Individual and Institutional Correlates of Health and Social Service Use in Schools

Background and Purpose:  Drawing on an ecological framework for adolescent service utilization (Cauce, et al., 2002) this study considers individual and institutional correlates of health and social service use in schools. Previous research in this area has collapsed all types of school-based services together into one measure, and/or failed to account for the dependence among observations that occurs when research subjects are nested within clusters like schools or districts. This study overcomes such methodological limitations by using multilevel modeling techniques to analyze nuanced administrative data from 15 school-based health and social service programs in one large urban school district. Specifically, we examine the relationship between student-level characteristics, school-level factors, and use of three different types of services: primary prevention (health and wellness education or promotion), medical care (first aid, reproductive health, dental or vision services), and psychosocial interventions (individual counseling, specialty mental healthcare or case management).

Methods:  The administrative data sample includes all high school students (n= 6,696) who participated in school-based services in this district during the 2008-2009 school year. We used multilevel models to estimate relationships between student characteristics (race, gender, English language fluency, zip-code level poverty rate, and age) school-level factors (demographic features of the student body and suspension rates as a proxy for negative school climate), and use of each service modality. All models controlled for these individual and institutional covariates.

Results:  A higher proportion of students received psychosocial interventions (68%) than prevention services (30%) or medical care (47%). (Students could access multiple types of services). Students’ race, gender, age, and English fluency were significantly associated with use of all three service categories, but zip-code level poverty rates were not. The strength and direction of these relationships depended on the service modality. For example, Black (OR=2.25, p<.001), Latino (OR=1.98, p<.001), White (1.70, p<.001) and Multiracial/Other youth (OR=2.20, p<.001) experienced greater odds of receiving psychosocial interventions than Asian students. This trend was reversed for prevention programs; Asian students were more likely to utilize this service modality. At the school-level, significant relationships were observed between students’ use of psychosocial interventions and the proportions of Black and Latino youth in the school (OR 7.15 p<.001), students receiving free lunch (OR 3.25, p<.000), and negative school climate (OR .001, p<.001). There were no significant relationships between school-level factors and participation in prevention services; results were mixed for medical care.

Conclusions and Implications: These findings extend ecological theories of adolescent service utilization by demonstrating significant relationships between contextual factors and service use. This study also replicates past research demonstrating the widespread use of psychosocial interventions targeting individual youth and the more limited deployment of prevention modalities in educational settings, particularly in schools serving low-income students of color. Given growing consensus among school social work scholars that the most effective service models reflect an ecological orientation and full continua of prevention to intervention activities, these findings may indicate a need for additional resources to support and sustain prevention programming in schools.