Yvonne Unrau, Michael Wells and Mary Ann Hartnett describe a US family-centred, needs-based foster care programme named Promise and present evaluation findings based on a comparison group design evaluation. Promise allowed for greater discretion among line-level workers to meet the unique service needs of families served, promoted greater team-oriented communication and involved more foster family involvement than the comparison group. An initial statistical comparison revealed that foster children in the Promise group (n = 380) experienced greater stability in their caseworker assignment and, to a lesser degree, greater placement stability over a 15-month period when compared to foster children served under the conventional model (n = 436). However, only the caseworker continuity effect remained when further analysis was undertaken. Similar rates of permanency achievement were reported for both models. Implications for foster care policy, practice and research are presented.