COSI Blog

For COSP Facilitators and Practitioners

Neil Boris, MD, holding the 2nd edition of What Works for Whom? A Critical Review of Treatments for Children and Adolescents

Circle of Security Parenting™ (COSP) was created in 2010 and trainings for COSP Facilitators began in 2011. That was 15 years ago! COSP was designed to be what is called a “preventive intervention.” Preventive interventions are programs that are designed for everyone (such programs are also called ‘universal’ interventions) and designed to prevent behavioral or social problems before they get ingrained or to detect and manage such problems at an early age.

When COSP was created, the only intervention our company had was what we now call the COS Intensive™ Intervention. COS Intensive is very different from COSP: as its name suggests, it is a more intensive intervention (typically 12 or more sessions) and designed for clinically-trained providers to systematically address already existing attachment disturbances. Programs like COS Intensive are called ‘targeted interventions.’ They're not universal and instead of prevention, they are designed to intervene with dyads in which there are already signs that the caregiver-child relationship is dysfunctional. Graphically, public health professionals distinguish these different levels of intervention as follows:

Figure 1: A Public Health Pyramid. Universal interventions are meant for any child or family; they are designed to prevent health or mental health problems from developing or to detect such problems early. Selective interventions are for high-risk groups who may already be manifesting health or mental health symptoms. Targeted interventions are more akin to treatment. For these families, measurable symptoms and limitations are already evident.

What works for whom?

So where does that phrase, ‘What Works for Whom?’ come in? Well, these varying levels of interventions likely have their own set of limits. In other words, you wouldn’t want to use COSP with a population that needs more intensive targeted intervention. Preventive interventions may well not be robust enough for families who qualify for targeted intervention. And, you wouldn’t want to use COS Intensive as a universal intervention. Training costs are high and you’d want those more capable professionals working with the families with more risk. This part about matching interventions to risk seems simple and clear. But there’s more to the ‘What Works for Whom?’ picture.

What the recent COSP trials found

Before getting into that more complex picture, it’s worth talking a bit about research on COSP, particularly recent important studies that have been published in the last few years and were bundled into what’s called a meta-analysis that was published in September, 2026. That meta-analysis had a fairly provocative title! Are you ready? Here it is: 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.²

The two largest and most recent studies that contributed data to this research review trained and supported providers in COSP, but had them implement the program with populations of caregivers that were already seeking intensive intervention for significant mental health conditions. Many of these families had infants that were younger than 6 months of age, a group for whom some of the key concepts presented in COSP are less relevant. In other words, both a study in Denmark³ and one in England⁴ effectively used a preventive intervention for a population that better fits the targeted category. While we’ve shared with COSP Facilitator trainees for years that COSP is not suited as a standalone intervention for caregivers at high-risk, these studies underscore that point.

Our team is grateful that both of these research teams tested COSP in this way. Where we might have hoped that COSP would impact outcomes for the caregivers and their infants recruited, we agreed with the conclusions of both studies that more intensive intervention is likely necessary for such populations.

Asking a different question

Still, when you read the title of that meta-analysis (“No robust effects…”), it’s hard not to be discouraged. Does this mean that COSP has no place as an intervention? I’ve been asked that very question from colleagues who’ve read the studies from Denmark and England and then read the meta-analysis as well. The answer, also backed up by research, is that COSP definitely has its place. When you ask the question a different way, changing it from ‘does COSP work?’ to ‘for whom might COSP be a well-matched intervention?’, there is important evidence available.

For example, an even larger study of COSP, this one from Australia,⁵ showed significant impact of COSP on parents and parenting, albeit in a sample that included caregivers who were not as high-risk as the samples from Denmark and England. In fact, the group of parents in the Australia study fit best in the middle portion of the public health pyramid (see Figure 1, “Selective”). If we take the COSP studies from Denmark, England and Australia and place them on that public health pyramid, Figure 2 emerges.

Figure 2: Linking Recent Research to the Public Health Pyramid. Circle of Security Parenting (COSP) was designed as a universal preventive intervention. Recent studies suggest that COSP may have impact not just with the general population but with families at moderate risk (Maxwell et al., 2021). However, families like those studied by Rosan et al., 2025 in England and Stuart et al., 2025 in Denmark seem to require more intensive attachment-focused parenting interventions.

And this is where the ‘What Works for Whom?’ quote comes in. Interventions have to be of the right intensity for the needs of the population. Not only that, research never tells a story that is simple or linear. For example, some of you reading this may object that I am comparing outcomes from randomized controlled trials (like the Denmark and England studies) with a non-randomized, controlled trial (the Australia study). It’s a reasonable criticism, though the authors of that ‘What Works for Whom?’ book would suggest caution here. For those who are really into research, here’s a quote from the Introduction to that 2nd and most recent edition of the book:

As you can tell, our team tracks all this research carefully. The evidence base matters! A couple of times per year, we update the bibliography on our website so that those of you interested in COS-related studies can look up the latest studies yourselves. We also meet as a team and think carefully about how we can improve our interventions. Perhaps the greatest gift that the studies from Denmark and England have delivered is the impetus for our team to consider ways to heighten the impact of COSP.

Where might we go from here?

In 2025, we reached a major milestone. We created, tested and revised the entire COS-Intensive training such that learners could take all the coursework needed online and in 2026, we added CEU credits to the mix. Now clinicians can learn COS-Intensive and prepare for case-based consultation while gaining 6.5 IACET credits. At the same time, we know not every professional has the funding or clinical setup to implement a targeted intervention like COS-Intensive. Sixty-five hours of training is a lot and that’s only to prepare for case consultation!

Our own analysis of the research literature on attachment-based interventions for families goes beyond tracking studies on COSP. When we take into account the lessons learned from recent COSP research and we analyze data on other attachment-based interventions, it seems clear that one useful option is to enhance COSP with personalized video review (which is a key part of the COS-Intensive intervention). In fact, research data from colleagues in Japan⁶ suggests that using video review in conjunction with COSP may enhance the impact as compared to COSP alone. We think there is room on the very same public health pyramid for an intervention that is not quite as intensive as COS-Intensive, but is more intensive than COSP alone. Our team has already started pilot testing an enhanced version of COSP with colleagues in Australia and we expect to learn much from this work.

We’re grateful for the thousands of COSP Facilitators around the world who use the program with all sorts of families. We will continue to warn against using COSP with caregivers whose current function suggests they are at high-risk for attachment disturbances or even already showing signs of such disturbances. For clinicians who work with such families we offer COS-Intensive training and are pursuing ways to enhance COSP with video review as well.

Matching support to each family

So how does a facilitator decide which level of support fits a particular family? To help answer that question, our team developed a decision tree, Individualizing COS Parenting™ Delivery, that helps practitioners match the depth of individualization to each caregiver-child dyad. You can download the decision tree here.

Free one-page handout (PDF, letter landscape)

Download the Decision Tree See the Research Behind It

It is important that all practitioners consider “What Works for Whom?” This question is never simple to answer when applied to each individual family. Not every intervention is robust enough for every family and yet using interventions that are too specialized is not advisable. Our decision tree provides an organized way to think about matching intervention intensity to client need. We’re using research to inform practice and are grateful to all the research teams from around the world who support those efforts. As we develop and test protocols for adding individualized video review to COSP for clients, we’ll update you on our findings and we look forward to more independent studies on ‘enhanced COSP’ as well. Stay tuned as we walk this journey together with you.

References

1. Fonagy, P., Cottrell, D., Phillips, J., Bevington, D., Glaser, D., & Allison, E. (2015). What works for whom? A critical review of treatments for children and adolescents (2nd ed.). Guilford Press.
2. Bakermans-Kranenburg, M. J., Galbally, M., Hawkins, E., & van IJzendoorn, M. H. (2026). 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. Journal of Child Psychology and Psychiatry.
3. Stuart, A. C., Smith-Nielsen, J., Egmose, I., Reijman, S., Lange, T., Wendelboe, K. I., Stougård, M., & Væver, M. S. (2025). Evaluating the efficacy of Circle of Security-Parenting as an addition to care-as-usual in families affected by maternal postpartum depression and/or infant social withdrawal: A randomized controlled trial. Development and Psychopathology.
4. Rosan, C., Alyousefi-van Dijk, K., Cornelius, V., Darwin, Z., Babalis, D., Phillips, R., Richards, L., Wright, H., Pilling, S., Fearon, P., Pizzo, E., & Fonagy, P. (2025). Clinical effectiveness of the Circle of Security-Parenting group intervention for birthing parents in perinatal mental health services in England (COSI): A pragmatic, multicentre, assessor-masked, randomised controlled trial. The Lancet Psychiatry, 12(11), 817-829.
5. Maxwell, A.-M., Reay, R. E., Huber, A., Hawkins, E., Woolnough, E., Sweller, N., & McMahon, C. (2021). Examining the effectiveness of Circle of Security Parenting (COS-P): A multi-site non-randomized study with waitlist control. Journal of Child and Family Studies, 30(6), 1454-1468.
6. Kitagawa, M., Iwamoto, S., Umemura, T., Kudo, S., Kazui, M., Matsuura, H., & Mesman, J. (2022). Attachment-based intervention improves Japanese parent-child relationship quality: A pilot study. Current Psychology, 41(12), 8568-8578.