Research That Matters (January 17 - 20, 2008)

Objectives: The present study explores the implementation of the SDMŪ model and its impact on child welfare decision making and outcomes. Background: Structured Decision MakingŪ (SDMŪ) is a child welfare decision making model that was adopted by the State of California in 1999, and subsequently implemented in several counties, including Los Angeles County. A multi-level study was conducted in conjunction with the Los Angeles County Department of Children and Family Services (DCFS) to explore the implementation of SDMŪ and its effects on child welfare decision making and outcomes. One of the sub-studies involved a survey of DCFS caseworkers and managers in Los Angeles County. The survey was designed to examine the social workers' experiences with the DCFS-led training and implementation of SDMŪ, their feelings regarding any changes in their decision making and the perceived effectiveness of the SDMŪ tools with respect to various child welfare outcomes (e.g., protecting children, reducing the rate of subsequence abuse/neglect substantiations). Method: A total of 485 DCFS caseworkers and managers who implemented the SDMŪ model completed a web-based survey. In order to understand the implementation of the SDMŪ model, descriptive analyses of the usability (e.g., ease of use, reliability, accuracy) and actual usage of each tool were conducted. To examine SDMŪ's impact on child welfare decision making and outcomes, latent path analyses were conducted. First, using AMOS 5.0, confirmatory factor analyses were performed to test the validity of two latent constructs of perceived effectiveness of SDMŪ. Once the measurement model was established (CFI=.99, RMSEA=.07), a latent path model was estimated to examine the effects of worker characteristics, DCFS-led training, and implementation of SDMŪ on worker decision making and service outcomes (CFI=.98, RMSEA=.04). Results: Findings reveal considerable variation in the implementation of the SDMŪ tools. Of all the tools, the family strengths/need assessment tool (86%) was most utilized, followed by the safety assessment tool (80%); the least utilized tool was the hotline/response priority tool (8%). Descriptive statistics show that child welfare workers had largely positive perceptions about the SDMŪ model and tools. For example, 94% of the participants found the SDMŪ tools easy to use, and 92% found the model to be accurate. Results of the structural equation modeling suggested that organizational tenure (â=-.28) and job position (i.e., being a frontline worker) (â=-.23) had a significant effect on perceived effectiveness of SDMŪ on decision making. In addition, social workers who were more satisfied with the SDMŪ training were more likely to think that SDMŪ had a positive impact on their decision making (â = .10). Finally, perceive effectiveness of the SDMŪ tools on worker decision making was found to have a strong positive impact on the perceived effectiveness on service outcomes (â = .83). Conclusion: Implications for child welfare policy and best practices are considered, particularly those related to training and implementation of the SDMŪ tools and model. Also considered are issues related to the conceptualization of structured decision making and implications of our findings for future research on SDMŪ.