Exploring Heterogeneity in the Penn Resiliency Program's Effect On Depressive Symptoms: Who Benefits & Who Does Not?

Introduction: The Penn Resiliency Program (PRP) is a cognitive-behavioral depression prevention program for adolescents. On average, youth who participate in PRP report significantly lower levels of depressive symptoms through 12 months of follow-up (Brunwasser, Gillham, & Kim, 2009). However, PRP’s effects have been inconsistent. Two randomized-controlled trials (RCTs) found that PRP yielded benefits for some youth, but not others (Gillham et al., 2007, 2012). If PRP is to be implemented broadly, it is crucial that we are able to distinguish between youth who are most (and least) likely to benefit. This poster will provide new insights into the characteristics of responders and non-responders.

Methods/ Results: Data from two large randomized-controlled effectiveness trials (Gillham et al., 2007, 2012) were aggregated to provide optimal statistical power (N = 874). Both studies evaluated the effectiveness of PRP when delivered in public middle schools by school personnel. Students (ages 10-14) completed 6 commonly-timed assessments across the two studies spanning 24 months of follow-up. Preliminary analyses found that the effect of PRP on depressive symptoms (Children’s Depression Inventory) did not differ across the two studies. Growth mixture (GM) analyses will be used to model heterogeneous growth in depressive symptoms (the primary DV) and hopelessness (a hypothesized mediator). First, we will evaluate whether there are distinct latent trajectory classes within the overall sample reflecting differential symptom development following the intervention. We expect multiple trajectory classes to emerge characterized by different growth patterns: sustained improvement (full-responders), non-sustained improvements (partial responders), and no improvement (non-responders). Second, we will summarize the characteristics of students within latent classes and evaluate predictors of class membership (demographics and baseline mental health variables). Third, we will evaluate whether PRP’s effects differ across latent trajectory classes, which could help us better characterize students who benefit most/least from the intervention. Finally, within all latent trajectory classes, we will examine whether or not PRP has an effect on hopelessness, and if so, whether hopelessness mediates PRP’s effect. Analyses will be conducted in accordance with recommendations provided by Muthén et al. (2002, 2009).

Conclusions: The findings from this study may greatly improve our knowledge of who is most and least likely to benefit from participating in PRP. This information could lead to screening and recruitment strategies that maximize the overall efficacy and efficiency of the program. Additionally, the findings may lead to revisions that improve efficacy for non-responders.