Re-Designing Community Mental Health Services for Urban Children: Supporting Schooling to Promote Mental Health

Improving the accessibility and effectiveness of community mental health services for children has been a national concern for more than a decade (National Advisory Mental Health Council, 2001). In a seminal study, secondary data analysis of three nationally representative household surveys indicated that nearly 80% of low-income youth in need of mental health services did not receive services in the preceding 12 months, with rates approaching 90% for uninsured families (Kataoka, Zhang, & Wells, 2002). Lack of access to services is especially problematic in urban, low-income communities with high rates of nonattendance at initial appointments and rates below 10% for attendance at as few as four sessions (see McKay & Bannon, 2004). Infrequent use of mental health services has been attributed to stigma (Dempster, Wildman, & Keating, 2013) and concrete obstacles, such as inaccessible locations, lack of information about services, and social isolation (Harrison, McKay, & Bannon, 2004). Concentrated urban poverty also is associated with high risk of substantial mental health difficulties for youth (Cappella, Frazier, Atkins, Schoenwald, & Glisson, 2008). A longitudinal analysis of a large nationally representative sample of youth indicated a robust relation between neighborhood disadvantage and conduct problems over and above a series of family and individual risk factors (Goodnight, et al., 2012). Relatedly, exposure to community violence, affecting almost 80% of urban children (U.S. Department of Justice, 2003), is associated with poor academic performance (McCoy, Roy, & Sirkman, 2013) mediated by depression and disruptive behavior (Borofsky, Kellerman, Baucom, Oliver, & Margolin, 2013). A public health framework offers promise for organizing the design and delivery of more accessible and appropriately targeted services to children living in urban poverty. Within a public health framework, universal intervention strategies are deployed to attenuate risk factors and related behavior problems, while targeted interventions are simultaneously deployed for high-risk cases (Stiffman, Stelk, Evans, & Atkins, 2010). If delivered in those contexts naturally inhabited by children and families – primarily school and home – and focused on specific aspects of those contexts affecting child learning and behavior, service models encompassing such interventions also could be more effective and sustainable (Atkins & Frazier, 2011). In this study, we examined a service delivery model, Links to Learning (L2L), which involved the integration and delivery of universal and targeted interventions focused on supporting schooling for children with disruptive behavior disorders living in urban low-income communities. School-Based Mental Health Services Schools are the de facto providers of mental health services for children and youth (Farmer, Burns, Phillips, Angold, & Costello 2003; Green et al., 2013), providing an estimated 70% to 80% of psychosocial services (Rones & Hoagwood, 2000). However, the primary modality for school-based services, individual counseling (Foster et al., 2005), is largely ineffective for children with disruptive behavior disorders, which comprise the majority of school referrals (Farmer, Compton, Burns, & Robertson, 2002; Foster et al., 2005). This is especially evident for children attending schools in low-income urban communities. In a recent meta-analysis examining school-based mental health programs for low-income, urban youth (Farahmand, Grant, Polo, Duffy, & DuBois, 2011), null effects were found for most outcomes (mean ES = .08), and negative effects were found for programs focused on externalizing behaviors (ES = −.11). The authors suggested that these findings reflect a lack of attention to the many stressors apparent in low-income urban schools and proposed that effective services would require an integration of the school ecology into program planning and implementation. A second meta-analysis of programs based in community mental health settings serving low income urban youth (Farahmand et al., 2012) found positive effects for programs that supported parents or provided other community supports (mean ES = .38), and null effects for programs focused on direct services to youth (mean ES = .03). These findings suggest that individually focused services are contraindicated for low-income youth with disruptive behavior and that interventions are likely to be more impactful when they can be deployed in, and alter, family and community contexts. In the present study, we implemented and examined a model in which community mental health staff worked directly with parents and teachers in low-income urban schools to enhance children’s school success. Predictors of Children’s Learning: Teachers and Parents Informed by evidence supporting the effectiveness of focusing mental health services on the empirical predictors of youth offending (Henggeler, Schoenwald, Borduin, Rowland, & Cunningham, 1998), we constructed a model focusing mental health services on the empirical predictors of children’s learning to impact children’s school success (see Cappella et al., 2008). An extensive literature documents that children’s academic learning is compromised in urban low-income schools, with profound and growing gaps between poor and nonpoor U.S. children (Reardon, 2011). This has important implications given that academic achievement is a hallmark of children’s sense of competence (Masten & Curtis, 2000), and critical to social and emotional adjustment (Roeser, Eccles, & Freedman-Doan, 1999). Academic achievement can operate as a protective factor for urban children (Freudenberg & Ruglis, 2007), and is associated with positive relationships with peers, teachers, and parents, and improved classroom behavior (Atkins, Hoagwood, Kutash, & Seidman, 2010). In addition, a direct focus on schooling by mental health providers could bridge educational and mental health systems, and provide additional resources to struggling urban schools (Ringeisen, Henderson, & Hoagwood, 2003). Reviews of the educational literature reveal that teachers and parents contribute uniquely to children’s learning. Specifically, three components of teacher practices most significantly impact children’s learning: Effective instruction, classroom management, and teacher outreach to parents (Stringfield, 1994). Similarly, parent communication with teachers, homework support, and reading at home are associated with improved learning (Jeynes, 2005). These classroom and family predictors of learning were the focus of the mental health service model developed for this study (Cappella et al., 2008). Diffusion of Innovation: Teacher Key Opinion Leaders and Parent Advocates Diffusion theory posits that key opinion leaders (KOLs) spread innovations as a function of their influential role within their social network (Rogers, 1995). In the first study applying diffusion theory to urban schools (Atkins et al., 2008), KOL teachers, working with mental health providers (MHPs), promoted higher rates of teachers’ self-reported use of recommended interventions than consultation from MHPs without KOL support. These results supported an expanded role for KOL teachers for the dissemination of school-based mental health interventions. Similarly, involving parents with similar characteristics and experiences as the parents of children referred for services can reduce stigma, enhance participation in services, and influence behavior change due to shared experiences and reduced social distance (Frazier, Abdul-Adil, Atkins, Gathright, & Jackson, 2007; Hoagwood et al., 2010). The Current Study We examined the extent to which a mental health model, Links to Learning (L2L), focused on the key predictors of student learning and delivered by community mental health providers aligned with parent advocates and KOL teachers, would lead to greater reductions in children’s disruptive behavior at home and school as compared to mental health services-as-usual (SAU). Moderators included baseline child and family characteristics. This three-year longitudinal study utilized a multi-method, multi-informant design consisting of classroom observations, teacher report, parent report, and direct assessment of academic performance with random assignment of schools to L2L or assisted referral to community based SAU.