The committee's discussion

Links to evidence sources are given in square brackets. See evidence reviews for details.

Background

This guideline uses the NSPCC definition of harmful sexual behaviour: 'One or more children engaging in sexual discussions or acts that are inappropriate for their age or stage of development. These can range from using sexually explicit words and phrases to full penetrative sex with other children or adults.' (Harmful sexual behaviour NSPCC).

Using the term 'harmful sexual behaviour' avoids labelling young children as sexual offenders, but it does not reflect the diversity of children and young people who engage in sexualised behaviours.

It is critical to differentiate between sexually abusive behaviour and behaviours that are detrimental to the child or young person's development. Sexually abusive behaviours are coercive and involve forcing others to comply with an action. This can include oral, anal and vaginal penetration. Whereas behaviours that affect individual development could include, for example, compulsive masturbation or addiction to online pornography. Such behaviours may result in stigmatisation and victimisation, as well as making others feel uncomfortable.

There has been significant debate about how to describe children and young people displaying sexualised behaviour without labelling them as sex offenders. Difficulties in defining such behaviour are compounded by a general lack of knowledge of childhood sexuality and what constitutes normal sexual development.

The focus of this guideline is on children and young people who are sole perpetrators of harmful sexual activities, directed either towards themselves or other individuals. It does not focus on activities such as child sexual exploitation and peer-on-peer or gang-related sexual violence.

The Department for Education's Working together to safeguard children definition of sexual abuse (which covers physical, emotional and sexual abuse, and neglect) acknowledges that children can commit acts of sexual abuse. But it does not acknowledge that children and young people's sexual behaviour is on a continuum. The NSPCC's definition of harmful sexual behaviour acknowledges that children and young people's sexual behaviour is on a continuum. It defines harmful sexual behaviour as acts that are inappropriate for age or stage of development and are important factors when assessing and intervening with this group.

Children and young people's sexual behaviour problems are diverse. Various terms have been used to refer to children and young people who engage in developmentally unexpected sexual behaviours. These include: abuse-reactive, sexually reactive, sexually aggressive, sexualised children, children who molest, sexually abusive children and young sexual offenders.

The committee agreed with evidence from expert paper 1, and members' own experience, that many children and young people's display of harmful sexual behaviour will naturally come to an end as they mature. But members also agreed that concerns about a child or young person's sexualised behaviour should always be followed up and assessed.

Little is known about the number of prepubescent children who continue with harmful sexual activities as they get older. What is known is that only a small number go on to commit more serious sexual offences. In addition, the committee recognised the need to distinguish between chronological and developmental age when deciding how to assess and then intervene with this group.

The committee also agreed that it is important to distinguish between prepubescent children and adolescents for assessment purposes. Some behaviours may be considered normal in prepubescent children, for example, but would be of concern if they continue into adolescence. Likewise, behaviours that would be considered normal in adolescents may be regarded as highly unusual in prepubescent children and therefore merit a need for referral.

In addition, the committee recognised the need to distinguish between these groups in the context of the criminal justice system in England, where 10 is the age of criminal responsibility. The committee also noted that the Children Act 1989 defines a child as someone who has not reached 18 and that young people enter the adult criminal justice system at age 18.

The committee noted that young people with ongoing or long-term health or social care needs may need to move into adult services at age 18 or 25, if they have special educational needs or disabilities. It agreed that these transitions need to be managed as part of the harmful sexual behaviour assessment and intervention process.

The committee discussed the fact that pornography is now more readily available thanks to new technology, but no evidence was found in the systematic search of the literature. Expert testimony confirmed a significant gap in the literature on how pornography influences sexual behaviour [EP11]. So the committee agreed to make a research recommendation on this.

See also gaps in the evidence numbers 10–12.

Section 1.1 Multi-agency approach

Multi-agency, multidisciplinary team

The discussion below explains how we made recommendations 1.1.1–1.1.3.

Current practice

The committee acknowledged a number of statutory arrangements are relevant for children and young people displaying harmful sexual behaviour. This includes:

  • Section 10 of The Children Act 2004. This makes local authorities responsible for promoting interagency cooperation to improve the welfare of children in need.

  • Ofsted's Early help: whose responsibility? and the Department for Education's Working together to safeguard children give local safeguarding children boards statutory responsibilities for children in need of protection and those with highly complex needs. Safeguarding boards are also responsible for developing thresholds for action.

The committee acknowledged that the involvement of health agencies will vary, because support is provided in a multi-agency context.

The committee discussed the role of children and adolescent mental health services (CAMHS) in this area. Members noted that, in their experience, referral thresholds to CAMHS vary and that not all CAMHS services would take referrals for children and young people displaying harmful sexual behaviour.

The committee also noted that there is no coherent national commissioning framework on harmful sexual behaviour for people living in secure accommodation and few secure children's homes or young offender institutions offer these services.

The committee recognised these recommendations were not based on evidence of effectiveness or cost effectiveness. Rather, they were framed by the Department for Education and Ofsted guidance on interagency working and were generated by consensus from their own experience and from 2 expert testimonies [EP9 and EP10].

Evidence for effectiveness

From the committee's experience, interagency assessments would be a suitable route for developing multi-agency approaches. These could use, for example, the Common Assessment Framework (see 'Working together to safeguard children'), the local children safeguarding board and the Department for Communities and Local Government's Troubled Families Programme.

There was no evidence of effectiveness on the multi-agency approach but the committee agreed with the conclusions from 8 qualitative studies (3 high quality, 3 moderate quality and 2 low quality). These stated that early assessment should be 'joined up' with any subsequent therapeutic interventions, to ensure continuity between assessment and intervention [ES2.11, ES2.12].

Based on their own experiences, members also agreed that a lack of effective interagency working is often a key factor in serious case reviews (see New learning from serious case reviews: a 2-year report for 2009–2011 Department for Education).

Evidence for cost effectiveness

The committee agreed that failures in interagency working could have a serious 'knock on' effect and lead to a limited response to referrals. This may also have implications for cost effectiveness: failure to provide services may leave the behaviour unchecked and eventual entry into the criminal justice system. The latter would be expensive and is also likely to result in harm to others.

Additional factors taken into account

The committee reflected on expert testimony that described regional arrangements for harmful sexual behaviour services as well as agreements across agencies [EP10 and EP13].

Practitioners working with children and young people with complex needs, such as autism spectrum condition or conduct disorder, need access to specialists working in harmful sexual behaviour services. It would not be possible to agree interagency care plans without a multidisciplinary team of professionals who meet regularly to agree and evaluate outcomes.

The committee agreed that the practice of closing a social work file on a child or young person once they have been referred to harmful sexual behaviour services is not helpful. Members described how, in some instances, referrals to harmful sexual behaviour services were declined because specialist services do not have the resources to deal with the child or young person's additional needs. They agreed there was a need for greater recognition that a range of agencies is usually needed to provide support.

Trade-off between benefits and harms

The committee agreed that if nobody took lead responsibility for assessment and referral this could have a serious impact on delivery of the care plan. The 'knock-on' effect could be further harm to the child or young person, or risk of harm to others.

Resource impact and implementation issues

Many local safeguarding children boards and child protection committees across the UK now include harmful sexual behaviours in their interagency procedures and policy documents. Many also offer short courses on young sexual abusers as part of their interagency training programmes.

The committee agreed that failure to provide local expertise using a coordinated, multi-agency approach will mean children and young people displaying more serious behaviour are likely to receive ad hoc assessments and interventions. Sometimes this might result in out-of-area placements that could prove expensive. It could also lead to delays in providing a more effective intervention in the child or young person's locality and at an earlier stage.

Multi-agency, multidisciplinary working

The discussion below explains how we made recommendations 1.1.4–1.1.8.

Current practice

Committee members discussed their experience of working across agencies and disciplines within the harmful sexual behaviour service sector. Their experiences were generally positive, but they knew that there was room for improvement in many locations.

Evidence for effectiveness and cost effectiveness

There was no empirical evidence of effectiveness or cost effectiveness.

The committee agreed that children and young people who display harmful sexual behaviour are likely to have complex needs that can only be met by numerous health and social care agencies working together. Members agreed that a well-established multi-agency response is likely to lead to multiple cost savings to society. This includes:

  • savings from otherwise lost educational and employment opportunities during the child or young person's life

  • savings from preventing harm to others.

The committee agreed that local safeguarding children boards should implement NICE recommendations in this area and should identify lead agencies to commission and develop harmful sexual behaviour services.

The committee noted that various services are needed to respond to this group of children and young people, and that a statutory response is not always necessary. It agreed that the safeguarding needs of children and young people referred for an early help assessment should be a priority, with the results used to determine whether a statutory or criminal justice response is needed. (For victims of sexual abuse see Reporting and acting on child abuse and neglect, Home Office and Department for Education.) It recommended that a variety of referral and care pathways should be in place.

The committee heard testimony on the NSPCC harmful sexual behaviour framework [EP12]. It aims to improve local interagency working and coordination while recognising the resource constraints that local areas face. Members agreed that local authorities should offer a range of care pathways based on the 5 core domains of the framework to create a consistent approach across services.

Currently there is no national strategy or overarching service delivery framework for harmful sexual behaviour. The NSPCC framework was developed by a wide range of partners and is being tested in a number of local authority areas.

The committee agreed that, in line with 'Working together to safeguard children', local safeguarding children boards should set thresholds for when to refer a child or young person for early help assessment or specialist harmful sexual behaviour services.

The committee saw local safeguarding children boards as a potential vehicle for ensuring a coordinated approach to meeting the needs of this group. This would have the benefit of making use of existing services to improve multiagency working.

The committee recommended that local safeguarding children boards should ensure local policies and procedures are in place to train staff in children's social services to deal with concerns about a child or young person's sexualised behaviour. Concerns may be raised by professionals working in universal services or members of the public.

Local safeguarding children boards should recommend resources for professionals working in universal services to consult when there are concerns over a child or young person's sexualised behaviour.

Resource impact and implementation issues

The committee agreed that these recommendations are informed by government guidance and form the basis of current health and social care practice. It also agreed that a brief consultation (or triage meeting) over the phone or face-to-face with a concerned parent or teacher and a specialist practitioner could sometimes be enough.

Information sharing

The discussion below explains how we made recommendations 1.1.9–1.1.12.

Current practice

The committee agreed that information sharing between agencies remains a contentious issue. Members also agreed that fears about sharing information should not stand in the way of the need to promote the welfare and protect the safety of children. (See Information sharing: advice for practitioners providing safeguarding services to children, young people, parents and carers.)

Evidence for effectiveness

The committee recognised that the recommendations were not based on evidence of effectiveness but reflected the committee's experience of serious case reviews (see New learning from serious case reviews: a 2-year report for 2009–2011 Department for Education). The issues are also highlighted in the Department for Education's report 'Working together to safeguard children.' Seven qualitative studies reported there is reluctance among practitioners to share assessments (2 high quality, 4 moderate, 1 low) [ES2.13].

The committee also discussed the need for clear, effective protocols that are regularly evaluated. In particular, these protocols should highlight the adverse consequences of not sharing information – a point repeatedly made in serious case reviews, including cases of violent or sexual assault and rape.

Evidence for cost effectiveness

There was no cost effectiveness evidence for this set of recommendations.

Additional factors taken into account

Using information-sharing protocols is regarded as good practice, as outlined in 'Working together to safeguard children'.

Resource impact and implementation issues

The committee considered that the resource impact of setting up and agreeing information sharing protocols would be negligible, particularly compared with:

  • the negative consequences associated with poor information sharing and interagency working identified in case reviews

  • the costs and quality-adjusted life year (QALY) losses attributed to sexual offences and rape in the economic modelling report.

In addition, information sharing using a locally agreed approach does not have a significant resource impact because it uses established pathways and protocols, as set out in 'Working together to safeguard children'.

Section 1.2 Named safeguarding leads in universal services

The discussion below explains how we made recommendations 1.2.1–1.2.4.

Current practice

Any professional working in universal services who is concerned about the sexualised behaviour of a child or young person has a responsibility to notify their organisation's named safeguarding professional. Named safeguarding leads should discuss their concerns with children's social services so a decision can be made about whether referral for an early help assessment is needed (see Early help: whose responsibility? Ofsted).

The committee noted, however, that members were aware of instances where such behaviour had been ignored and children had been harmed.

Evidence of effectiveness

No evidence of effectiveness was identified for this section. The recommendations are linked to the Department for Education's guidance 'Working together to safeguard children'. This requires professionals in universal services and those providing services to adults working with children to identify emerging problems and share this with other professionals involved in early identification and assessment of harmful sexual behaviour. It requires them to discuss concerns and referrals through agreed routes so appropriate referrals are made, to reduce risk of stigmatisation.

The committee agreed that professionals working in universal services should use locally agreed resources to identify if a child or young person's behaviour is a cause for concern [ES2.1].

Section 1.3 Early help assessment

The discussion below explains how we made recommendations 1.3.1–1.3.10.

See also gaps in the evidence number 2.

Current practice

The principles of early help assessment are discussed in the Department for Education's guidance 'Working together to safeguard children' (see current practice in the 'Multi-agency, multidisciplinary team' section of the committee discussion). This type of assessment is key to identifying and addressing needs not being met by universal services. For example, it can identify the need for:

  • targeted services

  • specialist harmful sexual behaviour assessment

  • referral to the child protection team

  • referral to the criminal justice system.

Other helpful tools include: The use of whole family assessment to identify families with multiple problems (Department for Education) and AssetPlus, previously known as the Youth Offending Asset assessment (Youth Justice Board for England and Wales).

In their experience, members said that early help professionals and professionals not trained to work with people displaying harmful sexual behaviour need more information and resources to identify such behaviour. That is because referrals to children's services are based on a generic practice model and not one that specialises in harmful sexual behaviour.

Evidence for effectiveness

The committee agreed that an early help assessment using, for example, the Common Assessment Framework, can help identify what additional help the child or young person and their family need, apart from that provided by universal health and social care services. The early help assessment provides a model for gathering and recording information about the child or young person's strengths and needs, based on discussions with the family. It can also form the basis for their care plan.

The committee agreed that the early help assessment should be done by a lead professional who supports the child and family, acts as an advocate on their behalf, and coordinates the delivery of services. The decision on who should lead should be made on a case-by-case basis, with input from the child or young person and their family. It could be a GP, family support worker, school nurse, social worker or health visitor.

The committee agreed that it should be made clear which agencies are responsible for children and young people who do not fall under the remit of child protection, or who do not need harmful sexual behaviour services but do have unmet needs.

The committee agreed that a locally agreed identification tool should be used as part of the early help assessment for a child or young person displaying sexualised behaviour that is a cause for concern [ES2.1, ES2.13].

The committee based these recommendations on government guidance on early help (see 'Working together to safeguard children'). It also included expert testimony that recognised the need for early help to prevent escalation [EP12 and EP13].

There is no evidence on how effective tools are at identifying harmful sexual behaviour, but the committee agreed that a tool should be used as part of the early help assessment. This will help to identify harmful sexual behaviour and improve decision making. The aim would be to screen for age and developmentally inappropriate sexualised behaviour and language to decide whether to refer on to specialist harmful sexual behaviour services.

Tools commonly referred to in the literature, or used in practice, include: the Brook Traffic Light Tool and Hackett's continuum model of children and young people's sexual behaviours, patterns and cycles [ES2.1].

Evidence for cost effectiveness

There was no cost effectiveness evidence for this set of recommendations. The committee did not make it a priority for modelling. It considered the cost impact to be negligible because it is widely considered to be good working practice (see resource impact section below).

Additional factors taken into account

The committee considered that such assessments could prevent the escalation of sexualised behaviour. This, in turn, could prevent the need for specialist harmful sexual behaviour services, a statutory assessment under The Children Act 1989 or involvement of the criminal justice system.

The committee agreed that interagency assessments should use a joined-up process that focuses on outcomes.

Assessments should acknowledge chronological age and developmental status, and what constitutes healthy sexual behaviour among children and young people. This is particularly true when discussing children and young people with a neurodevelopmental disorder such as autism or a learning disability.

The committee also agreed that often, as children and young people mature, they stop displaying harmful sexual behaviour. However, their life chances may still be impaired (because of their previous behaviour or related factors). So they may need ongoing assessment and input, particularly in relation to educational and employment opportunities [EP1].

The committee acknowledged that those working with children and young people displaying inappropriate sexualised behaviour need a greater understanding of the benefits of an early help assessment. However, members also wanted to ensure that unnecessary assessments and referrals do not occur that may be stigmatising, so they included caution around this area in the recommendations.

Sexual behaviours exist on a continuum that ranges from normal and developmentally appropriate to highly abnormal and violent [EP1]. There is little evidence on interventions that address behaviours that fall short of thresholds needing a response from specialist harmful sexual behaviour services or the criminal justice system.

Locating sexual behaviour on a continuum that is related to development age is an important part of the assessment process, and can help practitioners and families make distinctions between different sexual behaviours.

The committee acknowledged that using assessment tools with prepubescent age groups that were designed for older groups could be harmful.

The committee agreed with expert paper 7. This recommends that when assessing children and young people displaying sexualised behaviour, practitioners should distinguish between children and young people in general and those with special educational needs and learning disabilities, or with autism. It noted that although the latter form a significant minority, there is a lack of tools for assessing their behaviour.

The committee agreed that using a tool to identify the sexual behaviour of children and young people would help make practitioners aware that this behaviour exists on a continuum. It would also help determine whether a referral to specialist services is necessary. Members recognised that if the tool is not designed for the subgroup being assessed, the results may not be accurate.

Resource impact and implementation issues

Use of early help assessment is regarded as good working practice. It provides a shared assessment and planning framework for all children's services in England. Because of this, the committee does not consider these recommendations will have any additional resource impact, except in areas that are not following good practice.

The committee believed that early identification and intervention would be cost effective by preventing escalation of the behaviour and avoiding involvement of the criminal justice system.

Section 1.4 Risk assessment for children and young people referred to harmful sexual behaviour services

The discussion below explains how we made recommendations 1.4.1–1.4.7.

See also gaps in the evidence number 2.

Current practice

Risk assessment tools are used to assess specific risks and needs arising from a child or young person's harmful sexual behaviour. In the UK, different models are used depending on whether they come into contact with child welfare, mental health or the criminal justice system.

When children and young people are charged by the Crown Prosecution Service for harmful sexual behaviour, the offence cannot be discussed with them while the case continues. But an assessment can still take place.

The committee discussed the AIM assessment model, which was originally developed for practitioners in the criminal justice system, but could be used in the community [EP8]. The committee was told that current practice is dominated by AIM2 designed primarily for boys and young men aged 12 to 18. The AIM model considers the level of management and supervision needed, together with the person's development and intervention needs. Members were also told that although the AIM assessment model has led to a more standardised approach, it is unclear how it might be applied outside the criminal justice system.

Members noted that the AIM model for under‑12s is used to assess:

  • children under 12

  • children between 10 and 12 whose harmful sexual behaviour needs a criminal justice response (10 being the age of criminal responsibility in England).

Members noted that AIM2 is used mainly to assess males aged 12 to 18 and:

The committee also discussed whether AIM2 could be used with young females and young people with learning disabilities. Members agreed this should be on a more limited basis. The developers of AIM2 do not recommend using the 'level of supervision' scale for young females, as it is likely to misrepresent the level of risk. A degree of caution is also advised when using AIM2 to predict sexual reoffending in young people with learning disabilities. At this stage the committee agreed the evidence available was too limited to make recommendations in these areas. It noted that further research is needed on assessing risk in all children and young people (see research recommendation 5).

Evidence for effectiveness

The committee considered the evidence of effectiveness for various risk assessment tools in terms of predicting sexual and non-sexual re-offending. It noted that 10 of the 11 quantitative studies in the evidence review were based on adolescent boys with a mean age of 15 who had been convicted of sexual offences.

Only 1 study included girls and a younger age group (boys 12.3 years and girls 11.9 years) who had recently begun to display harmful sexual behaviour [ES2.3]. All the studies were from North America, which may limit their applicability in the UK. They were all at risk of bias from the methods used.

The committee considered:

  • J‑SOAP‑II (5 low to moderate quality studies on future sexual re-offending). The evidence was inconsistent: 3 predicted sexual re-offending, 2 did not [ES2.3].

  • ERASOR (4 moderate to high quality studies). Three predicted sexual re‑offending, 1 did not [ES2.5].

  • Adapted AIM and AIM2 (2 moderate quality studies on future sexual re-offending). Both studies predicted that adolescents with and without learning disabilities who were previously known to sexually offend would reoffend [ES2.2].

  • J‑SORRAT‑II (2 studies, 1 low and 1 moderate quality). One study found it was able to predict future sexual re-offending among adolescent males convicted of a sexual offence, the other found no effect [ES2.8].

The committee noted that although the evidence is contradictory for J‑SOAP‑II and ERASOR [ES2.3, ES2.4, ES2.5, ES2.6, ES2.7], the tools look promising for assessing young people's risk of sexual and non-sexual reoffending.

The committee agreed that although the evidence from adapted AIM and AIM2 studies is limited [ES2.2, EP8], they are promising tools and are relevant because they were developed in the UK. Only limited attempts have been made to test their predictive validity [ES2.2].

AIM2 is also now being used (with caution) for girls and for those with learning disabilities. So the committee recommended further research on AIM2.

So the committee was unable to make a strong recommendation for the use of AIM for under‑12s or AIM2.

The committee agreed that the J‑SORRATT‑II was still undergoing research and was not used outside North America, so it could not currently recommend its use as a risk assessment tool [ES2.8].

There was no evidence that tools focusing on strengths (BERS-2) enhance the accuracy of ERASOR to predict sexual re-offending among adolescent young men who have committed a sexual offence [ES2.7].

The committee also considered evidence from 2 quantitative studies (moderate quality) on the SAVRY and YLS/CMI tools. It noted that SAVRY was unable to predict sexual or non-sexual reoffending for adolescent males convicted of a sexual offence [ES2.9]. The YLS/CMI tool did not predict sexual reoffending but did predict non-sexual violence, and any potential for non-sexual re-offending [ES2.10].

The committee considered 11 qualitative studies: 3 papers were rated high, 6 moderate and 2 low quality. Two moderate quality studies stated that AIM2 offered a more standardised approach to assessment, and encouraged better cooperation between young offender teams and social care departments in the UK. But practitioners reported frustration because they were not properly trained to use it. In addition, there was some confusion about its purpose and how the findings might be applied in practice [ES2.14].

The committee discussed the literature on risk assessment tools to predict future sexual violence. Members agreed it is limited because the number of re-offenders recruited into research studies is too small for the research designs needed to validate tools. This means that most UK agencies are using largely under-tested models to underpin their assessments of risk and need. So further research is urgently needed (see research recommendation 5).

Evidence for cost effectiveness

There was no cost effectiveness evidence for this set of recommendations. The committee agreed that a good initial assessment is vital when making decisions about therapeutic interventions, treatment placements and care plans.

Additional factors taken into account

There are no fully validated models or frameworks to suggest what core elements should be included in risk assessment tools.

The quantitative evidence on sexual abuse was largely drawn from North America. It reported on small clinical populations of relatively high-risk young people referred for specialist treatment. The assessment models used were adapted from models used for male adults convicted of a sexual offence.

Assessing the risk of sexual re-offending among young people is particularly challenging because of the enormous changes they undergo at this age. The committee also noted a key finding from research that indicates that many young people who engage in offending behaviours stop them as they mature[7].

The committee noted 2 specific risk trajectories evident in samples of young sexual abusers: general antisocial behaviours and harmful sexual behaviour. Most young people charged with sexual offences do not re-offend sexually, although the rate of non-sexual re-offending is substantially higher than average.

The committee agreed that risk assessment tools should consider a range of key elements, including the factors that led to the behaviour. The tools should also address the need for ongoing support and re-assessment [ES1.6, ES1.28].

The committee discussed risk assessment tools for different subgroups and acknowledged the lack of tools and models for different population groups.

The committee agreed that risk assessment tools and models designed for adolescents convicted of a sexual offence should not be used with prepubescent children displaying harmful sexual behaviour.

There are few empirical studies of assessment tools and interventions directed at the small proportion of girls and young women who sexually abuse others [EP4]. Research has indicated that females convicted of a sexual offence differ from males in various ways. For example, harmful sexual behaviour in girls is more likely to be motivated by aggression against them. The committee acknowledged the valuable work being done in this area by Barnardo's Cymru Taith project.

Evidence paper 4 discussed how boys and girls with harmful sexual behaviour are treated differently. For example, boys are more likely to be removed from mainstream school.

Trade-off between benefits and harms

Benefits include the adoption of a consistent approach to assessment. In addition, using locally agreed tools allows practitioners from different agencies and professional backgrounds to share information.

However, potential harm could come from the fact that the assessment of the level of risk is not accurate. On the one hand, this could lead to an over-punitive or over-restrictive approach. On the other, it could sometimes mean the child or young person doesn't get the support they need to prevent further harmful sexual behaviour, so exposing them to risk – to themselves and others.

Resource impact and implementation issues

The committee noted that AIM for under‑12s and AIM2 were developed for the UK, but have to be paid for and involve specialist training. In comparison, J‑SOAP‑II, and ERASOR are free and specialist training is not needed, so they would have less impact on training needs and resources. But their applicability in England is unknown.

Overall, the committee could not recommend 1 tool over another and noted that most effectiveness evidence came from North America. Internationally, the 2 tools with the highest degree of empirical support are ERASOR and J‑SOAP‑II, although the evidence for predicting sexual re-offending is not consistent across studies. Further studies are needed on larger samples. Also, studies are needed to compare the use of different models with the same samples.

In the absence of more consistent evidence, the committee agreed that it might be best if practitioners use AIM2, ERASOR, or J‑SOAP‑II for assessing risk. If time allows, the committee recognised that there may be benefits if the ERASOR and J‑SOAP‑II are used together to compare the use of these two models over time. In each case, the developers also recommend that practitioners use their own clinical judgement.

Only the most promising tools were included in the recommendations as examples of what was available. But the committee agreed that, given the uncertainty in the evidence base more research is needed (see research recommendation 5).

The committee also agreed by consensus that in their expert opinion children's social services and NHS England are best placed to identify who should undertake a risk assessment and these names were added to recommendation 1.4.1.

Section 1.5 Engaging with families and carers before an intervention begins

The discussion below explains how we made recommendations 1.5.1–1.5.4.

Current practice

Not all practitioners meet with families and carers before an intervention begins.

Evidence for effectiveness

There was no quantitative evidence of effectiveness on the role of practitioners. Six qualitative studies and 2 expert testimonies identified key features and approaches that practitioners could use to reduce barriers to services and improve communications between the practitioner and children, young people, parents and carers [EP9 and EP10].

The committee agreed that before an intervention begins, practitioners must consider whether the child or young person has been abused within the family or the victim is another family member. (See also NICE's guidelines on when to suspect child maltreatment and child abuse and neglect.)

The committee agreed with 2 high quality qualitative studies that offering families and carers the opportunity to meet the programme practitioner before an intervention starts may help to overcome any fears about getting involved in and continuing with the programme [ES1.23, ES1.24].

The committee agreed with 2 qualitative studies (moderate to high quality) that family and carer participation and support is crucial to getting young people involved with interventions. It also helps reinforce intervention messages in the home [ES1.20].

The committee agreed with 2 qualitative studies of moderate to high quality on the need for practitioners to accommodate a child or young person's changing needs and offer a flexible service to accommodate their social activities to maintain their interest [ES1.16, ES1.22].

The committee agreed with 3 high quality qualitative studies and expert papers 9 and 10 that the therapist's relationship with the child or young person is vital if an intervention is to be effective [ES1.21]. Members also agreed that, in their experience, interventions were only as good as the person providing them.

Evidence for cost effectiveness

There was no cost effectiveness evidence for this set of recommendations, but the committee agreed that encouraging practitioners to meet beforehand is likely to improve the outcome of the intervention. So they are likely to be cost effective and potentially cost saving.

Trade-off between benefits and harms

The committee agreed that the main benefit would be greater involvement with the intervention and improved outcomes for the child or young person, their family and carers.

Resource impact and implementation issues

The committee agreed that this recommendation would have a resource impact, particularly in terms of arranging meetings that are not part of the therapeutic intervention. But members agreed that increasing attendance and improving the relationship between the child or young person and the practitioner could lead to better outcomes and offset any resource implications.

Section 1.6 Developing and managing a care plan for children and young people displaying harmful behaviour

The discussion below explains how we made recommendations 1.6.1–1.6.5.

Current practice

Practice may vary. But good practice involves using a care plan based on the results of the assessment of the child or young person's risks and needs.

Evidence of effectiveness

Based on members' own experience, the committee agreed that care planning should be based on the results of the needs and risk assessment and should include the use of recognised resources.

Resource impact and implementation issues

The committee believed that these recommendations would not result in increased costs but would probably improve outcomes [ES1.15].

Section 1.7 Developing interventions for children and young people displaying harmful sexual behaviour

The discussion below explains how we made recommendations 1.7.1–1.7.15.

See also gaps in the evidence (number 1).

Current practice

Current practice is often based on cognitive behavioural therapy models used to treat adult men who have sexually offended. Developed originally in the US, these models came to prominence in the UK probation and prison services from the late 1980s.

Evidence for effectiveness

The evidence of effectiveness was from North America and may be only partially applicable to a UK population. The interventions reviewed mainly focused on those convicted of a sexual offence in treatment settings and will have limited applicability to children and young people outside the criminal justice system.

Many types of intervention are used to help children and young people displaying harmful sexual behaviour but not all of them have been evaluated.

The committee considered evidence statements covering 13 quantitative studies (4 randomised controlled trials, 3 controlled studies and 6 before-and-after studies). It noted that although the studies were grouped for analysis according to type of intervention, many included elements drawn from a range of approaches. This included cognitive behavioural therapy (CBT) and multisystemic therapy. It also considered qualitative evidence from 26 studies (11 low, 9 moderate, 6 high quality studies).

Of the 13 quantitative studies, 9 (2 randomised controlled trials, 1 controlled study and 6 before-and-after studies) of variable quality reported on the effectiveness of CBT-based approaches. These comprise a range of components delivered to both individuals and groups and focus on the sexually abusive behaviour.

Four papers reported on 3 studies of multisystemic therapy. Two randomised controlled trials and 1 controlled study (ranging from low to moderate quality) reported that multisystemic therapy significantly reduced the risk of adolescent sexual re-offending compared with CBT or usual care.

One controlled study (moderate quality) using adventure-based therapy (Legacy) for adolescent boys convicted of a sexual offence, reported no difference between the intervention and control group for re-offending rates for violent sexual offences.

The committee considered the evidence on CBT interventions from 4 low to moderate quality quantitative studies. These were abuse-focused and targeted the sexual behaviour of juveniles convicted of sexual offences using 1 or several components of CBT. This included:

  • satiation therapy, a method for reducing deviant sexual arousal

  • verbal satiation – repeatedly talking about deviant sexual fantasies to reduce sexual arousal from such fantasies

  • vicarious sensitisation, a form of conditioning used to treat teenage boys who have displayed harmful sexual behaviour towards younger children

  • cognitive restructuring therapy to help people to think differently about a situation, event, thought, or belief.

The committee noted the positive direction of all 4 studies but agreed that, on balance and from members' expert opinion and experience, it could not recommend these types of interventions [ER1, ES1.1, ES1.2, ES1.3, ES1.4, ES1.5].

The committee agreed with evidence from 3 low to moderate quality qualitative studies that stigma and ostracism may arise if a child or young person is labelled as a sex offender. It was keen to highlight that children and young people with harmful sexual behaviour are not 'mini adult sex offenders' and that offering interventions that are abuse-focused is potentially stigmatising [ES1.25].

The committee considered a study of moderate quality that compared CBT with dynamic play therapy with boys (61%) and girls (39%) aged 5 to 12. This targeted a range of harmful behaviours and included families and carers. It reported no significant difference between the 2 approaches. Both improved the children's ability to socialise while reducing their behavioural, affective and sexual behaviour problems [ES1.7].

The committee agreed the positive outcomes were likely to have resulted from the types of components that were included in each approach. This included: behaviour modification and psychoeducational principles in the CBT group; and client-centred and psychodynamic play therapy principles in the play therapy group [ES1.7].

The committee also considered evidence on 2 CBT programmes for young people displaying a range of harmful behaviours and personality disorders: SAFE‑T (Sexual Abuse, Family Education and Treatment Programme) and Thought Change System. Both interventions included family members and carers. (The evidence comprised 2 low to moderate quality quantitative studies.)

Both reported a decrease in harmful behaviours, with the SAFE‑T programme reporting a 72% reduction in re-offending rates for sexual assault [ES1.6 and ES1.8].

The committee considered the evidence of effectiveness for multisystemic therapy compared with CBT-based usual care for adolescents convicted of sexual offences.

Two moderate quality quantitative studies found that significantly fewer people from the multisystemic therapy group had been re-arrested for sexual offences at follow-up than from the comparison group [ES2.9].

One moderate quality quantitative study found that multisystemic therapy for adolescents charged with sexual offences led to a reduction in deviant sexual interests when compared with CBT-based usual care [ES1.10]. Two moderate quality quantitative studies reported improvements in problem sexual behaviour, psychiatric symptoms, antisocial behaviour, family and peer relations and school performance among adolescents charged with sexual offences, compared with CBT-based usual care [ES1.11, ES1.12, ES1.13].

Multisystemic therapy focuses on the family, which means its use will be limited because a significant number of children and young people who display harmful sexual behaviour are in out-of-home placements. Its main goal is to reduce the risk of re-offending by enhancing family and peer relationships. A big benefit is that carers become better at identifying friends who were having a negative influence on their adolescents and advising their children to stop associating with them.

Research suggests, however, that multisystemic therapy may not be as effective with all subgroups of young people who display harmful sexual behaviour. For example, there is a strong link between antisocial peer groups and young people whose harmful sexual behaviour is often directed towards peers and accompanied by other non-sexual criminality. This group is different from those whose harmful sexual behaviour targets younger, prepubescent children. The latter are less likely to have a social life or strong peer friendship groups.

The committee heard expert testimony on the ongoing trial of Multi-systemic therapy – problematic sexual behaviour in the UK and agreed that this may, in future, offer more conclusive results [EP14]. It noted that previous evaluations of the programme in the US were positive, and had been carried out by its designers.

The committee noted the results from 1 moderate quality study that evaluated an adventure-based programme (Legacy). This reported no difference for re-arrest rates for violent sex offences between groups but appears to be beneficial in reducing future risks of non-sexual reoffending [ES1.14].

Drawing on expert papers 2 and 5, members agreed that the duration and intensity of interventions should be adapted for those with learning disabilities. (For example, by having more frequent, shorter sessions, or longer sessions as necessary, or fewer participants in group sessions.)

The committee agreed with the evidence from 1 moderate quantitative study and 3 moderate to high quality qualitative studies that understanding the factors that lead to harmful sexual behaviour is an important part of relapse prevention and making future plans [ES1.6, ES1.17].

The committee agreed that victim empathy is a contested component of harmful sexual behaviour interventions [ES1.18].

The committee also agreed with the results from 6 qualitative studies (3 low, 1 moderate, 2 high) that interventions involving children and young people in supervised social activities helps promote self-esteem and socially appropriate behaviour [ES1.15].

The committee noted the evidence from 1 moderate quantitative study and 2 low to moderate qualitative studies highlighting the concerns of families and young people about not getting support to maintain their progress. The committee agreed this was an important component of services [ES1.6, ES1.28].

The committee agreed with 14 qualitative studies (11 low, 1 moderate, 2 high quality) that reported that communication skills, social skills training and anger management or 'emotional regulation' are important components of any intervention [ES1.19, ES1.27].

The committee noted the results from 3 qualitative studies (1 high, 1 moderate, 1 low) that showed that group interventions (for both the child and young person and their family and carer) can reduce their sense of isolation and provide valuable support. But members also noted that it may be problematic for those who find it difficult to talk in front of others [ES1.26]. In addition, they noted the difficulties involved in treating perpetrators of harmful sexual behaviour alongside their victims, as highlighted in 4 qualitative studies (2 moderate, 2 low) [ES1.25].

Evidence for cost effectiveness

The committee made the recommendations on cognitive behavioural therapy, multisystemic therapy and play therapy a priority for economic modelling. The model results showed a cost per QALY of under £20,000, but the committee questioned these estimates and thought that not all these therapies would in fact be cost effective. This is particularly true for children and young people who did not need a custodial sentence. That is because the studies that underpinned the modelling were from North America, where comparators are different.

Given that the multisystemic therapy trial in the UK has yet to report, the committee suggested that it would be prudent to continue with current approaches – but make them work better by following the recommendations outlined in this guideline. Getting better results at the same cost would automatically be cost effective.

If the current trial shows that more expensive methods are more effective than current methods, the approach advocated here (continuing with current practice) could be revised.

Additional factors taken into account

The sexual behaviour of children and young people exists on a continuum that ranges from normal and developmentally appropriate to highly abnormal and violent. Various approaches are needed to address these different behaviours. But there is little evidence on interventions that address behaviours that fall short of thresholds needing a response from the criminal justice system.

The qualitative studies identified programmes offering relapse prevention, anger management (emotional restraint), victim empathy, communication and social skills training. They also documented the emergence of family-level interventions and the role of the therapist as important components. The committee noted that mode of delivery (such as face-to-face or in groups) should be based on the needs and circumstances of the child or young person, as highlighted in the assessment and using clinical judgement.

The qualitative studies also highlighted the components of an intervention that participants, their families and professionals feel have value. But it is not clear which components are most effective for different groups.

The committee agreed that having to choose between cognitive behavioural therapy and multisystemic therapy is not realistic and that there are advantages to both (this includes the fact that effectiveness can depend on the family circumstances).

In the absence of clear evidence, the committee recommended continued use of these therapies. The evidence reported was based on small numbers of participants (and conducted by developers of the intervention in the case of multisystemic therapy). Members agreed that more research, with a low risk of bias and relevant to UK practice, is needed (see research recommendation 3).

Members acknowledged that multisystemic therapy is a more complex approach that needs the child or young person to be living in the family home or in a stable foster family situation (for at least 18 months in the case of multisystemic therapy Problematic Sexual Behaviour). As a significant proportion of children and young people displaying harmful sexual behaviour may not be living in a family situation this approach may not work. In that respect, cognitive behavioural therapy might be a more pragmatic solution but, the downside is that it needs more follow-up once the intervention has ended.

The committee did not put the list of interventions in order of priority, because the results of the child or young person's assessment should help practitioners decide what type of intervention to offer. Members agreed that what was needed was a 'toolbox' of approaches that could be tailored to individual needs. From their own experience, they agreed that comprehensive, multicomponent interventions that focus on the child or young person's family and background are more promising than those that focus solely on the abusive behaviour.

The committee also discussed looked after children and young people displaying harmful sexual behaviour and the need for foster carers to be adequately trained. It heard testimony from expert paper 10 that young people in the care system with harmful sexual behaviour often experience multiple placement moves. This, in turn, can affect the child's willingness to form attachments and makes therapeutic interventions more challenging.

Section 1.8 Supporting a return to the community for 'accommodated' children and young people

The discussion below explains how we made recommendation 1.8.1.

Current practice

The Glebe House model, a specialist children's home, is an example of residential practice in this area. It is based on a therapeutic community model for adolescent males with a known history of harmful sexual behaviours. The committee noted that the Glebe House model is not usual practice in this area, and that the types of interventions offered at Glebe House are very different from those offered by custodial services.

Young people in a young offender's institution serving a custodial sentence for a sexual offence do not always receive harmful sexual behaviour services. Local youth offending teams should provide this service but it can take months to arrange. That's because it may involve transferring the person to a young offender's institution that offers specialist services. It is not uncommon for a transfer from one custodial setting to another to take place a few months before the release date.

The committee noted that even where harmful sexual behaviour services are commissioned, the threshold for provision varies and is occasionally too high. For example, young people who receive a custodial sentence for harmful sexual behaviour may not be offered these services if their sentence is under 6 months.

Youth Justice Board statistics for 2014/15 show that the average time, from the date of an offence being committed until completion of court proceedings, is 66 days. But for sexual offences this rises to 295 days – so many would not be eligible for support by the time they are sentenced. In addition, young people aged between 12 and 17 who receive a 12‑month detention training order would not be eligible. That is because half the sentence will be spent in custody and the other half will be supervised by the youth offending team in the community.

However, the Wakefield harmful sexual behaviour model allows all young people displaying harmful sexual behaviour, whether or not it is part of the offence, to be referred. This is also regardless of the length of time they spend in custody and whether or not they receive a custodial or community sentence.

In this model, any agency involved with the young person can refer and self-referrals are also accepted. Everyone is offered a consultation plus a transition package, regardless of whether they are discharged into the community or transferred to adult prison.

Evidence for effectiveness

The committee noted that a small number of children and young people displaying harmful sexual behaviour may warrant placement in a specialist residential or secure setting. It drew on evidence from expert paper 6 as an example of a specialist children's home that uses a therapeutic community model.

The committee agreed that, if possible, residential settings should draw on the values and approaches of a therapeutic model originally developed in the field of social psychiatry by Rapoport and Roscow[8]. This is based on 5 social psychology principles: attachment, containment, communication, involvement and agency.

The committee agreed that interventions in residential settings should be based on the principles outlined in this guideline, including the principles and approaches set out in sections 1.6 and 1.7.

The committee agreed that residential settings should also provide a range of services, including ongoing support, to enable a child or young person to successfully integrate back into the community. In addition, if it is in the best interests of the child or young person, out-of-home care should not undermine relationships with their family. The committee referred to evidence previously noted for sections 1.7 and to ES1.6 and ES1.28.

Evidence reviews

Details of the evidence discussed are in evidence reviews, reports and papers from experts in the area.

The evidence statements are short summaries of evidence. Each statement has a short code indicating which document the evidence has come from.

Evidence statement (ES) ES1.1 indicates that the linked statement is numbered 1 in review 1. ES2.1 indicates that the linked statement is numbered 1 in review 2. EP1 indicates expert paper 1 'Definitions, epidemiology and natural history of HSB'. EP2 indicates expert paper 2 'Developmental pathways towards sexually harmful behaviour and emerging personality disorder traits in childhood'. EP3 indicates expert paper 3 'Harmful sexual behaviour of children'. EP4 indicates expert paper 4 'Girls who display harmful sexual behaviour – developing assessment tools and intervention resources'. EP5 indicates expert paper 5 'Glebe House'. EP6 indicates expert paper 6 'Turn the page'. EP7 indicates expert paper 7 'Harmful sexual behaviour – children and young people with learning difficulties who display harmful sexual behaviour'. EP8 indicates expert paper 8 'AIM project'. EP9 indicates expert paper 9 'Service user expert testimony'. EP10 indicates expert paper 10 'Practitioner and advocate expert testimony'. EP11 indicates expert paper 11 'Pornography and its impact on harmful sexual behaviour'. EP12 indicates expert paper 12 'The development of an operational framework for children and young people who sexually harm'. EP13 indicates expert paper 13 'An overview of policy and practice'. EP14 indicates expert paper 14 'MST PSB trial'.

If the committee considered other evidence, it is linked to the appropriate recommendation below.

Section 1.1: ES2.9; EP1, EP9, EP10

Section 1.2: ES2.1, ES2.13; EP1, EP7; Department for Education's 'Working together to safeguard children.'

Section 1.3: ES2.1; EP7, EP12

Section 1.4: ES1.6, ES1.28, ES2.2, ES2.3, ES2.4, ES2.5, ES2.6, ES2.7, ES2.8, ES2.10, ES2.14; EP3, EP4, EP7, EP8

Section 1.5: ES1.16, ES1.18, ES1.20, ES1.21, ES1.22, ES1.23, ES1.24; EP1, EP9, EP10

Section 1.6: ES1.6, ES1.17, ES1.28, ES2.2, ES2.3, ES2.4, ES2.5, ES2.6, ES2.7, ES2.11, ES2.13; EP3, EP4, EP8

Section 1.7: ES1.6, ES1.7, ES1.8, ES1.10, ES1.11, ES1.12, ES1.13, ES1.14, ES1.15, ES1.16, ES1.17, ES1.19, ES1.25, ES1.26, ES1.27, ES1.28; EP2, EP3, EP5, EP9, EP10, EP14

Section 1.8: ES1.6, ES1.15, ES1.28; EP5, EP9, EP10

Gaps in the evidence

The committee's assessment of the evidence on harmful sexual behaviour and stakeholder comments identified a number of gaps. These are set out below.

1. A comparison of the effectiveness of therapeutic approaches such as cognitive behavioural therapy and multisystemic therapy for children and young people who display harmful sexual behaviour and their family and carers.

(Source ER1)

2. Evidence of effectiveness for recognised assessment and treatment models, such as the Good Lives Model and AIM2, for children and young people who display harmful sexual behaviour.

(Source ER1)

3. Empirically evaluated tools to assess need and predict the risk of harmful sexual behaviour among children and young people in the community including:

  • different age groups (that is, children under 10 and children and young people aged 10 and older).

  • those with neurodevelopmental or learning disabilities

  • those from black and minority ethnic communities

  • those at the less severe end of the harmful sexual behaviour spectrum.

(Source ER2)

4. Evidence on interventions aimed at younger children (prepubescent or under 10) who display sexualised behaviour that is of concern.

(Source ER1)

5. Evidence on actuarial models used to assess children and young people who display harmful sexual behaviour.

(Source ER1)

6. Rates for continued problematic sexual behaviours following prepubescence.

(Source EP1)

7. Factors that encourage children and young people to go on to commit more serious sexual offences.

(Source EP1)

8. Evidence of the impact that pornography and new technologies have on harmful sexual behaviour such as sexting, the posting of sexual images and grooming.

(Source EP11)



[6] This guidance was subsumed into 'Working together to safeguard children' and is no longer current.

[7] Moffitt T (1993) Adolescence-limited and life course persistent anti-social behaviour: a developmental taxonomy. Psychological Review 100: 674–701.

[8] Rapoport R, Roscow I (1960) Community as doctor. New York: Arno Press.

  • National Institute for Health and Care Excellence (NICE)