A Longitudinal Analysis of Trajectories and Predictors of Fidelity Using the SafeCare Parenting Model

Evidence-based interventions are increasingly used in public health, mental health, and child welfare settings in an effort to improve services and outcomes for families. Even if an evidence-based intervention has been demonstrated in clinical trials to have a positive impact on consumer outcomes, it must be implemented appropriately to be fully effective. There is a growing body of literature studying the factors that impact implementation success or failure, as well as identifying various measurable domains of implementation. One such domain is fidelity, which has been defined as, “the degree to which an intervention was implemented as it was prescribed in the original protocol or as it was intended by the program developers” (Proctor et al., 2011). Through ongoing implementations, the National SafeCare Training and Research Center (NSTRC), which disseminates the SafeCare model, has collected 14,778 fidelity observations of SafeCare sessions by some 868 providers in 172 agencies. This data set, collected over 10 years, contains longitudinal observations on implementation fidelity among SafeCare providers, as well as a range of individual provider variables that may be related to fidelity, including demographics, work history, and a key attitudinal measure, the Evidence-Based Practice Attitude Scale (Aarons, 2004). The current project uses data collected from ongoing SafeCare implementations to address several research questions. In addition to a broad exploration of whether fidelity tends to increase, decrease, or stabilize over time, I analyze the relationships between provider level variables (demographics and attitudes toward evidence-based practice) and change in fidelity over time, with a particular focus on how attitudes toward evidence-based practice impact fidelity over time. A multilevel modeling approach was used to explore fidelity trajectories, as well as predictors of those trajectories. Growth curve models were used to analyze trajectories, and time by predictor interaction terms were included to explore the relationships between predictors and fidelity over time. Overall, SafeCare fidelity was stable and high. Several variables, including the key EBPAS variables, were related to fidelity on average, though the effect sizes were small. These small effect sizes are an indication that NSTRC’s efforts to systematize training processes are effective, at least with regard to program fidelity. Fidelity, however, is a necessary but not sufficient condition for program effectiveness, and directions for future research which incorporates fidelity data with other program outcomes are discussed