Reflections on an article by Nancy Thomas: Reactive Attachment Disorder

I was reading an article online by Nancy Thomas called “What is Reactive Attachment Disorder” and find her pre – suppositions rather disturbing. Particularly, when she has no formal training in the assessment of attachment that I can discern from the attachment and child development experts in the field. When I reviewed the list of individual she has trained with, none of them that I was able to research had any formal training in the assessment of attachment quality of children. I may have to dig deeper with whom she has trained and with her trainers have trained with…

1) She presumes attachment is an “all or none” construct. This is a false. She claims on her site ” Attachment is defined as the affectional tie between two people. It begins with the bond between the infant and mother. This bond becomes internally representative of how the child will form relationships with the world. Bowlby stated “the initial relationship between self and others serves as blueprints for all future relationships.” (Bowlby, 1975)”

Her citation by Bowlby is now an outmoded concept. Attachment researchers such as Sroufe, Carlson, Waters, Marvin, Siegel, Soloman, Bretherton and others clearly state that that attachment occurs along a developmental path, and should one have an insecure form of attachment early on, this does not preclude them from developing a secure attachment over their early life or later life with a trusted and sensitive caregiver, spouse, or therapist. For example, let’s assume a mother depressed during the first year of an infants life but during the second year, receives and responds positively to treatment and becomes sensitively attuned to her child – the attachment security and quality can be developmentally altered onto a more secure path. “Blueprints” can be changed. Very, very, very rarely does a child lack an attachment relationship to any caregiver.

2) Her statement: Attachment Disorder is defined as the condition in which individuals have difficulty forming lasting relationships. This is false. This is clearly know as a “hasty generalization” in logic and a fallacy. She has taken a very complex topic as attachment and made an over generalization about it.

3) She then launches into a discussion of “non-attached”. This is very, very rare and false and evidence of all or none thinking. Research is clear that children who have experienced maltreatment can and do often have an attachment relationship with there caregiver. It is often a form of insecure attachment with evidence of what is usually an organized pattern having disruptions to it (known as disorganization). A more accurate reference is a “child with distorted attachment that initiate a maladaptive pathway…(Alan Sroufe, personal communication 9/20/11)

4) Children with “non-attachments” are doomed. This is false. The University of Minnesota has done extensive work with children in orphanages in Romania who exhibited “no attachment pattern” who are now exhibiting increasing security in their caregivers. This is very promising. And, again, lack of attachment to any caregiver is extremely, extremely rare.  A more accurate reference is a “child with distorted attachment that initiate a maladaptive pathway…(Alan Sroufe, personal communication 9/20/11)

5) Attachment Disorder is accepted as a disorder by professionals with a clear definition and she describes the symptoms. This is false.   There is and continues to much debate about the diagnoses of RAD and while progress is being made in the upcoming DSM-V; there remains much work to be done.

6) She lists “causes” of attachment  “disorder.”  Again, the pre-supposition of disorder is rather disturbing. Attachment is a description of the quality of the relationship and how well a child has confidence in his/her caregiver to meet his/her needs consistently and sensitivity. The list she provides are certainly correlated with forms of insecure attachment, however; to claim  a cause is false. Also, children who exhibit secure attachment patterns can later develop an insecure pattern for various reasons: death of a parent, severe illness, divorce, etc.   Patterns of attachment security/insecurity occurs over a lifetime.

What does this all mean…be careful what you read on the internet. Do your research. Be critical. Ask questions. My friend is an author and said, “Writing a book has made me an expert.” I appreciated his comment on this – but writing a book makes someone a writer, not an expert. Only proper knowledge and proper training and proper ongoing research and experience make someone an expert. Clearly, Nancy may have some thoughts and skills on parenting techniques that are helpful, but an attachment expert, she appears very misinformed regarding what attachment is and  not. This is no fault of her own. It is a very misused and abused concepts among treatment professionals who lack specific training from experts in the field.

One more thing, she discusses her upcoming study on her effectiveness by reducing cortisol levels in parents proving her techniques work. The proper research based on her claims would be assessing the cortisol levels of the children and seeing if those are reduced combined with other measures of assessing attachment security in the children to determine their attachment security. After all, she claims on her website to “provide help for each wounded child….” See below for a direct pull from the site:

“We offer information on adoption / attachment and bonding issues, and Reactive Attachment Disorder (RAD) to families and professionals. We are also the official home of Families by Design providing educational materials and seminars.The goal of this site is to provide help for each wounded child with attachment disorderPTSD (Post Traumatic Stress Disorder)ODD (Oppositional Defiant Disorder)….”

After receiving a posting, I reviewed this article. It appears that the pages that were cited from the Nancy Thomas website are no longer available. I am hoping that the site has been updated to reflect what is currently known from research. Research can provide us with sound information. It allows us to take what we know and make it into models that can hopefully help – as my mentor for the past two year, Dr. Bruce Perry states, “Essentially all models are wrong…but some are useful!”


  1. Clearly, you are selling something. I have raised an unattached child. Nancy Thomas was our guru and savior. He is now an adult, and although he has had some success navigating a world in which people are fairly feelings-connected, he still has a disconnect. As a soldier, he was able to clean up the remains of a suicide with no emotional response because, “It’s my job.” Superficially, he is a comical guy, and very sweet, but there is no depth to his relationships. Have you raised an unattached child? Have any of the people whose research you are using? I am a mother, a step-mother, a grandmother and a licensed special educator. I trust the words of a dedicated parent/step-parent/adoptive parent far more than the research-educated person reading/writing in scholarly journals. You can’t understand RAD unless you are the parent–or the child.

  2. Reprinted from Woolgar and Scott (2013) The Negative Consequences of Over-Diagnosing Attachment Disorders in adopted children: The Importance of Comprehensive Formulations
    There has been increasing interest in the area of attachment disorder in the light of findings from the series of studies of children adopted from orphanages in the former Eastern Europe (e.g. Rutter, Kreppner, & Sonuga-Barke, 2009; Zeanah & Smyke, 2008). However, there is a growing mismatch between the rigour of academic studies and the application of the attachment disorder construct in everyday practice. There is widespread use of the term ‘attachment disorder’ by services for adopted
    and fostered children, but there is little scientific or diagnostic consistency in the way in which the term itself is used in these contexts (Nilsen, 2003). Frequently, the constructs used bear little resemblance to the diagnostic categories that exist in DSM-IV-TR and ICD-10 (Prior & Glaser, 2006) and there is particular concern about some of the ineffective and even dangerous treatments that have followed from its use (AACAP, 2005; Chaffin, et al., 2006; Hanson & Spratt, 2000), which raises issues about how clinicians should use them in practice (DeJong, 2010).

    The use of recognised diagnostic terms offers reliability and common understanding but is made more complicated for attachment disorder because the proposed aetiology forms part of the diagnosis (Zeanah, 1996). Both subtypes of reactive attachment disorder (RAD) in DSM-IV-TR and the equivalent diagnoses of RAD and disinhibited attachment disorder (DAD) in ICD-10 highlight pathogenic care occurring before the age of 5 (see Prior & Glaser, 2006). So there is a risk of diagnosing attachment disorder in samples of children who have been subject to early loss or trauma based primarily upon their caregiving history, without giving due attention to their current presentation. Indeed, the requirement for evidence of pathogenic care before the age of 5 is often ignored in diagnoses, but this is essential given the importance of very early experience in the development of the attachment system. The risk of misdiagnosis is further compounded by the availability of online checklists and questionnaires that claim to identify attachment disorder using lists of symptoms that span several recognised alternative diagnoses, including conduct disorders, autistic spectrum disorders, learning disability, attention deficit hyperactivity disorder (ADHD), as well as bizarre items such as ‘a fascination with fire, blood, gore, weapons, evil’ (Prior & Glaser, 2006). We argue that these complexities with the diagnosis often lead to the over-identification of attachment disorders at the expense of more common problems that may often be treatable by standard means. This over-diagnosis occurs even though clinicians have been advised to avoid the ‘allure of the rare disorder’ (Haugaard, 2004) and to consider more everyday disorders such as conduct problems, mood and anxiety in samples of maltreated and neglected children (Chaffin, et al., 2006).
    To preserve the integrity of the attachment disorder construct it is important for clinicians to keep the distinctions between attachment disorder, attachment patterns, disorganised attachments and attachment difficulties intact, together with their respective evidence bases (see O’Connor & Zeanah, 2003; Prior & Glaser, 2006). Then there is also the confusing and inappropriate description of a parent’s relationship to their child being referred to in attachment terms, especially in assessments for Court (e.g. ‘What is the carer’s attachment to their child?’), or indeed for the child’s relationship to non-attachment figures, pets or even, in our clinical experience, to inanimate objects. It is important to recall that Bowlby described attachment security arising from the child feeling safe and protected by their tie to its carer, not vice versa (Bowlby, 1958, 1982). Within the developmental psychology literature, the attachment construct is precisely operationalised, to refer to accessing protection and safety, which is quite distinct from the much wider notion of ‘relationship’ (Goldberg, Grusec, & Jenkins, 1999); however, in our experience the term attachment is often used in clinics and children’s services to refer to relationships in general. Similarly, the meaning of attachment disorder in the diagnoses of RAD and DAD refers to a severe, pathological breakdown in the normal attachment system that an infant or very young child develops with its caregiver. This means something much more substantial than problems within the child-carer relationship or difficult early experiences and, consequently, attachment disorders are thought to be very rare, even in high-risk maltreated samples (Zeanah & Smyke, 2009).

    To some extent the problems with the construct of attachment disorder that we will illustrate in this paper have probably developed in the context of an absence of effective terms to describe the sometimes bizarre and atypical presentations in children, including looked-after and adopted children, who have been subject to early disruptions in their attachment relationships or exposure to significant and persistent traumas within the caregiving relationship. The research literature has recognised this difficulty. For example, Zeanah and colleagues have proposed a range of putative clinical presentations that could indicate degrees of disordered attachment, including role reversal and compulsive compliance (Boris, et al., 2004; Zeanah & Boris, 2000). While there have been some attempts to think about the assessment of the relationships that drive secure attachments (Zeanah & Benoit, 1995) and an attempt to categorise problems at the relationship level (Zeanah & Smyke, 2008), such relationship disorders are not the same as the child’s specific attachment to a particular carer or the presence of an attachment disorder within the child. In a further complication of the picture, while the distinction between the two forms of attachment disorder in ICD-10 is well made, recent research suggests that the disinhibited form of attachment disorder is characterised by widespread deficits in brain and social function that may be best construed as a neurodevelopmental disorder and not as a malfunction of normal attachment processes (Rutter, et al., 2009).
    There is clearly cutting edge research going on in this field to further refine various attachment constructs but in the meantime we are left with diagnostic categories within the ICD-10 and DSM-IV-TR systems that can be hard for clinicians working with adopted, fostered and maltreated children to make consistent and reliable use of (Boris, et al., 2004). Fortunately, there have been some helpful guidelines suggested by Prior and Glaser, consistent with recommendations made by O’Connor and Zeanah (2003) and two sets of American practice parameters (AACAP, 2005; Chaffin, et al., 2006) to guide the clinical application of attachment disorder in a way that is also consistent with Bowlby’s original proposal. Concerning ICD-10, Prior and Glaser suggest that there needs to be evidence that the attachment system has failed and that no discriminated attachment figure has been achieved, e.g. the absence of an effective attachment figure to whom the infant or young child seeks proximity and comfort (RAD) or the lack of a specific, discriminated attachment figure (DAD). Whatever else, there needs to be good evidence that it is the attachment system itself that is not functioning, rather than the presence of behaviours that could be explained by non-attachment-specific factors, such as the result of exposure to trauma, other common psychiatric disorders or neurodevelopmental problems. Moreover, it is unhelpful to think that there is an ‘attachment disorder’ outside and distinct from one of its recognised forms of either DAD or RAD in ICD-10 or RAD-inhibited or RAD-disinhibited in DSM-IV-TR. Again, the scarcity of RAD and DAD diagnoses, even in groups at high risk of early pathogenic care such as looked-after and maltreated children, means that more common disorders should be considered first (e.g. Boris, et al., 2004; Meltzer, Gatward, Corbin, Goodman, & Ford, 2003).
    Large-scale research into the health and well-being of looked-after children in the UK has identified the types of mental health disorders commonly seen. While there are bound to be some differences between adopted and looked-after samples, most adopted children in the UK will have come through the looked-after system, for at least some part of their lives and, as such, the data from these studies can guide us to the likely range of disorders and diagnoses to consider for adopted children in the UK (Ford, Vostanis, Meltzer, & Goodman, 2007; Meltzer, et al., 2003). For example, looked-after children have significantly elevated rates of conduct problems and ADHD, compared with both high-risk and normal controls in birth families (e.g. 39% vs. 10% and 4%, for behavioural disorders respectively and 8%, 1% and 1% for hyperkinetic problems respectively; Ford et al., 2007). These are disorders for which well-established care pathways using evidence-based treatments already exist. Indeed, there is no evidence that we know of to suggest that the conventional treatments are not effective in looked-after children. The same research study also looked at the prevalence of attachment problems. Using a narrow definition that maps well onto the ICD-10 disorders, only 2% of looked-after children were identified as having a possible attachment disorder, whereas about 20% were identified with a broader set of attachment-related problems that went beyond the ICD-10 diagnoses (Meltzer, et al., 2003). It is important to note that this report specifically considered that both the narrow and the broad definition are likely to be unreliable estimates and that they should not be used to estimate attachment disorder prevalence. Indeed, because the instrument did not assess the requirement for pathogenic early experiences (Meltzer, et al., 2003, pp. 138–139) and nor could it make use of the recommended structured observations of the child’s behaviour with carers and strangers, these figures are likely to be over-estimates of attachment disorder.