Adolescent-to-Parent Violence in Adoptive Families.

Adolescent-to-parent violence (APV) has received little attention in the social work literature, although it is known to be a factor in families whose children are at risk of entry to care. The behaviour patterns that characterise APV include coercive control, domination and intimidation. Crucially, parental behaviours are compromised by fear of violence. This article discusses the unexpected findings from two recent adoption studies of previously looked after children in England and Wales. The studies exposed the prevalence of APV in the lives of families who had experienced an adoption disruption and those who were finding parenting very challenging. Two main APV patterns emerged: early onset (pre-puberty) that escalated during adolescence, and late onset that surfaced during puberty and rapidly escalated. The stigma and shame associated with APV delayed help seeking. The response from services was often to blame the adoptive parents and to instigate child protection procedures. There is an urgent need for a greater professional recognition of APV and for interventions to be evaluated with children who have been maltreated and showing symptoms of trauma.

Introduction

There has been a silence in the social work literature and in UK government policy on young people who are the instigators of violent behaviours in the home (Gallagher, 2008; Holt, 2012). These behaviours are known as child-to-parent violence (CPV), adolescent-to-parent violence (APV) or mother abuse (Hunter ). There is not one specific behaviour that marks out APV, but a pattern of habitual coercive behaviours that reverse the usual parent/child power dynamic (Wilcox, 2012). Even though the behaviours that mark out APV are familiar to social workers (and are often referred to as ‘challenging behaviour’), the CPV/APV behaviour patterns have gone unrecognised (Nixon, 2012). Consequently, interventions are often inappropriate or inadequate and are implemented in the absence of policy and practice guidelines (Kennair and Mellor, 2007; Holt and Retford, 2013).

The silence may be because it creates difficulties for social workers who are used to working with children who are the victims of violence and not the instigators. Raising the topic of APV is difficult for a profession that seeks to counteract the demonisation of young people and does not fit easily into the gendered/power theories and interventions (Wilcox, 2012) that underpin much of the work on intimate partner violence (IPV).

Most of the APV literature has been published in the last ten years and comes from academics working in the criminal justice field (e.g. Holt, 2012) or from practitioners outside the UK (e.g. Gallagher, 2008; O'Connor, 2007). There is surprisingly little UK social work literature, Biehal (2012) being an exception. In an emerging field of study, there is a lack of consensus on every aspect of APV including definitions, terminology, causes and interventions.

In this article, we focus on our own unexpected discovery of APV when researching adoptions that had disrupted or were in difficulty. Our study found that APV was responsible for most of the adoption disruptions.

Background

Whilst there is a growing literature (Helen Bonnick has usefully collated the available evidence on her website, ), APV remains an under-theorised and under-researched area of family violence (Gallagher, 2008; Wilcox, 2012; Holt, 2012). It has been described as the last taboo about violence in the home (Condry and Miles, 2014). There is no agreed APV definition, as it describes a wide variety of physical and psychological behaviours designed to control, coerce and dominate the parent and family members. In this article (and in our analyses of interview data), we applied Paterson and colleagues' definition of APV, as it highlights not only behavioural patterns, but also the impact on parenting, and places the behaviours within the context of relationships. They described APV as:

Behaviour considered to be violent if others in the family feel threatened, intimidated or controlled by it and if they believe that they must adjust their own behaviour to accommodate threats or anticipation of violence (Paterson , p. 92, emphasis added).

The lack of an agreed definition has resulted in the estimates of APV varying between 3 and 27 per cent (Gallagher, 2008; Holt, 2012). In the UK, Parentline Plus (renamed Family Lives) reported that 17 per cent of the 30,000 telephone calls to their helpline in 2008 were from parents concerned about their adolescents' verbal aggression and a further 8 per cent were concerned about physical aggression. Inevitably, surveys that ask about any violent incidents in the past six months produce high prevalence rates whilst research that has focused on patterns of controlling habitual behaviours produce much lower rates (Holt, 2012). Some surveys count any type of violence in the family whereas others exclude sibling violence. Prevalence rates are also affected by parents' unwillingness to report APV. This form of family violence carries multiple stigmas and shame. It stigmatises a parent as a ‘bad parent’ and usually this means mother blaming (Downey, 1997; Edenborough ).

The evidence is very mixed on the socio-demographics of instigators and their victims (see reviews Cottrell and Monk, 2004; Kennair and Mellor, 2007; Walsh and Krienert, 2007; Hong ; Holt, 2012). General population surveys tend to find no significant difference by the gender of the instigator, whilst studies that have used police reports do find a gendered phenomenon. For example, Condry and Miles (2014), using Metropolitan Police reports of 2,336 APV incidents between April 2009 and March 2010, found that 87 per cent of instigators were sons and 77 per cent of victims were female and usually mothers. Whilst this reflects the gendered profiles seen in IPV, there is a lack of research on whether mothers are more likely to reports assaults than are fathers, girls are more likely to use other forms of control and whether the police take more seriously the assaults by boys on their mothers.

Causes

There is a recognition that APV is more likely to occur in families where children have been exposed to IPV (Cottrell and Monk, 2004; Hunter ). Not all children exposed to IPV go on to become violent and theories of the ‘cycle of violence’ have been heavily critiqued for being too deterministic (Baker, 2012). Nevertheless, most studies of APV show an association between children who are aggressive and exposure to IPV, but the mechanisms are unclear. It may be that children learn that controlling behaviours are the way to deal with conflict, they may idealise the perpetrator of the violence and copy the behaviours, or absorb the messages that aggression is an acceptable way to treat women or be angry with the mother for not protecting herself or the children (Nixon, 2012; Cottrell and Monk, 2004). In much of the literature, there is an assumption that, by changing the family environment (i.e. removal of the violent partner), the problem of APV will diminish. Some writers view APV as another form of domestic violence and state that its gendered nature should be recognised and referred to as ‘mother abuse’ (Hunter ).

Specific parenting styles have also been examined. There is some evidence that aggression is transmitted across generations and that harsh paternal parenting has a stronger effect on children's levels of aggression than does harsh maternal parenting (e.g. Marler ). Other writers (e.g. Gallagher, 2008; Kotch ) have linked aggression with laissez faire or neglectful parenting styles resulting in young people growing up with an exaggerated sense of entitlement.

Early difficulties in the relationship between the child and their primary attachment figure have also been associated with the development of violence. Tremblay (2000) describes how children's physical aggression peaks at around two years of age: commonly known as the terrible twos. Children learn to control aggression and impulsivity and develop the capacity to recognise and reflect on emotions and behaviours in themselves and others through their parents' sensitive and attuned responses to their behaviours. It is these capacities often referred to as mentalization that are thought to inhibit aggression. Fonagy (2004, 2012) argues that it is the failure of normal developmental processes to control aggression that signals violence and that violence happens in the absence of mentalization.

The development of aggressive behaviour has also been associated with child maltreatment and the subsequent stress experienced by children. There is evidence that chronic stress biologically alters the stress response resulting in hyper-vigilance, changes to normal cortisol pattern, alterations to reward processing and errors in identifying emotions correctly (e.g. McCrory ; Jaffe and Christian, 2014). Others (e.g. Zeanah, 2009; Corriveau ) have focused on the impact of maltreatment on the child/parent relationship and the development of insecure avoidant styles of relating, resulting in children who fear intimacy, lack trust in adults and whose focus turns away from relationships to controlling their environments. From a different perspective, Kids Company (2009), through its extensive work with abused children and concern at rising levels of violence, has proposed a new clinical concept of a ‘violence adapting syndrome’ characterised by abused children in persistent states of hyper-arousal.

The focus of much of the research has been on the development of aggression and violence and developmental models have rarely been applied to the problem of APV (see Cottrell and Monk, 2004; Hong ). It is likely that multiple factors increase the risk of CPV/APV behaviours. Children who are adopted from care are likely to carry all or most of the identified risks and we now turn to our research findings to consider aggression and APV in the context of adoptive family life.

The research studies

Two studies on adoption disruption were recently completed: one funded by the Department for Education (Selwyn ) and the other funded by the Welsh Government (Selwyn and Meakings, 2015). Adoption disruption was defined as a child who had been adopted out of care and who had left their adoptive family under the age of 18 years old. The aims of both studies, using similar methodology, were to: calculate the rate of adoption disruption in England and in Wales, and to explore the experiences of those involved in or at risk of disruption. Ethical permission was obtained from the School for Policy Studies ethics committee at the University of Bristol.

Using national data, we found that the post-order adoption disruption rate was very low: only 3.2 per cent over a twelve-year period in England and 2.6 per cent over an eleven-year period in Wales. Adoption disruption is not a common event. Nevertheless, an adoption disruption is a traumatic event and so, to understand more about the experience of disruption, a sample of adoptive parents was recruited. Detailed methodology can be found at . In brief, a survey was sent out by thirteen English local authorities to parents who had adopted a child between 1 April 2002 and 31 March 2004. Parents were asked how the adoption was faring and whether the child was still living at home. The same survey was replicated on the Adoption UK website and could be completed by any parent who had adopted a child from care. Surveys were returned by 390 parents caring for 689 adopted children. A quarter of the parents whose child was still living at home stated that they were finding parenting very challenging and were struggling. From the 390 survey responses, all the parents (n = 35) who had experienced an adoption disruption and thirty-five parents who described parenting a child living at home as very challenging were selected to form the English interview sample. In addition, twenty Welsh families (ten disruptions and ten who were currently finding parenting challenging) were recruited using information from local authorities and snowballing techniques. Face-to-face interviews (average length of interview three hours) were completed with ninety adoptive families: forty-five parents who had experienced a disruption (the ‘Left home’ group) and forty-five parents who were finding parenting challenging (the ‘At home’ group). Numerical data were analysed in SPSS using non-parametric tests. Qualitative data were analysed using a framework approach (Ritchie and Spencer, 1993) with the themes, concepts and ideas identified prior to data collection. Analysis used five key stages of familiarisation with the data and the context, identification of themes, indexing, mapping and interpretation.

At the time of the interview in 2013/14, the children whose adoptions had disrupted were on average seventeen years old (SD = 2.93) and those living at home were on average fifteen years old (SD = 2.43). Most (87 per cent) of the disruptions had occurred during the young people's teenage years. The majority (n = 73) of the parents were part of a couple, including two same-sex couples, and seventeen were single adoptive parents. Thirteen households contained no children but most (85 per cent) had other adopted children or birth children (range from zero to four children) still living at home.

It was during the analysis of the interview data that CPV/APV emerged as the main reason adoptions had disrupted and prevalent in the families who described parenting as very challenging. APV was a factor in thirty-eight of the forty-five adoption disruptions. Before exploring the behaviours that challenged the parents, it is important to consider what was known, by the parents, about their child's pre-adoption history.

The adopted children

All of the ninety adopted children (forty-eight boys and forty-two girls) were carrying risks of poor developmental outcomes as they entered care. The children whose adoptions later disrupted were carrying even greater risks. Table 1 highlights the older age of the ‘Left home’ group at key time points in the adoption process. Children whose adoptions disrupted were, on average, three times older at entry to care compared to the average age (fourteen months old) of most adopted children (Selwyn ).

Table 1

The ages of the ninety young people at entry to care, at adoption and at disruption

Children	Age at entry to care	Age at adoptive placement	Age at the Adoption Order	Age at disruption
Adoptions disrupted (n = 45)	Average 3.6 years (SD = 2.00)Range 0–10 years	Average 5.3 years (SD = 2.18)Range 0–11 years	Average 6.0 years (SD = 2.54)Range 1–12 years	Average 14 years (SD = 2.45)Range 6–17 years
At home but very challenging (n = 45)	Average 2.1 years (SD = 1.88)Range 0–6 years	Average 3.7 years (SD = 2.56)Range 0–10 years	Average 4.5 years (SD = 2.77)Range 0–12 years

Age at entry to care usually reflects the length of exposure to maltreatment and most of the children had entered care because of abuse and neglect (Table 2). Many children had suffered multiple forms of abuse. Maternal alcohol and /or drug misuse (55 per cent) was common, resulting in eleven children being removed by Children's Services at birth. Often, little was known about birth fathers apart from the violence they had inflicted on their partners and children. Over a third (34 per cent) of birth fathers had served a prison sentence for crimes involving violence.

Table 2

Children's exposure to maltreatment in the birth family prior to adoption (n = 90)

	Children who had left home (disruptions), n = 45 (%)	Children at home, n = 45 (%)
Neglect	42 (93)	36 (80)
Domestic violence	38 (84)	25 (55)
Physical abuse	23 (51)	18 (40)
Sexual abuse	22 (49)	7 (15)
Rejection/abandonment	12 (27)	15 (33)

The children had also faced additional adversities pre-adoption: sixty-one (68 per cent) had experienced more than one move in foster-care (range zero to thirteen moves) and forty-two (46 per cent) of the adoptive parents had concerns about the quality of the foster-care shown to their child with adopters expressing concern at ‘cold’, ‘clinical’ or ‘professional’ foster-care. Nearly one in four (23 per cent) of the children had experienced poorly managed introductions and a difficult transition into the adoptive family. It was also apparent from adoptive parents' accounts that many of the children had not been well prepared and did not understand why they were being adopted.

The children whose adoptions had disrupted had faced more adversities in the birth family and whilst they were looked after compared with the children who were still at home. They were statistically more likely to have experienced domestic violence (Fisher's exact = 0.05), sexual abuse (Fisher's exact = 0.00) and to have had longer exposure to these adversities (Mann-Whitney U = 279.00, z = –2.458, p < 0.014). They were also five times more likely to have had two or more moves in foster-care prior to joining their adoptive family in comparison with the children in the ‘At home’ group (Fisher's exact = 0.001).

Violent, controlling and coercive behaviours (n = 62)

Parents were asked about the nature of the difficulties they had faced in adoptive family life. Seventy-five of the ninety parents stated that children's physically aggressive behaviour had been or was very challenging and of these sixty-two parents described sustained patterns of behaviours that met Paterson's definition of APV. All but seven of the forty-five families who had experienced a disruption involved APV, as did seventeen of the forty-five families who were currently finding parenting very challenging.

Parents were also asked when serious difficulties first began. The majority of parents (80 per cent) described an early onset of aggressive and controlling behaviours shown by their child whilst still in primary education. Whilst children were young, parents could manage outbursts by physically removing children to prevent harm but, even so, some of the parents had been hurt (e.g. broken nose). During puberty, the power balance shifted. Adolescents were more independent and their physical maturity meant that parents could no longer remove the adolescent or escape themselves. It was this shift that tipped the aggression into APV. For example, parents said:

He trapped me in his room. He took the mattress off his bed and pushed the mattress up against me. I was rammed against the wall. He pinned my shoulders against the wall and was snorting and growling at me … I was really under threat, and I was really scared. I got out of there and I thought, ‘I shouldn't have gone in his room. I shouldn't have let him get me in this position’ (At home).

She's physically attacked me …. When she was smaller I could pick her up and put her in a room. I've not been able to do that since she was 9 or 10 …. She's verbally aggressive, destroys things … made allegations against my brother. What she's done is sabotaged all my support networks as well …. She just wants to control me …. She really bullies me, stops me going into her room, puts her foot in the door, finger in my face, telling me to f*** off and laughs because I can't do anything about it, because she's so much physically bigger (At home).

Not all aggressive and controlling behaviours were evident from early on. Twenty per cent of parents described a sudden onset of APV associated with puberty. Parents often likened the onset of difficulties to a switch being flicked, with young people's behaviour changing suddenly and accompanied by a very rapid escalation of severe violence. One parent explained:

He just turned from this lovely little kid into this very angry person … it was like a switch (Left home).

All of the families where there had been late onset APV had disrupted whereas 54 per cent of the early onset children were still living at home.

Parents described being assaulted, intimidated, threatened, locked in rooms, money stolen, mobile phones hidden, property damaged and even taken hostage in their home. Weapons were used by some young people: knives most frequently but also scissors or other implements that came to hand. Most (n = 38) of the violent controlling behaviour was sons to their mothers, but twenty-four girls also engaged in these behaviours and fathers and siblings had been attacked, as had pets and grandparents. The following interview extract illustrates the different behaviours intended to control and intimidate:

From the minute he got up to the minute he went to bed he just terrorised us … threatening us with knives … throwing stones at us, throwing buckets of water at us, squirting us with bleach … the TV was locked in his bedroom …. You would be walking along and he would suddenly just punch you in the back for no reason …. You couldn't even leave the dogs with him. If they were laying in here and Freddie [study child] walked in they would leave and I've known one of them to wet herself [in fear]. [Husband] was beaten round the head with a broom. I can remember one night … we went to bed and lay there and I can remember crying …. He came in and he punched me in the back and he said, ‘Yes, you cry you bitch’ (Left home).

Parents gave graphic accounts of the violence they endured and frequently spoke of how they had had to change their parenting behaviour to protect themselves and their other children. For example, parents said that they had to put locks on bedroom doors to prevent attacks whilst they were asleep. Some couples had split parenting tasks: one parent parenting the adolescent who was the instigator of the violence and the other parent focusing on the other children in the household. Mothers especially were afraid of being left alone with their child and parental patterns of work and leisure were altered to ensure that this did not occur:

If he goes to school in the morning and he's angry, as much as I love him, I dread him coming home … it's got to the point where I've said to [husband] that he's not to go out down the gym [and leave me alone with child] (At home).

Parents said they felt helpless and defeated because young people refused to acknowledge or accept parental authority. Yet, most spoke of their love for the child and wished help could have been found.

In some families, the violence was predictable and would erupt if the adolescents' wishes were not immediately met, such as asking the young person to come home at a certain time. Parents were fearful and avoided confrontations. For example, one mother said:

I would never have been surprised if she had stabbed me in my sleep, never. I had to change lots of things … back down on things for fear that she would trash the house and break everything (Left home).

In other families, parents could not predict when the intimidation and violence would surface. Parents spoke of ‘walking on eggshells’ and ‘psychological warfare’, as they discovered knives and other weapons hidden around the house, believing that they were intended to be found. Parents felt intimidated and said, for example:

If I thought he was going to go off into a mood I'd turn around and walk away, but he would follow. He wouldn't let me walk away …. He liked to invade your personal space, get up really close and intimidate …. He would grab me round the throat. I was really quite scared … I didn't know if he was going to stop or not (Left home).

Fathers who were the focus of the violence felt particularly shamed at being overpowered or controlled by their child. Some fathers had retreated to avoid conflict. For example, one father avoided contact with his son by going to bed at 8.00 p.m. every night when he came home from work. One mother described the attacks:

Then he started turning on my husband at the age of 15 …. It was physical violence and intimidation …. My husband's been in the corner of the kitchen on the floor, trapped, crying, physically crying, he's so scared (Left home).

More frequently, fathers expressed fears for their wives or for other children's safety and felt that they were reneging on their duty to protect their family. They wondered what was happening at home whilst they were out at work and struggled to talk about these fears to the often young female social worker who visited the family.

Having a child who was controlling and violent was very shameful and especially so for adoptive parents who felt they had failed, were inadequate and felt total hopelessness about the situation. Many parents had become isolated, as either friends had stopped visiting because of what they witnessed in the household or the adoptive parents felt such shame that they had stopped inviting. Home was a not a safe place, as one mother explained:

I'm not safe in my own house. I can't breathe, I can't relax (At home).

Feelings of failure were exacerbated by the response from the agencies that adopters turned to for help. Previous research on APV (Holt and Retford, 2013; Nixon, 2012) has consistently found a very poor response from services with some interventions making the problem worse and this was also the case for the adoptive families. The shame and stigma of living with violence in the home meant that many of the parents did not ask for help until they were desperate but most parents received a poor response. Parents reported that they were often told that either the local authority (LA) or Child and Adolescent Mental Health Services (CAMHS) did not provide services for adopted children or the behaviours did not fit the criteria for services. In quite a few cases, parents were told by social workers to ring the police if they needed support. Three-quarters of the families had had some police involvement and some parents described police officers as being like their social worker and their best source of support.

A quarter of the families did rate social workers as their best source of support. Good support to adoptive families was provided by LAs who were able to provide multidisciplinary support. Parents appreciated LA interventions that were provided by a ‘team around the family’. These teams comprised social workers, psychologists, mentors and occasionally educational psychologists and occupational therapists. A few parents also received good support from CAMHS that had an adoption specialism.

Rather than receiving a responsive service, parents felt that social workers often blamed them for the child's behaviour. Indeed, most of the blame was heaped on adoptive mothers who said that they were made to feel as though they had ‘failed’ and had ‘let down’ the child. Professionals did not recognise APV, but instead framed the difficulties in the context of poor (adoptive) parenting and/or saw the child behaviours as a problem of anger management. For the most part, the controlling elements of the behaviours went unrecognised, as did the battle for the ‘parental space’. Here, we are using the term ‘parental space’ to refer to the parents' position within the family and in particular nurturing and controlling roles. By the time agencies were involved, many parents had given up that space and it was occupied by the young person. Attempts to regain the space, by trying to insist on boundaries, were thwarted by some young people who made allegations of abuse that led onto child protection investigations. Most of the parents vehemently refuted the allegations, although a few parents (both fathers and mothers) admitting to having ‘lost it’ or having fought back in self-defence. For example, a mother said:

She beat her dad up. She just started punching and kicking and punching him, absolutely going berserk, I mean unhinged berserk. He's a very gentle giant, never ever laid a finger on her … he wouldn't do, it's what she wanted. She always used to bully him quite a lot and something inside me just snapped … and I just got in the middle and I slapped her round the face, which I regret, but I did slap her (Left home).

Other child protection investigations were begun by children's social workers as a response to the difficulties in the family. Several parents who reported symptoms associated with secondary trauma described how they were investigated when social workers suggested that they were not showing their child sufficient emotional warmth. None of the children was removed because of Section 47 investigations but the process of the investigation left the parents feeling undermined and traumatised. Parents thought it left the young person in a more powerful position, as parents were fearful of doing anything that might result in another allegation with the real possibility that they might lose their jobs (about half of the parents worked in health, education and social care). Allegations also often created splits and conflicts between children and adopters' social workers, with adoption workers being accused of colluding with the adoptive parents and the children's social workers accused of not understanding the adoption context. The aftermath of an investigation left workers feeling metaphorically bruised and parents unsupported and did not result in meaningful support for the young person or the family. Parents whose child had ‘Left home’ were statistically more likely to say they felt blamed by professionals in comparison with parents whose child was still at home (Fisher's exact = 0.001).

The pattern of professionals failing to recognise controlling behaviours and support parents continued for the families whose children re-entered care. Young people usually insisted that their parents should be kept in the dark and excluded from case conferences or reviews. Generally, social workers went along with these demands and parents described how professionals had seemed to take sides with the child, thus exacerbating the dysfunctional power dynamics within the adoptive family. Parents were frustrated that they had been excluded from decision making, despite their child being accommodated under a Section 20 arrangement.

At the time of the interview, some young people had returned, of their own accord, to live with or near their adoptive family. Other young people were being parented at a distance, with parents acting as financial guarantors on flats, doing their washing/cooking or being at the end of a phone when help was needed. Only two adoptive parents were no longer in touch with their child.

Parents wanted social workers to recognise that the adoptive family was likely to offer the best hope and resource for the young person as they moved into adulthood. They wanted to be treated with kindness and compassion and for their parenting to be supported, even if that had to be at a distance. Most of the parents in this study did not have that experience.

Discussion

There is a body of evidence that demonstrates adopted children's substantial developmental recovery after maltreatment (e.g. Van IJzendoorn and Juffer, 2006; Lloyd and Barth, 2011). The focus of this study though was not on the majority of adoptions, but on when adoptive families had faced great difficulties. Forty-five parents who had experienced a disruption and forty-five parents who were currently finding parenting very challenging were interviewed in England and in Wales. We were surprised to find that APV was a factor in all but seven of the disruptions and was present in seventeen of the forty-five families where the parents were struggling. It was also surprising to find an absence of discussion about APV in the social work literature, although social workers are very familiar with the behaviours and have much to contribute to the debates. Perhaps the term ‘challenging behaviour’ has disguised the behaviours that mark out APV and a generic term has been an obstacle to developing a deeper understanding of the difficulties.

The ninety children in this study were carrying many risks to healthy development as they moved into their adoptive families such as abuse and neglect by the birth family. They were older at entry to care than most other children who go on to be adopted and a significant proportion had accumulated more risks though poor experiences whilst they were looked after. As in other studies of APV, domestic violence was prevalent in the study children's birth families. However, the presence of APV in adoptive families challenges the view that, if the child is no longer living in direct contact with domestic violence, aggressive behaviours will reduce. The different patterns of onset suggest that resolution is not simply a matter of changing the environment.

Two patterns of APV emerged: early onset (pre-pubescent) with the severity gradually escalating into adolescence and late onset (puberty onwards) with a sudden appearance of aggressive controlling behaviours with the frequency and intensity increasing rapidly. Findings from this study would suggest that aggressive controlling early behaviours are unlikely to go away and social workers need to recognise the patterns and ensure there is early intervention.

The second late onset pattern was much more difficult to predict and highlights how adoption support plans made at the time of the placement need to be flexible enough to take account of changing circumstances. Despite parents reporting early happy memories, the late onset APV families had all disrupted in comparison with 43 per cent of the early onset families. The APV late onset pattern did not follow the patterns identified in studies of the development of aggression. In those studies (e.g. Patterson, 1982), early starters were at greater risk of long-term anti-social behaviour than later starters. Late onset APV had a different trajectory of severe rapidly escalating violence inside and outside the family.

As an emerging area of study, there is much to learn about the relationship between APV and violence. Violence is commonly defined as aggression intended to cause extreme physical harm (e.g. Baron and Richardson, 1994). In this study, the intention behind violent acts did not seem to be primarily about physical harm, but as a means to control and dominate. There was also an element of young people gaining pleasure or reward from the domination. However, it was the threat of violence as well as its expression that made many parents fearful. Indeed, all the parents experiencing APV talked about ‘living in fear’ and this element should be added to definitions of APV to differentiate it from one-off acts of aggression. Much more work is needed to understand the different APV patterns, as the routes and causes may differ. This study was unable to explore whether the birth mother's substance misuse during pregnancy or genetic vulnerabilities/epigenetic changes were factors.

In the course of family assessments, we would urge all professionals working with adoptive families to ask whether the parents are experiencing CPV/APV. Parents are unlikely to volunteer the information unless asked directly. CPV/APV carries multiple stigmas—parenting a badly behaved child and experiencing violence by a child in the home. For adoptive parents, there is further shame, as they feel they have been assessed and approved as adoptive parents and entrusted with a child. They felt failures and this view was often reinforced by the response from professionals who told parents they had let the child and them down, blaming parents for the child's behaviour and using terms such as ‘failed placement’.

One of the difficulties with APV is that adults and professionals outside the family may struggle to understand how it is possible for a young person to exert such control. Children and adolescents are not usually in a powerful position. So what kind of power was being exploited by adoptive children? We would suggest that children exploited the very characteristics that are sought by agencies in their adoptive parents such as commitment and investment. In addition, adoptive parents often bring to the parent/child relationship two other characteristics. First, parents have often experienced unsuccessful fertility treatment and do not want to be thought of as ‘failing’ again. Second, parents are acutely aware of their responsibility and the trust placed in them by social workers. These latter characteristics ensure that parents often do not ask for help until the situation is desperate.

We were surprised to find allegations being used by some children to threaten their adoptive parents and exert control. A similar finding was found in a US study of APV in the general population (Eckstein, 2004). Allegations made by adopted children raise very difficult practice issues. At a time when social workers are under intense media scrutiny and accused of not listening to children, the challenge for professionals is to listen to the voice of the child and to understand the control issues that underlie APV. Evidence from this study would suggest that investigations should preferable be joint worked (children's and adoption social worker), and that parents are fully informed of the allegations and enabled to participate in the process. Social workers need to consider the impact on the whole family (including siblings) of child protection procedures.

The Home Office definition of the perpetrators of family violence has been expanded to include young people aged sixteen years or older (Home Office, 2013). The reduction in the criminal age assumes that, from sixteen years of age, young people can make choices about their actions and be prosecuted for family violence. Yet, an early history of maltreatment and trauma can place children's emotional development far behind their chronological age. One of the main differences between APV and IPV is that, in APV, the instigator usually continues to live at home and parents have to go on parenting. In this study, parents were sometimes asked by the police if they wished to prosecute following violent assaults, but none wanted to give their child a criminal record. Parents wanted help.

Timely interventions are urgently needed, as the behaviours had not stopped post disruption. Fourteen of the forty-five young people who had left home had committed serious violent and/or sexual offences including, for example, aggravated burglary, actual bodily harm and rape; thirty-five young people had had multiple foster-care/residential placements with moves occurring because of assaults on carers or staff and others were taking the behaviours into their adult relationships. The link between APV and later IPV has been noted (LaPorte ) and the pathways into adulthood for most of the young people who had left their families was concerning. Parents feared that their child's behaviour would result in a prison sentence or an early death.

There are a number of interventions for APV, but none to our knowledge has been evaluated with traumatised or maltreated children. Programmes such as Break4change (Ministry of Justice, 2009), e-imagination and restorative justice (Sherman and Strang, 2007; Fonagy, 2012), Non Violent Resistance (Omer, 2004; Jakob, 2010: Newman ), Step-up (Anderson and Routt, 2004) and Who's in Charge (O'Connor, 2007) are used. We need to test out these approaches with looked after and adopted children, as CPV/APV is not just an issue for some adoptive families, but is also responsible for placement instability in foster and residential care.

It has long been recognised that the impact of maltreatment and trauma does not disappear simply because a child is adopted. However, that recognition has not been matched with appropriate, evidence-based support services for adoptive families. It is time for that situation to change.