Characteristics of Effective School-Based, Teacher-Delivered Mental Health Services for Children Using Meta-Regression

Background

Research on school-based services demonstrates that mental health services are provided to vulnerable children and adolescents who may not otherwise have access to such services. Given the accessibly and integration of mental health in schools, it is important to understand various factors impacting the effectiveness of school-based mental health services, like providers’ background or delivery format. Although previous research has explored the benefit of using teachers in school-based mental health services, the findings do not specify what characteristics of teacher-delivered services are effective or ineffective. To inform the implementation of teacher-delivered services and how social workers consult with teachers, this meta-analysis evaluated various program characteristics of teacher-delivered, school-based mental health services in relation to the clinical effectiveness of these services.

Methods

This meta-analytic study used data that searched across 9 electronic databases and 19 intervention websites for studies published from 2000 to September 2016. The primary search terms were: “school*”, “intervention*”, “random*”, and “teacher*.” The data consisted of randomized control trial studies that examined the clinical effectiveness of teacher-delivered school-based mental health interventions for students’ internalizing and/or externalizing behavior problems. To be included in the data set, a study needed to include sufficient statistical data to calculate at least one effect size. The moderators evaluated from this data set were: internalizing and externalizing outcomes, intervention context, tiers of intervention, duration of intervention and type of measurement used in primary studies. The data used in this analysis included 123 effect sizes across 24 primary studies. We used robust variance estimation in meta-regression to estimate an overall treatment effect as well as for moderator analysis.

Results

Using a random-effects model, the overall intervention effect for internalizing disorders was statistically significant (d=.13, 95%, confidence interval [CI]=[.03, .22], p=.02). Using the same model, the overall intervention effect for externalizing behaviors was not statistically significant (d=.03, p=.23). Further regarding internalizing disorders, tiers of intervention significantly moderated intervention effect. To demonstrate, classroom-based interventions reported significantly greater intervention effect for internalizing behaviors than small-group-based interventions (b₁=0.30, p=.02). Subgroup analysis indicated an overall significant intervention effect for internalizing behaviors across classroom-based interventions (d=.22, 95% CI [.11, .33], p< .001). In comparison the overall effect of small group-based interventions for internalizing behaviors was not significant (d= -.08, p=.25). Intervention length was a moderator for internalizing disorders. The overall effect for “medium” length treatment was statistically significant (d=.19, 95%, CI [.02, .36], p< .05). However, the effect was not found to be statistically significant for “brief” interventions (d=.24, p=.09).

Conclusion

Results of this study align with previous literature, showing tier of intervention as a moderating variable for internalizing disorders and specifically building a case that teachers are more effective in delivering classroom-based Tier 1 interventions for internalizing disorders than other intervention tiers. This could be because many Tier 1 interventions align more with teachers’ professional skills. Tier 1 interventions are often taught from pre-set curriculum and contain psycho-education. This study also has implications on length of treatment, that “medium” dosages are more affective in the context of teacher-delivered interventions.