Study 1
The Henggeler and colleagues (1992) study reported the results of a randomized control trial of Multisystemic Therapy (MST) in South Carolina. The Family and Neighborhood Services project randomized juveniles to a treatment-as-usual (n = 41) or the MST intervention (n = 43). They were referred primarily on the basis of a determination by the Department of Youth Services (DYS) of their imminent risk of out-of-home placement due to their involvement in serious criminal activity. The young persons in the sample had on average 3.5 previous offenses and 9.5 weeks of previous incarceration. Fifty-four percent of the sample had at least one arrest for violent crime, with 71 percent having been previously incarcerated for at least 3 weeks. The mean age of the sample was 15.2 years, and 77 percent were male. Of the sample, 56 percent were African American, 42 percent were white, and 2 percent were Hispanic. Twenty-six percent lived with neither biological parent, and families had on average 2.7 children. There were no significant differences between groups.
The MST treatment for the 43 families lasted an average of 13.4 weeks, with 33 hours of direct contact and therapists providing 24-hour case coverage. Sessions typically lasted between 15 and 90 minutes and were held weekly or as frequently as daily, depending on the treatment stage and the severity of the crises experienced. The service-as-usual group received court orders (including curfews, school attendance, and program participation), which were monitored by probation officers who met with the youths at least once a month. Failure to comply with court orders was met with a court review, and possible referral to a DYS institution, with some receiving substantive services due to a combination of family difficulties and resistance associated with mental health issues.
The study measured participants’ criminal behavior, incarceration, self-reported delinquency, family and peer relations, social skills, and problem behaviors. Pretest and posttest surveys were administered, and archival records for arrests and incarceration were collected on average at 59.6 weeks post referral. One-way analyses of variance (or ANOVA) were used to test differences between groups in arrest and incarceration data, and one-way analyses of covariance (or ANCOVA) were used to test for differences between groups on self-reported delinquency and psychosocial measures at posttest, with pretest scores as covariates. The study authors did not conduct subgroup analyses.
Study 2
Borduin and colleagues (1995) examined the long-term effects of a randomized control trial of MST and an individual therapy in Missouri. Youths in the sample were eligible if they had at least two prior arrests, were living with at least one parent figure, and had no evidence of psychosis or dementia. Ninety-two juveniles were randomized to the MST condition, and 84 were randomized to a comparison individual therapy condition. The average age was 14.8 years and 67.5 percent were male. The majority (70 percent) were white; 30 percent were African American. The youths had on average 4.2 previous arrests, and all had been previously detained for at least 4 weeks. There were no significant demographic or criminal history differences between groups.
There were, however, significant differences in the duration of treatment for program completers, with MST participants averaging 23.9 hours, compared to 28.6 hours for the comparison treatment group. The study measured participant and family member psychiatric symptomatology, adolescent behavior problems (as reported by the mother), family functioning and interactions, peer relations, and criminal activity collected from State records at 3.95 years following release from probation.
Data was analyzed using multivariate analyses of variance (or MANOVA); survival analysis was used to assess the impact of the treatment conditions on criminal activity. The study authors conducted subgroup analyses on age, race, socioeconomic status, gender, pretreatment arrests for a violent crime, and number of pretreatment arrests.
Study 3
An independent randomized clinical trial of the effects of MST was conducted by Timmons–Mitchell and colleagues (2006). The study examined recidivism at 18 months and youth functioning at 6 months posttreatment. The study consisted of 93 youths who came before a family court in a Midwestern State between October 1998 and April 2001. To be included in the study, juveniles had to have a felony conviction, a suspended committal to the DYS facility, and parental consent to participate. Participants were then randomized into a treatment MST group (n = 48) and a treatment-as-usual (TAU) group (n = 45). The youths were all on probation or had previously been on probation at the time of the study.
The average age of the sample was 15.1 years at the time of enrollment; 22 percent of the sample was female. The participants were 77.5 percent white, 15.5 percent African American, 4.2 percent Hispanic, and 2.8 percent biracial. There were no significant differences between groups in age, racial or gender make up, or in previous criminal history (age of first offense and number of previous offenses, misdemeanors, and felonies).
The treatment group was enrolled in the MST program for an average of 144.84 days. Service delivery was provided to the 48 families with a high level of fidelity to the MST program. TAU participants’ access to services was monitored by probation officers and court services. While the TAU group did have referrals to anger management, drug and alcohol, and individual and family therapy services, records indicated that attendance was sporadic and that the group had an overall low level of service use.
Arrest data was collected for the MST group for the 19 months posttreatment. For the TAU group, who do not have a clear treatment end point, arrest data was counted if the arrest occurred between 6 and 24 months postrecruitment. For the same reason, adolescent functioning data was collected at baseline, immediately after treatment, and at a 6-month follow- up for the MST group, and at baseline, and 6 months and 12 months postrecruitment for the TAU group. The arrest data was collected from family court records. Child functioning was measured using six areas of the Child and Adolescent Functional Assessment Scale: school and work, home, community, behavior toward others, moods/emotions, and substance use. Offense data was analyzed with likelihoods and relative odd ratios using logistic regression and survival analysis. Adolescent functioning data was analyzed using general linear modeling. The study authors did not conduct subgroup analyses.
