Abstract
The purpose of this study was to examine differences in discharge outcomes between latent classes of youth in psychiatric residential treatment. The mediating effect of family therapy, behavioral management incidents, and length of stay on class membership and treatment outcomes were examined. The sample included 447 youth assigned to one of four classes. Guided by Thornberry and Krohn’s (2005) interactional theory of continuity and change, change in functional impairment was predicted based on the composition of risk versus protective factors that comprised the latent classes. A manual 3-step approach was used to fit a latent class mixture model and estimate conditional effects on impairment at discharge. A mediation model was used to examine indirect effects of treatment factors on outcomes between latent classes. The results showed that classes with lower-level risk factors and more protective factors experienced significantly greater reductions in impairment on average. Treatment outcomes were mediated by behavioral management incidents but not length of stay or the number family therapy sessions. The results demonstrate the usefulness of person-centered approaches for conducting subgroup analyses in residential care outcomes studies; highlighting differences in outcomes between groups and treatment factors that may mediate these differences.