The Relationship Between Common Factors and Program Factors in Parent Management Training, the Case of Working Alliance and Treatment Fidelity

Introduction: Common factors, as opposed to specific program factors, are non-specific treatment ingredients that are unrelated to method and therapeutic orientation. The extent to which common factors, or specific factors account for treatment effects is an issue of great debate. In particular, the role of alliance as a common factor is interesting in regards to evidence-based treatments, in which specific treatment factors are empirically identified and supported. The present study investigated the role of work alliance between parents and therapists in Parent Management Training- the Oregon model (PMTO) and how alliance relates to problem behavior in children.

Method: Participants were 331 Norwegian parents who rated client-therapist working alliance at three time points (session 3, 12, and 20). Competent adherence to the PMTO treatment protocol was assessed by PMTO specialists from evaluations of videotaped therapy sessions. Parents and teachers reported children’s problem behaviors at baseline and at the end of therapy. Structural equation modeling was used to analyze the repeated measures data.

 

Results: Parents reported high and stable levels of alliance and fidelity from pre intervention assessment to post intervention assessment, with no correlational or direct relations between the two. Treatment fidelity predicted reductions in parent reported externalizing behavior, whereas working alliance was related to less change in problem behavior. Alliance and fidelity were unrelated to teacher reported behavior problems.

 

Conclusions: Using a longitudinal design, we investigated the development and relation between treatment fidelity and parent-therapist alliance, and how these predicted externalizing problem behaviors in children participating in PMTO. In the present study, results suggest that statistically, treatment fidelity and working alliance seem to be separate processes, although being theoretically related process variables. Furthermore, fidelity and alliance both predicted parent reported problem behavior, but whereas fidelity predicted change in problem behavior, alliance was related to less change in problem behaviors. Both fidelity and alliance were unrelated to teacher reported problem behavior. More research is needed to investigate whether these findings can be replicated and extended beyond PMTO.