Children diagnosed with autism spectrum disorder (ASD) show attachment behaviour typical for their developmental level, whereas those diagnosed with reactive attachment disorder (RAD) rarely or inconsistently do so. Children with RAD or disinhibited social engagement disorder (DSED) are likely to have a history of repeated changes of primary caregivers, which can inhibit the formation of stable attachments, or of rearing in unusual settings that limit opportunities for selective attachments (American Psychiatric Association 2013).
Over 50% of children with ASD, particularly those with higher cognitive ability, develop secure attachments. However, in the wider context, intellectual disability itself does not preclude the development of secure attachment relationships; for example, children with Down syndrome show secure attachment rates similar to those in the general population (British Psychological Society 2017).
How to tell the difference?
The article by Davidson et al (Reference Davidson, Moran and Minnis2022, this issue) aims to help in differentiating between ASD and RAD or DSED, exploring and delineating attachment behaviour in each of these conditions.
To achieve this, the article reviews the evidence of several studies by the authors themselves and others, concluding that the best tools to differentiate between ASD and RAD/DSED are the Coventry Grid and unstructured observational assessment such as the Scottish Centre for Autism's ‘Live Assessment’.
The original paper on the Coventry Grid (Moran Reference Moran2010) provides a detailed discussion of the problems in differentiating between attachment problems and ASD. The paper highlights two reasons why this may be difficult: first, children can behave atypically in the clinical environment, which can lead clinicians to suspect RAD/DSED; and second, ASD and RAD/DSED both affect relationships with others, so there is some overlap between them.
However, in their article, Davidson et al note a third situation: when there is a history of abuse and neglect (leading to RAD/DSED) that might have occurred alongside suspected ASD further investigating is needed. They note that there is an overlap between abuse and ASD. Individuals with neurodevelopmental disorders and their families may also be targeted by abusers, who may perceive that abuse is less likely to be detected, prevented or reported by those with such disorders. The article discusses neglect that might be caused during parenting owing the challenges of caring for a child with ASD and the fact that parents may themselves also have neurodevelopmental disorders. Clinicians assessing for ASD should be aware of both possibilities.
If the parents of a child with ASD pick up on their child's attachment signals and respond well to them, secure attachments are likely to develop. Behaviours indicating insecure attachment may be due to social communication difficulties associated with ASD (National Institute for Health and Care Excellence 2015).
Well-recognised differentiating features include: restricted, repetitive or ritualised behaviours and selective impairment in social communication (for example, impairment in intentional communication targeted at influencing the behaviour of another person is commonly seen in ASD but is not always a feature of RAD; (American Psychiatric Association 2013). Stereotypical behaviours such as rocking or flapping are seen in ASD, RAD and DSED, and are therefore not always useful for differential diagnosis. Severe social neglect seen in RAD and DSED may lead to comorbidity, including developmental delays in cognition and language.
Social impulsivity seen in children with attention-deficit hyperactivity disorder, where it is associated with inattention and/or hyperactivity, is not a characteristic of DSED, so it can be a differentiating feature (American Psychiatric Association 2013).
The ETRADQ
In view of the dearth of validated tools for assessment of RAD and DSED, Monette et al (Reference Monette, Cyr and Terradas2022) developed the Early Trauma-Related Disorders Questionnaire (ETRADQ). Using factor analysis they reported that it showed significant internal consistency, test–retest reliability, convergent validity and known-group validity on the RAD subscale (three items: low selective attachment, low social and emotional responsiveness, and emotional unpredictability) and the DSED subscale (two items: interactions with unfamiliar adults and social disinhibition). The ETRADQ potentially could be used in addition to the Coventry Grid to help support a diagnosis of RAD or DSED in a person with or without ASD.
Conclusions
In people first presenting to mental health or intellectual disability services in adulthood, Davidson et al have not commented on whether a robust childhood and developmental history, if available, could be used instead of the observational methods referred to in their article for differentiating RAD or DSED from ASD.
Nonetheless, this article goes some way to addressing the differentiation of ASD from attachment disorders, specifically RAD and DSED, despite multiple challenges, including the conceptualisation of attachment; the use of appropriate terminology; study methodology; and using fit-for-purpose assessment protocols (those used for typically developing children might not be generalisable to children with neurodevelopmental disorders and vice versa).
Author contributions
A.B. and P.B. contributed equally to the conceptualisation, writing up and finalising of the manuscript.
Funding
This article received no specific grant from any funding agency, commercial or not-for-profit sectors.
Declaration of interest
A.B. is on the editorial board of BJPsych Advances but did not take part in the peer-review or decision-making process. P.B. has no competing interests to declare.
eLetters
No eLetters have been published for this article.