The Swedish Implementation of Multisystemic Therapy for Adolescents: Does Treatment Experience Predict Treatment Adherence?

Abstract

Treatment effects may not be observed if an empirically supported intervention is not administered with treatment adherence. This retrospective study investigated how therapists' adherence to Multisystemic therapy (MST) varied during a 6-year period in Sweden. Adherence was measured using the Therapist Adherence Measure, which was provided by caregivers. The associations between treatment adherence and therapist-reported posttreatment youth outcomes were also assessed. Retrospective adherence data were obtained for 973 families with youths between 12 and 17 years old who received MST from 68 therapists divided into 10 teams. Implementation of MST occurred in 2 waves between March 2003 and August 2009. Multilevel structural equation modelling was used to evaluate family- and cohort-level predictors and outcomes of treatment adherence. The results confirm previous research: Treatment adherence predicts MST treatment outcomes. With respect to the relationship between MST treatment experience and MST treatment adherence, the analysis showed mixed results. An increase in years that a team had been active predicted an increase in therapist adherence to MST. Therapists from the 2nd implementation wave had higher therapist adherence rates than therapists from the 1st implementation wave. Nevertheless, a therapists' experience with MST did not predict MST treatment adherence. The results suggest that the found links among treatment experience, treatment adherence, and treatment outcomes provide support for the hypothesis that the collective experience of the team members promotes treatment adherence. However, results are mixed and further research is needed.

Notes

Note: Parameter estimates were obtained using the robust maximum likelihood estimator. Estimates for posttreatment outcomes are logistic regression coefficients. Labels of hypothesized pathways are included in parentheses. Standardized effects for continuous predictors are path coefficients standardized with respect to both predictor and outcome. Standardized effects for dichotomous predictors are path coefficients standardized with respect to the outcome only.