General framework

The literature on specific needs and risk assessment in adolescents who are violent, have mental health problems and/or offend is developing. The issue of assessment is closely linked to the issues of both prevention and treatment.

Reviewing both prospective and retrospective research suggests that victimisation and loss at an early age have consequences for future violent acts, in particular the association of physical abuse and violence. Stiffman et al (1996), in their US longitudinal study, found that a combination of personal variables (gender, substance misuse) and environmental variables (history of child abuse, stressful and traumatic events, rates of unemployment) predicted almost a third of the variance in adolescent violent behaviour.

Sheldrick (1999) suggests a traditional adult forensic approach to risk assessment in young offenders who are violent, breaking down the current incident/offence and historical data (Box 1). It is essential that assessment of empathy for the victim be carried out in the context of the emotional and developmental status of the young person and with an understanding of the protective mechanisms of amnesia and clinical denial following a violent offence/incident.

Risk escalates when there is failed multi-agency risk assessment and management, which includes failure to respond to reported episodes of violence, poor record-keeping and communication, and taking a cross-sectional rather than long-term view of the young person and his or her behaviour.

In assessment of young people in whom there is a risk of violence and mental health problems, certain key factors must be considered; these are listed in Box 2. When a young person has been assessed, it often becomes apparent that there are a number of overlapping risks (e.g. risk of harm to self, especially in male young offenders, can be lost in the level of concern about risk of harm to others). Practitioners should tease out these overlaps and state each risk as specifically and separately as possible. Risk should be predicted only for the immediate/short-term future. Regardless of the type of risk assessment used by the clinician, the multi-agency team needs to know the situational factors. Will the behaviour occur in the public/domestic domain, with/without provocation, in the immediate or extended vicinity? Will the victim(s) be perceived by the young person as vulnerable? Will the young person hold hostile attributions towards the victim? Will the victim represent certain meanings for the young person, e.g. authority figure, abuser, bully? Will the victim be in the immediate or in an extended vicinity? Who are the potential victims? (always list them). Does the young person lack concepts of emotion and/or lack safe problem-solving skills. Does the young person misuse substances, engage in other high-risk behaviour and is he/she easily provoked? (see Fig. 1).

Fig. 1 Assessment framework

General principles

Standard clinical assessment tools used in child and adolescent psychiatry cover many of the areas considered in forensic child and adolescent risk assessments (Reference Gowers and GowersGowers, 2001). This is especially important as juvenile justice systems in particular are not always equitable. In choosing between the many scales available it is important to question not just their proven scientific properties but also their feasibility for practitioners to use (given that resources available to CAMHS practitioners so often fall short of the ideal).

It is important to consider the purpose for which the scale be used (see Table 1). Scales that measure psychopathology may not be good ways of assessing the risk that the psychopathology poses. Measures used to map out types of symptom must have good content validity. An instrument required to pick out one group of symptomatic people from the rest of the community (e.g. mental health screening of young people in custody) needs to have good criterion validity. A related issue is the extent to which the scale is intended to measure change.

Table 1 Grid for specifiying requirements of a structured scale in a juvenile forensic population

	Assessment required (yes/no)
Purpose of assessment	Psychopathology	Need	Risk
Screening of all juveniles coming into contact with an agency
Detailed assessment, e.g. for sentencing, planning treatment
Measuring change, e.g. during treatment or sentence

Affective disorders

Depression is commonly present in delinquents. Clinic attenders with conduct disorder have rates of depression of 15–31% (Reference Goodyer, Herbert and SeckerGoodyer et al, 1997). As antisocial youths are generally viewed as difficult and disruptive, the risk is that depression will go unrecognised by non-mental health agencies. Where depression and conduct disorder coexist, risk of substance (especially alcohol) misuse is high in young male and female offenders. Young offenders with comorbidity for depression and conduct disorder are particularly vulnerable to suicide attempts. Mental health screening in the young offender population may also be important to ensure early detection of sub-threshold cases of mania, where there can be chronic and rapidly fluctuating symptoms of irritability, emotional lability, increased energy and reckless behaviours.

Early-onset psychosis

A prodromal phase of non-psychotic behavioural disturbance occurs in about half of cases of early-onset schizophrenia and can last between 1 and 7 years. It includes externalising behaviours, attention-deficit disorder and conduct disorder. Mental health screening should include a focus on changes in social functioning (often from an already chaotic baseline level) to a state including perceptual distortion, ideas of reference and delusional mood (Reference ClarkClark, 2001).

As in adult life (Reference Taylor and GunnTaylor & Gunn, 1999) most young people with schizophrenia are non-delinquent and non-violent. Nevertheless, there may be an increased risk of violence to others when they have active symptoms, especially when there is misuse of drugs or alcohol. The risk of violent acts is related to subjective feelings of tension, ideas of violence, delusional symptoms that incorporate named persons known to the individual, persecutory delusions, fear of imminent attack, feelings of sustained anger and fear, passivity experiences reducing the sense of self-control and command hallucinations. Protective factors include responding to and compliance with physical and psychosocial treatments, good social networks, a valued home environment, no interest in or knowledge of weapons as a means of violence, good insight into the psychiatric illness and any previous violent aggressive behaviour and a fear of their own potential for violence. These features require particular attention but the best predictors of future violent offending in young people with mental disorder are the same as those in the general adolescent population (Reference Clare, Bailey and ClarkClare et al, 2000).

Autistic spectrum disorders and learning disability

Autistic spectrum disorders are being increasingly recognised but are often overlooked in forensic groups. Their identification is critical to the understanding of offending, especially in adolescents with a learning disability. This is particularly so if an offence or assault is bizarre in nature, the degree or nature of aggression is unaccountable and/or there is a stereotypic pattern of offending. O'Brien (1996) and Howlin (1997) proposed four reasons for offending and aggression in autistic persons:

1. 1. their social naivety may allow them to be led into criminal acts by others;

2. 2. aggression may arise from a disruption of routines;

3. 3. antisocial behaviour may stem from a lack of understanding or misinterpretation of social cues;

4. 4. crimes may reflect obsessions, especially when these involve morbid fascination with violence – there are similarities with the intense and obsessional nature of fantasies described in some adult sadists.

Knowledge about criminal behaviour in young people with a learning disability is still developing (Reference HallHall, 2000). The clinical issues that arise concern the additional care needed in interviewing at assessment, as much as the possible special implications with respect to antisocial risk that arise from their cognitive level or profile. As with anyone else, associated psychopathology, life experiences and personal background should constitute part of the assessment.

‘Prodromal’ personality disorder

Models used to describe severe personality disorder in adults (Royal College of Psychiatrists, 1999) may also be applied to young people. Hare (1993) differentiated between ‘psychopathic personality’, characterised by an unemotional and callous interpersonal style, and ‘antisocial personality disorder’, in which the key elements are poor impulse control and antisocial behaviour. O'Brien & Frick (1996; Reference FrickFrick, 1998) argued that these two groups have a different aetiology. Christian et al (1997) described a group of children with a callous, unemotional interpersonal style, who lacked guilt and who did not show empathy or emotions. Box 4 lists various factors thought to direct the pathway to serious antisocial behaviour and interpersonal violence.

Violence witnessed by or perpetrated against the child

The theoretical work of the 1970s provided a framework for understanding the multiply determined nature of abuse and neglect (Reference Ammerman, Hersen, Ammerman and HersenAmmerman & Hersen, 1992). Ecological models were said to interact and converge to bring about family violence. This belief led to an increased interest in clinical assessment and treatment. The epidemiological research of the 1980s documented the widespread prevalence of child abuse, neglect, spouse battering, elder mistreatment and psychological abuse. This in turn fuelled the development and evaluation of interventions for both victims and perpetrators.

Predominant factors within families that contribute to longer-term aggressiveness and risk of violence have been clearly identified in child rearing and parenting styles (Reference Farrington and BoswellFarrington, 2000). Against a backdrop of multiple deprivation three key clusters emerge. The first is the presence of criminal parents and siblings with behavioural problems. In the second, the day-to-day behaviour of primary caregivers is one of parental conflict, inconsistent supervision and physical and emotional neglect, with little or no reinforcement of pro-social behaviours. The child learns that his or her own aversive behaviour stops unwanted intrusions by the parents/caregivers. Young people who assault others have lower rates of positive communication with their families. A third cluster of family factors linked with later violence in the child include cruel authoritarian discipline, physical control and shaming and emotional degradation of the child.

Where young people live in a culture of violence in their homes and neighbourhoods, school has the potential to offer positive outlets to satisfy needs for belonging and recognition and acceptance through non-violent means. Failure to achieve academic success, peer approval and satisfying interpersonal relationships within the school context creates additional stresses and conflicts, with increased likelihood of aggression and violence linked to competition for status and status-related confrontations with peers (Reference Laub, Lauristen, Elliot, Hamburg and WilliamsLaub & Lauristen, 1998).

Domestic violence remains one of the most pervasive social problems, affecting most of the population directly or indirectly (Reference BlacklockBlacklock, 2001). One in four women experience domestic violence at some time, and a study by the National Children's Homes Action for Children in 1994 found that 75% of mothers subjected to domestic violence said that their children had witnessed it (Reference BlacklockBlacklock, 2001).

Abuse of parents by children

There is remarkably little known or documented on children violent to their parents, which is surprising given the level of knowledge and concern about antisocial and violent behaviour committed by children in school or in the community. The abuse of parents (with the exception of the elderly and the extreme act of parricide) can at best be seen as underresearched and at worst as a taboo subject associated with a sense of shame that parents cannot protect themselves against the violence of those with whose care and protection they are charged.

Seminal work by Strauss, Gelles and Steinmetz dating back to 1980 (Reference Strauss and GellesStrauss & Gelles, 1990) estimated that 18% of children between the ages of 3 and 17 years carried out one or more acts of violence towards parents. The victim was most often the mother; the violence ranged in severity from 10% of children hitting or striking a parent, 7% striking a parent with an object, 1% severely beating a parent to 1% using a potentially lethal weapon.

What is known about children who are violent to parents? Adolescents who have assaulted peers, siblings, teachers and strangers are at increased risk of having assaulted or threatened to inflict physical violence on their parents. Children who inflicted physical violence on parents were described by Harbin & Madden (1979) as more often economically dependent, male and aged between 13 and 24 years. Adolescents who are violent to parents appear to straddle socio-economic groupings. Family dynamics provide the greatest insight into the evolution of this particular form of violence in children (Reference Harbin and MaddenHarbin & Madden, 1983). It is common to find that the parents of such children convey inconsistent values about aggression. Although severe child abuse in this group is unusual, parents who use harsh child rearing techniques are at increased risk of being victims of violence at the hands of their children. Role reversal between parent and child (parentification) takes place, to the extent that parental authority is given over to the child.

Parents tend to deny the seriousness of acts of aggression and violence displayed against them. In the clinical situation when a child is referred because of violent acts occurring in the community or school, parents will not always reveal a preceding or coexisting history of violence in the home.

Violent assaults against parents are most likely to occur in the context of parenting stress arising from parent–child disagreement. Failure to set consistent limits on unacceptable behaviour produces a dynamic in which the child experiences an evolving sense of empowerment and, indeed, a sense of justification in dealing with increasingly failing attempts at parenting by escalating both the frequency and severity of violent responses. The tyrannical child is rewarded and reinforced in this behaviour by the parents giving in to this graded continuum of physical aggression.

The evolution of parental inconsistency and loss of authority that includes fear of the child can be likened to the need to maintain relationships through co-dependency as seen in adult male-to-female domestic violence. In domestic violence, the violence and abusive behaviours are used to control. The abusive behaviour is also used to support the sense of entitlement that allows an adult male to see his behaviour as reasonable given his partner's ‘unreasonable’ resistance to his expectation. This further fuels the process of partner blaming (Reference BlacklockBlacklock, 2001). Whatever other factors place a child at risk of being violent, this sense of entitlement and experience of being in control can lead to aggression and violence becoming the child's normal form of social transaction in the home and then beyond.

Using a semi-structured approach to assessment of violence

Assessment of violence, in its various settings, degrees of severity and the presence or absence of mental disorder, must be approached systematically. Again the adult literature may be called upon, as it contains graded series' of questions that may be used with children and adolescents to ask about their violence (Box 5). By first giving the young person the opportunity to talk about violence done to him or her, whether at home, at school or in the community, the clinician will be able to gain an informed understanding of why this child is presenting at this time with this particular violent act (Reference Sheidow, Gorman Smith and TolanSheidow et al, 2001). The same structure can be used to take a history from parents, allowing them to explore with the clinician the depth and extent of violence displayed towards them behind the closed doors of home.

Interviewing a young person

Even with the prospect in the near future of a full set of gold-standard risk assessment tools for children and adolescents, the reality remains that the clinician often must interview a young person who is reluctant to be seen. The young person may have grown to view him-/herself as a constant focus of complaint, criticism and blame, ‘assessed out’ by a range of professionals, and skilled in the avoidance techniques of passive or overt hostility. This can make safe engagement a daunting prospect.

Complex cases take time, and regardless of requests for a ‘quick-fix’ risk assessment, the CAMHS team should carefully think through how to approach any assessment. Some general principles of risk assessment are outlined in Box 6and discussed below. The team should insist that all known information is provided and that the referrers are open and honest about the purpose of the assessment. Where possible, the team should provide the young person and the family with user-friendly information about the service and the assessment process. In an emergency, whether in a clinic, the home, a police station or residential open or secure facility, the team should insist that, as far as is possible in any crisis, its own emotional and physical safety and those of other professionals and the young person are assured. In no circumstances should an individual enter a situation alone without immediate back-up.

It seems likely that CAMHS teams are to be increasingly asked to undertake risk assessment. When the risk behaviour is that of violence, health service managers must themselves make an adequate ‘risk assessment’ of the physical circumstances in which the CAMHS team will work. They must also ensure that all staff have appropriate training not only in child and adolescent risk assessment but also in self-protection – as a minimum, in disengagement techniques.

In an emergency assessment it is vital to stress that any examination can provide only a ‘here and now’ snapshot of the young person, who is often highly aroused because of the emotional crescendo surrounding him or her. The latter is all too often a response to the fear and sense of hopelessness, helplessness, frustration and, at times, anger that the young person's behaviour has engendered in carers, who may be emotionally and physically exhausted and feel unsupported.

Allowing the young person to (a) state his/her own list of percieved problems and (b) state an opinion on the list of problem behaviours adults believe he/she is demonstrating, can, whether at interview resulting in a long silence or verbal abuse, allow the young person to have a voice in his/her own assessment. This can at least in some cases allow the young person to use the assessment constructively. At the outset it is essential to make clear the boundaries of confidentiality, which by the time a crisis is reached, may mean that much of what is said by the young person has become ‘need-to-know’ information (Reference Bailey and CordessBailey, 2001).

In most cases, the young people can sustain interview and start to share with the interviewer their account of the escalation of violent behaviour. At each stage ask the individuals what they think and feel and about accompanying physiological experience. Start with injustices and acts against them and then what they have done to others. A systematic approach avoids the pitfalls of stereotyped beliefs about types of violence committed by boys and girls and opens up issues of violence against and by children in the family home. It exposes those children who have genuine difficulty linking thoughts, feelings and the physical responses of their bodies when experiencing sadness and anger under varying levels of stress and/or when the end result is violent behaviour. It therefore offers a more complete picture of the young person's behaviour, and the assessment can then inform not only treatment choices but how delivery of treatment may need to be modified for the individual child.

This process takes time but it has the added advantage that during assessment the young people start to gain insight into themselves, which can enable them to be open about their fear, or indeed lack of fear, of their potential to commit more serious acts of violence in the future. This approach has been adopted by the Forensic Adolescent Community Team Service (FACTS) in Salford to address risk of violence, and similar systematic approaches have been used with children involved in substance misuse, criminal behaviours and fire-setting and with children who are sexually abusive. When combined with holistic needs assessment, complex cases can start to be broken down into constituent components, which can be addressed by different parts of the multi-agency services involved, utilising a Care Programme Approach (CPA) and social services care planning. The developing adolescent risk assessment tool derived from the HCR–20 sits well with this approach, as its risk component has similarities to the CPA. Key to safe assessment is bringing the young person actively into the process from the outset. CAMHS are skilled in the art of clinical engagement and this ultimately is most likely to increase the effectiveness of an informed-needs-led multi-agency response to risk.