An independent randomized clinical trial of multisystemic therapy with non-court-referred adolescents with serious conduct problems

Adolescent conduct problems exact substantial personal and social costs, making effective treatments essential. One of the most widely disseminated programs for treatment of adolescent conduct problems is Multisystemic Therapy (MST; e.g., Henggeler, Cunningham, Schoenwald, & Borduin, 2009). MST is a family-focused, home-based treatment for adolescents with serious antisocial behavior and emotional disturbance. MST is derived in part from Bronfenbrenner's (1979) theory of social ecology, and views individuals' behavior as embedded in a complex network of interconnected social systems. Treatment is multi-faceted, targeting the behavior of individuals, family, peer, and other key systems such as the school (Henggeler et al., 2009). In the first MST outcome study of youth with serious antisocial behavior, Henggeler et al. (1986) reported improved family relations and decreased behavior problems for youth receiving MST relative to a comparison group. Subsequent evaluations have led to MST being generally accepted as an efficacious treatment (e.g., U.S. Department of Health and Human Services, 2001; although see Littell, Popa, & Forsythe, 2005, for a contrary opinion) and widely adopted in North America and Europe (Henggeler, 2011). For a number of years, however, it has been suggested that results of evaluation studies may be unintentionally influenced by investigators' allegiance to the treatment. Several studies (e.g., Houts, Berman, & Abramson, 1994) have found that the magnitude of effects of psychosocial interventions is related to estimates of investigator allegiance even when allegiance is assessed in prior publications (e.g., Robinson, Berman, & Neimeyer, 1990). There have been several evaluations of MST for adolescent conduct problems in which the developers of MST have not been co-authors on study reports or directly involved in the evaluation. In general, these studies have found relatively modest effects. Ogden and Halliday-Boykins' (2004) study is generally considered one of the first independent evaluations of MST, although its second author has published extensively with the developers of MST. Ogden and Halliday-Boykins (2004) was conducted in Norway, with youth with serious antisocial behavior randomly assigned to receive MST or usual child welfare services. Compared to the control group, MST was associated with a significantly greater decrease in internalizing symptoms and a marginally significant greater decrease in externalizing symptoms. The magnitude of effects of Ogden and Halliday-Boykins (2004) as well as that of Sundell et al. (2008), an independent evaluation of MST conducted in Sweden, was more modest compared to other trials of MST (e.g., Borduin et al., 1995). This may have been due to the comprehensive nature of the “usual services” provided by Norway and Sweden's child welfare system. An independent evaluation of MST was also conducted in the U.K. with court-referred juvenile offenders (Butler, Baruch, Hickey, & Fonagy, 2011). Youth were randomly assigned to receive MST or comprehensive services-as-usual that included substance abuse treatment, anger management training, social skills problem-solving, and other services. Butler et al. (2011) found that MST was associated with a significantly greater reduction in the number of nonviolent offenses at 18 months, and a non-significant trend favoring MST in regard to violent offenses. The MST group also showed significantly greater improvement in delinquent behavior as reported by parents and youth, and for aggressive behavior as reported by parents (but not youth). In the first independent trial of MST conducted in the U.S., Timmons-Mitchell, Kishna, Bender, and Mitchell (2006) evaluated 93 adolescents involved with the juvenile justice system. Re-arrest rates for adolescents who received MST were significantly lower than for adolescents who received treatment as usual (67% vs. 87%), although time to re-arrest did not differ significantly for the two groups. Post-treatment scores for 4 of 6 subscales on the CAFAS functioning measure (Hodges & Wong, 1996) were significantly better for the MST group compared to the treatment-as-usual group. However, the magnitude of these effects was lower than found in most previous trials conducted by the developers of MST. Moreover, outcome data were obtained from a single secondary source (i.e., court records coded by research assistants), which limits the generalizability of the results to broader adolescent conduct problems. The purpose of the present study was to conduct a randomized, independent evaluation of MST in the United States. Strengths of the study included a multi-informant, multi-system assessment focused on psychopathology, interpersonal functioning, the home and school environment, and criminal behavior assessed through court records. Participants were randomly assigned to receive MST or services as usual. The present study is the first randomized, independent assessment of MST conducted in the U.S. involving adolescents with conduct problems who were recruited from a source other than the legal system. In the present study, treatment fidelity was assessed through expert coding of audiotaped sessions as well as reports from parents and MST supervisors. In most studies of MST, treatment fidelity has relied on parent reports rather than coding of session audiotapes (e.g., Sundell et al., 2008), due to the costs. In the present study, we hypothesized that MST would be associated with significantly greater rates of improvement on our primary outcome measures than the control group, and that arrest rates would be significantly lower among adolescents who received MST than those who did not.