Heterogeneity in Response during Multisystemic Therapy: Exploring Subgroups and Predictors.

Multiple studies have shown that Multisystemic Therapy (MST) is, at group level, an effective treatment for adolescents showing serious externalizing problem behavior. The current study expands previous research on MST by, first, examining whether subgroups of participants who respond differently to treatment could be identified. Second, we investigated if the different trajectories of change during MST could be predicted by individual (hostile attributions) and contextual (parental sense of parenting competence and deviant and prosocial peer involvement) pre-treatment factors. Participants were 147 adolescents (mean age = 15.91 years, 104 (71%) boys) and their parents who received MST. Pre-treatment assessment of the predictors and 5 monthly assessments of externalizing behavior during treatment took place using both adolescent and parents' self-reports. Six distinct subgroups, showing different trajectories of change in externalizing problem behavior during MST, were identified. Two of the 6 trajectories of change showed a poor treatment response, as one class did not change in externalizing problem behavior and the other class even increased. The remaining 4 trajectories displayed a positive effect of MST, by showing a decrease in externalizing behavior. Most of these trajectories could be predicted by parental sense of parenting competence. Additionally, lower involvement with prosocial peers was a predictor of the group that appeared to be resistant to MST. Adolescents do respond differently to MST, which indicates the importance of personalizing treatment. Protective factors, such as parental sense of parenting competence and prosocial peers, seem to require additional attention in the first phase of MST.

Psychopathy as predictor and moderator of multisystemic therapy outcomes among adolescents treated for antisocial behavior.

OBJECTIVE: The aim of the present study was to determine whether psychopathic traits act as a predictor and/or moderator of the effectiveness of Multisystemic Therapy (MST). METHOD: The sample included N = 256 adolescents (188 boys and 68 girls) referred for conduct problems, randomized to MST or Treatment As Usual (TAU). The mean age was 16 years (SD = 1.31). Assessments were carried out before and immediately after treatment (6 months later). Three psychopathic traits (callous/unemotional traits, narcissism, and impulsiveness) were assessed with parent reports. Adolescents and parents were informants on externalizing problems. RESULTS: MST was more effective than TAU in decreasing externalizing problems for the "lower callous/unemotional" and "lower narcissism" group, but not for the "high callous/unemotional" and "high narcissism" group (moderators). Impulsiveness was found to predict more post-treatment externalizing problems rated by adolescents (predictor), but not more post-treatment externalizing problems rated by parents. CONCLUSIONS: These findings point out the clinical relevance of adequately assessing psychopathic traits in adolescents referred for treatment of antisocial behaviour, and identifying those adolescents who show high levels of these traits. It is important to tailor MST specifically to meet the needs of juveniles with high levels of callous/unemotional traits and high levels of narcissism to obtain the same level of effectiveness as with juveniles scoring lower on these traits.

Within-intervention change: mediators of intervention effects during multisystemic therapy.

OBJECTIVE: The present study tested the hypothesis that improvements in parental sense of competence during multisystemic therapy (MST) lead to positive changes in parenting, which in turn lead to a decrease of adolescent externalizing problems. Mediational models were tested separately for 3 dimensions of parenting (positive discipline, inept discipline, and relationship quality) that are targeted in MST. Each model included "3-path mediation," in which 2 mediators (i.e., changes in parental sense of competence and parenting dimension) intervene sequentially between the independent (i.e., intervention status) and dependent variable (i.e., change in externalizing problems). METHOD: Participants in this randomized controlled trial were 256 adolescents and their families who received either MST (n = 147) or treatment as usual (n = 109). In addition to pre- and postintervention assessments, 5 monthly within-intervention assessments took place. RESULTS: Both preintervention-postintervention comparison, through analysis of covariance, and comparison of trajectories during intervention, through latent growth modeling, showed that MST enhanced growth in parental sense of competence and positive discipline, led to no deterioration in relationship quality, and resulted in a decrease in adolescent externalizing problems. The results supported a sequential pattern of change for positive discipline: Changes in parental sense of competence predicted changes in positive discipline, which in turn predicted decrease in adolescent externalizing problems. No support was found for mediated effects of inept discipline and relationship quality. CONCLUSIONS: The results affirm the importance of directly targeting parental sense of competence and positive discipline in future interventions aimed at decreasing adolescent problem behavior.

Perceptions of problem behavior in adolescents' families: perceiver, target, and family effects.

Considerable research has focused on the reliability and validity of informant reports of family behavior, especially maternal reports of adolescent problem behavior. None of these studies, however, has based their orientation on a theoretical model of interpersonal perception. In this study we used the social relations model (SRM) to examine family members' reports of each others' externalizing and internalizing problem behavior. Two parents and two adolescents in 69 families rated each others' behavior within a round-robin design. SRM analysis showed that within-family perceptions of externalizing and internalizing behaviors are consistently due to three sources of variance; perceiver, target, and family effects. A family/contextual effect on informant reports of problem behavior has not been previously reported.

Level validity of self-report whole-family measures.

This article introduces an approach to testing the level validity of family assessment instruments (i.e., whether a family instrument measures family functioning at the level of the system it purports to assess). Two parents and 2 adolescents in 69 families rated the warmth in each of their family relationships and in the family as a whole. Family members' ratings of whole-family warmth assessed family functioning not only at the family level (i.e., characteristics of the family as a whole) but also at the individual level of analysis (i.e., characteristics of family members as raters), indicating a lack of level validity. Evidence was provided for the level validity of a latent variable based on family members' ratings of whole-family warmth. The findings underscore the importance of assessing the level validity of individual ratings of whole-family functioning.