Life course outcomes and developmental pathways for children and young people with harmful sexual behaviour

ABSTRACT Most outcome studies for children and young people who have displayed harmful sexual behaviour have focused on sexual recidivism as their primary outcome measure. Relatively little is known about broader life outcomes for children displaying such behaviours, nor about the processes involved with longer-term developmental success or failure. This paper examines long-term life course outcomes for 69 adults in the UK who presented with abusive sexual behaviour as children. Between 10 and 20 years after their childhood sexual behaviour problems, few in the sample had sexually reoffended, but general life course outcomes were much less positive. A range of individual, relational and social/environmental factors appeared to be associated with successful and unsuccessful outcomes. Successful outcomes were associated with stable partner relationships, wider supportive relationships, and educational opportunity and achievement. The findings highlight the importance of broad-based, developmental interventions in assisting those with childhood sexual behaviour problems to live successfully. Practice impact statement The findings outlined in this paper support a move away from a traditional focus on clinical programmes of “sex offence specific work” with young people towards more developmental approaches that focus actively on wider individual, relational, and socio-environmental factors associated with desistance and resilient life outcomes. This includes, but is not restricted to, interventions that are future-oriented and that promote the skills necessary for the development of supportive personal and intimate partner relationships. The importance of wider health promotion, general wellbeing and educational success is also emphasised.


Introduction
Existing empirical findings about children and adolescents with harmful sexual behaviours show that they are a highly heterogeneous population, reflecting a diverse range of backgrounds, motivations, age of onset, behaviours exhibited and victims (Hackett et al., 2013;Righthand & Welch, 2001).As awareness of the issue of sexually abusive behaviour perpetrated by children and young people has grown, so has recognition of the importance of seeing such behaviour not in isolation from other elements of a child's life, but as a critical developmental experience that sits alongside, and that may be the symptom of, other developmental challenges for children.Consequently, professional consensus on interventions for children and young people has moved away from the earlier influence of adult sex offender theories and practices (Hackett et al., 2006;Longo, 2003) towards more systemic interventions, for example Multisystemic Therapy (Letourneau et al., 2013), and towards rehabilitative approaches, such as the Good Lives Model (Ward et al., 2007).
This has resulted, among other developments, in a shift in thinking about the terminology that should be used to describe the problem itself.Whilst terms such as "juvenile sex offender" or "adolescent sexual abuser" have been in common use, they have been criticised for their stigmatising and labelling impact, their tendency to define the whole child by one aspect of the child's behaviour, and their associated implication that this behaviour constitutes a fixed identity trait (American Psychological Association, 2020; Willis & Letourneau, 2018).This is not reflective of either the wide diversity of sexual behaviours expressed by children and adolescents 1 (Hackett, 2014), or the evidence regarding low rates of sexual recidivism (Caldwell, 2016;Lussier, 2017).As a result, terms such as "children with harmful sexual behaviour" are now the preferred terminology for many organisations working in this field, especially in the UK.We reflect that terminology in this paper using Hackett et al.'s (2016) definition of harmful sexual behaviour as: sexual behaviour expressed by children and adolescents under the age of 18 which is developmentally inappropriate, may be harmful towards the child or adolescent themselves or others, or be abusive towards another child, adolescent or adult.(p.12) Despite the shifts in conceptualising the problem of harmful sexual behaviour in childhood, and the associated developments in intervention responses, little is known about the long-term developmental outcomes for this group through adolescence and into adulthood.

Developmental life course approaches
For almost two decades there have been calls for a paradigm shift towards developmentally informed approaches to harmful sexual behaviour of children and adolscents (Chaffin et al., 2002;Hackett, 2004Hackett, , 2014;;Lussier, 2005Lussier, , 2017;;Smallbone, 2006).A developmental life course criminology approach could be instrumental to this shift.Developmental life course criminology integrates theories of developmental criminology (Farrington, 2002;Le Blanc & Loeber, 1998) and life course (Laub & Sampson, 2003).Developmental criminology considers early life experiences and their impact on youth offending.A life course perspective argues that early life experiences are insufficient to explain later behaviour, that human development does not stop in childhood and that life events and turning points in adulthood are also important (Laub & Sampson, 2003;Sampson & Laub, 1993).A further example of life course theory is Sampson and Laub's (2005) age-graded theory of informal social control, which views crime as an emergent process reducible neither to the individual or the environment, but which suggests that social structures indirectly affect delinquency through the social bonds and controls they build.So, in combining the perspectives of developmental criminology and life course frameworks, developmental life course criminology considers: the development of offending and antisocial behaviour; risk and protective factors at different ages; and how life events effect the course of an individual's development (Farrington, 2003).
Although these theoretical perspectives were developed in relation to general offending they have since been considered and applied to understanding children and adolescents with harmful sexual behaviour.Lussier (2017) argues that traditional approaches to "juvenile sexual offending" focus on the identification of traits but neglect developmental considerations: The current state of knowledge on the origins and developmental course of sexual offending is limited to crude developmental parameters based on a limited number of studies conducted mainly with small samples of adjudicated juveniles that are not representative of the general population involved in sexual offenses.(Lussier, 2017, p. 71) Lussier (2017) argues that instead of continuing to focus on the between-individual differences, research needs to examine within-individual stability and changes across developmental stages and life transitions, and how they may affect harmful sexual behaviour.In other words, this emphasises the importance of adopting a developmental life course criminological perspective and shifting consideration from a variable-orientated approach to a person-centred approach, viewing the individual as a whole.Taking a holistic approach requires not only consideration of individual, but also relational and socio-environmental factors (Firmin & Knowles, 2020;Firmin & Lloyd, 2020;Smallbone et al., 2013).

Outcomes for children with harmful sexual behaviour
To date, much of the focus of outcome research on children and young people with harmful sexual behaviour has been on sexual recidivism rates and on the identification of risk factors implicated in both sexual (and to a lesser extent) non-sexual reoffending.Whilst knowledge about risk factors for sexual recidivism is important, it may not be enough to understand the developmental factors and challenges that follow harmful sexual behaviour in early childhood and adolescence.Likewise, stopping a child from sexually abusing others is a necessary but insufficient intervention goal.Despite the low reported sexual recidivism rates for children and adolescents with harmful sexual behaviour, many policy responses are still uniquely concerned with risk of reoffending and consequently much practice remains weighted towards this issue as a primary concern.Simply stopping a child from continuing something that, even without intervention, is not likely to continue, is hardly a triumph, especially if in the process that child's broader life chances are ruined or severely inhibited by the consequences of our interventions and policies.
How far the expression of harmful sexual behaviour in childhood is a marker for future developmental risk, as opposed to risk of criminality or sexual violence, is at present uncertain.Few studies have attempted longer term follow up to understand the nature of the lives of adults who in their childhood had displayed harmful sexual behaviour.Consequently, the extent to which harmful sexual behaviour is a risk marker for all kinds of developmental problems not connected with sexual or non-sexual violence or offending is unclear.Assuming some children do well and others badly in their broader development, the role of the harmful sexual behaviour and the chain of experiences that follow it, remains largely unknown.Crucially, the mechanisms that determine poor developmental outcomes following childhood harmful sexual behaviour may be unrelated to risk factors for sexual or non-sexual recidivism.If this is the case, then traditional "sexual offence specific" approaches for children and adolescents may very well miss the mark when it comes to supporting more general, positive life goals.
This research study seeks to address these important knowledge gaps by examining long-term outcomes for a sample of 69 adults who displayed harmful sexual behaviour as children/ adolescents and by considering them in relation to existing theories of desistance and resilience.

The study
The overall aim of our study was to describe and analyse the experiences and life circumstances of young adults who, in their childhoods and adolescence, were subject to professional interventions because of their problematic sexual behaviours, and to consider the implications of these experiences for future policy and service delivery.Within this overall aim, our research questions were: (1) What life course outcomes and developmental experiences (at individual, relational and socioenvironmental levels) are evident in individuals who displayed harmful sexual behaviours in their childhood and who had received intervention as a result of those behaviours?(2) What factors appear to contribute to whether longer-term developmental outcomes are successful or unsuccessful for these individuals?

Design
Our research used a range of qualitative methods including documentary analysis and interviews.Generating a sample for retrospective follow-up at a considerable time distance from original referral was a major challenge and required extensive, careful, and ethically sensitive preparatory work over four distinct stages.First, we sought ethical approval to access and review 700 historical case files that constituted all children under the age of eighteen years old referred to nine UK specialist services between 1992 and 2000.The cut off period for referrals ensured a follow-up period for the subsequent stages of the work of between 10 and 20 years in each case.The services in question specialised in work with children with problematic sexual behaviour and young people who had sexually offended (which, as noted above, we refer to throughout this paper as "harmful sexual behaviour", even though that term was not in use in the time period represented in the case files).The services were a mix of community-based and residential facilities across England and Wales in the statutory, voluntary, and private sectors ensuring the major types of service working in this field at the time were represented.Given the nature of the services as specialist resources, the children and youth referred to them were often those with very significant sexual behaviour problems.Masson et al. (2011) describe in further detail the methodological complexities encountered when looking back at these historical case files and their implications for future research of this type.
For each of the 700 cases, we collected data on: age at referral; ethnicity; gender; the nature of the problematic and abusive behaviours; convictions recorded at the time of referral; victim age(s) and gender; as well as the referred child family history, including their own experiences of victimisation (physical, emotional, and sexual).The data collected were entered into the Statistical Package for Social Sciences (version 16) to obtain descriptive statistics and undertake exploratory data analysis.Of the total sample, 97% were male and 3% female.Forty-two per cent had prior criminal convictions.A wide range of concerning sexual behaviours had been displayed by those in the sample.The most frequently occurring behaviour involved inappropriate touching of others' genitals, with just over four-fifths of all those referred across all sites having engaged in this type of behaviour.In addition, just over half of the sample had penetrated or attempted to penetrate another individual and 50% had engaged in non-contact sexual behaviours.Sexual violence or the use of physical force was a feature of the behaviour of almost one in five of individuals.
The majority of the sample (66%) were known to have experienced at least one form of abuse or trauma including physical abuse, emotional abuse, sexual abuse, severe neglect, parental rejection, family breakdown and conflict, domestic violence and parental drug and alcohol abuse.There was documented and clear evidence that 31% of the sample had been sexually victimised earlier in their childhoods, with a similar rate for prior physical abuse.For a full description of the whole sample of 700, see Hackett et al. (2013).
In stage 2 of the study, we identified a smaller proportion of around twelve cases from each of the nine participating sites with a view to trying to trace the ex-service users.We used a stratified purposive sampling strategy to identify cases which were representative of the full range of clients from each service over the whole period.As the services differed somewhat in nature, we had to do this at the level of each of the services.We used a matrix developed by the research team to select 117 cases for more in-depth analysis, (see Appendix 1 for the list of factors included in the matrix).The matrix was designed to capture the range of children and young people with whom each of the services had worked and included, for example, cases involving intra and extra-familial abuse, contact and non-contact offenders, male and female victims and those involving violent and nonviolent behaviours.We reviewed these cases in depth and produced a case profile for each individual, as well as extracting key contact information for each case.Approximately 95% of the 117 cases were male and "White".The majority, 63%, were between 13 and 16 years of age at referral.Over half of the sample were noted to have experienced a form of abuse other than sexual abuse, and just over half had either themselves been sexually abused or suspected of being sexually abused.
In stage 3 we searched for participants through a variety of means, including by contacting addresses on historical files for the individuals and using the knowledge of professionals who had worked with the individual concerned.We also used a range of publicly available data sources, including checking the UK electoral register, social network sites and the use of general internet search engines.We successfully traced 87 of the 117 individuals.Masson et al. (2011) provide more detail about the rationale for this approach to tracing participants and the work undertaken to do so, including the practical and ethical issues involved.
Stage 4 of the research comprised of contacting the traced individuals to invite them to participate in interviews.It was vital not to compromise the current living situation, safety, or wellbeing of the individuals who we were attempting to contact.We took an extremely cautious approach with individuals being contacted discreetly by the nine services that had originally worked with them, rather than directly by the researchers.When services contacted individuals initially, they did so in a neutral manner, without specifying the nature of the service or work that had previously been offered, in case any messages were opened, or calls overheard by someone else other than the intended recipient.Where contact was established, permission was sought by the specialist services for the researchers to approach the participants to discuss the project more fully.Informed consent to participation was then negotiated and participants were given a choice as to whether they would prefer a face-to-face interview with the researchers or other less direct means of contact, such as telephone calls.
A small number of former service users were not in a position to, or chose not to, talk directly to us, communicating in writing instead.This included, for example, one individual who was incarcerated.A number of people gave permission for the researchers to interview their parents or carers instead.This approach meant that, whilst we were also able to conduct in-person interviews with 24 of the former service users, in total, we were able to collect detailed outcome data (e.g. from phone and in-person interviews and written interview responses) on 69 of the 87 individuals we had traced.Our flexible approach meant that the outcome data was often comprised of multiple data sources, such as from the former service user themselves (however they chose to provide it to us), their parents, carers or professionals involved with them.This was complemented at times by written professional reports that were shared with us with the individual's consent.Case file data and other historical information provided by professionals who had worked with the participants as children helped to provide to broader context around the individual's circumstances at the time of their engagement with the services working on their harmful sexual behaviour.Information gathered from other sources such as the interviews with parents, carers and professionals working with the individuals after their original intervention ended or more recently, provided further detail on the life outcomes for participants.
Where interviews were held, a mix of narrative and semi-structured methods (Flick, 2006) was used to encourage participants (or their parents, carers, or other professionals) to recall and reflect on their key life experiences and the impact of their harmful sexual behaviour as a child or youth in later life.Whether they took place in person or by telephone, interviews were undertaken by the authors in mixed gender pairs and were audio recorded.Interviews lasted between one and three hours in length.The interviews aimed to elicit participants' self-constructed stories of their lives and experiences, which were also compared with data from the case files and any other information gathered, for example, from professionals who had known the young person during the original intervention, or from other family members or carers.

Sample
The majority of the sample was male (64 male, 5 female).We recognise our sample is overwhelmingly male, as is the focus of broader literature in this area, and that there are empirically verified differences between males and females in terms of anti-social, aggressive, and harmful sexual behaviour.However, we feel it is important to include the female participants in this study to give voice to their experiences and for these to be included in the overall understanding of the specific issues were are examining here.We also acknowledge there are different gendered understandings and meanings around relationships, however, we found that in this context participants' experiences were very similar in terms of the outcomes and the factors leading to these.
The sample of 69 was aged between eight and eighteen years old at the time of the harmful sexual behaviour that resulted in their referral to the services involved, and the average age of referral was fourteen.As stated above, the follow-up period ranged from between ten and twenty years in each case, meaning that at the time that the interviews were conducted, all participants were in their twenties and thirties.The average follow-up period was 13 years.Participants had displayed a wide variety of acts of harmful sexual behaviour when they were young, commensurate with the range reported in the larger sample of 700 as described above.

Data analysis
Based on a contextualist theoretical framework, thematic analysis (Braun & Clarke, 2006, 2012) was used to analyse the data.Thematic analysis is a flexible and practical method for identifying, analysing, and reporting patterns (themes) within data (Braun & Clarke, 2006).Using this method, members of the research team first read and re-read a sample of interview transcripts and written accounts independently, identifying initial codes.These initial codes were compared and discussed to arrive at an agreed coding structure for the full analysis which was subsequently applied to all transcripts and written accounts using the qualitative data analysis software package NVivo.Following initial analysis, the codes were categorised and labelled into key themes related to individual, relational and social/environmental factors.Key themes identified within these different levels were: living with the past; ambition and achievement and problem solving; relationships with family of origin; relationships with partners and being a parent and support and isolation.Adopting the kind of holistic, developmental approach outlined in the introduction to this article, we considered the individual, relational and socio-environmental factors associated with resilient or persistently problematic life outcomes.Based on these factors, following on from the coding process, members of the research team made informed judgements and classified cases as indicating "successful", "mixed" or "unsuccessful" life outcomes.Four members of the research team undertook this process independently, then comparing individual ratings with the whole team to ensure consistency of rating.Discrepancies in the ratings were rare, but when they occurred, the team discussed the factors present in the case and we were able in each case to reach consensus.These classification measures used were based upon Farrington et al.'s (2009) life success measures which comprise of nine criteria originally derived from interviews with adults who had been considered "delinquents" in childhood: satisfactory accommodation; satisfactory cohabitation; satisfactory employment; not being involved in physical fights; satisfactory alcohol use; no prohibited drug use within the preceding 5 years; non-self-reported offence within the preceding 5 years; a score of 0-4 on the General Health Questionnaire; not convicted in the previous 5 years.Farrington and colleagues (2009) classified participants as leading successful lives if they were successful on at least six of these nine criteria.
We based our classification on Farrington et al.'s (2009) factors but extended them somewhat in order to consider a broader set of evidence of successful outcomes aligned with those found in general desistance and resilience literature, both more generally and specifically regarding harmful sexual behaviour.For example, we included not just "satisfactory cohabitation" but stable partner relationships (whether co-habiting or not), pro-social friendships and networks, and positive carer and professional relationships.These factors were grouped at individual, relational and social/ environmental levels.The categories and factors used to classify individual cases were not fixed due to the complex nature and combinations of factors experienced by individual participants.Like Farrington et al. (2009) we used the presence of at least six success factors as a guide when determining whether overall life outcomes were successful, however, we were mindful that in individual cases more nuanced consideration was appropriate.For example, some participants experienced success in some areas of their lives but were unsuccessful in others, therefore their overall life outcome could not be appropriately classified as either successful or unsuccessful.Therefore, we introduced the category of "mixed" to address this and more accurately reflect the findings in individual cases.In doing so, we were able to acknowledge the areas of success individuals experienced in some parts of their lives.As a guide, we considered having between three and five success factors as determining overall outcomes as "mixed", with less than three success factors indicating unsuccessful outcomes.

Ethical issues
Ethical approval for the research was obtained from the authors' respective universities and from all the services who participated in the research.Great efforts were made to ensure that the study complied with ethical codes of conduct in research with human participants, such as those produced by the British Psychological Society (2010) and the British Sociological Association (2017).In addition, the research team was very aware of the need to conduct the research study in accordance with the principles and requirements of the UK's Data Protection Act (1998) which was the primary legislation governing the research during the data collection.

Results
As far as could be ascertained by self-report and official records, most participants had not reoffended.Only a small proportion had reoffended sexually, with three individuals reconvicted for sexual assault and one for making and distributing indecent images of children, giving a 6% sexual recidivism rate.However, general reoffending was more common, with a number of participants explaining that they had been in trouble with the police, but not convicted, for low level antisocial behaviour, such as being drunk and disorderly in a public place.Adjudicated non-sexual reconvictions were part of the trajectory of six of the 69 individuals (i.e.9% of the sample) who, between them, had multiple offences including drug crimes, theft and burglary, physical assault and domestic violence, arson and, in one case, murder.Although reoffending rates were low and in line with larger quantitative studies that have specifically focused on measuring both sexual and non-sexual recidivism, our specific focus on more general life outcomes revealed much more diverse, and in many cases, negative developmental trajectories.
In the cases with the most successful outcomes, former service users reported that they were living very stable and happy lives.These individuals reported no enduring negative consequences arising from their childhood behaviours and, given the data from interviews and information collected about them from multiple sources, their lives in adulthood appeared to be remarkably "normal".This did not mean that they were unconcerned about their previous sexual behaviour problems; indeed, many expressed their shame and distress about the harm they had caused their victims and others when they were children.However, these individuals also repeatedly made statements that they had put their previous behaviours behind them, turned over a new leaf, found a new positive identity, and no longer considered themselves as a risk to others.
Conversely, in the least successful cases, participants described their lives as highly negative, unhappy, and characterised by a continuing series of personal problems and challenges.Although many of these individuals had not gone on to reoffend, they reported high levels of discontent and distress about their lives.They were living in highly unstable and disadvantaged states, and they often told us about a series of personal disasters that had befallen them after the end of the interventions for their harmful sexual behaviours.For many, their childhood experiences appeared to remain an open wound, from which they had as adults never really succeeded in moving on.
Few of these outcomes, either successful or unsuccessful, appeared to be related to their previous harmful sexual behaviour.Some of those whose original assessments predicted the worst sexual recidivism outcomes were living the most positive lives and vice versa.Few variables relating to the harmful sexual behaviour itself (e.g.age of onset, nature of sexual behaviours, gender of victims) appeared to contribute to such a diversity of life course outcomes.It was, however, the case that amongst those reporting unsuccessful adult outcomes were the small number of individuals in the sample who had presented not only with harmful sexual behaviours, but also non-sexual violence and significant levels of general offending in their youth.Conversely, some of those presenting with the most positive long-term outcomes reported having sexually abused others after having been sexually abused themselves as children.A potential explanation for this could be that these individuals in particular were able to contextualise the reasons why they had harmed others sexually, understanding that their behaviours were influenced by their own abuse experiences.These individuals appeared to have a more coherent narrative about what they had done as children and felt more able to "forgive themselves" as adults for their earlier behaviours.Such an understanding could have helped these individuals to distance themselves from their past identity as an "abuser" as they entered adulthood.
So, the question remains, if there were substantially more varied developmental outcomes than either sexual or non-sexual recidivism outcomes, what accounted for this?What are the factors (especially if they are not related to harmful sexual behaviour) that accounted for such variance on general outcomes?To try to understand this, we sought to establish what was typical about the developmental trajectories of those participants who did well despite the marker of childhood harmful sexual behaviour, and those who did not.This understanding is important; if we can understand what factors contribute to successful long-term outcomes children and youth with harmful sexual behaviour, then we can focus responses and interventions on these factors and work to improve the lives of those who are otherwise at risk of doing badly.
As discussed above, to try to identify and differentiate the factors associated with the diversity of outcomes we were noting, we classified life outcomes as successful (26% of cases), mixed (31%) or unsuccessful (43%).Individuals with successful outcomes tended to report happiness and stability in their lives.Stable partner relationships or enduring carer and professional relationships were a feature of the lives of most participants with successful outcomes.Educational achievement and the ability to gain and maintain employment also constituted significant desistance factors.Unsuccessful outcomes, in contrast, were associated with unhappiness, poor relationships and unstable lifestyles as well as poor mental and physical health.Relationship failure, chaotic or unstable living conditions and drug and alcohol misuse were also common amongst those with unsuccessful outcomes.Life outcomes were classified as mixed outcomes in circumstances where the individual's young adult experiences featured some successful and some unsuccessful factors.The factors associated with successful and unsuccessful outcomes are detailed in Table 1 and are presented in terms of individual, relational and social/environmental levels.

Individual level factors
At an individual level, key themes were living with the past and ambition, achievement and problem solving.Successful developmental outcomes were associated with personal ambition, life purpose and optimism about the future.Participants with better outcomes also tended to possess selfbelief, a sense of individual agency and control over their own lives.Participants who had more successful outcomes were those who were ashamed of their harmful behaviour as a child or youth, but who were able to take responsibility for it.They were now able to separate themselves from that behaviour, recognising they were no longer that child and they considered themselves no longer at risk of similar behaviour in adulthood.In contrast, participants with poorer outcomes were unable to offer a coherent narrative of their past and demonstrated a lack of understanding of the impact of the label of "sex offender" on their lives.
For many participants, living with the legacy of their past harmful sexual behaviour was still a significant issue; an historical marker which affected their self-perceptions and their expectations of interactions with others.A number expressed ongoing guilt, remorse, and shame about what they had done so many years ago.For example, one participant commented: I do get bad thoughts, not bad thoughts like that [sexual thoughts] but they're down thoughts, because I know what I've done is wrong, there's nothing I can do to change it or anything, other than to get on with my life and try and make amends.(Aged 33 years, mixed outcome) Some participants reported feeling that they were still labelled as adults by their childhood behaviour or even, as one put it, "branded" like an animal by it.These participants reported living under constant fear that their past would seriously affect their lives and relationships in adulthood.One commented that "it's ready to bring you down, it's like you walk around with a secret" and another remarked "mud sticks".These participants reported occasions when historical information about them limited their adult life chances without consideration of their current life circumstances and in the absence any continuation of their childhood sexual behaviours.One man explained how professionals continued to view him through the lens of his childhood behaviour, despite the fact that he had displayed no such repeated behaviour throughout his later adolescence and into adulthood: The coppers [police] have connected something that happened … in a completely different location, not even in the same decade, we're talking a different century, I was thirteen, fourteen, so we're looking like half a lifetime later and they're still connecting the two.(Aged 28 years, unsuccessful outcome) In this context, a key question for many participants was who to tell about their past because of their worries about the negative reaction they might get if they told partners, other family members or employers.As one participant expressed it: You kind of feel you're hiding it from everyone, so you're not being yourself, you want to tell people, but you're worried about the consequences.(Aged 21 years, successful outcome) Feelings of life purpose and achievement were linked to individuals' perceptions of their own achievements in education, employment, relationships, and parenthood (all discussed further below).Success in these areas contributed to participants feeling that their lives had purpose, that they no longer had to define themselves as ex-abusers or as someone who had had a sexual behaviour problem; instead, they could re-define themselves as workers, providers or as individuals who had successfully achieved qualifications.These achievements and experiences seemed to be protective in helping participants move on from negative childhood identities associated with abuse, to new identities with life purpose and fulfilment.Participants' educational achievement was an important factor identified in those with more successful outcomes.This was both in terms of how educational achievement supported the development of self-esteem and feelings of life purpose and fulfilment, but also how it provided participants with the skills necessary for employment.Access to opportunities for educational attainment in young adulthood was important; this was true even if participants' school careers had been characterised by underachievement, truancy, and exclusion as a result of the risks that their sexual behaviours were seen to present to other pupils or staff.Later educational opportunities through training programmes, outward bound activities or academically focused post-16 education appeared key in providing skills and experiences for participants so that future employment and financial security became tangible and realistic goals.One respondent who had previously had a very poor school record, including persistently running away from a residential boarding school, took great pride in showing us his record of achievement including certificates for being part of the crew of a sailing ship and learning to drive a truck.
Overall, at the individual level, participants with the best outcomes had higher levels of self-confidence and could articulate a clear purpose to their lives.Those with unsuccessful outcomes, by contrast, lacked the ability to plan, both in the immediate and longer term, with most of them, seemingly, just surviving day-to-day.It was notable how often participants demonstrating the worst overall outcomes also presented with high levels of physical ill-health.For these individuals, a tendency to neglect self-care, longstanding unhealthy lifestyles and associated poor body image impacted on self-esteem, self-confidence, and individuals' ability to form relationships, including intimate partner relationships.In other words, poor overall health acted as a catalyst for, and intensified, other difficulties arising from their histories of sexual behaviour problems.
Physical health promotion is an area which is almost entirely missing from the orthodox literature on "adolescent sexual offender" treatment but could have a very significant role as a protective factor, allowing people to turn around their lives following harmful sexual behaviour.Other participants experienced enduring mental health problems including depression, suicide attempts and self-harm.Participants with intellectual disabilities also generally had poorer outcomes than those who had no such disabilities.There were a small number of examples of adults with learning disabilities seemingly lost in the system, stuck for many years in highly restrictive residential services without a clear plan for their futures.

Relational level factors
At a relational level, key themes were: relationships with family of origin; relationships with partners; and being a parent.Successful outcomes were more evident in those with stable partner relationships, as well as positive wider relationships and support systems.In contrast, many of those with less successful outcomes had experienced relationship failure and were isolated and lacking social support.

Relationships with partners and being a parent
Those participants with successful outcomes into adulthood were most often those in stable intimate partner relationships.In some cases, such relationships reflected positive outcomes in the individual's personal functioning and relationship skills and their success at the socio-environmental level.However, for others, their partners represented a crucial stabilising force in their lives; accepting them, encouraging them to turn their lives around and helping them to forge a new identity.In contrast, those with unsuccessful outcomes often reported difficulties in relating to others, underpinned by a lack of confidence, feeling scared about being hurt if they made any kind of relationship commitment or because their lives were generally chaotic (due to drugs, alcohol, lack of employment or continuing minor criminal behaviour).Such vulnerability regularly resulted in the formation of transient, and often volatile, partnerships with other disadvantaged or vulnerable individuals; relationships which might well be viewed with reservation by professional agencies.One respondent who had a long-term girlfriend and who was being supervised in the community understood that: Probation won't agree with it, obviously because they're classing it as a volatile relationship … basically they don't want me to be with her, I know full well they don't.(Aged 22, unsuccessful outcome) Even where relationships seemed happy and stable, the question about informing partners about their history was a significant issue for some participants.One man had a steady girlfriend who wanted to have a baby.He also wanted to have children, but the situation presented challenges about whether he should tell her about his past.He worried: What will happen if I tell her, a kick in the teeth, she spreads the word, I'm hated, that's it, I've got to start a new life all over again.(Aged 26 years, successful outcome)

Experience of parenthood
Some of the interviewees talked at length of their experiences of being parents or step-parents.Some were estranged from or had little contact with their children, but others reported positive experiences of being a parent that contributed to their self-perception.For example, one participant commented that: It made me realize that, you know, I'm a pure adult now, I've just created the most amazing thing in the world, it's the best feeling I've ever had in my life, it's the greatest achievement I'll ever have in my life.(Aged 28, mixed outcome) A few participants compared their hopes and aspirations for their own parenting identity with their experiences of their own fathers.One was clear that he wanted to be: … a father that's there, I grew up with a fatherless past and never really knew my dad.My dad apparently died before I was born and I don't want that to happen to my child.(Aged 21,mixed outcome) Another reported that being a parent was: … tiring, it's fun, it's enjoyable, I find that a lot of things which I always wanted to do when I was younger, like with my dad, I can now do with my little boy, for example he loves playing football and I liked playing football when I was younger, but my dad didn't, so I never got to play football.(Aged 32, successful outcome)

Family relationships
In addition to relationships with their partners and own children, negotiating and managing contact with their families of origin was challenging for many participants.As a result of their own experiences of abuse or trauma as children and their harmful sexual behaviour, many had been physically separated from other family members during their adolescence.In adulthood, they faced the dilemma of whether to reconnect with their family and how best to do this in the absence of any kind of professional mediation or restorative help.Whilst some had tried to do this, their attempts were not always successful and resurrected past tensions.Some kept family contact limited because of the abuse they had experienced at the hands of parents or other family members; others were aware that their own harmful sexual behaviour had badly disrupted the functioning of the family or had harmed their siblings and felt continuing shame about this.For example, one participant who had sexually abused his sister, told us: It was very important for me to sit down with my sister and talk about the issue, it really was, because I felt if I hadn't done that, I'd never have known how she felt towards me in that sort of sense.(Aged 30 years, successful outcome) He said his sister had forgiven him, but he went to report continuing tension with his mother.
In many cases though, many years later, participants had still not had any conversation with their birth families about their past behaviour and its impact on family members and family relationships remained an "open sore" that continued to be anxiety and guilt provoking: I think deep down, my mum still has regrets about what I did, which is understandable, but I don't really think she wants to say, she doesn't want to use the word "hate", but sometimes I get the feeling that she does hate me for what happened, which is fair enough, it's understandable, we do get along, I mean me and my mum do have a rough relationship, we really do.(Aged 33 years, mixed outcome)

Professional relationships
For most participants, the quality (positive or negative) of the relationships that they had developed with one or more professionals involved with them as a child had a specific impact on the future direction of their lives.One participant emphasised the importance of a good connection between professionals and children.He described the need for: A spark, it's like I've had a few social workers (but) some social workers get on better with some people and some get on better with others, I don't know, you have to have a certain spark … where you can actually honestly come up to this person and tell them anything on this earth or ask them for their opinion … you need someone that you can openly be yourself with, you can't just paint yourself up and act someone different when they're there, because then that's just a false relationship.(Aged 21 years, mixed outcome) We heard about examples of workers who were going beyond "the call of duty" (and possibly the guidelines of their agencies) by maintaining contact with their service users long after cases had been formally closed.Consequently, even in their twenties and early thirties, if some participants experienced difficulties of some kind, they could still ring up the person who had worked with them to ask them advice about relationships, parenting, work plans, and so on.This was highly valued by individuals who did not always have other supports in their lives.In this context, these professionals became "social anchors'; people who were there in the lives of young people to help them with critical decisions or at key developmental turning points.When we discussed this with the professionals concerned, they tended to be embarrassed and worried that by providing this level of ongoing support, they were acting against their agencies' protocols and would "get in trouble".However, these "social anchors" were typical of some of the most successful outcomes, even though orthodox child welfare and adult services systems (certainly in the UK) tend to mitigate against such positively focused and enduring professional supports.

Social/ environmental level factors
"Support/ isolation" was the key theme identified here.In addition to the supportive personal and professional relationships, those participants who were doing better in early adulthood had wider social support.For example, one participant who was doing well described the importance of his friendship network as one that "just help(s) each other through things, difficult horrible situations, to be a support network for each other".On the other hand, the extent of isolation of participants who continued to experience major problems in their everyday lives was striking.Many presented as lonely and unsupported, with very limited or no networks and friends.
Pro-social friendship and colleague networks were also important in terms of allowing individuals to experience a sense of prosocial "belonging" and linked to this was the validation of cultural identity and feelings of fit into their wider environment that was evident among those with more successful outcomes.Successful outcomes at this level were also associated with employment and stable housing.
Those with mixed or unsuccessful outcomes experienced chaotic lives and unstable living conditions.They faced a diverse range of practical problems, including very poor housing or homelessness, unemployment, chronic debt, and continuing offending (mostly of a low-level nature such getting into fights, vehicle offences and minor theft).It was noticeable, too, how many of these participants had experienced, or were still experiencing, drug and alcohol misuse problems.

Reflections on services
In addition to the data gathered about the individual, relational and social/environmental related factors influencing life outcomes, some participants we interviewed also reflected on the services they had engaged with as children and youth with harmful sexual behaviour.For those leading successful lives in young adulthood this feedback was typically positive.Indeed, some respondents were fiercely loyal about the community or residential service they had been involved with and spoke powerfully about their intervention being the key factor that helped them to change.For example, one participant commented that his residential placement, a therapeutic community, became: like a family to you, you know, you have your daily meetings, you have your everything else, you have your workshops and things like that, you have your therapy … (The home) to me will always have a place in my heart and may long it carry on doing the work it does.(Aged 30, successful outcome) In contrast, other participants, usually those with mixed or unsuccessful outcomes, felt that the services they had encountered had not substantially helped them to understand the reasons for their past behaviour and move forward positively.A few felt that they had emerged from significant periods of intervention without any real sense as to why they had behaved in the way they had done, one commenting that "you've got a problem with no origin, no root, no explanation, no hint of why it happened" and it appeared such participants were still searching for this understanding to move forward in their lives.
Notwithstanding the need to resolve past issues, several participants criticised the way in which the interventions offered to them appeared to be too focused on their previous harmful sexual behaviour, with little emphasis on the ongoing challenges they were experiencing in their lives at the time.They argued powerfully that the services they accessed should have helped them focus on the future in order to address the ongoing developmental challenges that they faced as children, adolescents, and young adults.For example, one young man who had experienced two years of intervention complained that: Every time you go in there, they got you talking about the abuse, but they're always referring back, to try to get you to understand more about what's happened, but I found that, after they got me talking about it, it was just the same, going back, so when I'm trying to go forward.Then I come to puberty, it's a big jump because it's something new that I've never experienced before, it's all emotions going up, and then to go back and it's just the same, you're going back and back, there's no forward.It's like it would have been better to probably, instead of going back and discussing what happened, maybe talking about how I felt in relation to my current situation in puberty, how I was feeling, how it was affecting me.(Aged 21 years, successful outcome).

Discussion
The overwhelming majority (94%) of this sample of adults had not sexually reoffended between ten and twenty years after the expression of harmful sexual behaviour in their childhoods.At the same time, overall developmental outcomes were less positive, with only just over a quarter of the sample (26%) classified as having successful developmental outcomes.This study has sought to understand the factors that might account for such a disparity in sexual recidivism and general life outcomes and has used a developmental approach to add to the small existing literature on desistance and resilience in young people with harmful sexual behaviours.We have found that, like Sampson and Laub (2005), later life events and social influence were important in fostering desistance, developing resilience, and separating off the participants' past histories from their present circumstances and behaviours (Farrington et al., 2009).In line with strengths-based approaches, we found that participants with better outcomes were those whose individual agency and social embeddedness were evident.
The life journey narratives shared by the young adults in this sample indicate that they were generally trying to deal with many of the same issues that other adolescents face, particularly those with vulnerabilities or from disadvantaged backgrounds.Other studies examining life outcomes for vulnerable youth also highlight very similar difficulties to those experienced by young people with harmful sexual behaviour (e.g.Gypen et al., 2017;Preyde et al., 2011;Shin, 2004).They face the same challenges in managing the transition to early adulthood with limited or no formal support or support from their families or communities whilst experiencing developmental difficulties.Like the adolescents in Shin's (2004) study of life after out-of-home care, participants in our study had to navigate adolescent development at the same time as experiencing a range of other issues, such as mental and physical health problems, childhood trauma, educational, financial and employment difficulties, lack of social support, negative peer influences and the social stigma associated with their identities.
What appeared to particularly contribute to individual success in later life were experiences that had occurred more randomly in the lives of participants, rather than as a result of the content and modality of the professional interventions they had experienced as children and adolescents (i.e.meeting a supportive partner; having ongoing support from trustworthy and respected professionals; finding opportunities to participate and forging a prosocial adult identity through employment, etc.) Largely, the interventions offered did not appear to plan for successful developmental outcomes in some of these important areas of individuals' lives.Those individuals who did achieve success in those areas found them more as a result of chance than planning.Intervention design and delivery needs to target these areas much more vigorously and proactively, for example by teaching and supporting safe intimate partnership relationships or teaching skills to find and maintain satisfactory employment.Leaving things to chance is not an adequate approach.In other words, critically, many of those in our study who achieved the best outcomes did so more because of random events and key developmental turning points, rather than as a result of planning on the part of the professional systems engaged with them.
Given that known sexual recidivism rates are so low, with non-sexual recidivism rates typically higher (McCann & Lussier, 2008;Prentky et al., 2010), then this further calls into question the utility of risk-based approaches in the treatment of harmful sexual behaviour.What are we aiming to do through intervention?Are we just trying to stop children and youth displaying harmful sexual behaviours or are we trying to help them develop in the same way as other adolescents?
As the life narratives of our participants show, interventions need to acknowledge the past and identified risk factors for individuals but then focus much more substantially on overall developmental need and on engaging the individual towards ensuring positive life trajectories.As Ward and Brown (2004) point out, interventions need to focus on the internal and external conditions necessary for an individual to realise his own particular "good life", and by doing so individuals will be motivated.In other words, the best way to lower sexual and non-sexual recidivism rates is by equipping individuals with tools to live more fulfilling lives rather than developing more risk management strategies (Ward & Brown, 2004).The situation is challenging in that many of the policy frameworks within which interventions are still currently offered, with their emphasis on restriction and punitive control, arguably deprive children and young people of their ability, and right, to meet these normative developmental goals.
We advocate going beyond even a "twin focus" of treatment (to reduce harm and promote welfare), as suggested by Ward and Brown (2004), to a "triple focus"; one that not only promotes longer term welfare and reduces harm to the young person and others, but which also actively supports normal child, adolescent, and young adult development.Such a triple focus would mean much more overt attention being paid to wider child development in interventions for children and adolescents with harmful sexual behaviour, rather than just on the management of risk and welfare as has been evident to date.This additional focus needs to be actively built into the design and delivery of interventions rather than being left to chance.
At times the "adolescent sexual offender" treatment field continues to emphasise offence specific or abuse specific work to the exclusion of meaningful, in-depth, and active interventions in the child's wider life context.The key implication of this study is a necessary shift from passive attention to social issues (e.g. a child recounting their problems to a therapist in a session) to a very active engagement in the child's social and environmental ecology, not as an add on to the real business of "therapy", but as the key determinant of long-term, sustained change.
We have described (in Table 1 above) some of the factors that we have found to be particularly prevalent in the lives of young adults who appear to have done well developmentally, as opposed to those who have evidently not.Accordingly, these may be considered to constitute "protective factors" and "risk factors", though they look very different from the orthodox "adolescent sex offender" risk factors which have been identified in the literature to date.We suggest that the achievement of these positive factors, and the countering of the negative factors, could be the focus of a revised paradigm for developmentally focused practice with this population.If so, we also need to better understand the mechanisms underpinning these various factors.In our study, it was not the presence of any one of these factors, but the cumulative impact of these factors across the three levels as described that made a difference.In other words, these influences were often catalytic, producing a chain of events and circumstances, both positive and negative, that influenced developmental trajectories.Rutter (1999) has discussed what he terms "risk chain effect", stating that: Psychosocial risk experiences play a substantial role in the development of conduct problems in childhood … and the carry-forward of ill-effects into adult life is much influenced by these negative chain reactions by which people's behaviour increases the likelihood that they will have further adverse experiences.(Rutter, 1999, p. 130) If this is true for psychosocial risk experiences, then we consider that, conversely, psychosocial protective experiences play a substantial role in the development of healthy behaviours in childhood with substantial carry-forward of positive effects into adult life influenced by these protective chain reactions.This would mean that an individual young person's positive behaviour would increase the likelihood that they would have further positive experiences.The emphasis in terms of practice then, should be on influencing and building these positive, protective chain effects which would help signal to developing children that their future is likely to be largely stable and in a relatively risk-free environment, thereby discouraging risk taking and encouraging the development of pro-social cognitions, behaviours and relationships (Durrant, 2017).There could be several key practice implications, as suggested below.

Development, development, development … .
General youth development literature emphasises the importance of developing the social and emotional capabilities of young people in achieving positive life outcomes and demonstrates the strong relationship between sustained personal and social development and such outcomes (McNeil et al., 2012).Social and emotional capabilities are not static traits, they can be learned and developed with appropriate support (McNeil et al., 2012).
Services for children and adolescents with harmful sexual behaviour should therefore focus on sustained personal and social development to improve wellbeing and future success.Services should focus on building on young people's strengths and provide opportunities for their educational achievement.As well as providing general and practical assistance around employment, health, and housing it is necessary to provide particular support to develop their self-worth and confidence and to engage in healthy relationships like other children and adolescents.Doing so will increase the likelihood of successful outcomes in their lives beyond solely reducing sexual, or indeed non-sexual, recidivism.
Health promotion has been especially absent from standard models of treatment, yet the findings of this research demonstrate the implications of poor health on wider outcomes for these young people.Consequently, the promotion of good health and self-care should be an important developmental goal.

Involvement, involvement, involvement …
To develop the most effective interventions for children and adolescents with harmful sexual behaviour, the theory of developmental intentionality (Walker et al., 2005) might offer a helpful framework.It posits that young people are more likely to achieve desired developmental outcomes when they are actively engaged in their own learning and development.Such programmes aim not to shape young people but rather provide learning opportunities to help them shape themselves (Walker et al., 2005).When there is a good fit between young people and the intentional supports and opportunities they take part in, engagement is high, and the chances of positive outcomes are greatly improved (Walker, 2006).To achieve this, attention to long-term developmental outcomes for adolescents needs to permeate every aspect of programmes and professional practice must be anchored in the ethos of youth development (Walker et al., 2005).Just as treatment goals in strengths-based adult sex offender treatment are agreed between offender and therapist to increase motivation and engagement, so too are opportunities tailored to individual needs and wishes in youth programmes designed with developmental intentionality principles.

Relationship, relationship, relationship …
Establishing and maintaining consistent and supportive working relationships between young people and service staff during interventions and maintaining ongoing professional support thereafter is important.Potentially a similar system to that of circles of support and accountability (used in the community to support adults who have perpetrated sexual offences) may provide important ongoing support and assistance to those with problematic sexual behaviour in childhood as they enter young adulthood and navigate the challenges this can bring.Services also need to educate and work with children and young people to help prepare them for the specific challenge of disclosing their past behaviour to others, supporting them to make good choices around who to tell, when and in what circumstances, as well as how to best share that information.

Focusing unrelentingly on the future …
In addition to the factors identified as relevant for obtaining successful outcomes for these young people, participants' reflections on the services they received as children and young people also provide helpful insight into how practice could be improved to better support children and young people with harmful sexual behaviours.It was clear that the quality and nature of their relationships with professionals was viewed as central to the future direction of participants' lives and that good interventions were also felt to be a key factor in helping change their thinking and behaviour.Services need to help children look forward and plan for their futures, both practically and emotionally, and build resilience as much as looking backwards to understand their past behaviour and its harmful nature.

Supporting the development of intimate partner relationships …
To date the field of treatment for children and youth with harmful sexual behaviour has struggled with addressing intimate partner relationships during adolescence.Although a modest amount of sex education in the therapy room is often part of "treatment", limits are routinely placed on young people developing actual relationships due to concerns about the perceived risks they may present to intimate partners.Controls are often implemented on the engagement of these young people in the kinds of activities that would typically allow the development of such intimate relationships.However, in doing so the result is that these young people often become adults with almost no experience of sex other than that of their own abusive behaviours in childhood and their own victimisation experiences.In other words, in many cases we are not helping to prepare young people well enough to foster the interpersonal skills required for the successful relationships which have been shown to be so important in the achievement of successful outcomes in adulthood.Most adults with successful intimate partner relationships do not just arrive in adulthood and then miraculously have these skills; they negotiate and practice relationships and sex during adolescence.Without affording young people who have displayed harmful sexual behaviour the opportunity and support to develop healthy intimate relationships as they mature, we are hindering their chances to obtain intimacy with others, the very foundation of many other successful outcomes in adulthood.

Mentoring young parents …
A similar issue arises regarding parenthood.This is often seen as an area of risk, whereas for many adults in this study having children provided a key developmental turning point, addressing a need for many to develop a positive identity, one separate from their abusive histories.Services need to consider how best to support this and prepare young people for parenthood, potentially by mentoring or parenting programmes for young parents.

Limitations of the study
The strengths of the study are that it involves a relatively large qualitative sample which is reflective of a much larger population from which it is drawn.It also represents one of the first attempts internationally to look at the long-term developmental outcomes for those with problematic sexual behaviour as children and young people.In contrast to other studies in this arena, the sample is not restricted to participants drawn from forensic or incarcerated populations.Nonetheless, there are some limitations to the research.
First, the retrospective nature of the data is a limitation.Given that the data were collected as part of a follow-up study, they do not represent current cases open to the services involved and this may limit how applicable the findings are to the range of children now being referred because of their sexual behaviours.
Second, while our follow-up period of between ten and twenty years since referral is significantly longer than most previous attempts to follow up this population, the majority of our sample was at a broadly similar life stage at follow-up, i.e. typically in their mid-to-late twenties and into their thirties.It may be that some of the processes and issues highlighted in this study either remit or are exacerbated as a function of time.For example, developmentally, it can be hypothesised that the stability of intimate partner relationships typically increases as people progress through their teens, into their twenties and beyond that into their thirties, etc.Although in the present study we did not find that length of time from referral to follow-up distinguished the sample in terms of success of life course outcomes, the very point of developmentally focused practice is to be sensitive to the whole life course.This highlights, of course, the need for further research that goes beyond the developmental periods captured in this study.
A third limitation is that, despite the relatively large sample size for a qualitative study, participants essentially self-selected to take part, therefore the accounts of those willing to take part may not reflect the full range of circumstances and outcomes experienced by young adults with problematic behaviours in childhood.
Fourth, we also recognise that our own positions, as in any qualitative research, will be influenced by our pre-existing understandings, values and experiences and, in this case, this includes our previous work which examines these issues in a holistic way.

Conclusion
This article has examined the long-term developmental outcomes of a sample of young adults who demonstrated harmful sexual behaviour as children and has provided empirical support for a move away from a traditional risk management focus to more developmental approaches in work with children and adolescents with harmful sexual behaviours.The findings indicate the need for services to take a holistic and developmental approach in working with such clients, focusing actively on wider individual, relational and socio-environmental factors associated with more resilient life outcomes.Taking a developmental approach will not only help reduce the likelihood of further sexual and non-sexual offending but, importantly, will assist those with problematic and harmful sexual behaviour in their past to go on to lead more pro-social, fulfilling, and meaningful lives.

Table 1 .
Factors impacting life course outcomes for participants who displayed harmful sexual behaviours in childhood.