Differentiating Responder and Nonresponder Patterns of Participant Engagement and Their Link to Internalizing Symptom Reduction in a Universal Preventive Intervention

Introduction: We examine associations between patterns of participant engagement and response and ‘nonresponse’ to the Bridges to High School program (Gonzales et al., 2011). Bridges is a family-focused, universal group intervention for middle school students and their parents and aims to prevent a broad range of youth problem outcomes, including youth internalizing symptoms. We hypothesize families will demonstrate diverse patterns of attendance and that intervention response, assessed by one-year post-intervention mother and youth report of youth internalizing, will differ across patterns. We also hypothesize family, parent, youth, and program variables will predict attendance patterns and corresponding intervention response.

Methods/Results: The sample is low-income, Mexican-origin mothers and youth (n=295). Attendance is binary. The CBCL and YSR (Achenbach; 1991) assessed internalizing. We used latent class growth modeling (Nagin, 1999; Muthén & Muthén, 2010); preliminary results support a 4-class (C) solution. C1 (33%) and C2 (40%) had similar trajectories with probability of attending each session 85% or higher. Predictors did not differentiate C1 and C2; however, C2 youth had higher internalizing than C1, as well all other classes. C1 and C2 were less acculturated and had higher parent involvement in education, participation intentions, and group cohesion than other classes. C3 (11%) had a high probability of attending sessions 1-5 but initiated a progressive decline in attendance mid-program. C3 families were acculturated; mothers reported high familism values, group cohesion, and education involvement; youth were high on internalizing at pretest. C4 (16%) had a high probability of attending sessions 1-3 followed by a very steep decline in attendance at session 4. C4 had high familism values and education involvement but low participation intentions and group cohesion. Follow-up repeated measures ANOVA showed all classes decreased on internalizing from Time 1 to Time 4.

 

Conclusions: All classes showed a decline in internalizing post-intervention, but one class was higher than all others on internalizing, despite strong attendance, suggesting that this class had the poorest response to intervention components that targeted internalizing symptoms. Mothers in the poor response class reported strong participation intentions and intervention group cohesion and were highly involved in their child’s education, which may have motivated continued participation. Low acculturation differentiated the 2 high-attending classes, one of which showed the poorest response to the intervention. Forthcoming analyses will examine youth and family characteristics that may explain why these two descriptively similar classes had different intervention responses.