Is There a Scale up Penalty? Attenuation of Program Effects in Large Scale Implementation of the Oregon Model of Parent Management Training (PMTO)

PRESENTAION TYPE: Individual Paper

CATEGORY/THEME: Prevention research/ Large-scale dissemination

TITLE: Is There a Scale up Penalty? Attenuation of program effects in large scale implementation of the Oregon model of Parent Management Training (PMTO)

ABSTRACT BODY:

Introduction: Desirable effects of intervention programs may attenuate when they are scaled up from the effectiveness phase to broad scale dissemination. Research studies on the cost-benefit of preventive parenting programs, have shown a scale up penalty from a low 25 % to a high of 50 %. After the introduction of PMTO in Norway in 1999, an empirical study was conducted in order to examine the attenuation of program effects between the effectiveness and the large scale implementation phase.

 

Methods: Data in this study was collected from the PMTO Effectiveness study and the Dissemination/fidelity study, partly funded by the US National Institute for Drug Abuse (NIDA). Participants were clinically referred children with conduct problems, aged 3 to 12 years, and their caretakers. Relevant covariates, known to affect parent training interventions, were entered into the Analysis of Covariance ITT models (ANCOVA). The outcomes of the study were effect size differences between the two phases of implementation.   

 

Results:

In the baseline analysis (ANOVA) there were greater heterogeneity in both service providers and target population in the dissemination phase compared to the effectiveness phase. The ANCOVA models for parent reported outcomes gave the following results: externalizing behavior problems F(4,198) =.562 , p= .454, =.003, internalizing behavior problems F(1, 207) = 3.479, p = .064, =.017 , and social competence F(1,215) = 3.604, p = .59, = .017. The corresponding teacher reported outcomes were; externalizing behavior F(1,190) = 1.394, p = .239, = .007, internalizing behavior F(1,190) = .435, p = .510,  = .002, and social competence F(1,182) = .793, p = .374,  = .004. All effect sized differences between the samples were small or only marginally significant.  

Conclusions:

Despite the larger heterogeneity in both service providers and in the target population, effect size differences between the effectiveness and national implementation phase were small, ranging from 0.2 % to 1.7 %. The minor loss of program effects may be explained by the Norwegian political and social context, and the National Implementation Team who planned, carried out and monitored the scaling up process, combining a top – down and bottom –up organizational approach.