Too attached to attachment?
Having recently been given the task of writing the content of an online course in attachment theory, I read Elizabeth Meins’s article (‘Overrated: The predictive power of attachment’) in your January issue with great interest. Although I came across the concept through social work colleagues when working in the field as an educational psychologist in the 1980s and 90s, I had never previously studied the subject either as part of my initial training or subsequently. Starting from a naive perspective, I therefore researched the subject from first principles, and formed the impression that the evidence on which attachment theory is based is rather flimsy.
For example, whilst we can be pretty sure that there are critical periods of development in certain species as a result of animal experiments, ethical considerations mean that these can never be reproduced in humans. The existence of critical periods in human development was a popular essay topic in my undergraduate days, and remains unresolved some 50 years later.
Mary Ainsworth’s strange situation procedure identified 40 per cent of her sample of 100 children who showed either insecure-avoidant behaviour, or insecure-ambivalent/resistant behaviour. If such behaviours are thought to be problematic for subsequent mental health, how are we supposed to contrast this with Helen Minnis and colleagues’ 2013 estimate of 1.4 per cent of a deprived Glaswegian community displaying ‘Maltreatment-Associated Psychiatric Problems’?
However, it is the DSM-5 criteria themselves that place the nail in the coffin of attachment disorder. The criteria list a number of behaviours that are indicative of attachment disorder, which differ for Reactive Attachment Disorder (RAD) and Disinhibited Social Engagement Disorder (sometimes referred to as DAD). However, both include a further criterion that ‘The child has experienced a pattern of extremes of insufficient care…’.
At first sight it seems odd to include the postulated cause of a disorder as a diagnostic criterion, and on closer examination it seems to be logically contradictory. The argument goes like this – if it is necessary to include the causal agent in the definition, then there must be a group who meet all the other behavioural criteria but do not have an attachment disorder because they have experienced ‘good enough’ parenting. Their behaviour must therefore have a different foundation, but if this is true, then there must also be a group whose behaviour also arises from this different foundation, but who have experienced ‘not good enough’ parenting, yet they are identified as having an attachment disorder. The definition does not, therefore, distinguish between these groups, and this undermines the whole concept. How do we know that the causal agent in any child diagnosed as having an attachment disorder is not that other, unidentified agent, that the definition implicitly acknowledges?
It follows that the only thing we can be sure that the DSM criteria define is ‘ …a pattern of extremes of insufficient care…’
Van IJzendoorn and colleagues’ response to Meins’s argument (Letters, March 2017) acknowledges that children can display the behavioural criteria of disorganised attachment without having been abused, while conversely about half of abused children seem not to develop these behaviours. Some other causal agent – personality traits perhaps? – must surely be at work?
It is not that long ago that autism spectrum disorder was believed by many in the field to result from poor parenting, but that once widely held view has now been dismissed. Does attachment theory give us another rod with which to beat parents? The RAD criteria acknowledge the similarity with autism spectrum disorder, whilst the DAD criteria acknowledge the similarity with attention-deficit/hyperactivity disorder. Any parent would struggle with a child with either of these conditions, but throw in any number of additional circumstances that many disadvantaged parents face, and it is perhaps not surprising that ‘a pattern of extremes of insufficient care’ may be the result. Whilst there may be confusion in diagnosis with these conditions, I am not, however, suggesting that this is a single alternative explanation, but that a personality trait such as resilience, linked with concepts of childhood trauma, provides a much more logically consistent framework for understanding these behaviours.
I have no doubt that children’s development can be adversely affected in all sorts of ways which may result from poor or abusive parenting, bereavement, displacement as a result of war, etc., resulting in childhood trauma, and that we need to configure our services to meet their needs as best we can, whether through parent education and support, or through other therapeutic pathways. However, I am left with the feeling that a process of groupthink has emerged around the concept of attachment disorder. There are man professionals and practitioners whose careers have been built around this concept, who are heavily invested and will no doubt be resistant to any alternative view. In response, I would point out that I have come to this conclusion from a neutral viewpoint, and would be interested in understanding how my logic may be flawed.
Hugh Clench CPsychol
BPS Members can discuss this article
Already a member? Or Create an account
Not a member? Find out about becoming a member or subscriber