiver perceived need for (77.6%; sensitivity = .77; specificity = .78) and utilization of (71%; sensitivity = .71; specificity = .71) adolescent MHS. Elastic net identified adolescent clinical externalizing and internalizing problems, and parental stress related to child(ren)'s behavior as high value classifiers of both outcomes. Youth living with non-kin caregivers were also significantly more likely to utilize MHS. Findings highlight the importance of assessing clinical need, stress related to child(ren)'s behavior, and caregiver kinship in understanding the likelihood that at-risk families will seek adolescent MHS.