Abstract
Persistent conduct problems with onset in childhood are a high-risk factor in the development of serious social and emotional adjustment problems in adolescence and adulthood [1, 2]]. In terms of prevention, the early identification of a subgroup likely to have a poor prognosis is a high priority. Cruelty to animals is one of the symptoms forming the diagnostic criteria for conduct disorder in the DSM-IV. It is of relevance to note that there were anecdotal reports of notorious serial killers found to have a history of cruelty to animals in childhood [3]. It raises the question of whether being cruel to animals (CTA) may be used as a marker of a subtype of persistent conduct problems predicting a poor outcome.
In the 1960s and 1970s, there were several studies using a case report method that pointed out the link between cruelty to animals in childhood and adolescence and recurrent aggressive behaviour in adulthood [4]. The three symptoms of enuresis, fire-setting and cruelty to animals were considered a set of strong markers of early antisocial personality development. According to psychoanalytic theorists, these markers implied an arrested infantile character whose symptoms were expressions of deviant sexual impulse [5]. These concepts were refuted in the 1970s [6]. Fire-setting and cruelty to animals were considered as part of a whole spectrum of symptoms in the diagnosis of conduct disorder. Enuresis is generally no longer considered a symptom of conduct problems. Fire-setting has been given more research attention in the past 20 years [7, 8, 9]] and yet a search of the literature found surprisingly no controlled clinical study on children who are cruel to animals. All the controlled clinical studies on cruelty to animals were carried out in an adult population. However, these studies did suggest a link between a childhood history of being CTA and a pattern of recurrent aggressive behaviour in adulthood [10].
Cruelty to animals was thought to be associated with adverse psychosocial factors. In a study of 152 criminals and non-criminals in Kansas and Connecticut, the family background of these adults and their history of being CTA in childhood were retrospectively examined [11]. Aggressive criminals have a high occurrence of history of being CTAand domestic violence, especially paternal violence and alcoholism. Adverse family situations and exposure to violence may explain cruelty to animals in childhood.
Apart from psychosocial factors, being CTA is also considered to be associated with biochemical abnormalities in the brain. Kruesi reported a case of low cerebrospinal fluid (CSF) concentration of a serotonin metabolite, 5-hydroxyindoleacetic acid (5-HIAA), in a girl who was CTA and had marked aggression [12]. Subsequently, he and his colleagues reported that physical aggression directed toward people and animals was associated with lower CSF concentration of 5-HIAA both at baseline and at 2 years follow-up of children younger than 12 years [13, 13, 14]]. Cruelty to animals may be a marker of a subtype of conduct disorder which has a biological link.
When considering the subtypes of conduct disorder, the usual comorbidities include attention deficit hyperactivity disorder (ADHD), depression and anxiety disorder. The relationship between being CTA and these subtypes is unclear. Cruelty to animals may be a result of impulsive behaviour which could be linked with ADHD. It may also occur when a child is affected by mood problems linking cruelty to animals with depression and anxiety. A related psychological characteristic which is important to investigate is that of self-esteem. It is generally reported that children with conduct disorder experience low levels of self-esteem. As there has been no controlled study investigating children who are CTA, it is unclear whether children with conduct problems and who are CTA differ in the way they perceive their self-worth and abilities.
We decided to revisit the area of cruelty to animals by re-analysing a data set of children presenting with persistent conduct problems to child and adolescent mental health services which was collected in the course of studying the outcomes of these children [15, Luk E: unpublished data]. We examined a group of children who were CTA as reported by their parents on an item of a commonly used parent self-report questionnaire, the Child Behaviour Check List (CBCL) [16]. As there has been no comparison study of children who are CTA in a clinical population, we think this will be a useful first step to research this issue. A number of questions are considered in this study: (i) do children with persistent conduct problems who are CTA have a more severe psychopathology than those who are not; (ii) is being CTA linked with adverse psychosocial factors; (iii) is being CTA more common in the male gender which is a possible link to biological and psychosocial aetiology; (iv) is being CTA linked with ADHD symptoms or the internalising disorder symptoms; and (v) is there any difference between the children who are CTA and those who aren't in terms of their self perception?
Method
As mentioned, this is a reanalysis of previously collected data. A previous study examining the outcomes of children presenting with persistent conduct problems were carried out between 1993 and 1996 [15, Luk E: unpublished data].
Subject selection criteria
Primary school children referred to a mental health service with symptoms suggestive of oppositional defiant/conduct disorder were screened. They were included if they had at least one definite conduct symptom, apart from cruelty to animals, as shown in either a parent or a teacher questionnaire, Child Behaviour Checklist Questionnaire (CBCL) or Teacher Report Form (TRF), respectively, [16] with a duration of at least more than 6 months. Our inclusion criteria was broad because even subclinical levels of persistent conduct problems were associated with a significant increase in psychological morbidities in adult life [17]. Exclusion criteria included evidence of neurological disorder, mental retardation or psychosis, child abuse under investigation, and custody issues awaiting court appearance.
For the purpose of this study, we recruited a sample of children from the community as a comparison group. With the permission of the education authority, two primary schools were contacted for recruitment of subjects in the community. A total of 37 children and their families agreed to participate.
Only one item on the CBCL and no item on the TRF refers specifically to cruelty to animals. Children referred to the mental health service who were rated on the CBCL as either sometimes (1) or definitely (2) cruel to animals were grouped together. Those who received a rating of not present (0) were treated as the clinical control group. Children recruited from the community formed the community control group.
Assessment procedure
All children referred to a regional mental health service who might fulfil the criteria of the study were screened by the research assistant. Parents were sent letters of introduction, an information sheet about the study, consent form and the screening instruments. Following that, they were contacted by telephone to explain the study further. With their consent, the teacher of the child was sent rating scales and contacted for a telephone interview. An interview with the parents and the child was arranged within 2 weeks to carry out the pre-treatment assessment. The interview was approximately 90 min in duration. Although we intended to interview both parents, only in 30% of the cases were fathers available for interview. In cases where both parents were available, only the mother's reports were used for analysis.
After obtaining consent from the parents and the school, the children in the community were interviewed at school and the parents filled in the questionnaires at home. Teachers were also asked to complete the questionnaires.
Measures
Parents’ report
The measures used for parents include: (1) the Child Behaviour Checklist (CBCL) [16], a well-established instrument providing a comprehensive assessment of the psychopathology of the child; (2) the Eyberg Child Behaviour Inventory (ECBI) [18] to assess the severity of the oppositional defiant/conduct symptoms (the ECBI has good psychometric properties and is a good instrument to measure change of conduct problems in primary school children); (3) the General Health Questionnaire (GHQ) [19] to assess parent/s’ mental health; (4) the Family Assessment Device (short version; FAD) [20], a 12-item scale that assesses the general family functioning; (5) the Spanier Dyadic Adjustment Scale (SDAS) [21] to assess the relationship between the parents; (6) demographic information including age, sex, parent's occupation and education, ethnicity, family structure and income. A social adversity factor was then constructed based on the family size, unemployment, education, separation, single parent and income with unitary weight on each factor [22]. The interviewer would explain to the parent if a parent could not understand any part of the questionnaires.
Child's report
Measures used for children include: (7) the Harter pictorial/normal self-perception profile for children (SPPFC) [23], a well-established instrument for measuring self-perception which has 36 items (for children who are not able to read, usually under 8, a 24-item pictorial version is used); (8) the Birleson Depression Self-rating Scale (BDSRS) [24], which measures negative mood in children and adolescents (for children who could not read, the 18 items were read to the child. The interviewer would make a judgment whether the child could understand the items).
Teacher's report
The measure used for the teacher include: (9) the Teacher Report Form (TRF) [16], a well-established instrument providing a comprehensive assessment of the child's strengths and difficulties from the teacher's perspective; (10) a teacher telephone interview in which the teacher was asked over the phone the symptoms of the child based on DSM-III-R criteria of attention deficit hyperactivity disorder. This measure was not carried out in the community sample.
Statistics
The two clinical groups and the community groups were compared according to four areas of interest: demographic; symptoms; self-perception; and parent's mental health and family functioning. MANOVA was used in the last three sections to account for the inflated chance of significance.
Results
Screening
As mentioned, this study is a re-analysis of data collected for outcome studies [15, Luk E: unpublished data]. Three hundred and nineteen children were screened: 60 did not give consent; 118 did not fulfil the criteria and were not interviewed. In total, 141 clinically referred children and 37 children from the community were included. Within the clinically referred group, 40 were rated as CTA on the CBCL item. Within the community group, one child was rated as such. This child was not included in further analysis in order to ensure a clear contrast between the comparison groups.
Comparison of the two clinic-referred groups and the community group
Demographic
As shown in Table 1, the community group had a significantly lower social adversity score than both clinical groups. The clinical groups have a higher male to female ratio and younger parents than the community group. There is, however, no significant difference between the clinical groups in terms of gender ratio and social adversity scores.
Symptoms
In Table 2, the symptoms as reported by the parent, the teacher and the child are shown. The difference between the community group and the two clinical groups are as expected.
Interestingly, the CTA group have greater problems and severity on the Eyberg Child Behaviour Inventory than the non-CTA, but there is no significant difference between the two clinical groups for the ADHD and depressive symptoms.
Self-perception
The results of the Harter Self Perception Profile For Children was unexpected. The CTA group has the highest score among the three groups: CTA group 111.0 (SD = 14.2), non-CTA clinic group 103 (SD = 17.4) and community group 107.4 (SD = 13.8) (F = 3.2, p < 0.05; Sheffe post hoc analysis showed no significant statistical difference). Because the younger age group used the 24 items pictorial version while the older age group used the 36 items version, a further analysis was carried out according to the two age groups. As shown in Table 3, the difference was only found in the older age group. The CTA group reported higher self-esteem than the non-CTA group with respect to scholastic and athletic competence.
Demographic data
MANOVA comparison between the three groups: symptoms and competence of the child
Parents’ mental health and family functioning
As shown in Table 4, the expected difference between the community group and the clinical group was again found. There was a trend suggesting that the CTA group has poorer family functioning.
Discussion
The results confirm that the clinical groups differ from the community group both in terms of symptoms and psychosocial aspects. However, the more interesting comparison is between the two clinical groups. Referring to the five questions we raised in the introduction, some of the answers are consistent with previous findings and others are surprising.
First of all, children who are CTAseem to have more severe conduct symptoms and it is possibly a marker linked to a subgroup of serious conduct problems. Second, there is a trend which suggests that children who are CTA come from families with more difficulties in family functioning. Third, there is also a slight trend that being CTA is linked with the male gender. A larger number of subjects are required to further clarify these differences in family functioning and gender. Fourth, cruelty to animals occurring in persistent conduct problems seems to be not explained by a link with ADHD or internalising disorder symptoms. Fifth, there is evidence that the older children who are CTA have a highly elevated self-perception. In particular, they rated themselves as significantly better at scholastic tasks than the non-CTA group, which contrasted with their mothers reporting them as significantly poorer than the non-CTA group on scholastic performance according to the CBCL. This is an interesting finding and obviously needs investigation.
Comparison of the three groups on the Self-Perception Profile for Children
Cruelty to animals and personality development
The link between persistent conduct disorder in childhood and social maladjustment in adulthood is well-established [25]. In the more severe cases, the social maladjustment presents in the form of personality disorder, notably antisocial personality disorder. The unexpected findings of a elevated self perception in the older children who are CTA may be related to the difficulties these subgroups of children experience during the process of personality development.
Recently, Frick and his colleagues proposed that the concept of psychopathy in adulthood can be extended to childhood [26]. They identified a callous and unemotional (CU) trait in the children of their sample which consisted of a dimension of behaviour characterised by lack of guilt, lack of empathy and superficial charm. These characteristics are considered primary in clinical descriptions of psychopathy in adults [27]. It is possible that this trait is consistent with the findings of our study which reported significantly elevated levels of self-worth in children with conduct disorder who were CTA compared to those who were not. In this respect, not only does our study provide data on the first controlled study of children with CTA but it also supports the link between childhood CTA and adult psychopathy. It should be emphasised that such an interpretation must be made cautiously and within the appropriate context. Such a link is of course not a direct linear relationship; many other factors come into play which leads to the development of adult psychopathy. Nonetheless, it is important to investigate such markers in our endeavour to improve our capacity for prevention of long-term antisocial behavioural problems.
MANOVA comparison between the three groups: parent and family measures
Limitations of the study
First, as our study is a re-analysis of existing data, we are unable to provide data on the detailed nature of being CTA. Clearly, we need to clarify parents’ interpretation of being CTA. In so far as CBCL is a commonly used assessment instrument, the validity of the item of CTA is important to establish. We are currently carrying out a validity study on parents’ reports of children being CTA by using a semistructured interviews to further explore the context and specifics surrounding being CTA in children. Second, our sample has not included cruelty to animals occurring in other clinical groups such as children with intellectual disability, psychosis and pervasive developmental disorder.
Clinical implications
In this study, being CTA was reported in 28% of the children with persistent conduct problems presenting to a mental health service. If this prevalence is confirmed, then this appears to be not an uncommon phenomenon which is under-recognised. Considering that it causes enormous suffering to animals [28], it is therefore also a serious social issue. A study of cruelty to animals in the community is required to examine its prevalence and correlates. Such studies are important for both clinical reasons as well as to protect the suffering of animals.
Footnotes
Acknowledgements
The research was supported by grants from the Royal Children's Hospital Research Foundation and the Psychiatric Branch, Human Services, Victoria, Australia. We would like to thank all the children and families who participated, the staff of the Royal Children's Hospital Mental Health Service and the Maroondah Hospital Child and Adolescent Mental Health Service for their assistance, our research assistants Melanie Davern and David Field, and the support of Professor Robert Adler and Dr Peter Birleson.