Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

There has been much interest in sexual behaviour, sexual deviance, and the treatment of sexual offenders for more than a century (for reviews see Laws & Marshall, 2003; Marshall & Laws, 2003, and Yates, 2002). Researchers, clinicians, and philosophers have long been intrigued by human sexual behaviour and have offered various theories regarding the origins of, and treatment for, sexual deviance. Indeed, perspectives have been based in psychodynamic theory, behavioural theory, and cognitive-behavioural theory, among many others (Laws & Marshall, 2003; Marshall & Laws, 2003; Yates, 2002, 2003). Moreover, numerous approaches to the treatment of sexual deviance have been proposed and implemented, most without empirical support at the time of implementation. For example, early treatment focussed on medical or pharmacological interventions which, while these seemed promising, were not based on research at the time of implementation. Early behavioural approaches that focussed on extinguishing deviant sexual arousal were not very effective, as these were based on the assumption that addressing deviant sexual arousal was sufficient as a complete intervention on its own (Marshall & Laws, 2003; Yates, 2002). Similarly, early perspectives that viewed sexually deviant behaviour as being based in anger or lack of social skills did not demonstrate anticipated results in terms of outcome.

Over time, cognition came to be recognised as important in understanding sexual deviance and in the treatment of sexual offending, and more comprehensive cognitive-behavioural/social learning approaches were adopted. More recently, the relapse prevention approach, which was considered the most respected intervention with sexual offenders, (cf. Laws, 1989, 2003), has been discredited due to a number of major shortcomings, not the least of which is its lack of demonstrated effectiveness (see below; Hanson, 1996, 2000; Laws, 2003; Yates, 2005, 2007; Yates & Kingston, 2005). However, although its use in treatment is diminishing, many adherents continue to use this model (McGrath, Cumming, Burchard, Zeoli, & Ellerby, 2010) because of its appeal to clinicians (Laws, 2003; Yates, 2007).

Currently, a multimodal/multiple component model of sexual offender treatment is the norm in most jurisdictions, although again, research support is equivocal. Various researchers have attempted classification systems of sexual offenders (e.g. Knight & Prentky, 1990) and of the sexual offence process (e.g. Ward & Hudson, 1998). While some of this research has proved informative with respect to risk and prevention of reoffending (e.g. Kingston, Yates, & Firestone, 2012; Kingston, Yates, & Olver, 2014), most has not. Newer models such as the good lives model approach (Ward & Stewart, 2003) that have been proposed have not yet been demonstrated in research to influence the ultimate outcome of reduced recidivism and victimisation, despite having been in existence for some time. Furthermore, developments in various jurisdictions, most notably the United States, such as the use of polygraphy, restrictions related to residency, and the containment approach, have similarly not been shown to be effective, despite substantial human and financial investment.

In this chapter, we review and comment upon current approaches to the treatment of sexual offenders. We offer a critical analysis and commentary on current approaches to intervention with sexual offenders, the effectiveness of these approaches, and recommendations for future directions.

Sexual Offender Treatment in Past Practice: Did We Go Wrong Somewhere?

A full review of the history of sexual offender treatment is beyond the scope of this chapter. As such, the interested reader is referred to Laws and Marshall (2003), Marshall and Laws (2003), and the numerous texts available on this topic.

Research and theory regarding the basis of sexual deviance and the best approach to the treatment of sexual offenders is typically described as having evolved considerably over the last 30–40 years. However, when examining the research literature, we submit that this cannot accurately be described in either research or practice as an evolution—defined as “a process of change in a certain direction (i.e. unfolding) and “a process of continuous change from a lower, simpler, or worse to a higher, more complex, or better state (i.e. growth; Merriam Webster, 2014). Instead, we argue that a variety of different methods of implementing treatment, based typically on the dominant philosophies, models of behaviour, and/or political influences of the time, have each been attempted as methods of intervention in a relatively random, ad hoc manner and/or with a narrow, unidimensional focus on one aspect of sexual offending behaviour. For example, both behavioural interventions and pharmacotherapy were intended to achieve such outcomes via the control of sexual arousal, and these approaches initially showed promise and achieved some of their desired results (Laws & Marshall, 2003; Yates, 2002, 2003). However, it was soon realised that behavioural approaches and pharmacological interventions targeting sexual arousal alone were insufficient as these did not address cognitive and emotional aspects of sexual offending and because deviant sexual interests are present in only a minority of offenders.

Later, with the advent of a feminist perspective on sexual violence, it was thought that sexual offenders were motivated by anger towards women and/or the sociological phenomenon of rape resulting from systemic male privilege (i.e. patriarchy) and women’s inequality (see Yates, 1996 for a review). Manuals targeting anger were developed and applied as a model of treatment. However, research and clinical practice is lacking in this area but suggests that the broader target of sexual, emotional, and behavioural self-regulation is more appropriate. In addition, sociological approaches relating to the patriarchy of sexual aggression have not been validated or demonstrated to influence recidivism, and all that remains of this model in actual practice are treatment exercises that attempt to promote understanding and challenging of “rape myths” (Burt, 1980) or, as these are known currently, cognitive distortions. Many other similar examples exist in the treatment literature and upon critical examination of the various treatment programmes available around the world.

One approach to sexual offender treatment where this phenomenon is perhaps most evident is the adoption of the relapse prevention (RP) approach . Originally developed within a medical model to assist alcoholic patients to maintain abstinence following treatment for alcohol addiction (Marlatt, 1982; Marlatt & Gordon, 1985), the RP model assumes that individuals are underregulated with respect to problem behaviours and that they lack adequate coping skills to control behaviour. Treatment within this approach is based on assisting clients to develop an understanding of those situations which place the individual at risk for recurrence of the problem behaviours, developing strategies to avoid these situations, and instilling skills and “adaptive” mechanisms to cope with high-risk situations. Within RP, clients are not viewed as self-directed and are assumed to be continually attempting to abstain from the problem behaviour, to set themselves up to encounter situations which will inevitably lead to failure, and to subsequently experience negative emotional states associated with this failure as a result of deficits in the ability to cope with life events, thereby leading to relapse. Despite this approach not representing the dynamics of sexual offending, its problematic focus on avoidance goals, constructs and methods that are not applicable to many sexual offenders, and lack of empirical support, RP continues to be widely implemented in current interventions. One such example is California’s Sex Offender Treatment and Evaluation Project (SOTEP ; Marques, Wideranders, Day, Nelson, & van Ommeren, 2005), a programme that was in operation at Atascadero State Hospital between 1985 and 1995 and was subjected to a rigourous outcome evaluation (see below).

To return to the question that is the title of this section, “did we go wrong somewhere?”, we submit that the answer is “probably (although not intentionally)”. In other words, the manner by which science advances in any field is through trial and error of theoretically based good ideas—methods, approaches, and models that show promise at some level and that are then empirically tested for their effectiveness in practice. So, while we are not prepared to state that earlier efforts were incorrect, we submit that the problem lies in the adherence to models or approaches to intervention that failed empirical testing and the continued application of these approaches in current day intervention. In brief, we have failed to learn lessons from our previous efforts and continue to make the same mistakes. As we discuss below, following a review of the extant treatment outcome literature, this continues to the present day.

What Does the Research Tell Us?

The utility of sexual offender treatment is a contentious issue, and there is considerable debate about how best to evaluate treatment programmes and how effective programmes are in reducing risk for sexual offending (Marshall, Marshall, Serran, & O’Brien, 2011; Rice & Harris, 2003; Seto, 2005).

In one of the first, large-scale narrative reviews regarding the efficacy of sexual offender treatment, Furby, Weinrott, and Blackshaw (1989) examined 42 studies of treated and untreated sexual offenders and concluded that sexual offender treatment has no demonstrable impact on sexual offender recidivism. Researchers have since disputed Furby et al.’s (1989) conclusion and have identified a number of problems with the review (Marshall & Pithers, 1994; Yates, 2002). For example, the interventions were conducted prior to 1980 and, as such, failed to meet contemporary standards of effective intervention. Additionally, very few of the selected studies compared treated and untreated offenders, and, therefore, differences between the studies on issues such as length of follow-up, sample size, and pretreatment levels of risk to reoffend made it difficult to draw firm conclusions (Yates, 2002).

Given the methodological problems inherent in qualitative narrative reviews, several meta-analyses have been conducted to better determine the cumulative effect of treatment outcome studies. Two early meta-analyses (Gallagher, Wilson, Hirschfield, Coggeshall, & MacKenzie, 1999; Hall, 1995) were conducted that examined treated and untreated sexual offenders. Both quantitative reviews reported a positive but small effect of treatment, and, importantly, those certain types of treatment (e.g. cognitive-behavioural) were superior to strictly behavioural interventions. The results of these early quantitative reviews were limited, however, as selected studies incorporated significant bias (e.g. comparing treatment completers versus treatment dropouts). The inclusion of treatment dropouts in the control condition is problematic, and this likely increases this group’s recidivism rates, given that such individuals likely possess characteristics related to risk and recidivism. Not surprisingly, when these biased studies were removed from subsequent analyses, the treatment effect was no longer significant (see Rice & Harris, 2003).

In 1997, the collaborative outcome data project committee was established to organise the existing outcome literature and report on treatment effectiveness. Hanson et al. (2002) conducted a meta-analysis of 43 published and unpublished English language studies on psychosocial treatments comprising 9454 sexual offenders. Results indicated that treated sexual offenders had lower sexual recidivism rates (12.3 %) than sexual offenders in comparison conditions (16.8 %). Studies with the strongest methodological design (i.e. random assignment) showed no effect of treatment, whereas studies described as incidental assignment (i.e. studies with no a priori reason to suspect group differences between treated and untreated sexual offenders) showed a positive effect of treatment. Rice and Harris (2003) have criticised the selection and categorisation of the studies in the Hanson et al. (2002) meta-analysis, and they noted problems ranging from cohort effects to treatment versus control group comparability. After a reanalysis of six studies that met stricter methodological criteria for promoting group comparability, Rice and Harris (2003) concluded that there was no positive effect for treatment and, in fact, that treated sexual offenders had a higher recidivism rate than the comparison group, although this difference was not significant.

Lösel and Schmucker (2005) have since provided the largest and most comprehensive meta-analytic review of sexual offender treatment. Their review consisted of 80 comparisons derived from 69 studies, comprising 22,181 sexual offenders. Overall, results showed that treatment reduced recidivism rates compared to control conditions, but again studies employing more methodologically rigorous designs revealed no group differences in recidivism. Most recently, Långström et al. (2013) conducted a systematic review of psychological, educational, and pharmacological interventions intended to reduce recidivism among sexual offenders against children. Among the original 167 articles selected for review, only eight met minimal methodological quality representing low or moderate risk of bias (three randomised control trials (RCTs) and five controlled observational studies). Results demonstrated some minimal evidence for multisystemic therapy, an intensive approach that targets environmental systems such as schools and families (Borduin et al., 1995), for adolescent sexual offenders (based on one RCT). With regard to adult sexual offenders against children, the authors noted that there was insufficient evidence for medical and psychological interventions from which to draw firm conclusions.

A number of Cochrane reviews have been conducted that focus specifically on RCTs. Briefly, the Cochrane collaboration comprises a number of centres and specific specialities, which conducts systematic reviews on a number of topics and provides access to such reviews within a comprehensive database. The Cochrane collaboration restricts its evidence included in their reviews to RCTs. Most recently, Dennis et al. (2012) conducted a comprehensive search of articles that were published up until 2010. Ten studies were ultimately selected representing 944 sexual offenders. Five studies involved cognitive-behavioural-type interventions, four described behavioural interventions, and one involved psychodynamic treatment. The authors concluded that the evidence for the effectiveness of sexual offender treatment is weak and they advocated for additional RCTs, emphasising methodologies that minimise risk of bias.

In summary, results from meta-analytic reviews of sexual offender treatment have failed to provide strong empirical support for positive treatment outcome. Moreover, effect sizes are generally small but meaningful (see Cohen, 1992), particularly when compared against the effect sizes produced among treatment options for other medical and behavioural disorders (Marshall, 2006). Although there has been some debate about the practical and procedural utility of RCTs (Marshall, 2006), this approach is considered the “gold standard” in programme evaluation, and studies employing this approach have not shown treatment to be associated with reduced recidivism rates. Results from California’s Sex Offender Treatment and Evaluation Project (Marques et al., 2005), a programme evaluation in which an RCT was used, failed to show a treatment effect and have often been cited as strong evidence against treatment effectiveness. However, closer inspection of this particular programme demonstrated that, because of the date of implementation, few known criminogenic factors were targeted in treatment. Moreover, finer-grained analyses suggested that there was a certain proportion of offenders (i.e. those who “got it”) who may show a greater treatment response than other offenders within the programme. Such findings suggest that perhaps treatment can be effective, but that it must be designed and implemented based on the literature, and continued effort needs to be placed on rigorous evaluations. Specific criteria for establishing evidence-based therapeutic approaches have been provided, which focus on at least two evaluations incorporating appropriate control groups (see Chambless et al., 1998). Such criteria can assist in the implementation and interpretation of outcome evaluations.

Sexual Offender Treatment in Current Practice: Are We Doing Something Wrong?

In response to this question, we submit that, while some progress has been made, the answer is an equivocal “yes”.

The most effective approach to sexual offender treatment at present is the risk/need/responsivity (RNR) approach (Andrews & Bonta, 2010), shown to be effective in reducing recidivism among many offender groups, including women, youth, violent offenders, and sexual offenders (Andrews et al., 1990; Dowden & Andrews, 1999a, 1999b, 2000, 2003; Hanson, Bourgon, Helmus, & Hodgson, 2009). This model utilises evidence-based methods to tailor treatment to individual offenders who vary in the risk they pose to reoffend, the factors that lead to offending behaviour, and clients’ capacity to respond to our interventions (and our capacity to respond to their needs and individual particularities). It is based on a comprehensive theory based on an empirically based understanding of the reasons for which individuals engage in criminal behaviour (the psychology of criminal conduct; Andrews & Bonta, 2010). The RNR approach demonstrates that treatment is most effective when programmes: (1) target offenders who are at moderate to high risk to reoffend (i.e. the risk principle), (2) target changeable risk factors that are empirically linked to recidivism (i.e. the need principle), and (3) vary methods of delivery in such a manner as to ensure maximum benefit for individual offenders depending on their own circumstances and capabilities and doing so using a cognitive-behavioural/social learning approach (i.e. the responsivity principle). While the RNR also includes the principle of professional discretion, we submit that it is this element of the model, for which research support is lacking, to which clinicians most often adhere.

Regrettably, in spite of decades of empirical support, the RNR approach has not been widely adopted in the treatment of sexual offenders. Indeed, research has shown that adherence to RNR principles produces larger and more positive treatment effects among violent offenders (Dowden & Andrews, 2000) and sexual offenders (Hanson et al., 2009) compared to programmes that do not adhere to this model of treatment. Research has also found that delivery of appropriate service (i.e. that which adheres to the RNR) is not more expensive than inappropriate service and is cost-effective. For example, costs for appropriate service for a 1 % reduction in recidivism range from $0.25 to $9.40, compared to the costs of inappropriate service and traditional punishment, costing $19.67 and $40.43, respectively, for a 1 % reduction in recidivism (Romani, Morgan, Gross, & McDonald, 2012). While some jurisdictions explicitly adhere to this model and include it as a matter of policy (e.g. the Correctional Service of Canada), few other organisations and/or jurisdictions utilise this approach. For example, as McGrath et al. (2010) illustrated in their comprehensive survey of sexual offender treatment in North America, use of this model with adult sexual offenders ranges from 0 to 37 % of organisations providing treatment.Footnote 1

The reluctance to implement treatment using such an evidence-based approach is perplexing at best. While criticisms of the RNR model suggest that it is insufficient to engage offenders and to motivate them to explore changing their behaviour (e.g. Ward & Stewart, 2003), these criticisms ignore the substantial body of evidence and the fact that the majority of treatment programmes utilising this model also explicitly take a motivational enhancement approach. (In fact, some clinicians and researchers have mistaken the responsivity construct as constituting wholly or predominantly motivation when it actually encompasses many different internal and external characteristics and circumstances.) In the authors’ experience (and according to research), the RNR model applies to treatment regardless of setting (e.g. prison, residential treatment, in the community) and can easily be broadened to include motivational, self-regulation, and positive psychology approaches (Yates, Prescott, & Ward, 2010; Yates & Ward, 2008). As noted earlier, the RNR model has been demonstrated to result in reduced recidivism and can be enhanced via effective therapeutic practice. In fact, studies have shown that adherence to the risk, need, and responsivity principles has been associated with a 10, 19, and 23 % difference in recidivism, respectively (Bonta & Andrews, 2007). Among sexual offenders specifically, research indicates that adherence to the principles of the RNR results in incremental effectiveness with adherence to none, one, two, or all three principles (odds ratios of 0.21, 0.63, 0.64, and 1.17, respectively) (Hanson et al., 2009). Given the above, it is perplexing at best how this approach, despite its long-standing existence and empirically demonstrated effectiveness, is absent from the majority of treatment programmes.

Compounding this problem is the continued reliance on implementation approaches that are not supported by research. For example, research has consistently shown that treatment of low-risk offenders (those who are assessed using actuarial measures and determined to be of low risk to reoffend and demonstrating few dynamic risk factors) is, at best, ineffective and, at worst, can have the iatrogenic effect of increasing risk and resulting in increased recidivism. Conversely, appropriate treatment provided to higher-risk offenders has been shown to reduce recidivism while also finding that many programmes continue to provide intensive treatment services to low-risk offenders (Andrews & Bonta, 2010; Hanson & Yates, 2013; Lowenkamp & Latessa, 2002, 2004; Lowenkamp, Latessa, & Holsinger, 2006), yet intensive treatment of these offenders continues at present. Furthermore, substantial sanctions, such as sex offender registries and notification, are utilised with this group of offenders (see Chapter x this volume). Continuing to apply intensive treatment and sanctions, driven by ideology, philosophy, and political factors, can only serve to increase the risk of recidivism and future victimisation. In attempting to understand the rationale for this phenomenon, the authors have found that this appears to be fourfold: (1) denial on the part of clinicians that low-risk offenders actually exist; (2) the belief that low-risk offenders are actually undetected higher-risk offenders; (3) clinicians’ and organisations’ personal philosophies, such as the belief that all sexual offenders require treatment regardless of risk; and (4) political influence and attendant jurisdictional policies that dictate treatment requirements regardless of research findings (e.g. all offenders must be heavily sanctioned, treated, and managed). It is noted that there is also the potential for loss of livelihood in some cases—for example, in some jurisdictions, clinicians have noted that refusal to treat such offenders would result in financial penalties due to termination of contracts to treat offenders (which also represents a serious ethical concern).

Another approach to which clinicians adhere in spite of the absence of research support is, as noted above, the RP approach. In their survey, McGrath et al. (2010) found that this model continues to be used in as many as 85 % of North American treatment programmes—a figure we find disturbing given research findings that this approach does not address the dynamics of sexual offending and an absence of research demonstrating its effectiveness.

Taken together, the relative absence of adherence to the RNR model, the continued treatment of low-risk offenders, and the high rates of utilisation of RP suggest a strong reluctance, or perhaps an aversion, to applying evidence-based practices to the treatment of sexual offenders.

In addition to the above, various practices are being adopted on a regular basis in the field of sexual offender treatment that are not based on a theory of the causes of sexual offending, research pertaining to effective intervention practices, or research demonstrating effectiveness. As indicated above, while science evolves on the basis of good ideas that are subject to investigation and evaluation, some of these more recent practices (some of which are exceptionally intrusive, not to mention expensive) have not been evaluated or continue to be implemented in spite of early research indicating a lack of effectiveness. Notably, many of these practices, such as the containment approach (a multi-agency collaborative approach that explicitly takes a victim-centred philosophy, that aims to exercise control of risk in the community using treatment, probation, and polygraph, and that does so regardless of the risk level of the offender; English, 1998), sex offender notification and public registries, civil commitment, and the use of the polygraph in treatment, are not driven by research or even theory pertaining to the aetiology of sexual offending behaviour or basic principles of effective therapeutic intervention. In fact, investments continue to be made in these methods, the likely outcome of which is increased recidivism, reduced community safety, and the diversion of scarce treatment resources to these nontherapeutic activities.

What is perhaps most disturbing, in the authors’ experience, is the acceptance by organisations and clinicians of such methods as valid clinical practice, with the attendant risk that any hope for the establishment of a therapeutic relationship with clients will be absent or impossible. To provide an example, the use of the polygraph to establish a full sexual history (the value of which is undemonstrated) and/or to evaluate the implementation of therapeutic tasks in the community (i.e. “maintenance” polygraphs) is becoming well entrenched in some jurisdictions, most notably the United States. Recent research, however, does not support the effectiveness of this tool in reducing recidivism and victimisation (Meijer, Verschuere, Merckelbach, & Crombez, 2008; Rosky, 2012) and runs the risk of eventually leading to a deterioration in clinical skills necessary to gather information from clients and to establish an effective working alliance, which is shown in research for various problems to account for a substantial amount of the variance in positive outcome (e.g. Marshall et al., 2003; Witte, Gu, Nicholaichuck, & Wong, 2001). While it is acknowledged that, in some jurisdictions, the use of tools such as polygraphy and containment is a legislative or other requirement for sexual offenders, the extent to which clinicians have adopted and embraced their use as clinical tools is disturbing. It is further acknowledged that this is a problem both at the individual clinician level and at the organisational/jurisdictional level, and that organisational or political influence can have a substantial undue impact on clinical practice. It is the authors’ hope that clinicians recognise these practices and are provided with appropriate training and opportunity to separate legislative or jurisdictional requirements from clinical practice in the delivery of treatment.

In a related vein, several newer approaches to sexual offender treatment have been proposed, as examples, trauma-informed approaches such as eye movement desensitisation and reprocessing (EMDR; e.g. Ricci & Clayton, 2008; Ricci, Clayton, & Shapiro, 2006), a victim-centred approach (e.g. English, 1998), the self-regulation model (an adaptation of self-regulation theory [Baumeister & Vohs, 2004] to the sexual offence process; Ward & Hudson, 1998), and the good lives model (Ward & Stewart, 2003). In the context of this analysis, it is important to recognise that these approaches have not yet been demonstrated in research to reduce recidivism and victimisation. Yet these have now been in existence for a sufficient period of time that research to assess their impact and effectiveness on ultimate outcome (i.e. recidivism) should have been conducted but has not likely due to reluctance to change on the part of clinicians and organisations. While some research has been conducted on the validity and impact with respect to intermediate treatment targets, such as motivation (e.g. Yates & Kingston, 2006; Yates, Simons, Kingston, & Tyler, 2009), and the extent to which clinicians like the model (in the case of the good lives model; Ware & Bright, 2008), the time has come for the ultimate test of effectiveness (i.e. reduced recidivism). In light of the current status, the authors implore caution in the application and utilisation of these approaches, lest we be destined to repeat the past.

In conclusion, regardless of the specific tool or method or the personal, philosophical, or political approach to sexual offenders, we view as essential to the effectiveness of current and future practice the ability to critically evaluate models and approaches and their application, to do so with a sound understanding of the research basis of each, and to resist practices that will result in the degradation of clinical skills and effective intervention.

Content and Process of Treatment: Have We Got It Right?

Our answer to this question is an unequivocal “no”. Treatment of sexual offenders, as indicated above, continues to adhere to models and approaches (new and old) that have not been demonstrated to be effective or that have been demonstrated to be ineffective and to ignore approaches and models that are effective. Nowhere is this more evident than in the examination of the specific content and process of current treatment programmes.

To begin, in many jurisdictions, treatment is delivered entirely without structure or an overarching model and approach based on research and without quality review of adherence to the approach and so of unknown and questionable fidelity, resulting in a lack of information pertaining to content and process of treatment that is implemented with clients. It is no surprise that outcome results are inconclusive and that research is inconsistent given the current state of the field. In this section, we explore a few specific examples as illustrations.

Regrettably, in spite of research to the contrary, many theorists and clinicians continue to insist that treatment manuals create restrictions on clinical practice (Levenson & Prescott, 2013; Marshall, 2009; Gannon & Ward, 2014). It is perplexing and disturbing that our discipline discounts research indicating that the most effective correctional programmes are those that adhere to specific standards, including the use of manuals, which creates consistency, ensures that treatment is evidence based, and ensures treatment integrity and fidelity (Gendreau & Goggin, 1996, 1997; Gendreau, Little, & Goggin, 1996; Hanson et al., 2009; Hanson & Yates, 2004). Those who adhere to this view argue that adherence to treatment manuals is incompatible with the development of an effective therapeutic alliance with clients, a well-established element of treatment in general (e.g. Marshall, Burton, & Marshall, 2013), although research support for this assumption is absent. The notion that structure and content are incompatible with therapeutic process, with the attendant conclusion that content and process are dichotomous constructs that cannot be reconciled, is indeed perplexing. When this argument is presented, reference is ironically made to the work of Andrews and Bonta (2010), in which it is stated that intervention needs to be individualised. What is missing, however, is that this tailoring of treatment must be based on risk, need, and responsivity, as well as structure. However, the reference to the structure of this model, for which there is extensive research support, that involves cognitive-behavioural intervention to target specific criminogenic needs empirically demonstrated to be linked to recidivism, is typically ignored in this argument.

One major problem with current sexual offender treatment is that it is far too long in duration. This appears to be based on the belief that “more is better”, as well a negative effect of the amount of time in treatment that is dedicated to factors not demonstrated to be empirically related to sexual offending or recidivism and to the use of extensive exercises of questionable value to treatment (e.g. autobiographies, extensive analyses of the offence process, victim empathy or “clarification” letters, and overcoming denial/minimisation). Admittedly, there is little research evidence pertaining to the effective dosage of treatment required for offenders presenting with varying levels of risk to reoffend and various criminogenic needs or dynamic risk factors. Regarding dosage, recommendations have been made (Bourgon & Armstrong, 2005; Hanson & Yates, 2013), yet treatment in most jurisdictions does not adhere to such risk-based recommendations.

Research has clearly delineated those factors known to be associated with increased risk for recidivism, such as intimacy deficits, sexual and general self-regulation, and the presence of sexual deviance/preference (Hanson, Harris, Scott, & Helmus, 2007). What we are only beginning to learn is how to weight these various factors and their relationship to static risk factors. Regardless and in spite of perplexing academic criticism that these factors have their basis in research (e.g. Gannon & Ward, 2014), factors that place offenders at risk to reoffend, and that can be targeted in treatment, are known and must be targeted if treatment is to be effective. Yet many programmes continue to target treatment goals that are unrelated to recidivism reduction, such as denial, self-esteem, personal distress, empathy, and individual accountability (Hanson & Morton-Bourgon, 2005; Yates, 2009). In the authors’ opinion, this continues as a result of individual, societal, and legal values, which emphasise such constructs as remorse and taking responsibility for one’s actions. For example, a fundamental premise of the criminal justice system is to hold individuals accountable for their actions, and it is a societal expectation that one experiences remorse when harm to others has been caused. Necessarily, this societal expectation influences organisations and individuals, including clinicians delivering sexual offender treatment. However, while these are laudable goals and are an essential element of punishment (i.e. sentencing) within the criminal justice system, research either does not support their inclusion as treatment targets that will reduce recidivism or the considerable amount of time taken in treatment to address these issues. A similar problem exists with the currently emerging “victim-centred” approach to sexual offender treatment. While it is inarguable that victims’ experiences are important and deserving of attention and intervention, their application in the treatment of offenders (e.g. in the form of understanding victims’ perspectives and making amends) is undemonstrated. In addition, because the focus is to raise awareness of harm caused (i.e. empathy), this approach is unlikely to influence treatment outcome, thus representing another instance in which treatment continues to absorb practices that are not empirically supported. Clinicians and organisations need to be able to differentiate between the goals of the legal system, societal expectations, and public policy and what works in sexual offender treatment in order to reduce risk and promote community safety.

The above also leads to an artificial dichotomy between protection of the public via reduced recidivism and victimisation and enhancing the psychological and community well-being of the offender. Many treatment programmes and some newer treatment models focus on the well-being of the offender as an essential part of treatment. This is rightfully an important goal of human service providers in all fields—clinicians wish to reduce distress and enhance individuals’ lives. However, what is absent is the problematisation of this approach within criminal justice systems and its potential impact on the fundamental human rights and liberty of citizens (which includes offenders). In brief, as a field we need to examine the fundamental ethical violation of incarcerating individuals or applying (sometimes long-term) sanctions such that we may make individuals’ lives better. We cannot imagine a profession outside the criminal justice system that would condone restrictions on liberty and freedom in order to improve well-being in the absence of evidence of risk to oneself or others. Despite our legitimate desire to improve people’s lives, we do not believe this should be a condition of treatment or a requirement to retain or reacquire freedom in the absence of risk or its reduction, and we view this as unethical.

Much research has been done pertaining to effective therapist characteristics and therapeutic approaches that influence the outcome of treatment (Beech & Fordham, 1997; Marshall et al., 2002; Shingler & Mann, 2006; Yates, 2002, 2014; Yates et al., 2000). Andrews and Kiessling (1980) introduced several dimensions of effective correctional practice, termed core correctional practice, that were intended to promote treatment outcome in offender populations. Arguably, the most important principle was the quality of interpersonal relationships, which denotes the specific therapist characteristics that are associated with treatment success (Dowden & Andrews, 2004). Specific therapist characteristics that have been shown to enhance treatment effectiveness include demonstrating such features as empathy, respect, sincerity, confidence, and interest in the client. Being a prosocial model, being “firm but fair”, reinforcing and encouraging clients, creating opportunities for success, dealing appropriately and effectively with resistance, being appropriately challenging without being aggressively confrontational, and creating a secure treatment atmosphere all contribute to treatment outcome (Fernandez, 2006; Marshall et al., 2002). For example, research indicates that establishing a positive therapeutic relationship with the client accounts for a significant proportion of the variance in treatment outcome (Fernandez, Shingler, & Marshall, 2006; Hanson et al., 2009; Witte et al., 2001; Mann, Webster, Schofield, & Marshall, 2004; Marshall et al., 2003).

Behavioural rehearsal and practice, designed to inculcate new skills into individuals’ behavioural repertoires, are also essential elements of treatment yet are methods that are insufficiently utilised in current insight-based approaches (Fernandez et al., 2006; Yates et al., 2010). Similarly, using motivational enhancement techniques and creating a positive and safe treatment environment lead to improved compliance with treatment, progress, and enhanced motivation and prevent termination or dropout from treatment (Beech & Fordham, 1997; Kear-Colwell & Pollack, 1997; Marshall, Anderson, & Fernandez, 1999; Prescott, 2009).

Given that research clearly indicates that offenders who do not complete treatment reoffend at significantly higher rates than offenders who complete treatment (Hanson & Bussière, 1998; Hanson et al., 2002), it is essential that treatment is delivered in a manner that is motivating and engaging for clients. However, the authors are aware of few programmes or organisations that explicitly select or screen therapists for such characteristics and that train treatment providers in these effective techniques or that deliberately train service providers to respond to the characteristics of the offenders with whom they work (i.e. attending to responsivity). In fact, unlike other professional practice areas such as psychology, psychiatry, or social work, there are no universally standard professional practice requirements for therapists delivering sexual offender treatment, who range in training and experience from prison officers to probation officers to other professionals with various levels and type of education and training. While there is no research suggesting the superiority of one discipline over another in the delivery of treatment, practice requirements and training are highly variable across jurisdictions, clinicians are not typically preselected for the essential characteristics that enhance treatment success, supervision and quality review of treatment implementation is not consistently utilised, and training received by clinicians is often absent or inconsistent.

As a final note, we note the lack of tolerance for harm reduction in the field of sexual offender treatment, which has long been proposed as having potential as a measure of treatment outcome or success (Laws, 1996). While from a clinical perspective we as a field purport to advocate for the reduction of harm through intervention and treatment, in practice, the goal of sexual offender treatment appears to remain one of complete abstinence (while simultaneously holding the belief that sexual offending is a life-long problem from which one cannot recover). However, if we were working in the area of addictions or general mental health, we as clinicians would be at least minimally satisfied with a level of progress that reduced symptomatology and risk to self or others. If we were treating a patient suffering from depression, we would be satisfied with reducing active suicidal intent, even if the client remained feeling hopeless. In such a case, we would not restrict freedom and liberty (i.e. we would not [and could not legally] commit the patient while we continued to work with the client for depression in the absence of imminent risk of harm). In sexual offender treatment, we do not work this way. To provide a parallel example, sexual offender treatment aims to reduce the potential for sexual victimisation and harm. However, even if we effectively treat the client to manage deviant sexual arousal or to effectively manage risk, if such a client continues to present with deviant sexual fantasy or urges, we continue to incarcerate or otherwise restrict the offender, until he or she is able to demonstrate that such urges no longer exist—an outcome that is impossible to assess and determine definitively, in addition to being an unlikely occurrence.

Sexual Offender Treatment in the Future: Where Do We Go from Here?

In this chapter, we have reviewed current and previous processes of the development and implementation of sexual offender treatment. We have noted that various theories of sexual offending behaviour, models of intervention, and elements of specific practice have been developed in an ad hoc fashion based on good ideas at the time, which may or may not have had a basis in research and theory. We further argued that current approaches to the treatment of sexual offenders apply and retain elements of intervention that are undemonstrated in research to be effective or that have been demonstrated to be ineffective while omitting approaches known to be effective and that this state of affairs is unduly influenced by political, organisational, and personal bias regarding the objectives of sexual offender treatment and the manner in which it should be implemented. While research supporting the effectiveness of sexual offender treatment is equivocal at best, we nonetheless implore those responsible for the delivery of sexual offender treatment to attend to the research basis for the content and structure of treatment, its specific implementation and delivery methods, and the requirements for effective therapy, in order to maximise the probability of success via reduced recidivism and victimisation and increased community safety. We need to do so while continuously examining and challenging our own biases and those of our organisations, jurisdictions, and political circumstances while differentiating between ideology, legal requirements for practice, and political influence, so we can deliver empirically supported interventions in an ethical and effective manner.