2 - (SYM 63) Assessing Profiles of Risk Using Self-report and System Records Among Adolescents Involved in the Juvenile Legal System

Background. Adolescents in the juvenile legal system (JLS) experience especially high rates of mental health and substance use difficulties, adverse childhood experiences, and interpersonal difficulties (Kemp et al., 2025). Within the broader JLS, adolescents are often screened for mental health symptoms and other risk factors several times using different methods (e.g., self-report rating, professional/provider rating). Ratings completed by system professionals, such as probation or correctional officers, may be particularly useful in predicting problematic outcomes among JLS-involved adolescents; however, such ratings have rarely been included in the extant literature to date. This study utilized ratings of risk factors completed by probation officers to examine whether distinct clusters of JLS-involved youth could be identified, after which clusters were examined for differences in mental health, substance use, and system-level variables.

Method. Retrospective data collected via chart review were used for the current study. Data included several different types of self-report and system variables. Data for the present analyses included records from 133 adjudicated youth in the Northeastern United States. The overwhelming majority of youth were male (89%) and youth were on average 16.97 years old (SD = 1.19). The sample was racially and ethnically diverse. To determine the presence of distinct clusters, items completed by probation officers on the Structured Assessment of Violence Risk in Youth (SAVRY) were were assessed using k-means cluster analysis. Upon identifying the best-fitting cluster solution, chi-square and t-tests were used to assess cluster differences.

Results. The two-class k-means cluster analysis was the best fitting model consisting of a High Needs Class (59.4%) and a Moderate Needs Class (40.6%). Notably, the High Needs Class were more likely to have a history of child maltreatment, early caregiver disruption, and been exposed to violence in the home. The classes did not evidence statistically significant differences in age, race/ethnicity, sex, or self-reported substance use severity. However, there were several other differences between the classes on various system variables.

Implications. Specific to JLS-involved youth, ratings and questionnaires completed by system professionals may be particularly useful in predicting negative symptom and mental health outcomes.