Evaluating the Impact of the My Life Intervention for Older Youth in Foster Care: Moderation of Proximal Outcomes By Risk

Background/Purpose: The My Life model is a year-long weekly coaching/mentoring intervention for older foster youth that uses structured didactic, experiential, and relationship-building activities to help youth develop skills that can increase engagement in transition planning and improve transition outcomes (Powers, Geenen, Powers, et al., 2012). Randomized trials demonstrated hypothesized effects on youth self-determination skills and self-efficacy. This analysis specifically reports on the moderation of these proximal intervention outcomes by selected risk factors related to foster care placement.

Methods: 288 foster youth (mean age=17.31, SD=.61) were randomized to intervention condition, then re-assessed at 12 months (n=243) and at 24 months (n=226, with 78% retention overall). Participants are 53% female, 19% Hispanic/Latino, and 46% White, 16% Black/African American, 19% multi-racial, and 19% other. 59% received special education services, including those with developmental disabilities (22%), 45% were in non-relative foster homes, 23% in relative/kin placement, and 32% were in more restricted settings. Outcomes were analyzed with a two (treatment group) by three (time point) repeated measures design, using the GLMM growth curve model to test for main effects of group, time, and group-by-time interaction on measures including the Arc Self-Determination Scale, Career Decision Self-Efficacy Scale, and the My Life Activity Checklist (results reported separately). To understand moderation of intervention effectiveness on the outcomes, we analyzed variation by foster placement stability, restrictions on self-determination in the current placement, and youth report of symptoms of trauma-related stress. These baseline predictors were included in the repeated measures analysis of group-by-time effects, with additional contrasts for three levels of each predictor.

Results: The moderators had consistent statistical associations with group differences on the outcomes. Trauma symptoms were measured by the 25-item Child Report of Post-traumatic Symptoms (CROPS: Greenwald, 1999) (sample mean = 16.24, SD=8.71). Broadly, greater response to treatment was found for participants who reported low-to-moderate traumatic stress levels prior to participation. Placement stability was based on the days-per-placement for the current foster care episode (mean = 511.52 days [SD=616.31]) and is expected to impact youth and caregiver capacity to engage in the intervention. Overall, the intervention seemed to be more effective for young people who had high (about 1000 days) or average stability. Lastly, placement restrictiveness was measured with items from the Restrictiveness Evaluation Measure for Youth (Rautkis et al., 2009) (mean=5.05, SD=2.14, on 3-15 point scale). Moderation by restrictiveness showed that response to treatment is specifically observed among youth with low-to-average placement restriction at baseline. (Detailed results will be presented.)

Conclusion/Implications: Compared to the control group, youth in the treatment group who entered the study having experienced average-to-high placement stability, low-to-average restriction in their current placement, or low-to-average baseline post-traumatic stress, were most likely to show statistically significant improvement on outcomes related to self-determination skills and self-efficacy. These findings inform research and practice related to this specific intervention, but also offer evidence that may translate to other interventions about how foster care factors may inhibit youth skill-building during times of crisis or when there are above-average restrictions on youth opportunities to learn and apply skills.