Enhancement of Training As Usual in a Child Welfare Setting with a Change Agent

Background/Purpose: This study applied Roger’s (2003) diffusion of innovation theory to test the effect of adding interaction with a change agent to a workshop-based training as usual (TAU) on case managers’ and therapists’ use of a parent management training (PMT) intervention with foster parents.  Despite considerable investment in workforce development, the training models most commonly used in child welfare settings have little effect on practice behaviors, and few studies have examined the individual-level and social factors supporting use of a new practice (Proctor et al., 2009).  Rogers’ theory posits that innovations are diffused through social interactions as well as aspects of the practice and setting. A key role is played by “change agents,” who are similar to other providers but have personal qualities that increase their influence. This study used an experimental design to test the effect of change agent interactions on use of new practices after TAU. Attitudes about evidence-based practices (EBPs) and level of contact with change agent were expected to moderate the enhanced training effect.

Method: An experimental design compared use of the intervention after case managers and therapists (staff; N = 57) were randomized to receive either TAU or “enhanced training,” which followed TAU with change agent interactions. The change agent was housed at the agency part time and engaged in casual interactions with “enhanced training” staff about the intervention and held events where the intervention was discussed.  Use of the intervention was collected longitudinally with up to five in-person interviews over a period of 15 months (188 observations total).  Use was measured by the number of times staff used study materials (manuals or handouts) to talk with foster parents about PMT components. Staff contact with the change agent was assessed through fidelity measures maintained between T2 and T3. Attitudes about EBPs were assessed with Aaron’s (2004) EBPAS. Zero-inflated Poisson mixed regression models were used to test between-group differences at each time point and moderation by the EBPAS. Extent of contact with the change agent T2-T3 was expected to be correlated use at T3 but not at T2.  

Results: Use was significantly higher in the change agent condition at Times 2-4, as expected (B= 1.28, 1.67, and 1.56 at T2, T3, and T4 respectively, ps < .05). At T5, after the change agent had left the agency, use diminished and there was no significant difference between the groups.  Attitudes toward EBPs did not moderate the effect of the change agent on use. Contact with the change agent T2-T3 was associated with subsequent use at T3 (r = .57, p < .01), as expected, supporting that level of contact with the change agent was the active component of the experimental condition.

Conclusions:  Results support the positive effect of interactions with a change agent on use of a novel intervention. Attitudes about EBPs did not impact this effect, but level of contact was predictive of use, and ongoing contact was necessary for new practices to be sustained. Implications for the development of effective training models will be discussed.