We write as consultant psychiatrists at a national specialist clinic in the UK, which since the 1980s has taken referrals for children who have experienced abuse and neglect. In the course of this work we have learned that dissociation is a common and often overlooked symptom in traumatised children. In the past year, however, we have seen two children referred to the clinic already diagnosed with dissociative identity disorder (DID) and have been concerned about the diagnoses and advice parents have been given.
There is very limited research about DID in children, and there are developmental questions to be asked about making such a diagnosis at an age when identities are still forming. At the same time, repeated traumatic experiences of an interpersonal nature can have a profound effect on a child's identity, memories and self-organisation. It is common for children who have been removed from home to have a limited or confused understanding of why this has happened, and to have negative thoughts about themselves from which they may take refuge in fantasy. They may have had formative experiences, such as of sexual abuse, which have not been acknowledged by the adults in their lives. Such circumstances must necessarily be a breeding ground for identity confusion. Although the diagnosis of DID may be rare or even of doubtful validity in children, we are keen that this should not obscure an improved understanding of how abuse and neglect increase the risk of dissociative symptoms.
Defining dissociation
A review of dissociation is complicated by the lack of a clear scientific consensus as to its nature. It may, of course, be non-pathological. Some of the symptoms in the Dissociative Experiences Scale (DES) (Dubester Reference Dubester and Braun1995), such as driving and not remembering part of the journey, are a part of everyday life. The same applies to the fantasy, role-play and imaginary friends of childhood. Meta-analytic data have indeed shown gradual decreases in dissociation with age (van Ijzendoorn Reference Van IJzendoorn and Schuengel1996). Meanwhile, there is debate about what psychiatric symptoms may be treated as dissociative. Consider for example the criteria for post-traumatic stress disorder (PTSD). The DSM-5 (American Psychiatric Association 2013) has adopted a dissociative subtype of PTSD, featuring symptoms of depersonalisation and derealisation, but many authors argue that flashbacks are also dissociative (Dorahy Reference Dorahy and van der Hart2015). As vivid, sensory, fragmentary experiences, they certainly are not integrated in the way of typical autobiographical memories. Auditory hallucinations have also been treated as dissociative in nature (Moskowitz Reference Moskowitz, Read, Farrelly, Dell and O'Neil2009), and this may be recognisable to clinicians who notice how voices often reflect traumatic themes in a patient's life.
The DSM-5 criteria for DID refer to ‘recurrent gaps in the recall of everyday events, important personal information and/or traumatic events that are inconsistent with ordinary forgetting’. Amnesia for traumatic events, whether complete or partial, is, however, recognised as a common feature of PTSD in both the ICD-10 and DSM-5 classifications (World Health Organization 1993; American Psychiatric Association 2013). Historically, such amnesia has been seen as a defence mechanism, but more recent work on PTSD has looked at memory dysfunction, one observation being that affected people simultaneously experience excessive involuntary remembering (such as flashbacks) and impaired intentional recollection (Brewin Reference Brewin2007). This picture will, of course, be further complicated when traumatic events occur at a preverbal stage or an age when autobiographical memories are not yet as reliably encoded and stored.
Evidence showing a high level of comorbidity between borderline personality disorder and DID (Dorahy Reference Dorahy, Brand and Şar2014) should therefore be thought about in the light of which symptoms in the former (feelings of emptiness, disturbances in self-image) may consist of dissociative elements. The DSM-5 recognises this in its reference to severe dissociative symptoms as a possible feature. In what can become a polarised debate about the presence or absence of a disorder, we are keen therefore to keep an open mind about how dissociation may manifest trans-diagnostically and with a spectrum of severity.
The association between early-life trauma and dissociation
The research which informed the adoption of a dissociative subtype of PTSD in DSM-5 showed that dissociation predicted higher PTSD severity and levels of comorbid psychiatric disorders (van Huijstee Reference van Huijstee and Vermetten2018). Those with the dissociative subtype were more likely to have a history of early-life trauma; why might this be the case? This question has attracted a number of hypotheses. First, given that dissociation is more common in childhood, could traumatised children be more vulnerable to a deviation in developmental trajectory that leads to more persistent dissociative experiences? Second, does attachment play a role in a stress diathesis model (discussed further below)? Third, is it simply due to the fact that early-life trauma is highly correlated with increased chronicity and severity (Dorahy Reference Dorahy and van der Hart2015). We are also prompted to reflect that children are usually helpless in the face of maltreatment, meaning that dissociation may offer some relief from experiences that are otherwise inescapable. The evidence has tended to suggest that, although trauma exposure increases risk, there are other factors at play, and some people will experience dissociation without such exposure (Briere Reference Briere2006). Several longitudinal studies have shown family environmental factors contributing to the development of dissociative symptoms independently of, and more significantly than, trauma exposure; these include parental dissociative symptoms and maternal unavailability (Ogawa Reference Ogawa, Sroufe and Weinfield1997) and lack of involvement and contradictory, role-reversed or disoriented parental responses (Dutra Reference Dutra, Lyons-Ruth and Borelli2005).
The potential role of attachment style
A replicated finding in these studies and others (Carlson Reference Carlson1998) is that disorganised attachment styles in infancy predicted dissociative symptoms in adolescence and young adulthood. Children with a disorganised attachment will show contradictory behaviours (both proximity-seeking and avoidant) towards caregivers – presumed to be experienced as either frightening or frightened (Main Reference Main, Hesse, Greenberg, Cicchetti and Cummings1990) and producing a paralysing motivational conflict for the child. Disorganised attachment behaviour is often marked by moments of freezing, dazing or trance-like states that bear a phenotypic resemblance to dissociation. This inability to develop strategies for comfort and protection, with a failure to integrate conflicting and discrete emotional states may, it is hypothesised, increase vulnerability to dissociation as a response to future stressors (Liotti Reference Liotti2004).
Diagnosis of DID in children: the evidence and potential effect on parenting advice
When it comes to DID in children, however, the evidence base is very small. Boysen (Reference Boysen2011) summarises this and finds a total of 255 children described in the literature, with four research groups in the USA accounting for 65% of the cases, and 93% of the cases emerging from descriptions of children in treatment. With an evidence base of this nature, there can be no confident claims about the validity of the diagnosis, let alone its prevalence or what effective treatment might look like.
In spite of this, we have had several children referred to our clinic with diagnoses of DID. This has led to advice to parents to acknowledge and respond to their children's ‘alters’ and attribute feelings and behaviours to them. We are very concerned about this practice: children with traumatic backgrounds are especially in need of emotionally sensitive parenting to become more secure in their attachment relationships. They also need help gradually to develop a more integrated sense of self as individuals who are valued for stable traits and preferences and who feel responsible for their actions, all of which is likely to be impeded by attention paid to alters.
A call for further research
While wishing to advocate a high degree of caution about the diagnosis of DID in childhood we are keen to understand more about how parenting, other aspects of the family environment and traumatic experiences may, in additive ways, increase vulnerability to dissociation. There is a need for more research, which should include increased use of measures of dissociation in population and clinical samples and further longitudinal studies collecting data on risk influences such as attachment status over time. It is conceivable that DID is more common from adolescence, a time when psychological reworking of early trauma takes place, and identity formation is a key developmental task. Given the strength of recent evidence demonstrating an association between emotional abuse and neglect and a range of mental illnesses, increased risk of dissociation should also be researched (Humphreys Reference Humphreys, LeMoult and Wear2020). Evidence in some analyses that parental dissociation is a risk factor for dissociative symptoms in children (Ogawa Reference Ogawa, Sroufe and Weinfield1997) should prompt more genetically informed research. Furthermore, are there interventions that may reduce the future incidence of dissociation in those who are vulnerable? In the meantime, clinicians should be vigilant to the possibility of dissociation in traumatised children as well as to the wider impact their experiences have on identity, memory and self-organisation, and therapeutic work should be directed accordingly.
Author contributions
Both authors made contributions to the conception of this article and analysis of the research literature. S.W. wrote the first draft and both authors subsequently revised it, gave approval for its final form and take responsibility for the content.
Declaration of interest
None.
ICMJE forms are in the supplementary material, available online at https://doi.org/10.1192/bja.2020.35.
eLetters
No eLetters have been published for this article.