Research Review: No robust effects of Circle of Security-Parenting (COSP) on parenting, parents, or children - a systematic review and meta-analysis of randomized controlled trials.
Circle of Security-Parenting (COSP) is one of the most popular attachment-based preventive support programs for parents and children. More than 50,000 participants from different (para)professional disciplines have been trained in the formal COSP™ protocol.
In a preregistered meta-analysis and synthesis of all eight randomized trials (N = 1,094 families), we examined COSP's efficacy on caregivers' parenting (sensitive parent-child interactions, reflective functioning, and parenting attitudes), on parent mental health outcomes (stress, depression, and anxiety), and on child attachment security and behavior problems.
We found an overall nonsignificant effect size of Cohen's d = 0.142 (SE = 0.089; p = .154, 95% Confidence Interval -0.068, 0.353, 95% Prediction Interval -0.444, 0.729), in a homogeneous set of studies. The pooled effect sizes of the parenting, parent, and child domains did not differ from each other or from zero. Cluster-robust inference tests showed converging (nonsignificant) results. The majority of the randomized studies (82%) were identified with some concerns about risk of bias.
We conclude that the evidence for COSP efficacy does not keep pace with the uptake of COSP in professional practice and policy.
In a preregistered meta-analysis and synthesis of all eight randomized trials (N = 1,094 families), we examined COSP's efficacy on caregivers' parenting (sensitive parent-child interactions, reflective functioning, and parenting attitudes), on parent mental health outcomes (stress, depression, and anxiety), and on child attachment security and behavior problems.
We found an overall nonsignificant effect size of Cohen's d = 0.142 (SE = 0.089; p = .154, 95% Confidence Interval -0.068, 0.353, 95% Prediction Interval -0.444, 0.729), in a homogeneous set of studies. The pooled effect sizes of the parenting, parent, and child domains did not differ from each other or from zero. Cluster-robust inference tests showed converging (nonsignificant) results. The majority of the randomized studies (82%) were identified with some concerns about risk of bias.
We conclude that the evidence for COSP efficacy does not keep pace with the uptake of COSP in professional practice and policy.
Authors
Bakermans-Kranenburg Bakermans-Kranenburg, Galbally Galbally, Hawkins Hawkins, van IJzendoorn van IJzendoorn
View on Pubmed