(PS3-37) Mediation Analysis of School-Based Universal Depression Prevention for Japanese Children

Abstract Text:

Background:
While meta-analyses show that school-based universal depression prevention programs based on cognitive behavioral therapy are effective (e.g., Werner-Seidler et al., 2017), their mechanisms remain undertested. This study examines these mechanisms by reanalyzing data from an open trial with Japanese children (Sato et al., 2009) using mediation analysis.

Methods:
Participants were 310 Japanese children (10-12 years) allocated to either a school-based prevention program based on cognitive-behavioral therapy (CBT; n = 150) or a no-treatment control group (n = 160). The CBT group received nine 45-minute sessions of psychoeducation, social skills training, cognitive restructuring, and applied practice. Outcomes were self-reported depressive symptoms assessed by the Children Depression Inventory (CDI; Kovacs, 1985) and Depression Self-Rating Scale for children (DSRS; Birleson, 1981). Mediators included social skills (positive communication skills, assertive asking skills, and assertive rejection skills: Fujieda & Aikawa, 2001) and cognitive errors (Sato et al., 2005). These measures were assessed at pre- and post-intervention. Multiple mediation analyses were conducted with group (1 = CBT, 0 = Control) as the independent variable, pre-post change in either depression measure (CDI or DSRS) as dependent variables, and pre-post changes in social skills and cognitive errors as mediators.

Results:
For CDI, the direct effect of the intervention on depression was significant in the direct effect model (β = -.14, p = .035), but non-significant in the mediation model (β = -.12, p = .103), while the total indirect effect was significant (β = -.06, p = .047), suggesting full mediation. The intervention improved assertive asking skills (β = .22, p = .001) and assertive rejection skills (β = .20, p = .005). Both positive communication skills (β = -.21, p = .007) and assertive rejection skills (β = -.17, p = .024) predicted depression reduction.
For DSRS, the direct effect of the intervention on depression was not significant in the direct effect model (β = -.11, p = .065), nor in the mediation model (β = -.04, p = .601). The mediation model showed that the total indirect effect wasn’t significant (β = -.05, p = .062), though effects of intervention on assertive asking skills (β = .17, p = .010) and of assertive rejection skills (β = -.19, p = .016) on depression reduction were significant.

Discussion:
This prevention program works primarily by enhancing social skills, especially assertiveness. Assertive rejection skills were important mediators across both depression measures. Social skills improvement, rather than cognitive change, appears to drive effectiveness. This finding is consistent with other Japanese depression prevention programs emphasizing social skills training (Ishikawa et al., 2010), whereas U.S. studies report cognitive restructuring and behavioral activation as effective mediators (Kaufman et al., 2005; Seeley et al., 2018). Future research should examine how cultural factors, participant characteristics, and prevention levels influence program mechanisms to optimize depression prevention across different settings.