Multilevel Predictors of School-Based Mental Health Service Use Among Adolescents

Background: Adolescence is a critical time for growth and development, with many physical, social, and emotional transitions. Evidence suggests that youth progressively experience mental health issues and often discover mental health symptoms for the first time in adolescence. Despite increased vulnerability to mental health needs, there is a critical gap in mental health services for adolescents. Schools have come to the forefront of providing mental health services, as adolescents spend most of their day in the academic setting. Informed by Andersen’s Behavioral Model of Health Care Utilization as a framework, the current study aimed to examine individual- and school-level factors that relate to school-based mental health service use.

Methods: This study employed data from the National Longitudinal Study of Adolescent to Adult Health (Add Health), one of few national data including school-based mental health service use and school-level characteristics. The study recruited middle and high school students between 12-17 years old across 172 schools (N = 14,464). The dependent variable was the use of School-based Mental Health Services (SMHS) in the past year (0=no, 1=yes). Individual-level predictors included sociodemographic characteristics, enabling factors (social support, school connectedness, barriers to healthcare, location of SMHS), and need (anxiety and depressive symptoms, risk-taking behaviors). School-level factors included the school race ratio, school’s geographic location, and school size). We conducted a multilevel logistic regression model clustering at the school level..

Results: The sample was 15.17 years old on average, with even distribution of sex (50.3% male), and a majority of adolescents identified as White (55.7%), followed by Hispanic (18.8%) and Black (15.5%). Approximately 11% of variation in SMHS utilization is explained at the school-level. At the individual level, the odds of SMHS use was significantly associated with anxiety (OR = 1.11, p < .001), depressive symptoms (OR = 1.04, p < .001), and risk-taking behaviors (OR = 1.05, p < .001). The odds of SMHS was higher for female (OR = 1.68, p < .001), Black compared to White (OR = .75, p = .009), and public assistance recipients (OR = 1.35, p = .007). Interestingly, SMHS use was negatively associated with school connectedness (OR = 97, p = .03). At the school level, the odds of SMHS use was lower for adolescents attending a school with a majority of white students decreased one’s odds of using services (OR = .72, p = .04). Also, adolescents were more likely to use SMHS when their SMHS were available within their school or another place within school district, compared to referrals to outside agencies (OR = 1.51, p = .008).

Discussion/Implications: This study is among the first empirical studies providing important evidence regarding individual and school-level predictors of SMHS utilization. The findings corroborate the application of Andersen’s theoretical framework and suggest the importance of school-level contexts in school mental health. The study provides suggestions and implications for the continued need for school-based mental health professionals and school institutions’ efforts to meet students’ mental health needs and increase their access to services.