Resources

The standard delivery is appropriate for the general population, caregivers without clinical complexity.

• Maxwell et al. (2021), a 256-parent multi-site study with waitlist control, found significant improvements in self-efficacy and empathy, and reductions in hostility, helplessness, and depressive symptoms, with the largest gains among mothers with probable clinical depression.Maxwell, A. M., McMahon, C., Huber, A., Reay, R. E., Hawkins, E., & Barnett, B. (2021). Examining the effectiveness of Circle of Security Parenting (COS-P): A multi-site non-randomised study with waitlist control. Journal of Child and Family Studies, 30(5), 1123-1140.
• Kohlhoff et al. (2016), a small pilot of 15 mothers, found reduced stress, anger, and helplessness after standard delivery.Kohlhoff, J., Stein, M., Ha, M., & Mejaha, K. (2016). The Circle of Security Parenting (COS-P) intervention: Pilot evaluation. Australian Journal of Child and Family Health Nursing, 13(1), 3-7.
• Dexter and Wong (2024), a randomized waitlist-controlled trial with directly observed caregiver-child interaction, found that COSP produced measurable growth in reflective functioning, and that this growth is what explains the improvement in caregiving behavior.Dexter, C. A., & Wong, K. (2024). Circle of Security-Parenting randomized waitlist control study: Change in reflective functioning explains positive caregiver behavior. Journal of Child and Family Studies, 33(2).
• Cassidy et al. (2017), a randomized trial of 141 Head Start caregivers, found reduced unsupportive responses to child distress and better child executive function in the COSP group, but no shift in attachment security or behavioral problems.Cassidy, J., Brett, B. E., Gross, J. T., Stern, J. A., Martin, D. R., Mohr, J. J., & Woodhouse, S. S. (2017). Circle of Security-Parenting: A randomized controlled trial in Head Start. Development and Psychopathology, 29(2), 651-673.

For families with a specific caregiving concern or populations with added complexity (e.g., neurodevelopmental differences, foster carers, children with medically complex needs), COSP should be delivered with the needs of that specific population in mind. Facilitators should have dual expertise in COSP and the specific population or caregiving domain, and should tailor their delivery to the needs and differences of the families.

Neurodevelopmental differences

For example, when working with caregivers of children with neurodevelopmental differences, fluency in both the COS model and a working understanding of the child's neurodevelopmental condition, its communication profile, and its impact on the caregiver is important so that both bodies of knowledge can be integrated rather than delivered side by side. The facilitator needs to be confident in supporting caregivers to make sense of their child's needs on the Circle as expressed through behavior, which, naturally, is shaped by their neurodevelopmental profile.

• Kubo et al. (2021) found that standard COSP with added pre-assessment and follow-up sessions improved parenting self-efficacy, maternal mental health, and children's emotional and behavioral problems in mothers of children with autism spectrum disorder. The pre-assessment provided facilitators with an understanding of the dyad and the parent's state of mind regarding the relationship and their reactions to the child's diagnosis. This understanding informed facilitation throughout COSP.Kubo, N., Kitagawa, M., Iwamoto, S., & Kishimoto, T. (2021). Effects of an attachment-based parent intervention on mothers of children with autism spectrum disorder: Preliminary findings from a non-randomized controlled trial. Child and Adolescent Psychiatry and Mental Health, 15, 37.

Medical conditions and chronic illness

One study examined delivery of COSP to parents of children with cystic fibrosis. These findings plausibly generalize to other conditions that carry ongoing medical risk and treatment burden for a young child. Three things stand out.

First, a diagnosis made in infancy affects the ordinary bonding period itself and can impact caregivers' access to the informal support networks (mothers' groups, playgroups) that typically buffer early parenting, so isolation and lack of connection are part of the presenting problem.

Second, caregivers in this population find supporting the child's exploration disproportionately difficult, because ordinary risk-taking is filtered through a much higher baseline anxiety about the child's medical vulnerability, and this can look like overprotection when it may be a reasonable response to genuine risk. Supporting caregivers to differentiate between Shark Music and actual danger as it relates to the child's medical needs is essential, and requires a nuanced understanding of the medical condition. Thus, COSP facilitators need the same dual expertise described above, in this case fluency in COS alongside a working understanding of the specific condition's daily treatment burden and medical risk profile, so they can validate the caregiver's practical reality and support them to identify what is Shark Music.

Third, facilitators working with this population reported a strong pull toward giving medical advice or problem-solving practical care burdens rather than holding a reflective stance. This potentially reflects the complex and competing needs of caregivers. Additional support to address caregivers' urgent, concrete concerns relating to the medical condition may be needed to allow the COSP sessions to focus on the relationship. The caregiving burden of multiple and competing medical appointments is also a necessary consideration.

• Lowthe et al. (2026), a qualitative study of telehealth COSP with parents of children with cystic fibrosis, is the source of all three findings above.Lowthe, S., You, D., Mahmood, D., & Li, S. (2026). Circle of Security-Parenting in cystic fibrosis: Parent and facilitator experiences. Australian and New Zealand Journal of Family Therapy, 47(1), e70050.

Group composition

Qualitative evidence from the cystic fibrosis study suggests that grouping caregivers who share the same population-specific experience, rather than folding them into a general group, is a form of tailoring in its own right. Shared identity fostered openness and participant-led discussion, reduced the facilitator's need to draw people out, and supported peer validation and connections that outlasted the program. This is a different lever from dual expertise or bridging facilitation, both of which concern what the facilitator brings. This one concerns group composition, and it is worth weighing deliberately when deciding whether to run a population-specific group for neurodevelopmental, medical, or foster care populations, or fold these families into a mixed one.

Individual delivery of the standard COSP program is ideal for caregivers who may need more focused attention and support or whose interpersonal patterns make group participation difficult. Caregivers who are consistently withdrawn, devaluing of others, or monopolizing of group time can be supported through individual delivery. The COSP program content can be paced to the caregiver's needs rather than the group schedule. It can also enable more time to be spent responding to the caregiver's own Shark Music and relationship history and allow for more responsiveness to the unique caregiver's patterns.

Couples may also benefit from completing the program outside of a group, and may need different pacing to allow for couple and co-parenting issues to emerge and be addressed responsively.

• Zimmer-Gembeck et al. (2022)Zimmer-Gembeck, M. J., Rudolph, J., Edwards, E.-J., Swan, K., Campbell, S. M., Hawes, T., & Webb, H. J. (2022). The Circle of Security Parenting Program (COS-P): A randomized controlled trial of a low intensity, individualized attachment-based program with at-risk caregivers. Behavior Therapy, 53(2), 208-223.

For families presenting with complex psychosocial circumstances or facing multiple problems, COSP may be delivered within or alongside other wraparound or treatment services. For these families, parents often benefit from extended delivery time or additional coaching or intervention beyond the standard COSP facilitation. Treatment or intervention in other factors impacting parent capacity and availability is necessary.

Recent randomized trials reinforce this point. Two large trials, one in Denmark and one in England, delivered standard COSP to caregivers already seeking intensive help for significant mental health conditions, many with infants under six months old, an age for which some core COSP concepts are less relevant. Neither found meaningful effects. A 2026 meta-analysis that drew heavily on these trials reported no robust effects of COSP. These populations sit in the targeted part of the pyramid, though, and the findings are consistent with the long-standing guidance that COSP should not be used as a standalone intervention for high-risk caregivers. They do not show that COSP lacks value for the universal and moderate-risk families it was designed for (see Maxwell et al., 2021).

• Gerdts-Andresen (2021), a systematic review of seven studies of COSP with multi-problem child protection families, found some improvement in parental stress and self-efficacy across studies but no conclusive evidence that standard delivery increases parent-child attachment security in this population.Gerdts-Andresen, T. (2021). Circle of Security-Parenting: A systematic review of effectiveness when using the parent training programme with multi-problem families. Nordic Journal of Social Research, 12(1), 1-26.
• Risholm Mothander et al. (2018), a Swedish randomized trial adding standard COSP to treatment as usual in infant mental health clinics, found significant improvements in parent representations of the child and more emotionally available parent-infant interaction than treatment as usual alone, sustained to 12-month follow-up.Risholm Mothander, P., Furmark, C., & Neander, K. (2018). Adding "Circle of Security-Parenting" to treatment as usual in three Swedish infant mental health clinics: Effects on parents' internal representations and quality of parent-infant interaction. Scandinavian Journal of Psychology, 59(3), 262-272.
• Horton and Murray (2015) found that mothers in residential substance-abuse treatment who completed COSP showed improved emotion regulation and less lax or overreactive discipline, with the greatest gains among those with more education and no personal history of maltreatment.Horton, E., & Murray, C. (2015). A quantitative exploratory evaluation of the Circle of Security-Parenting Program with mothers in residential substance-abuse treatment. Infant Mental Health Journal, 36(3), 320-336.
• Belanger et al. (2023), a practice paper rather than a trial, describes delivering COSP before Parent-Child Interaction Therapy to build engagement ahead of PCIT's structured coaching, for families with a history of maltreatment.Belanger, K., Gennis, H., Ottenbreit, N., & Racine, N. (2023). Enhancing attachment-based aspects of PCIT for young children with a history of maltreatment. Frontiers in Psychology, 14, 1229109.

Recommendations for practice with multi-problem and child protection families

The Gerdts-Andresen (2021) review identified several concrete recommendations for services working with families where there are multiple psychosocial problems.

• Assess whether a family has the bandwidth for COSP before assuming it is the right starting point. The review raises a specific, practical question: whether caregivers managing several compounding stressors (financial debt, housing instability, psychiatric illness, substance use, involvement with the criminal justice system) have the emotional and cognitive availability the COSP reflective process requires. Where several of these are active and unresolved, reducing them, or at least stabilizing the most acute, may need to happen before or alongside COSP rather than being treated as separate from it.
• Do not treat "multi-problem" as a single category. The families in the reviewed studies varied widely in which risk factors were present and how severely, and none of the studies examined whether particular combinations or severities of risk changed how much families benefited. The practical implication is to make an individual judgment about accumulated risk and its severity for each family, rather than applying one standard of "enough support" to every family who meets a general multi-problem threshold.
• Do not assume the standard model transfers across cultural context without deliberate attention. Most of the studies in the review included participants from ethnic minority backgrounds, but none reported which groups were represented, whether language was a barrier, or whether the program was adapted for cultural fit. This is a genuine gap in the evidence rather than a reassurance that no adaptation is needed. Where other family-support literature has looked at this question directly, longer-term, home-based, and community-consulted approaches have fared better than imported group programs for Indigenous families specifically, which is a reasonable prompt to consult with the relevant community before assuming a standard group format is the right vehicle.
• Expect a reliable benefit on parental stress even where the picture on the relationship itself is unclear. A reduction in parental stress was the one consistent finding across every study in the review, regardless of design quality. Even in a population where the review found no conclusive evidence of attachment-level change, COSP-based work reliably did something for the caregiver's own burden, which is itself a legitimate reason to offer it as part of a wraparound plan.

Some families need more than standard COSP but cannot access, or do not need, a full clinical intervention. For them, personalized video review is a way to deepen COSP. It borrows a core component of COS Intensive without that model's training and resource demands. This tier sits on the pyramid between COSP and COS Intensive.

This is the tier with the clearest evidence that adding something to standard delivery produces a change that standard delivery alone does not. In video review, caregivers see their own child's needs on the Circle and their own responses in real moments of interaction, not only through the general examples in the program. This helps them move from understanding the model to applying it in their own relationship.

The evidence here is still emerging. The Kitagawa study is a small pilot, and an enhanced version of COSP with video review is now being pilot tested with colleagues in Australia. This tier is best treated as promising rather than established, and findings from that work will inform future versions of this guidance.

• Kitagawa et al. (2021) found that individualized video review added to completed COSP produced an increase in securely classified dyads and no remaining disorganized classifications. This effect appeared only at six-month follow-up. It was not visible immediately after COSP or immediately after the video sessions. For practice, this suggests that relationship-level change may take time to consolidate, and that measuring outcomes only at the end of delivery may underestimate benefit.Kitagawa, M., Iwamoto, S., Umemura, T., Kudo, S., Kazui, M., Matsuura, H., & Mesman, J. (2021). Attachment-based intervention improves Japanese parent-child relationship quality: A pilot study. Current Psychology, 41, 8568-8578.

COS Intensive is a targeted clinical treatment for caregiver-child relationships in which attachment difficulties are already present. It is individualized from the outset. An assessment of the dyad's interaction and the caregiver's state of mind regarding the relationship identifies the central struggle for that caregiver-child pair, and tailored video review sessions then address it directly. Delivery typically runs 12 or more sessions, and the original protocol ran 20 weeks. It is designed to intervene in the caregiving patterns and caregiver defensive processes (Shark Music) that maintain insecure or disorganized attachment.

COS Intensive is delivered by clinically trained providers who complete extended training and case-based supervision and consultation. This tier is indicated where caregivers' current functioning suggests a high risk of attachment difficulties, or where relationship difficulties are already evident. Recent trials of standard COSP with such populations suggest a more intensive, attachment-focused intervention of this kind is needed. Because of the training and resource investment, COS Intensive is not intended for universal use. It is best reserved for the families who most need it.

COS Intensive supports change at the level of the attachment relationship itself, including increases in security and reductions in disorganization. These findings come largely from pre-post clinical samples rather than controlled trials. Even so, they have been replicated across settings in the United States, Australia, and New Zealand, across individual and group formats, and among high-risk populations.

• Hoffman et al. (2006), the original trial of the 20-week Intensive intervention, found increased numbers of children classified as securely attached and fewer classified as disorganized.Hoffman, K. T., Marvin, R. S., Cooper, G., & Powell, B. (2006). Changing toddlers' and pre-schoolers' attachment classifications: The Circle of Security intervention. Journal of Consulting and Clinical Psychology, 74(6), 1017-1026.
• Huber, McMahon, and Sweller (2015) found significant improvements in parent-rated child behavior and emotional functioning, with the most severe presentations showing the greatest improvement.Huber, A., McMahon, C., & Sweller, N. (2015). Improved child behavioural and emotional functioning after Circle of Security 20-week intervention. Attachment & Human Development, 17(6), 547-569.
• Huber, McMahon, and Sweller (2015a) found increases in caregiver reflective functioning, caregiving representations, and child attachment security, and decreases in disorganization for those with high baseline levels.Huber, A. T., McMahon, C. A., & Sweller, N. (2015a). Efficacy of the 20-week Circle of Security intervention: Changes in caregiver reflective functioning, representations, and child attachment in an Australian clinical sample. Infant Mental Health Journal, 36(6), 556-574.
• Huber, McMahon, and Sweller (2016) found improved parental emotional functioning and reduced parenting stress following the same 20-week intervention.Huber, A., McMahon, C., & Sweller, N. (2016). Improved parental emotional functioning after Circle of Security 20-week parent-child relationship intervention. Journal of Child and Family Studies, 25(8), 2526-2540.
• Huber, Hicks, Ball, and McMahon (2020) found large, sustained improvements in supportive caregiving behavior and parenting self-efficacy at one-year follow-up, with no difference between individually delivered and group delivered Intensive intervention.Huber, A., Hicks, A.-M., Ball, M., & McMahon, C. (2020). Postintervention and follow-up changes in caregiving behavior and representations after individually or group delivered hybrid Circle of Security-Intensive intervention with New Zealand caregiver-child dyads. Attachment & Human Development.