Background

The contemporary concept of psychopathy was first introduced by Hervey Cleckley in 1941 where he gave prominence to features such as arrogance, charm, low levels of guilt, recklessness and antisocial behaviour. ^{Reference Cleckley1} Science on psychopathy trailed behind following Cleckley's contribution. However, when Hare developed the Psychopathy Checklist (PCL), research on adult psychopathy promptly advanced. ^{Reference Patrick2,Reference Hare3} Roughly 5 years after Hare developed the PCL, Forth began research on psychopathy with adolescents and shortly afterward studies with children emerged. ^{Reference Forth, Kosson and Hare4}

Curiosity in child psychopathy intensified over the ensuing decades and correspondingly so did the science for youth psychopathy. This interest and research effort resulted in a proliferation of journal articles and books specific to child psychopathy. As a result of this attention, the information base for psychopathy was strengthened to where sounder conclusions could be drawn. For instance, research showed that child psychopathy was a multi-faceted syndrome underpinned by grandiose–manipulative, callous–unemotional and daring–impulsive traits. ^{Reference Salekin and Lynam5,Reference Dong, Wu and Waldman6} Research also showed that childhood psychopathy was associated with meaningful correlates such as fearlessness, reduced skin conductance, risk-taking, a specific temperamental and neurocognitive profile, and genetic transmission. ^{Reference Salekin and Lynam5}

Given these findings, there was a compelling case for recognising the existence of psychopathic traits in childhood. It was therefore, not unexpected that the DSM-5 incorporated psychopathic symptoms into its newest manual ⁷ and that the ICD-11 is considering a similar modification. Even though psychopathy is increasingly recognised as an important child concept, only one dimension of psychopathy is incorporated in the DSM-5, and the ICD-11 may harmonise with the DSM, suggesting that psychopathy will continue to be underrepresented in both diagnostic manuals.

Problems with the current conduct disorder criteria

There are several points to consider in the appraisal of the conduct disorder criteria. Although limited prosocial emotion, which contains affective (callous–unemotional) traits, has been added to conduct disorder in the DSM-5, the extent to which this new specification will contribute to conduct disorder is unknown. A key question is whether the limited prosocial emotion specifier will be redundant with existing criteria, or merely identify a severe group of youth with conduct disorder. This sentiment is reflective of Spitzer and colleagues who believed that callousness is already tapped, indirectly, via serious behavioural items (such as bullying/threatening, armed robbery, forcing sexual activity, physical cruelty to animals/people and fire-setting). ^{Reference Spitzer, Endicott and Robins8} A second question pertains to the degree to which the symptoms improve representation of psychopathy in classification systems. By using seminal models of psychopathy, one can begin to determine the overlap from traditional conceptualisations, such as those of Cleckley and Hare, ^{Reference Cleckley1–Reference Forth, Kosson and Hare4} with that of DSM-5 ⁷ and ICD-10 ⁹ conduct disorder.

Regrettably, this type of contrast demonstrates low overlap. Of the 15 items that constitute DSM conduct disorder symptoms, only one (i.e. often lies) directly corresponds to the 16 Cleckley items. ^{Reference Cleckley1} After adjoining the specifier of limited prosocial emotion (4 items), only 4 of the 19 conduct disorder items (15 conduct disorder items plus 4 limited prosocial emotion items) correspond to Cleckley. When examining conduct disorder + limited prosocial emotion, alongside Hare psychopathy, ^{Reference Hare3} only 6 of 20 Hare items are captured by DSM conduct disorder. This overlap is roughly the same for ICD-10 conduct disorder. Thus, the agreement between psychopathy and DSM, and possibly ICD specifiers, is low.

A three-factor model for psychopathic traits in children

The aforementioned findings suggest that the DSM-5 and ICD-11 could profit from a wider range of psychopathy items. In order to establish a well-validated conduct disorder diagnosis, or conduct disorder subtype(s), that is representative of child psychopathy and its various dimensions, it would need to capture a wider array of psychopathic traits. ^{Reference Cleckley1,Reference Patrick2} This larger representation would acknowledge the factor-analytic work on psychopathy that has consistently yielded three- and four-factor models in child, adolescent and adult referred, and non-referred, samples. ^{Reference Dong, Wu and Waldman6} Noting the robustness of these models, Frick stated that factor analyses have ‘consistently identified three dimensions [interpersonal, affective, impulsive], in addition to antisocial behavior, similar to those identified in adult samples’. ^{Reference Frick10} Similarly, Dong et al stated their results ‘strongly support the robustness of the three-factor model of psychopathic traits in children’. ^{Reference Dong, Wu and Waldman6}

So what are the problems holding us back from accepting a more representative set of psychopathy symptoms relevant for conduct disorder? Three problems, although poorly validated, stand out. First, there is an assumption that child psychopathy has a single phenotypic core (for example, callous–unemotional). However, this assumption has been advanced without a clear definition of core and without scientific support. Before core traits can be determined, there would have to be a definition for core and data to support the core proposition. Data on the core of psychopathy have been surprisingly absent.

Second, it is assumed that grandiose–manipulative and daring–impulsive traits are not as well researched, observable or as stable in children, as callous–unemotional traits. However, the literature on the trait elements of psychopathy, show that the syndrome and its underpinnings can be detected early. For instance, research on grandiose–manipulative traits has shown that very young children excessively dominate others, tell calculated lies and purposely mislead others. Studies have also shown that egocentric traits, such as the desire to be the focus of attention, are detected in early childhood and span into adulthood. Beliefs of superiority, have led to proactive aggression, ringleader bullying and violence. ^{Reference Baumeister, Smart and Boden11}

The importance of daring–impulsive traits is similarly recognised from its early observability, stability and association with adverse outcomes. Studies on inhibitory control show that between the ages of 2 and 3 years children develop simple skills, such as quashing a motor response. Excitement seeking has been investigated in children who are 2 years old and variation in excitement seeking is observable at a very young age with a disproportionate number of children being identified as ‘dare devils’. ^{Reference Morrongiello and Lasenby12} Daring–impulsive traits have been linked with substance use, aggression and other problems in adolescence. ^{Reference Salekin and Lynam5}

Finally, there is the assumption that the core traits provide sufficient information to negate the need for the assessment of other dimensions. From available research, a case can be made for considering the entire syndrome, as well as recognising its dimensions. In a study with 2000 pre-schoolers it was found that the combination of high levels of all three child psychopathy dimensions most strongly related to behaviour problems, even more so than callous–unemotional traits, ^{Reference Colins, Bijttebier, Broekaert and Andershed13} replicating findings discovered in other studies with children and adolescents. ^{Reference Salekin and Lynam5}

The child psychopathy symptoms are underrepresented in the DSM and ICD and there is no compelling research to suggest that callous–unemotional traits alone should be specified. Instead, there are reasons to add the specifiers of grandiose–manipulative and daring–impulsive traits. First, clinicians would benefit from being able to accurately identify youth with conduct disorder who exhibit the entire syndrome. Most research on child psychopathy has been conducted on the entire syndrome and clinicians are often faced with treating youth who exhibit the entire disorder. ^{Reference Salekin and Lynam5} Second, children with varying levels of psychopathy dimensions present quite differently from one another. By incorporating these three dimensions, clinicians can accurately describe and treat youth with psychopathic traits. Finally, research has shown that the psychological and biological processes that underlie each facet are different, and thus the strategies to intervene will vary.

Implications

Several questions are important to consider in further specifying conduct disorder. First, if grandiose–manipulative and daring–impulsive traits are incorporated, consideration will be needed to determine whether they should be integrated as categorical or dimensional models. One option would be to have the new specifiers modelled after the limited prosocial emotion specifier. Under this framework, key traits would need to be identified. Four key traits for each dimension would be needed and a minimum of two of four traits could be required as well as a time duration (for example, 12 months). Published factor-analytic and item-response theory studies would help in determining the traits to be implemented. Dimensional models might also be considered as they increase reliability. Second, although sufficient information exists to consider adding the specifiers, additional research may be needed to examine each specifier's interaction with conduct disorder. Third, there has been research to suggest that psychopathy is not necessarily accompanied by antisocial behaviour (conduct disorder) and that a broader shift is worth consideration. However, Quay has shown that child psychopathy symptoms can be helpful in understanding conduct disorder. Moreover, behavioural traits assure some behavioural problems are present. Finally, research is needed to determine whether the use of specifiers alone is justified. Kumsta and colleagues have shown that callous–unemotional traits can be present in the absence of conduct disorder and may require intervention. ^{Reference Kumsta, Sonuga-Barke and Rutter15} A hybrid model, where child psychopathy dimensions are first considered as specifiers and secondarily as separate potential treatment targets, requires consideration.

Although research is necessary to determine stigmatising effects, ignoring the mounting science illuminating the multidimensionality, early observability and unique correlates of child psychopathy could prove harmful. The inclusion of child psychopathy traits will result in greater resolution for understanding conduct disorder and inspire new effective treatment programmes tailored more specifically to the disorder.

In closing, diagnostic classification systems of disorders are an integral part of healthcare delivery. As such, we need to be precise in our description and treatment of youth with conduct disorder. Although there are remaining issues regarding implementation, research indicates that child psychopathy dimensions are highly relevant and recognition of their relevance is necessary to advance clinical care and world science on the topic of conduct disorder, thereby further improving the well-being of youth with conduct disorder.