Research That Matters (January 17 - 20, 2008)

Background and Purpose: Team Decisionmaking (TDM), one of the four core strategies of the Family to Family initiative sponsored by the Annie E. Casey Foundation, is an approach to child welfare placement decision-making that places family members and community supports at the center of the process. TDM operates in 60 locations across 17 states, but no systematic multi-site evaluations have yet been conducted. This study is a preliminary assessment of TDM's impact on the likelihood of placement change recommendations, and the restrictiveness of recommendations for children in foster care in three sites: Denver, CO (n=711); Anchorage, AK (n=1,039); and Wake County, NC (n=916).

Methods: This study uses administrative TDM data pertaining to the types and numbers of attendees, child demographics, meeting location, and recommendations. Within each site, a binomial logistic regression model was conducted to measure the likelihood of a placement change recommendation (compared to no change in placement), as predicted by child characteristics, meeting location, whether family and community participation meets or exceed participation from agency staff, the presence of a foster parent and the presence of a caregiver. For each site, a multinomial logistic regression model was constructed to compare a recommendation to change the restrictiveness of placement (less, same-level, more) with not changing placements.

Results: From the binomial models, if a caregiver attended, TDM meetings in Denver were less likely to result in a placement change recommendation (OR=0.526, p<.01), as were meetings in Anchorage (OR=0.677, p=.064) and Wake County (OR=0.555, p=.058). From the multinomial logistic models, foster parent attendance in Denver resulted in a lowered likelihood of a less restrictive recommendation, versus no change in placement (OR=0.578, p<.05). This pattern was also evident in Anchorage (OR=0.639, p<.05) and Wake County (OR=0.271, p<.001). Caregiver attendance resulted in a lowered likelihood of less restrictive placement recommendations in Denver (OR=0.276, p<.001) and Anchorage (OR=0.431, p<.01), and a lowered likelihood of same-level recommendations in Wake County (OR=0.108, p<.001), compared with no change in placement. If family and community supports attended in greater or equal numbers than agency staff, meetings in Denver were more likely to result in recommendations for less restrictive placements (OR=1.700, p<.05), but meetings in Anchorage were less likely to recommend a more restrictive placement (OR=0.463, p<.01), versus no change in placement.

Implications: Based on findings in these three sites that control for child demographics, foster parent and attendance appears to have a strong effect in discouraging recommendations to change placements. Given the underlying premises of TDM, these findings are encouraging in that they suggest that encouraging communication among foster parents, caregivers and agency staff, especially during a crisis, may help preserve a placement. Yet, TDM's impact appears to differ across sites in regard to the effect of the number of family- and community-based participants. These initial findings add to our understanding of the decisionmaking process as it relates to placement changes in child welfare and also provide a perspective agencies can use in their self-evaluation efforts.