Abstract
Despite a continued evolution of the field of sexual abuser treatment toward a distinct professional discipline with clinicians using an increasing variety of treatment approaches, there is no consensus regarding the strength of our various clinical interventions as evidence-based practices (EBPs). This article provides a brief history of the development and goals of EBPs in medicine and mental health, and a review of the earnest efforts of researchers within the field to establish treatment approaches with sexual abusers as evidence-based. An appraisal of the current status of EBP’s with sexual abusers is provided. Although there have been improvements in the methodological quality of treatment outcome research with sexual abusers, divergent opinions about treatment effectiveness remain, and the field has not yet agreed on a system or set of criteria for what constitutes “evidence.” We contend that clinical practice has been influenced as much, or more, by new paradigms that are intuitively meaningful and perceived as needed than it has by what has been determined to be scientifically efficacious. This trend and other processes in our field that seem to be slowing the development of EBPs with sexual abusers are discussed. Recommendations for conducting evidence-based reviews and moving the field of sexual abuser treatment toward the use of a true EBP model are provided.
The assessment, treatment, and community supervision of persons with a history of sexual offending has increasingly been established as a distinct field within corrections and mental health within the last 30 years. As such, there has been a corresponding increase in research, assessment tools, treatment approaches, manuals, books, client workbooks, conferences, and professional organizations related to this emerging field (Laws & Marshall, 2003; J. Levenson, 2014; Marshall & Laws, 2003). However, despite the growing use of evidence-based practices (EBPs) in medicine and mental health (Hamer & Collinson, 2014; Pilecki & McKay, 2016), the field of sexual abuser treatment has not yet provided its researchers, clinicians, policy makers, nor its clients with a clear identification of those practices that can be endorsed by the profession as evidence-based.
Although efforts in the field to establish an evidence base are earnest and ongoing, the speed at which treatment models have been advanced and quickly adopted in clinical practice has far outpaced the efforts of researchers to validate their effectiveness for use with persons that have sexually abused. As a result, the specific treatment interventions that a person with a history of sexual offending ultimately receives can vary widely from location to location, and between programs and clinicians (Marshall & Marshall, 2010; McGrath, Cumming, Burchard, Zeoli, & Ellerby, 2010). In addition, given that sexual abuser research is so remarkably diverse and covers such a wide variety of clinical activities, there is a clear need for systematically organizing and synthesizing the research as a means of establishing EBPs with this client population. The field possesses a sizable literature of empirical research findings and has professional bodies that can be used to help direct efforts toward establishing and implementing authentic EBPs relating to the assessment, treatment, and community supervision of individuals with a history of engaging in sexual abuse. Toward that end, we propose a series of positive and practical recommendations to further advance that process within our field.
EBPs: Background, Definition, and Goals
The term evidence-based practice first appears in the medical and mental health literature of the 1990s and, along with a variety of related terms and ideas (e.g., evidence-based treatment, empirically supported treatments [ESTs], etc.), developed out of the concept of evidence-based medicine (EBM; Sackett, Rosenberg, Gray, Haynes, & Richardson, 1996). Archibald Cochrane (1972) is typically cited as having influenced the development of EBM after his book Effectiveness and Efficiency: Random Reflections on Health Services criticized the absence of evidence behind many health care approaches and in which he advocated for the use of randomized trials to help identify and verify the effectiveness of health care interventions. Shortly following his death in 1988, the Cochrane Collaboration was established in his honor. This international organization consists of thousands of volunteers who are dedicated to conducting systematic reviews (referred to as “Cochrane Reviews”) of randomized trials on a variety of medical treatments.
The evolution of EBPs from EBM in the 1990s was a process that yielded two new features. First, rather than focusing solely on the field of medicine, EBPs became an approach to improving outcomes that was ultimately advanced in multiple fields, including mental health (Spring, 2007), education (Odom et al., 2005), management (Rousseau, 2006), nursing (Melnyk & Fineout-Overholt, 2005), and corrections (Lowenkamp, Holsinger, Robinson, & Cullen, 2012). Second, EBP is a broader concept than EBM, which not only looks at the empirical basis for determining the most effective treatment approaches but also considers important idiographic variables in treatment outcome, such as the expertise of the clinician and the preferences of the client.
In 2005, American Psychological Association (APA) President Ronald Levant appointed the APA Presidential Task Force on Evidence-Based Practice. The APA Task Force on EBPs has defined evidence-based practices as “. . . the integration of the best available research with clinical expertise in the context of patient characteristics, culture and preferences” (APA Task Force on Evidence Based Practice, 2006). These three components—empirical research, clinician expertise, and client characteristics—have been described as a three-legged stool that provide support to the implementation of EBPs. The APA definition of EBPs is widely accepted in the mental health field, and there is general agreement that the starting point for implementing EBPs is empirically validating treatment interventions and other clinical activities.
Ultimately, the goal of establishing EBPs in behavioral health is to apply them in clinical practice to improve client outcomes. The scope of their application with persons that have engaged in sexually abusive behavior could comprise a wide variety of clinical activities, including assessment and treatment intervention, as well as activities relating to community support and supervision, and even public policy issues.
The First Leg of the EBP Stool: Establishing “Evidence”
The unique methodological challenges associated with conducting research on persons with a history of sexual offending have been well documented, and a short list would include the heterogeneity of the abuser population, low base rates of recidivism, long follow-up periods while clients are “at risk” in the community, and questions about how to best measure treatment progress and/or outcome (Abracen & Looman, 2004; Barbaree, 1997; Becker, 1990; Beech et al., 2007; Grossman, Martis, & Fichtner, 1999; Hanson, Morton, & Harris, 2003). As such, there are strongly held and often divergent opinions about treatment effectiveness, the quality of the research, and whether the evidence base that currently exists supports our various clinical methods (Abracen & Looman, 2004; Duggan & Dennis, 2014; Yates & Kingston, 2016). Public attitudes about persons who sexually abuse pose an additional challenge in our efforts to conduct research and communicate our results. The widely held belief that treatment does not work and that all or most of these individuals will recidivate has a significant impact on political opinion, public policy, financial and administrative support for research projects, and acceptance of our research findings (J. S. Levenson, Brannon, Fortney, & Baker, 2007). Nonetheless, much of the disagreement about our interventions and outcomes actually emanates from within our field and relates to the amount of evidence deemed necessary to declare that treatment is effective (Duggan & Dennis, 2014).
What amount of research is sufficient to establish “evidence” in the development of EBPs with sexual abusers? Equally important, how is research quality assessed when establishing an evidence base? Although there are many sensible approaches that can be used to evaluate the quantity and quality of research support for a given clinical intervention or activity, there is no single method for doing so. While the APA Task Force on EBPs, in discussing the “best available research” component, acknowledges that multiple types of research designs, evidence, and methods can contribute to EBPs, they do not provide specific recommendations or methods for distilling research from potentially numerous sources to arrive at a conclusion about the level of support that should be afforded to a given intervention (APA, 2006).
Efforts to Date to Establish an Evidence Base With Sexual Abusers
There are, in fact, a variety of designs and methods that are being used to conduct research with sexual abusers in an attempt to establish an evidence base. Harkins and Beech (2007) provide an informative summary of the various research methods that can be used for examining the effectiveness of treatment for persons with a history of sexual offending, including the advantages and shortcomings of each approach. Some research methods have greater empirical power than others, and this hierarchy of research methods is often conceptualized and graphically represented as a pyramid (Berlin & Golub, 2014). In this conceptualization, the base of the pyramid is frequently represented by research methods such as case studies or cohort studies. The apex of the pyramid, represented by empirically stronger research methods, typically includes either the randomized controlled trial (RCT) or approaches that synthesize primary research such as the systematic review or meta-analysis. Although the concept of a research methods hierarchy is helpful for providing a broad understanding of the relative empirical strength of various research approaches, many factors can influence the strength, quality, and validity of research findings, even for those methods at the top of the pyramid.
RCTs with sexual abusers
The RCT is often referred to as the “gold standard” because of its strength as a research approach (APA, 2006; Sackett et al., 1996). To date, there have been approximately 20 RCTs conducted with adult sexual abusers, most of which have focused on treatment outcome. These RCTs have evaluated both psychological and pharmacological treatments. Many of them, however, were conducted more than 20 years ago, are known to have serious methodological shortcomings, and included treatment programming that would be considered of questionable quality by current standards. These studies have demonstrated generally mixed outcomes regarding the effectiveness of treatment for adult sexual abusers (Duggan, 2014). For example, a systematic review of 10 RCTs that evaluated psychological interventions with persons who sexually abuse found that only a few of these studies provided information on re-offense rates, usually because of short follow-up periods, and relied instead on other outcome measures such as changes in social skills and anger management (Dennis et al., 2012).
There has also been a small number of RCTs that have evaluated multi-systemic therapy (MST) with juvenile male sexual abusers. These have provided more consistent positive outcomes, although, to our knowledge, the results have not yet been subjected to a Cochrane Review (Borduin, Schaeffer, & Heiblum, 2009; Letourneau et al., 2009). In addition, Duwe (2018) conducted an RCT that compared recidivism rates between individuals who participated in the community support program Minnesota Circles of Support and Accountability (MnCOSA) and those in a control group. Although the total sample size was only 100, the follow-up periods in the community averaged 6 years. Those individuals who participated in the MnCOSA program had significantly lower recidivism rates for new sex crimes and for four additional measures of general recidivism.
The time and financial costs often associated with conducting RCTs are well known (Duggan, 2014), and Marshall and Marshall (2007, 2010) have argued that other serious ethical, practical, and scientific problems make it difficult to use the RCT design in sexual abuser research. For example, the RCT design can produce a circumstance that denies treatment to those in the “no treatment” control group, potentially creating risk for sexual abuse to occur in the community. In addition, the Marshalls point to the difficulties of meeting RCT methodological demands for randomization, sample matching, and the fact that many individuals with a history of sexual offending cannot freely “volunteer” for treatment because of criminal justice mandates. In the case of juveniles with sexual behavior problems, resistance by the juvenile justice system to the use of random assignment to control groups has also been noted by Borduin, Henggeler, Blaske, and Stein (1990).
Others in our field have provided an equally cogent case for reasons that RCT designs in sexual abuser treatment outcome research can and should be used. Duggan and Dennis (2014) provide a thoughtful, item-by-item, response to each of the concerns raised by Marshall and Marshall, including the concern relating to recidivism risk for persons in a no-treatment control group. They point out that the use of ineffective, or perhaps harmful, treatment approaches creates an equally unethical circumstance.
As a means of overcoming some of the ethical and practical objections to RCTs, Seto and colleagues have suggested using random assignment to alternative treatment components, particularly where one component has shown, or is thought to have, a treatment effect (Seto et al., 2008). Also, rather than relying on recidivism as the only outcome measure, Hanson (1997) has suggested measuring in-treatment changes on dynamic risk factors that are known to be correlated with recidivism. This could potentially save time and reduce the costs associated with the RCT design. With recent advances in the recidivism risk assessment of sexual abusers, measuring intermediate treatment targets relating to dynamic risk factors, while waiting for actual recidivism data, could improve the use of RCT designs in our field. Research using the Violence Risk Scale–Sex Offender Version, evaluating the relationship between risk and change, may be applicable in this regard (Lewis, Olver, & Wong, 2013; Yang, Guo, Olver, Polaschek, & Wong, 2017).
Of note, the Association for the Treatment of Sexual Abusers (ATSA; 2010) has issued a policy statement in support of the use of RCTs as a method for evaluating treatment outcome, and their recommendations relating to group equivalence, treatment outcome evaluation, and well-defined treatment interventions are particularly helpful because they speak directly to some of the dilemmas and confounding variables that are specific to research with this population.
Meta-analyses and systematic reviews of sexual abuser treatment outcome studies
Meta-analyses and systematic reviews are two additional research approaches that are considered to have good empirical strength when used appropriately (Petrosino & Lavenberg, 2007), and the sexual abuser literature contains a substantial number of these reviews. Both of these approaches involve the synthesis of data from primary research studies and, as such, are commonly referred to as secondary research. Although both meta-analyses and systematic reviews have been conducted on a variety of clinical activities and research variables with sexual abusers, such as online sexual offenses (Seto, Hanson, & Babchishin, 2011), recidivism in female sexual abusers (Cortoni, Hanson, & Coache, 2010), and recidivism risk assessment tools (Hanson & Morton-Bourgon, 2009), their primary employment within our field thus far has been in assessing the findings of treatment outcome research.
Table 1 provides a summary of the most significant meta-analyses and systematic reviews that have examined adult, and in some cases juvenile, treatment outcome studies with sexual abusers. As can been seen in Table 1, there is a great deal of variability among these studies with respect to the treatment approaches being evaluated, the number of studies included, and the methods used for assessing the quality of studies to be used in the analysis. Also noteworthy is the relative inconsistency of findings regarding the effectiveness of the various treatment approaches commonly used with individuals who have sexually abused.
Summary of SR and MA on the Effectiveness of Sexual Abuser Treatment.
Note. SR = systematic reviews; MA = meta-analyses; RCTs = randomized controlled trial; MST = multi-systemic therapy; COSA = circles of support and accountability; CBT = cognitive behavioral therapy; SO Tx = sex offender treatment; RNR = risk, need, responsivity model.
In summary, the field has long sought to establish an empirical basis for our clinical activities and has worked hard to use research methods that are best suited to achieve that goal. Sexual abuser research that has used methods from the top of the research methods pyramid hierarchy has focused primarily, though not exclusively, on treatment outcome. Earlier research was often of questionable quality and provided either inconclusive or inconsistent findings regarding sexual abuser treatment effectiveness. The debate about the relative merits of using the RCT design with sexual abusers has yielded valuable discussion in our field, and potential methods for overcoming obstacles associated with its use have been proposed. Researchers are increasingly using quality control mechanisms in their research, such as the Maryland Scientific Methods Scale (Farrington, Gottfredson, Sherman, & Welsh, 2002) and the guidelines from the Collaborative Outcome Data Committee (CODC; 2007). A number of more recent meta-analytic studies and systematic reviews have reported positive treatment effects, particularly for cognitive-behavioral treatment approaches and those that adhere to principles of the risk, need, responsivity model (Hanson, Bourgon, Helmus, & Hodgson, 2009; MacKenzie, 2006; Schmucker & Lösel, 2008; Schmucker & Lösel, 2015).
While recent research on treatment outcome is encouraging, it is noteworthy that most of the research being conducted on persons with a history of sexual offending does not use the research methods listed at the top of the pyramid hierarchy. As a means of approximating the relative proportion of RCT, meta-analytic, and systematic review designs in relation to other research design methods in the sexual abuser research literature, we surveyed the contents of Sexual Offender Treatment, the official journal of the International Association for the Treatment of Sexual Offenders (IATSO), and Sexual Abuse: A Journal of Research and Treatment, published by the ATSA. Our survey counted the number of both quantitative and qualitative research articles but did not include book reviews or editorial notes. We reviewed each article published in both journals across an 8-year period, between January 2010 and December 2018, inclusive (for Sexual Abuse, Volume 22, March 2010 through Volume 30, December 2018, and for Sexual Offender Treatment, Volume 5, Issue 1 through Volume 12, Issue 2). Of the 56 issues of Sexual Abuse reviewed, a total of 251 scholarly articles, we found only one article in which the researchers used the RCT methodology, seven articles that used meta-analysis, and three studies that conducted systematic reviews. Our review of Sexual Offender Treatment included 68 articles, of which there were no RCTs, three studies that used meta-analyses, and one that conducted a systematic review.
Our review of 319 published articles in Sexual Abuse and Sexual Offender Treatment found that less than 5% of the studies used RCTs, meta-analyses, and systematic reviews as the main research methodology. We do not consider this to be symptomatic of a deficiency in the field of sexual abuser research. Rather, our survey suggests that researchers are examining a variety of clinical activities using an assortment of research methods that are best suited to their research questions. As such, we contend that an evidence-based review (EBR; described later) of the research from all levels of the research pyramid hierarchy, including research on clinical issues that do not directly address the broader question of treatment outcome, would be beneficial and could provide a means of establishing more definitive benchmarks regarding the strength of evidence on a wide range of clinical topics within our field.
The Absence of EBPs in the Treatment of Sexual Abusers
What then is the current status of EBPs relating to sexual abuser assessment, treatment, and community supervision? To date, we can find no clear consensus from within our field that speaks to the current strength of evidence (e.g., “strong,” “moderate,” “weak”) for any single treatment approach, intervention, or clinical activity as an EBP for use with individuals who have a history of sexually abusive behavior. Based on their systematic review of 11 studies, and using the criteria provided by the Promising Practices Network (PPN; 2007), Kim, Benekos, and Merlo (2016) conclude that sexual abuser treatments can be considered “proven” or at least “promising.” However, only the treatments for adolescents, and for surgical and chemical treatments, appear to have reached the effect sizes required by PPN criteria.
The Society of Clinical Psychology (Division 12, APA) has developed a website (https://www.div12.org/psychological-treatments/) relating to research-supported psychological treatments to help disseminate research evidence on EBPs. At the time of writing, under the section heading “disorders,” there were no sexual disorders (e.g., pedophilia) listed that would support their treatment as an EBP. Similarly, under the section heading “treatments,” there were no treatments listed for use with persons who have a history of sexual offending or for diagnoses that are common among this population.
Similarly, in 2007, the U.S. Substance Abuse and Mental Health Services Administration (SAMHSA) launched the National Registry of Evidence-Based Practices and Programs (NREPP; www.nrepp.samhsa.gov). The SAMHSA registry does list Multi-Systemic Therapy for Youth With Problem Sexual Behaviors as an EBP but has no other listing for topics or populations relating to the treatment of persons with a history of sexual offending.
The National Institute of Justice’s Crimesolutions.gov is a web-based clearinghouse that assesses the strength of evidence on programs and practices in the field of criminal justice (https://www.crimesolutions.gov/). Crimesoluations.gov provides a standardized and fairly objective process, including scoring instruments, for rating programs and practices. Based on scoring, programs and/or practices are rated as “effective,” “promising,” “inconclusive,” “ineffective,” or “no effects.” Crimesoluations.gov rates a cognitive-behavior therapy program for children with problematic sexual behaviors as “effective,” and four other sexual abuser programs as “promising.” In addition, the clearinghouse rates the practice of adult and juvenile sexual abuser treatment as “promising.”
Given the disposition of EBPs in the field of sexual abuser treatment, it seems important to ask how our research findings inform our decision making with respect to clinical activities and the content and delivery of treatment programming. Is the current process effective and sufficient, or does it prevent the field from advancing toward a more empirically informed process? We have identified three observable trends in sexual abuser research and clinical practice that provide insight into the current processes at work in our field and that could inform our ongoing efforts toward the use of a true EBP model with persons who have sexually abused.
Lack of Consensus in the Field Regarding an Evidence Threshold
The first trend that appears to be slowing movement toward the use of EBPs with sexual abusers is characterized by a general lack of consensus regarding the threshold for what constitutes sufficient “evidence.” This is actually one of several particularly vexing issues related to the development and implementation of EBPs that is not unique to our field.
For example, in evaluating the research to determine the evidence base for a specific clinical approach (e.g., Good Lives Model), how do factors such as the quantity of studies on that topic, the types of research methods that have been used to investigate it, and the quality of each study weigh into the decision to deem it sufficiently validated as evidence-based? How many RCTs, of good quality, should be considered adequate to claim an intervention or clinical activity as evidence-based? How should a large pool of research, with studies of various methodological designs and quality, be assessed to determine the evidence base for a given clinical activity? Furthermore, should the evidence threshold decision be dichotomous (e.g., sufficient evidence vs. insufficient evidence), or should nominal categories be created and used to denote a relative level of evidence (e.g., strong, modest, weak)? Our field has not yet embraced a system or set of criteria for making decisions about an evidence threshold.
Models for establishing an evidence threshold have been proposed by various writers and/or organizations. The PPN was developed in 1997, by the not-for-profit RAND Corporation, with the purpose of improving the lives of children and families by providing a website that features summaries of programs and practices that are shown to improve outcomes with children (PPN, 2007). The PPN website lists “evidence criteria” that are used to rate programs as either “proven” or “promising.” These criteria include considerations of research sample size, effect size, level of statistical significance, and whether or not a comparison group was used. The PPN does not require research to have been published in a peer-reviewed journal, or to have been replicated by other researchers, to be listed as either proven or promising. The PPN has not examined the topics of treatment programs or practices relating to youth or adults with a history of sexual offending using their evidence criteria framework. The PPN was discontinued in 2014 due to funding constraints.
The California Department of Social Services, in coordination with the Chadwick Center for Children and Families and the Child and Adolescent Services Research Center, has developed the California Evidence Based Clearinghouse for Child Welfare (CEBC; 2017) Scientific Rating Scale. The CEBC approach requires the clinical practice being rated to have a written manual that describes how its interventions are administered and to have outcome studies that have been published in a peer-reviewed journal. Based on the number of RCTs showing a practice to be superior to a comparison group, and by showing demonstrated clinical effectiveness over time, practices are rated as “Well Supported by Research Evidence” “Supported by Research Evidence,” or “Promising Research Evidence.” Conversely, when multiple controlled outcome studies have found that the practice has not resulted in improved outcome, or has had a negative effect, the practice can be rated as “Evidence Fails to Demonstrate Effect” or “Concerning Practice.” The CEBC has rated Multi-Systemic Therapy for Youth with Problematic Sexual Behaviors as “Well Supported,” and the Sexual Abuse: Family Education and Treatment Program as “Promising.”
Evans (2003) has also provided a unique framework for ranking evidence in health care interventions. In this model, clinical interventions are rated on two dimensions. First, research is rated as excellent, good, fair, or poor based on the methodology researchers have used in their study. As such, this model recognizes the generally superior empirical strength of RCTs, meta-analyses, and systematic reviews. However, Evan’s model also recognizes the multidimensional nature of evidence that can be drawn from a wide range of research methods, and evaluates studies across the three additional dimensions of effectiveness, appropriateness, and feasibility. These dimensions are intended to assess whether the intervention works as intended, is appropriate for its intended recipient, and the potential costs-benefits of actually implementing the intervention. We were unable to find any instances of the application of this model to research relating to individuals with a history of sexual offending.
Use of Imported and New Treatment Models That Lack Empirical Support for Use With Sexual Abusers
The second trend that could be slowing our movement toward the use of a true model of EBPs is the ongoing practice of integrating and importing clinical approaches and interventions, which have been used successfully with other mental health populations, into the field of sexual abuser treatment without sufficient evidence of efficacy with this population. During the past 30 years, a long list of treatment approaches have gained popularity for use with sexual abusers. Imported approaches that are now commonly used in the field include modifying cognitive distortions (Murphy, 1990), relapse prevention (Laws, 1989), motivational interviewing (Prescott & Porter, 2011), dialectic behavior therapy (Shingler, 2004), humanistic and experiential therapies (Bauman & Kopp, 2004; Longo, 2004), mindfulness (Gillespie, Mitchell, Fisher, & Beech, 2012), and, most recently, trauma informed care (J. Levenson, 2014). Similarly, several “new” or “original” models of sexual abuser treatment and supervision such as the Containment Approach (English, 2003), Self-Regulation (Ward & Hudson, 2000), and Good Lives (Ward, 2002) have also been advanced and, despite no initial evidence or systematic evaluation of their effectiveness with individuals who have sexually abused, many were widely embraced by treatment providers almost immediately after their emergence in the literature (McGrath et al., 2010). Given the lack of, or at best tenuous, empirical support for the use of these approaches with persons who have sexually abused, this trend suggests that imported approaches and new models of treatment that are seen, intuitively, as theoretically meaningful and clinically needed have had the greatest impact on clinical practice and program content.
We assert that the trend of importing and adopting new approaches, without adequate research to support their use with persons who sexually abuse, reveals at least two important insights about our field that need to be openly acknowledged, discussed, and managed. First, there is a natural tendency for paradigm shifts to occur, both in theoretical and clinical approaches, during the growth and development of a new field in behavioral health. Second, clinicians within our field experience a unique pressure to deliver effective treatment services to prevent sexual abuse and protect the community. This, coupled with the need to remain current with the most recent and most progressive treatment approaches, appears to result in a willingness to adopt clinical activities without proper empirical vetting prior to their wide-scale use. Paradigm shifts in an emerging field are understandable, and often lead to positive and needed change. However, the excitement generated by these new ideas and theories must be moderated by the ethical obligation to provide treatment interventions that have been sufficiently researched and found to be effective.
The Need for Training and Education Regarding EBPs
A third trend we have observed is the need for more training and education in the field of sexual abuser treatment about what “evidence-based” and EBP really mean, including a better understanding of the processes necessary for establishing and implementing EBPs. This is exemplified by the general lack of training on the topic of EBPs with sexual abusers at local, national, and international conferences, and by the paucity of published articles in our field that explicitly discuss concepts such as EBPs, best practices, ESTs, and related issues.
When the topic of EBPs with sexual abusers does appear in the literature, it is sometimes used to imply that clinical activities and interventions with this client population are evidence-based where no basis for this assertion is provided or currently exists. This trend can be found in the titles of books and journal articles, and occasionally in conference workshops relating to sexual abuser treatment. For example, the title may contain the wording “evidence-based” or “evidence-based practice,” but no definition of “evidence-based” is provided in the text. In some cases, references are made to specific research studies relating to the treatment approach being discussed in the article, but without a meaningful evaluation of the quality, validity, volume, and the overall strength of the broader pool of research on that topic.
The continued scarcity of training and education relating to the use of EBPs with sexual abusers, and the sporadic misuse of the term EBP, is of some concern. It could perpetuate a belief in the field that a more detailed and objective synthesis of the research is an unnecessary step toward determining what clinical activities are truly evidence-based. In addition, it creates a noticeable breach in the important link between research and clinical practice, particularly with regard to the process for critically reviewing, distilling, and implementing empirical findings into actual practice.
The Second and Third Legs of the EBP Stool
Although the process for establishing research “evidence” in the development of EBPs typically receives the greatest attention, the other two legs of the EBP stool (clinical expertise, and client characteristics, culture, and preferences) are of great importance and have yet to be adequately elucidated in the field of sexual abuser treatment.
In their description of the “expertise” component of EBPs, the APA (2006) lists eight clinical competencies that promote positive therapeutic outcomes: diagnostic judgment, treatment implementation, interpersonal expertise, self-reflection, use of research evidence, understanding of individual differences, seeking out of resources, and having a rationale for clinical strategies. These competencies are discussed in some detail, but are provided as generalized concepts in a broader professional context, and are not described with specificity to the clinical expertise necessary for the treatment of specific populations, such as persons with a history of sexual offending.
Ethical and practice guidelines for helping to determine whether a person has sufficient training to provide sexual abuser treatment currently do exist, and have provided some help in clarifying the EBP concept of expertise. For example, ATSA (2014) provides recommended qualifications (in the areas of education, training, and experience) to its members that provide clinical services to sexual abusers. These recommendations form an excellent starting point for addressing the EBP issue of clinical expertise for treatment providers, particularly the recommendations under the section for “Advanced/Specialized” providers.
Yet additional clarification by the field relating to what defines expertise seems warranted. What are the key clinical skill sets that need to be mastered by a therapist treating individuals with a history of sexual offending before certain interventions or approaches can be said to be established competencies? For example, some research has indicated that therapist qualities such as warmth, respect, and empathy are positively associated with treatment progress and outcome in sexual abusers (Drapeau, 2005; Marshall et al., 2002). If this is shown to hold true, then we should be asking how training and continuing education of therapists can be best structured to assist treatment providers in developing this skill, and how to assess when these skills have been mastered. Similarly, if interventions specific to any of the models that have been advanced for the treatment of sexual abusers are found to be evidence-based (e.g., relapse prevention, risk-need-responsivity, Good Lives, etc.), then training that teaches and assesses competencies in using these methods would be vital to developing clinical expertise in treatment providers as part of the larger approach to the use of EBPs by sexual abuser treatment programs.
The third leg of the EBP stool actually comprises three distinct concepts, “client characteristics, culture, and preferences.” These concepts, individually and as a whole, speak to the significance of two fundamental ideas in the application of EBPs. The first is the importance of considering idiographic factors in the delivery of clinical services. Individuals have unique histories and needs, which should be addressed and incorporated into their treatment. The second idea concerns the role of the client in decision making about their own care. Individuals should have sufficient information and education to choose their preferences in accordance with their own values and beliefs.
With respect to the consideration of idiographic factors in the delivery of clinical activities, there is a growing awareness in our field of the importance of assessing and considering client characteristics. Recidivism risk instruments, such as the Static-99 and Stable 2007, reflect the role of client characteristics (e.g., age, emotional identification with children) in recidivism risk, and several models of treatment that have been advanced for use with sexual abusers also give consideration to client characteristics and needs (e.g., Good Lives, Risk-Need-Responsivity).
The idea of shared decision making, although present in some recent approaches with sexual abusers such as the Good Lives Model (Ward, 2002), SOTIPS (an assessment scale) (Lasher, McGrath, Wilson, & Cumming, 2015), and motivational interviewing (Shingler & Mann, 2006), is a concept that has not been fully explored or realized in the field. One possible reason for this is that our clients are often mandated to participate in treatment. Moreover, as Spring (2007) has indicated, the role of shared decision making in EBPs is a complex process that requires a departure away from paternalistic models of care toward one that is more culturally informed. This requires clients to have information that allows them to weigh the risk and benefits associated with difficult decisions. In the field of sexual abuser treatment, these decisions often have significant legal and social risks to the client, such as the decision to acknowledge unreported sexual offenses, to register oneself as a “sexual offender,” or where to live in the community when local ordinances or parole stipulations restrict the best and/or healthiest options.
The Implementation of EBPs With Sexual Abusers
The tireless efforts of researchers in our field to determine the empirical basis of our clinical approaches and activities have been mentioned earlier. In addition to those efforts, there are distinct organizational structures that assist treatment programs and clinicians in adhering to standards of practice that are based on empirical findings in the sexual abuser or correctional research literature. Treatment program accreditation standards, for example, have been established in both the United Kingdom and Canada. Based on large-scale meta-analytic reviews of the research literature on correctional treatment for general offenders, researchers and clinicians working in Her Majesty’s Prison Service (England and Wales) developed 10 treatment program accreditation criteria (Lipton, Thornton, McGuire, Porporino, & Hollin, 2000). In addition to their use with treatment programs for the general correctional population, these 10 criteria have also been implemented for use within Britain’s Sex Offender Treatment Programme (SOTP; Perkins, Hammond, Coles, & Bishopp, 1998). Correctional Service Canada (CSC) subsequently modeled their accreditation standards for sexual abuser treatment programs on those of Her Majesty’s Prison Service (Hanson, Broom, & Stephenson, 2004). Among other criteria, the CSC accreditation standards were used in an effort to affirm that treatment approaches and intervention methods were grounded on an empirically based model of change, targeted criminogenic treatment needs, matched the learning styles and abilities of the offenders, were delivered at appropriate and sufficient dosage, provided for continuity of care and follow-up services, and included mechanisms for monitoring program integrity over time (Hanson & Yates, 2013).
Closely related, and also from the field of corrections, tools for evaluating the practice standards of treatment programs exist, such as the Correctional Programs Assessment Inventory (CPAI; Gendreau & Andrews, 1994) and the Evidence-Based Correctional Program Checklist (CPC; Latessa, 2012). Both the CPAI and the CPC are tools that can be used to assess how closely a correctional treatment program meets those standards that are believed to be effective with the correctional population.
In the United States, administrative boards that oversee practices and policies regarding sexual abuser assessment, treatment, and community supervision exist in many states. These boards, variously known as sex offender management boards, sex offender advisory boards, sex offender management bureaus, and so on, can be found in at least 10 states (California, Colorado, Delaware, Idaho, Illinois, Massachusetts, New Mexico, Oregon, Pennsylvania, and Washington). Most of these boards provide guidelines, standards, or certification relating to performing court-ordered evaluations, qualifications and standards of practice for treatment providers, and standards for the management and supervision of individuals on parole and/or probation. Similarly, professional organizations such as ATSA (2014) and IATSO (Miner et al., 2006) also provide practice guidelines to their members relating to the assessment, treatment, and supervision of individuals with a history of sexual abusive behavior.
Importantly, while treatment program accreditation standards, correctional program assessment tools, and management boards provide much needed structure and parameters for the oversight of treatment delivery systems with sexual abusers, it is important to recognize that they are ultimately informed by the extant empirical literature. As such, many of the structures that have been put in place by these organizational entities could also be used to assist in the implementation of an EBP model for persons with a history of sexual offending.
The process for actually implementing EBPs with individuals who have sexually abused, at a programmatic level, which would logically follow an identification of treatment approaches that are evidence-based, is only briefly discussed here. Yet it is vital that program directors, administrators, and organizational leaders in the field of sexual abuser treatment begin to give consideration to the myriad of issues involved in this process. Even with an established base of empirical evidence regarding clinical activities with sexual abusers, and a firm understanding of how clinical expertise and client characteristics influence outcomes, the implementation of EBPs within the treatment and supervision systems where our clients receive services will be challenging.
Several models and approaches for ensuring the effective implementation of EBPs have been advanced (Fixsen, Naoom, Blase, & Friedman, 2005; Houser & Oman, 2010; Kaper et al., 2015). The implementation of EBPs typically occurs at several levels: at the individual or case level, at the program or agency level, and within the larger organizational system or professional field (Bogue et al., 2004). Sexual abuser treatment and supervision services are somewhat complex in that they can be found across a spectrum of delivery systems such as correctional facilities, inpatient forensic mental health, civil commitment institutions, outpatient private practice, community mental health centers, and more generally under the broader umbrella of the criminal justice system’s community corrections, parole, and/or probation supervision programs. As a result, effectively implementing EBP’s with sexual abusers will require coordination and communication across the various levels, and within the variety of systems, within which clinical services are delivered. Toward this end, empirical research evidence related to the three legs of the EBP stool must get accurately communicated to program administrators. Training of end-users (e.g., clinicians, parole officers) will be needed, and fidelity to the EBP model should be evaluated on an ongoing basis to ensure proper use and effectiveness (Kaper et al., 2015).
Recommendations for Moving Toward the Use of a True EBP Model With Sexual Abusers
With the increased use of methods for assessing quality in treatment outcome research, there appears to be growing empirical support for the effectiveness of treatment, and a number of recent meta-analytic studies and systematic reviews have reported positive treatment effects. Yet even treatment outcome research that uses CODC standards to improve methodological quality continues to produce mixed results (Grady, Edwards, & Pettus-Davis, 2017). Although no clear consensus has been reached to affirm the strength of our various treatment approaches as EBPs, MST for youth with problematic sexual behavior has been endorsed by SAMHSA as an EBP and by the CEBC as “Well Supported,” and cognitive behavioral theapy (CBT) and risk-need-responsivity (RNR) approaches with adults appear very promising.
In addition to treatment outcome, there is a wide variety of other topics on which sexual abusers are being researched and on which a clear consensus regarding the strength of evidence has not yet been reached (e.g., the role of empathy in treatment, use of the polygraph, the relationship between denial of offense and recidivism). Although the current standard for determining “evidence” in health care research is the Cochrane Review, both a strength and limitation of the Cochrane Review is that the eligibility criteria for inclusion of research often limits these reviews to focusing exclusively on RCTs (Higgins & Green, 2011). While there are presently a limited number of RCTs in the sexual abuser research literature, a large and valuable trove of primary and secondary research on a variety of subject matters is available. As such, we contend that the status of EBPs in our field is not entirely due to the absence of research evidence but is also due to the absence of a systematic organization and evaluation of the research that currently exists, particularly with respect to research outside the treatment outcome literature.
Given the growth and development of our field, we believe that it is time for researchers, clinicians, and professional organizations to make more purposeful and careful movements toward the use of EBPs (see Figure 1). As our first recommendation, we encourage researchers and clinicians to come together to discuss the most pressing issues in the field and to prioritize research goals. A clear priority should include assisting researchers and program administrators in finding the means for conducting more RCTs, while making use of the quality control mechanisms and alternative methods for measuring outcomes described earlier (e.g., intermediate treatment targets, random assignment to alternative treatment components, and measurement of in-treatment changes on dynamic risk factors).
Recommendations for advancing the use of EBPs with sexual abusers.
The importing of treatment approaches from other areas of mental health into the field of sexual abuser treatment was understandable during its infancy. We now possess a wealth of knowledge and research relating to the characteristics of individuals with a history of sexually offending, recidivism risk factors, and treatment and supervision approaches on which an evidence-base could be established. As we move forward, new models of treatment and the importing of approaches for use with sexual abusers should be investigated as part of ongoing research trials with sexual abusers prior to their adoption and broader use within the field.
Concomitant with ongoing efforts in the field to conduct more RCTs of better quality, our second recommendation is that the field should begin using an “evidence-based review” process that would synthesize available research findings and develop clearly defined evidence thresholds as a means of identifying and communicating the level of empirical strength of research on our clinical activities. The process of attempting to infer evidence, based primarily on the results of meta-analyses and systematic reviews of treatment outcome studies without establishing evidence thresholds, has shown to be problematic in communicating the level of confidence in our findings. We contend that this process has essentially forced researchers and clinicians to perceive the issue of “evidence” as dichotomous (evidence vs. lack of evidence) rather than one that exists on a spectrum with varying levels of support (e.g., weak to strong). Given the variety of methodologies being used to conduct sexual abuser research, the diversity of topics under investigation, and the importance of assessing research quality as a critical component of the process for determining evidence, we believe that a formal EBR process that gives consideration to the unique challenges of our research is a necessary step toward advancing a true EBP model.
For a variety of reasons, including scientific objectivity, an EBR process cannot be accomplished by a single clinician or researcher reviewing the literature and then forming his or her own conclusions. It requires a broad-based effort on multiple levels, including participation and leadership from our professional organizations. Toward this end, we suggest the following objective and empirical three-step process for establishing the strength of research evidence for our clinical activities:
EBR Step 1: Organization of Research by Domain and Clinical Activity
First, with respect to organizing and synthesizing the research, we suggest the establishment of domain areas (e.g., community supervision/support), within which specific clinical approaches or activities (e.g., Circles of Support and Accountability [COSA]) would be planned for review. Search protocols and parameters should be clearly defined. The search engine(s), database(s), range of years, languages, whether or not gray literature will be reviewed, and any other criteria used for including or excluding studies should be determined prior to initiating the literature search for research articles relating to the clinical activity receiving the EBR. For any given clinical activity, both the primary and secondary research that has been conducted on that activity should be included in the EBR. Although both meta-analyses and systematic reviews are located at the apex of the research methods pyramid hierarchy, they remain subject to many of the same threats to internal and external validity as primary research approaches (Berlin & Golub, 2014; Sterne, Egger, & Smith, 2001), and therefore should also be evaluated for quality in the process of an EBR.
EBR Step 2: Scoring the Research for Strength and Quality
Second, a valid and reliable scoring tool that can rate the strength and quality of individual research studies is required. This tool should be capable of scoring the wide variety of research methods and analyses that exist in sexual abuser research, including but not limited to true experiments, quasi-experimental designs, correlational studies, and systematic reviews and meta-analyses. The scoring tool should also provide detailed scoring criteria, across multiple indices of research design and statistical analysis, which allow for clear and objective scoring of the research. Ideally, an organizational entity consisting of trained professionals who can score the research would be used. These evaluators should have an advanced understanding of the field of sexual abuser treatment, be knowledgeable in research design and statistical methods, and should receive standardized training in the use of the research scoring tool.
Several systems for scoring and rating research studies of diverse designs and methodologies have been developed and described (Chambless & Hollon, 1998; Kratochwill & Stoiber, 2002; Sirriyeh, Lawton, Gardner, & Armitage, 2012). Each approach recommends somewhat different methods or systems for determining the overall strength of research, but none have been developed specifically for use with research relating to individuals with a history of sexual offending. Most of these systems require a rating of several components of a research study’s design and methods on variables such as sample size, tools used for measuring the dependent variable, and statistical analyses used. As mentioned earlier, the CODC developed guidelines for researchers designing new studies specific to sexual abuser treatment outcome, and for reviewers critically examining research (CODC, 2007; Hanson et al., 2002). Since the initial development of these guidelines, several researchers have used them to address the issue of research quality in conducting treatment outcome studies on sexual abusers (Grady et al., 2017; Wilson, Cortoni, & McWhinnie, 2009). However, there has been no use of these guidelines as a scoring tool for evaluating studies as part of a formal three-step EBR process.
Most recently, Deming, Yates, and Barbaree (2016), as part of their ongoing work with the International Project for Evidence-Based Practices With Sexual Abusers (IPEPSA), have developed a scoring tool for use with research relating to sexual abusers. The IPEPSA tool, adapted from a system described by Sirriyeh et al. (2012), has a total of 22 items that score research articles across multiple areas of internal and external validity in an effort to determine their empirical strength. The items in the IPEPSA tool also differentiate between primary and secondary research methods, and have scoring criteria that recognize the unique nature of this population. Thus far, it has shown very good inter-rater reliability in initial testing, and is being used as part of a broader effort to evaluate the empirical strength of clinical activities currently being used with sexual abusers and to determine their standing as EBPs (Deming et al., 2016). IPEPSA is currently organizing the sexual abuser research literature into categories based on the type of clinical activity (assessment, treatment, community supervision, and public policy) and, with use of their scoring tool, has begun the process of synthesizing the research related to treatment outcome.
EBR Step 3: Developing Evidence Thresholds
Finally, the development of nominal categories (e.g., insufficient evidence, weak evidence, modest evidence, or strong evidence) is needed to convey the strength of the evidence-base for the research on each clinical activity under review. These categories, or evidence thresholds, should be based on objective criteria and should be capable of reflecting both the quality and quantity of the pool of research on a given clinical topic. The systems used by the PPN or the CEBC, discussed earlier, are examples of how evidence thresholds could potentially be structured.
Once these three steps are in place, all relevant research on a given treatment approach or clinical activity can be taken through the EBR process, and the first leg of the EBP stool, the strength of the “empirical evidence” in support of that activity, can be determined. Further development and use of an EBR process will require careful consideration and decision making within the field of sexual abuser treatment, and input from an international delegation of researchers, clinicians, and program administrators would be vital to its success. We believe this effort should come from within our field, but should also be conducted by an entity that can engage in this process with sufficient objectivity and independence.
As our third recommendation, we believe that additional theory, research, and training are needed relating to the second (clinical expertise) and third (client characteristics, culture, and preferences) legs of the EBP stool. A true model of EBPs with sexual abusers requires not only a firm base of research evidence regarding effective clinical activities but the ability to deliver these interventions with system-wide agreement as to the clinical expertise required by treatment providers and with respect to those ideographic factors in our clients that affect their ultimate use and effectiveness.
Finally, we recommend field-wide training and education regarding the development, goals, and implementation of EBPs. It is apparent that increased education and training are needed to help clarify the differences between EBPs and other similar terminology to professionals in our field and to other stakeholders (including clients). Authors, editors, and conference presenters should be more cautious and more precise in using the term evidence-based practice in their publications and workshops. For clarification, EBPs should be distinguished from a few related terms often found in the literature. For example, evidence-based programs is a term describing an integrated set of coordinated activities that may include more than one professional discipline and that incorporate EBPs in the delivery of those services (Fixsen et al., 2005). ESTs is a term used to describe specific treatments that have been shown to be effective for certain disorders or diagnoses, under specified circumstances (APA, 2006). Finally, “best practices” is a frequently used term in health care for which multiple and often divergent definitions can be found, and in some cases, the definition can appear similar to that of EBPs. The term originated in the organizational management literature in the context of quality improvement, where “best practices” are defined as the preferred technique or approach for achieving a valued outcome (Mullen, Bellamy, & Bledsoe, 2013). As Bogue and colleagues (2004) have reported, best practices do not necessarily imply attention to outcomes, evidence, or measurable standards, and are often based on the collective experience, expert opinion, and wisdom of the field rather than on scientifically tested knowledge.
We suggest that the IATSO and ATSA, and similar organizations, intentionally devote sessions at their conferences for research, workshops, and educational sessions relating to the development and use of EBPs with individuals with a history of sexual offending. It is hoped that through increased education and training, a new era in the field emerges in which a true EBP model can be built on clearly defined research priorities, EBRs of empirical research findings, and field-wide understanding and support for EBPs.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.