Keywords

14.1 Psychopathy and Therapeutic Pessimism

The diagnosis of psychopathy does not promise much good in the eyes of lay people and professionals alike. The therapeutic nihilism that used to be characteristic of the offender rehabilitation literature more generally, the so-called “Nothing Works” doctrine (Farabee, 2005; Martinson, 1974), has in recent decades been replaced by the “What Works” approach to offender treatment. The latter development should be credited to the groundbreaking scholarly work of Andrews and Bonta, who developed the Risk-Need-Responsivity (RNR) model of offender rehabilitation (Andrews & Bonta, 1994, 2010; Andrews et al., 1990). The RNR model purports that high-risk offenders should receive the most intensive treatment and risk management interventions, to maximize (violent) crime reduction (Polaschek et al., 2016). Psychopathic offenders are, by definition, high-risk offenders, consequently in need of the most intensive treatment (Skeem & Polaschek, 2020). In reality, high-risk psychopathic offenders are often the least likely to receive the most intensive treatment because they are usually assumed to be the most hardened and unlikely to respond to treatment (Skeem & Polaschek, 2020). Because of their dangerousness as a subgroup of the offender population, they are more likely to receive the death penalty in the US (DeMatteo et al., 2020), partly due to the pejorative connotations this diagnostic label holds (Edens et al., 2018).

Treatment approaches to psychopathic offenders can be distinguished into rehabilitating, risk-reducing treatment and treatment of psychopathy in its essence (Polaschek & Skeem, 2018). In the former type of treatment, the focus lies on reduction of risk factors for reoffending, such as substance use problems, criminal thinking styles, and poor angermanagement. A recently developed treatment that belongs to this type is Wong and Hare’s psychopathy treatment program (Wong, 2013; Wong & Hare, 2005). The second type of treatment is directed at changing the core features of psychopathic personality disorder (PD), such as limited affect and affect dysregulation (Chakhssi et al., 2014b; Galietta & Rosenfeld, 2012; de Ruiter et al., 2016).

In this chapter, we will provide a review of the existing research base on the treatment of psychopathy. We will take a historical approach, starting with early treatments and empirical studies into their effects, up until more recently developed interventions. The knowledge base on treatment effects on psychopathy in adults is limited and the research methodology applied is often less than optimal. The scarcity of empirical research into psychopathy treatment stands in great contrast to the database on the treatment of other PDs, for instance, borderline PD. A recent Cochrane meta-analysis and review on psychological treatments for BPD (Storebø et al., 2020) included 75 randomized controlled trials with a total of 4507 participants, which tested 16 different kinds of psychotherapy, dialectical behavior therapy (DBT) and mentalisation-based treatment (MBT) being the most frequently used. Our review is necessarily much more limited in scope. Our goal is to separate some wheat from the chaff, so this bit of wheat might be useful in generating future crops which hopefully result in more effective treatment for individuals with psychopathic PD in the future.

14.2 Early Treatment Approaches (1990–2006): Therapeutic Communities

The therapeutic community (TC) has been one of the experimental treatments for psychopathy (Harris & Rice, 2006). The TC is based on the premise that a milieu or environment that is therapeutic can be created which is useful to effect positive behavior change. However, how this is implemented varies dramatically and not all TCs are created equal. Some early reports, prior to the 1980s (e.g., Copas et al., 1984; Copas & Whitely, 1976; Kiger, 1967; see also Dolan, 1998), reported positive results regarding the effectiveness of the TC at reducing violence and other disruptive behavior in psychopathic patients. However, methodological limitations, including poor definitions of psychopathy and lack of control groups (e.g., no comparative data for untreated psychopaths), made findings very difficult to interpret (Lösel, 1998).

Ogloff et al. (1990) were the first to explore the impact of contemporarily defined psychopathy (i.e., defined by the Psychopathy Checklist [PCL]; Hare, 1985) on treatment behavior in a TC treatment program in a Canadian forensic hospital (N = 80). In this often-cited study, the primary treatment modality was a large therapeutic group that met on weekdays for approximately two hours. The group was described as unstructured and relied strongly on the input of patients. PCL scores were used to divide patients into high (total score of 27 or more), moderate (18–26), and low (score of 17 and below) psychopathy groups. The outcome variables included: length of time spent in the TC program; ratings (on a 4-point scale) of degree of motivation/effort put into the program; and ratings (also on a 4-point scale) of degree of clinical improvement shown during treatment. Note that degree of motivation and degree of clinical improvement were coded from clinical and institutional files. Patients diagnosed as psychopathic (PCL ≥ 27) performed significantly poorer on all three outcome criteria than patients with moderate or low PCL scores. On average, psychopathic patients remained in the program for a shorter period, showed less motivation and clinical improvement than each of the other groups (Ogloff et al. 1990).

The now classic study conducted by Rice et al. (1992) of a TC operated through the Penetanguishine Mental Health Centre in Ontario, Canada, undeniably has had the greatest impact on the idea that TCs for psychopaths are set up for failure. The authors retrospectively evaluated the 1960s Oak Ridge Social Therapy Unit, a hospitalization program for mentally disordered offenders thought to be especially suitable for psychopaths. It operated for over a decade and drew worldwide attention for its novelty.Footnote 1 Treatment was intensive and highly unconventional, to say the least. It was largely peer operated and involved intensive group therapy for up to 80 hours per week, in which the men would be locked in groups in small rooms, left to discuss their issues and to confront each other on their behavior. There was little input from staff, with the patients left essentially to run their own treatment. Hallucinogens and sedatives were administered to lessen defenses or to augment disclosure, often at the direction of other patients. The men were subjected to group pressures, nude encounter groups, and deprivation in various forms. Thus, the key components of the treatment were highly experimental, based on principles of brainwashing in a Chinese prison camp (Harris et al., 1994) and judged by the authors themselves as violating patients’ rights by 1992 standards (Rice et al., 1992).

In the Rice et al. (1992) study, 146 treated offenders who had spent at least two years in the program were matched with an equal number of untreated offenders (controls). The offenders were matched for age, criminal history, and index offense. All offenders were scored on the Psychopathy Checklist-Revised (PCL-R; Hare, 2003) based on file information (gathered in the 1970s) and a cutoff score of 25 or higher was set for classifying offenders as psychopathic. The results of a follow-up roughly 10.5 years posttreatment showed that there was very little overall difference between the two groups (i.e., treated and untreated offenders), despite the rather lengthy and intensive treatment program that the treated offenders had received. However, when the groups were divided into psychopathic (PCL-R ≥ 25) and nonpsychopathic (PCL-R < 25), it was found that more treated psychopathic offenders recidivated with a violent offense compared to untreated psychopaths (77% versus 55%). The opposite was true for nonpsychopathic participants—that is, more untreated than treated nonpsychopathic offender failed (39% versus 22%). Thus, treatment was associated with a reduction in violent recidivism among nonpsychopathic offenders but with an increase in violent recidivism among psychopathic offenders. In addition, with regard to possible differences in treatment responsiveness between treated psychopathic and nonpsychopathic offenders, it was found that those classified as psychopathic showed poorer adjustment in terms of problem behavior while in the institution than the nonpsychopathic individuals, although they were just as likely as nonpsychopathic offenders to receive positive staff recommendations and achieve positions of trust. The authors speculated that the treatment provided a learning opportunity (e.g., learning about the feelings of others, behaving in socially skilled ways) for both psychopathic and nonpsychopathic participants alike. Whereas the nonpsychopathic individuals used the information to behave prosocially, the psychopathic individuals used it to manipulate and exploit others (Harris & Rice, 2006; Rice et al., 1992). According to Rice et al. (1992), “The results strongly suggest that the kind of therapeutic communitydescribed in this article is the wrong program for serious psychopathic offenders” (p. 408). “Community treatment programs that generally seek to cultivate pro-social empathic and caring qualities might inadvertently make psychopaths better equipped to ‘facilitate the manipulation and exploitation of others,’ and such treatment efforts could, therefore, be ‘associated with novel ways to commit violent crime’” (p. 409). Notably, during their stay in the TC, psychopathic offenders were significantly more likely than nonpsychopathic offenders to be referred to a ‘disciplinary subprogram’, to remedy noncompliance and to be written up and placed in seclusion for disruptive or violent behavior (Rice et al., 1992). These indices of misbehavior and punishment were, in turn, significantly predictive of recidivism. However, the effect of the TC on recidivism, after statistically controlling for these disciplinary sanctions (which may have resulted in a lower treatment intensity), apparently has not been examined.

Results of other studies conducted in TCs also suggested that TC approaches are not likely to benefit psychopathic offenders. Hobson et al. (2000) evaluated a TC in Grendon prison, England, and found a significant relationship between PCL-R Factor 1 scores and negative behaviors in therapy groups and on the ward. The authors found a particularly strong relationship between negative treatment behaviors and the PCL-R items ‘glibness/superficial charm’, ‘grandiosity’, and ‘failure to take responsibility’. Thus, Hobson et al. (2000) concluded that PCL-R Factor 1 scores should be considered when assessing an offenders’ suitability for participation in a therapeutic community. Finally, in a study of a TC for female substance abusers, Richards et al. (2003) found that, although ‘true’ psychopaths (PCL-R score > 30) were excluded from the treatment program, psychopathy scores were still significantly associated with poor treatment response (in terms of avoidance of urine tests, violent and disruptive rule violations, sporadic attention, failing to stay in the program), and upon release, fewer days in the community prior to receiving a new criminal charge. Factor 1 scores in particular were associated with increased risk for general recidivism.

Overall, the preponderance of evidence indicates that TC treatment approaches are not likely to benefit psychopathic offenders in terms of recidivism reduction. Other than by means of official recidivism data, the TC programs were not evaluated. For example, there was no measure of change in clinical outcomes, most importantly, there was no measure of change in psychopathic traits or other dynamic risk factors for violent recidivism. The early optimism regarding the effectiveness of the TC in treating psychopathy obviously diminished. However, the TC studies reviewed above, which suggested support for the thesis that psychopathy is not amenable to treatment, had significant methodological and conceptual flaws that bring into question the validity of their results.

14.3 Cognitive-Behavioral Treatment Approaches

Besides TCs, treatment programs based on cognitive–behavioral theory have been recommended for psychopathic offenders (e.g., Andrews & Bonta, 1994; Brown & Gutsch, 1985; Serin & Kuriychuk, 1994). Below, we will review studies that examined the effectiveness of cognitive–behavioral treatment approaches in the treatment of psychopathy.

A number of studies have examined the effectiveness of cognitive-behavioral treatment of sex offenders with high vs. low scores on psychopathy. Seto and Barbaree (1999) were the first to examine the association of PCL-Rpsychopathy, behavior during treatment, and recidivism among a sample of 216 sex offenders in a cognitive–behavioral and relapse prevention program at the Warkworth Sexual Behavior Clinic (WSBC), located in a medium secure federal penitentiary in Ontario, Canada. The treatment involved daily 3-hour group sessions over a period of 5 months. The treatment focused on the identification and understanding of individual offense chains by sequencing the thoughts, feelings, and behaviors preceding the commission of a sexual offense. In addition, a relapse prevention plan was developed for each individual offender (for a detailed description of the treatment program, see Barbaree et al., 1998). Notably, the treatment was not designed to target psychopathy. Offenders were assigned to one of four groups based on their scores on the PCL-R and a measure of treatment behavior (including attendance, participation in group sessions, disruptive behavior, global clinician ratings of motivation and change achieved in treatment) based on a median-split for each measure (median PCL-R score = 15). Results revealed that offenders scoring 15 or higher on the PCL-R who behaved well in treatment were much more likely to commit a new offense during an average follow-up period of 32 months than offenders in the other three groups. Among the more psychopathic offenders, those rated as participating well in treatment were five times more likely to commit a new serious (violent or sexual) offense as those who were rated as participating poorly. No such ‘paradoxical’ pattern was found for offenders scoring low (i.e., < 15) on the PCL-R. Based on their results, Seto and Barbaree (1999) suggested that “good treatment behavior should not be considered when making management decisions, especially for men who score higher on the PCL-R” (p. 1245). At the time this study was published, it caused considerable concern (Barbaree et al., 2006) because the findings echoed with Rice et al.’s (1992) evaluation of the Penetanguishine TC program. Even a treatment program that followed the principles of best-evidence correctional treatment (Andrews & Bonta, 1994; Andrews et al., 1990)―e.g., highly structured and cognitive-behavioral, matching the learning style of most offenders―could possibly make some psychopathic offenders worse.

However, subsequent follow-up research of the same sample provided a different and more complex picture. Barbaree (2005) examined the same sample using a longer follow-up period (mean of 62 months instead of 32) and more comprehensive (and less biased) recidivism data. In addition, he used a PCL-R score of 25 to split the sample, which is obviously more appropriate to identify a high psychopathy group. With the extended follow-up period and new outcome data, Barbaree (2005) found that there was no significant difference in serious recidivism rates between psychopathic offenders who showed good in-treatment behavior and psychopathic offenders who showed poor in-treatment behavior (34% versus 30%). Neither treatment behavior nor the psychopathy—treatment behavior interaction was a significant predictor of recidivism at any of the fixed follow-up times (follow-up periods of 3, 5, and 6 years). Barbaree (2005) concluded there was no evidence that justified Seto and Barbaree’s (1999) earlier conclusion that treatment made psychopathic sex offenders worse. He also stressed the importance of awaiting an accumulation of evidence over a number of studies before making major changes in policy and practice regarding treatment of psychopathy (Barbaree, 2005).

Langton et al. (2006) expanded the WSBC sample to 418 treated sex offenders and followed the sample up for 5 years post-release. The authors used a cut-off of 25 (instead of 15) to characterize PCL-Rpsychopathy in this sample. In contrast to the previous findings of Seto and Barbaree (1999) and Barbaree (2005), Langton et al. (2006) found a significant psychopathy x in-treatment behavior interaction effect, such that psychopathic offenders who displayed poor treatment behavior had a significantly higher and faster rate of sexual recidivism over the follow-up period than psychopathic offenders who behaved well during treatment. Finally, Looman et al. (2005) used a similar design as Seto and Barbaree (1999) to examine recidivism outcomes for a sample of 102 sex offenders who had participated in an institutional treatment program for sex offenders at the Regional Treatment Centre Sex Offender Treatment Program in Kingston, Ontario, Canada. The program was a 7-month high-intensity treatment program providing both group therapy and individual therapy to sexual offenders who are assessed as being high risk for reoffending and having high treatment needs. Looman et al. (2005) examined two indicators of treatment change: (1) had the offenders made good vs. poor treatment progress, and (2) were the offenders evaluated as having lowered their risk level at the end of treatment or not? The authors found that psychopathic offenders (PCL-R score > 25) with ratings of good progress in treatment reoffended seriously (i.e., violently or sexually) at a significantly faster rate than did either of the groups with lower PCL-R scores. High psychopathy participants with ratings of poor treatment behavior did not differ from either group of low PCL-R participants with regard to sexual and violent recidivism. The recidivism rate for the two high PCL-R groups did not differ, indicating a main effect for the PCL-R rather than an interaction effect. In addition, among high PCL-Roffenders, those rated as lower risk at posttreatment in fact reoffended at a lower rate (30%) than those whose risk was rated as unchanged (50%), although this difference failed to reach significance, probably due to the small cell sizes and concomitant low statistical power. Thus, both Langton et al. (2006) and Looman et al. (2005) found that one of their two groups of psychopathic offenders—split into good and poor treatment behavior—reoffended at a statistically equivalent rate to the two low psychopathy groups — also split by quality of treatment behavior. According to Polaschek and Daly (2013), this result could be interpreted as indicating that some psychopathic offenders might have benefitted from the treatment program, since in each study one group of psychopathic offenders had a comparable outcome to non-psychopathic offenders, although this improved result was found for the better treatment progress group in the Langton et al. (2006) study, and for the poorer treatment progress group in the study conducted by Looman et al. (2005). This difference may be explained by the use of different operationalizations of treatment process/treatment progress in both studies. Nevertheless, according to Polaschek and Daly (2013), the contradictory results in these studies provide some support for the idea that the heterogeneity of PCL-psychopaths extends to their response to treatment.

Further doubts regarding the efficacy of cognitive-behavioral treatment programs for psychopathic offenders emerged from outcome studies conducted outside Canada. For example, Hughes et al. (1997) described preliminary results for a very small sample of offenders who had participated in a program for mentally disordered offenders in an English high-secure forensic psychiatric hospital. The authors found that PCL-R scores (specifically, Factor 1 scores) were significantly inversely correlated with overall clinical change, even though patients with PCL-R scores over 30 were excluded from the study as it was assumed that they would not benefit from treatment. The authors concluded that, “the degree of therapeutic change was strongly mediated by level of psychopathy” (Hughes et al., 1997, p. 524). However, since ‘true’ psychopaths were excluded from treatment, this study does not provide an adequate test of the effect of treatment for psychopathic offenders. In another study, Hare et al. (2000) evaluated cognitive-behavioral prison programs for psychopathic and nonpsychopathic offenders in several English prisons. Offenders participated in a short-term angermanagement program involving social skills training. Two-year general reconviction rates for 278 offenders released into the community were determined for those who took part in at least one of the programs and for those who did not. A strong treatment effect was found for offenders in the high PCL-R Factor 1 group, but in the wrong direction. Treated offenders who scored high on Factor 1 had significantly higher rates of recidivism compared to high PCL-R F1 offenders who did not receive treatment (i.e., 85.7% vs. 58.8%). According to the authors, Factor 1 psychopaths may have increased their manipulative skills while in treatment. Hare et al. (2000) also called into question the appropriateness of the interventions because these programs showed very little benefit, even for non-psychopathic offenders. In fact, no clear description of the treatment in this study is given, and it is possible that it varied across settings (Salekin et al., 2010).

14.4 Salekin’s Meta-Analytic Reviews of Research on the Treatment of Psychopathy

Taken together, the early treatment research that used the PCL(-R) to define offenders low versus high in psychopathy not only supported the long held notion that psychopathic individuals were untreatable, but further suggested that treatment might actually make them worse. As noted by Polaschek and Daly (2013), the study by Rice et al. (1992) effectively “slammed the lid shut for many on the advisability of even attempting treatment” (p. 195). In fact, a number of treatment programs began denying treatment to people with high PCL scores based on the evidence that treatment would fail to help them or make them worse (e.g., D’Silva et al., 2004; Hughes et al., 1997; McCarthy & Duggan, 2010; Richards et al., 2003).

In 2002, Salekin published a meta-analysisFootnote 2 that challenged the general pessimism about the treatment of psychopathy. He reviewed 42 studies that he identified as evaluating the effectiveness of some form of therapy for psychopathic patients. The designs of the studies varied from large-scale evaluations of programs featuring several hundred patients to single case studies. Salekin (2002) found that, on average, approximately 62% of patients benefited from treatment, and 60% when case studies were removed. The most effective treatment modality was cognitive-behavioral (average success rate = 62%), followed closely by psychoanalytic psychotherapy (success rate = 59%). Additionally, treatment was found to be especially effective when delivered for a longer period (success rate = 91% for treatments > 12 months; 61% for treatments < 6 months), and for youths (96% success rate for juveniles vs. 63% for adults). The results also showed that many studies included in the meta-analysis contained serious methodological flaws. For example, many had very small sample sizes (i.e., 10 were case studies with 1 to 2 participants; 11 had sample sizes of 10 or less), only four of the 42 studies used the PCL-R as the objective measure of psychopathy, many treatment programs used approaches that were not supported by evidence, few studies had control groups, few studies followed up their clients post-treatment or used recidivism or violent behavior as an outcome variable, and most relied on clinical impressions to determine treatment effectiveness (Harris & Rice, 2006). Indeed, Salekin (2002) noted the poor quality of much of the research, “Though the studies in the current review may be less than optimal in scientific rigor, their inclusion is considered to be both necessary and important given our current state of knowledge on psychopathy” (p. 106). There were also serious methodological issues with the meta-analysis itself, such as questionable estimates used to determine treatment outcomes of the control groups (Harris & Rice, 2006). All these limitations notwithstanding, Salekin concluded that “therapeutic pessimism with regard to the psychopathy–treatment relation is not warranted” (p. 105).

It is legitimate to question the optimistic conclusion of Salekin’s (2002) review. For lack of well-designed and adequately powered effectiveness studies (D’Silva et al., 2004; Harris & Rice, 2006), a more appropriate and accurate conclusion could have been that the jury is still out on whether psychopathy can be effectively treated. A more recent review by Salekin et al. (2010) provided at least partial support for the treatability of psychopathy. Using 13 studies (eight treatment studies on adult samples and five on child and youth samples) that all used a contemporary and coherent operationalization of psychopathy (e.g., PCL-R) and that employed a contemporary model of intervention, such as cognitive-behavioral therapy, Salekin et al. (2010) investigated treatment outcome in terms of forensically relevant criteria. The authors found that treatment for adults showed low to moderate success with three out of eight studies demonstrating treatment benefits. Treatment of youth appeared to be more promising with six of eight studies showing treatment gains. However, as noted by Olver (2016), not all treatment programs were equally evidence-informed, and some published studies of well-known treatment programs were not included in the review (e.g., Looman et al., 2005), or only the earlier versions (e.g., Seto & Barbaree, 1999) of subsequently updated studies (Barbaree, 2005; Langton et al., 2006) were reviewed. Although there were less than optimal success rates with adults, the authors concluded that “bright line distinctions” regarding the treatability of psychopathic individuals from non-psychopathic individuals cannot be determined at this time (Salekin et al., 2010, p. 235).

14.5 More Recent Treatment Studies

Several quasi-experimental studies conducted in Dutch maximum secure forensic psychiatric hospitals concur with Salekin et al.’s (2010) conclusion that there is no clear-cut evidence for the nontreatability of psychopathic individuals relative to nonpsychopaths. The Netherlands has a long history of treating severely personality disordered offenders deemed diminished responsible for the crimes they committed. Dutch criminal law has permitted mandated treatment and secure confinement of mentally disordered offenders under the so-called TBS order since 1928 (van Marle, 2002). More than two-thirds of the patients committed under TBS legal statute have a PD without a concomitant major mental disorder, in contrast to forensic hospitals in North America (de Ruiter & Trestman, 2007).

Dutch forensic psychiatric hospitals mostly offer cognitive-behavioral treatment with a focus on behavioral chain analysis of the moment-to-moment experience of the individual during the offense, and relapse prevention (e.g., Laws et al., 2000). Chakhssi et al. (2010) investigated change during CBT treatment delivered in forensic psychiatric center De Rooyse Wissel to personality disordered offenders high vs. low on PCL-R diagnosed psychopathy. Seventy-four personality disordered offenders were divided into high-psychopathic and low-psychopathic cases (high-psychopathic traits was defined as PCL-R total score ≥ 26; 26 is the common European cutoff criterion; Cooke et al., 2005). Over a period of 20 months of forensic treatment, all offenders were assessed repeatedly by psychiatric nurses on risk-related behaviors. Group- and individual level analyses showed few significant differences between patients scoring high vs. low on psychopathy, in terms of treatment effect. Both high and low PCL-R scorers showed significant improvements in adaptive social behavior, communication skills, insight and taking responsibility. A subgroup of high PCL-R scorers (22%) got worse on nurse ratings of physical aggression during treatment, whereas none of the low PCL-R patients did. Post hoc analyses did not reveal differences on the four PCL-R facet scores between the psychopathic offenders who deteriorated and those who improved (Chakhssi et al., 2010, p. 675).

A somewhat comparable study was reported by Hildebrand and de Ruiter (2012) who examined change during forensic CBT treatment in 87 forensic patients, all mandated under the TBS-order, with different degrees of psychopathy [a median split (PCL-R = 22) was used to create the two groups]Footnote 3 in another Dutch forensic hospital, the Van der Hoeven Kliniek. The outcome measures used were different from the nurse-rated tool in the Chakhssi et al. (2010) study. Nurse ratings of interpersonal behavior, as well as self-report inventories and the Rorschach Inkblot Method were used upon admission and after 20 months of treatment. Findings showed no significant differences between patients high on psychopathic traits compared to those low on psychopathic traits in degree of change between the two time points on any of the indicators of dynamic risk (e.g., impulsivity, egocentrism, distrustful attitudes and hostility) as measured with self-report, performance- and observation based assessment tools (Hildebrand & de Ruiter, 2012).

Two earlier North-American studies had also shown that psychological treatment may be helpful to psychopathic offenders. For example, in an evaluation of treatment in 871 civil psychiatric patients (Skeem et al., 2002), psychopathy, defined as a score of 18 and higher on the PCL:SV,Footnote 4 did not moderate the effect of treatment involvement and subsequent violence during a post-discharge follow-up of one year. Similar findings were reported for a sample of 156 PCL-R assessed sex offenders. After a 10-year post-treatment follow-up, sex offenders who demonstrated positive therapeutic responses during a cognitive-behavioral program with a relapse prevention component were less likely to recidivate in violent and sexual crimes (Olver & Wong, 2009), regardless of their psychopathy scores.

To date, no randomized controlled trials of treatment effectiveness studies for psychopathy have been published in the literature. Most existing treatment approaches are cognitive-behavioral and focus on reducing psychopaths’ recidivism risk by addressing their antisocial cognitions, teaching them more effective coping skills and enhancing their motivation towards pro-social goals and behaviors (Polaschek & Skeem, 2018).

14.6 Treatments Designed for Psychopathic Offenders

As already mentioned previously, two types of psychopathy treatment can be distinguished: risk-reduction and PD focused. In this section, we will briefly review more recently developed treatment models that reflect these two types of treatment.

14.6.1 Risk-Reduction Approaches: Wong & Hare (2005) Psychopathy Treatment Program (PTP)

Wong et al. (Wong & Hare, 2005; Wong et al., 2012) proposed a two-component model for psychopathy treatment. Component 1 is termed the Interpersonal Component and entails managing the Factor 1 traits as a responsivity factor, while Component 2 is termed the Criminogenic Component and involves treating the criminogenic needs associated with Factor 2, per the risk and need principles of the RNR model. The rationale behind this treatment rests on the assumption that the primary objective in the treatment of psychopathic offenders is to reduce their risk for violence or other serious antisocial behavior.

The PTP is more a strategy for behavioral self-management rather than a cure for psychopathy. Participants of the PTP should be assisted in developing deeper insights into their lifelong psychopathology and to accept the fact that they will require long-term and continuing self-management for most aspects of their lives to keep them from recidivating. Community support upon re-entry into the community is crucial to help them refrain from a return to a criminal lifestyle (Wong & Burt, 2007).

To the best of our knowledge, there are no studies that provide a direct test of the effectiveness of the PTP program. Wong et al. (2012) presented three studies on the effectiveness of two RNR-based risk reduction programs (one for violent offenders and one for sexual offenders) that they deem consistent with the two-component model. The results showed that, for both violent and sexual offenders with significant psychopathic traits, risk reductions assessed during treatment by means of the Violence Risk Scale (Wong & Gordon, 2006), a file-based risk measure, were linked to significant reductions in sexual and violent recidivism after release into society. For one of the studies, the effect of the treatment was only revealed in the severity of reoffending, not in its frequency (Wong et al., 2012).

Sewall and Olver (2019) tested the two-component model in a long-term follow-up study (17.6 years post release) among a sample of 302 sex offenders, who had participated in an 8-month high-intensity sexual violence reduction treatment program, based on CBT and relapse prevention. They conducted many different analyses to examine interaction effects between psychopathy level, treatment completion, and therapeutic change (measured by scores on the Violence Risk Scale-Sexual offense version). High-psychopathy men (PCL-R ≥ 25) had significantly higher rates of treatment noncompletion (30%) compared to low psychopathy men (6%), but they did not show less therapeutic change. The authors also found support for the interpersonal/affective facet as a responsivity factor, in that the Affective facet correlated with decreased treatment progress, although they also found a significant correlation between the Lifestyle facet and treatment noncompletion, which was not predicted by the PTP model. Interestingly, there were no significant differences in sexual recidivism rates as a function of psychopathy and treatment completion status, even after controlling for pretreatment sexual reoffending risk score. Perhaps the most interesting finding from this study was that men who were high in psychopathy, high risk, and showed large therapeutic change, had a modest rate of sexual, but also violent, recidivism (Sewall & Olver, 2019).

Olver (2016) and Wong et al. (2012) do not see a role for personality change in psychopathy treatment: “Attempting to do so would be akin to an attempt to transform these individuals into warm, empathic, considerate beings who experience the normal range and intensities of human emotion. Not only are such attempts likely to fail, but there is little evidence that targeting the psychopath’s personality in treatment is linked to reductions in violence and other forms of recidivism” (Olver, 2016, p. 79).

14.6.2 Personality Disorder Approaches: Galietta & Rosenfeld (2012) Dialectical Behavior Therapy (DBT) for Psychopathy

Galietta and Rosenfeld’s (2012) motivation to develop an adapted version of DBT (Linehan, 1993) for psychopathic patients arose from their observation that effects (in terms of recidivism reduction) of most risk-reduction programs were rather modest (Tyrer et al., 2009). In their opinion, the focus on changing cognitions and behaviors in risk-reduction programs leaves the crucial problem of emotion dysregulation (both over- and underregulation) among psychopathic individuals untreated. Their choice of DBT, as opposed to other treatment models, was based on DBT’s proven effectiveness with borderline PD (Linehan et al., 1999, 2002, see also the recent Cochrane review by Storebø et al., 2020) and the conceptual overlap between borderline PD and psychopathic PD.

In Linehan’s (1993) treatment model, aversive childhood environments are viewed as important in explaining the etiology of borderline PD. Caregivers are seen as “invalidating” their child when they ignore the child’s emotional distress and punish emotional expression and emotionally driven behaviors. Empirical research provides support for the role of childhood trauma in the development of psychopathy: in offender samples, self-reported childhood trauma is associated with higher PCL-R scores (Graham et al., 2012; Kolla et al., 2013; Marshall & Cooke, 1999; Poythress et al., 2006; Weizmann-Henelius et al., 2010). The evidence appears strongest for Factor 2psychopathic traits, but some studies also find associations between childhood trauma and Factor 1 traits, such as blunted affect and lack of empathy (e.g., Graham et al., 2012; Marshall & Cooke, 1999).

Galietta and Rosenfeld (2012) share a number of relevant experiences in adapting DBT for psychopathic patients. First, creating commitment to treatment appears to be a crucial component. Second, a focus on the complete range of emotions, not just anger and hostility, is needed, because they observed that reactive anger was often a secondary response

“to a brief flash of fear or vulnerability, particularly in individuals who have a history of childhood trauma” (p. 328). Similar to DBT for borderline PD, the treatment consists of weekly individual and group sessions, and telephone coaching in between sessions. Tailored, individualized treatment targets are based on a behavioral chain analysis (BCA) of the index offense and if necessary, prior violent offenses. Compared to Linehan’s original model for borderline PD, the skills group is modified to include simpler language, more emphasis on problem recognition and problem solving, and mindfulness techniques to recognize and develop emotions as well as compassion for others. Telephone coaching is much more structured and scheduled in advance, instead of the on-demand set-up for borderline PD. The case study reported by Galietta and Rosenfeld (2012) points at the importance of the therapeutic alliance in fostering treatment motivation and commitment to DBT for psychopathy. The challenging nature of the population requires active emotional and practical support for DBT therapists (Galietta & Rosenfeld (2012).

14.6.3 Schema Therapy for Forensic PD Patients, Including Those with Psychopathy

Schema Therapy (ST) was developed for patients with severe PDs who are considered difficult to treat with traditional cognitive-behavioral therapy (Young et al., 2003). ST builds on the cognitive-behavioral approach developed by Beck et al. (1990), but places more emphasis on the processing of childhood origins of mental health problems, on experiential techniques, on the therapeutic alliance, and on maladaptive coping styles (Young et al., 2003). ST has already shown effectiveness for borderline PD (for a recent review, see Storebø et al., 2020), and Bernstein et al. (2007) designed an ST adaptation for use with forensic patients with severe PD. They stated explicitly that a high PCL-R score is not an exclusion criterion for treatment with ST (Bernstein et al., 2007). The forensic ST model hypothesizes that criminal and violent behavior can be explained by an unfolding sequence of maladaptive schema modes, or moment-to-moment states, that comprise emotions, cognitions, and behavior.

Schema Mode Work is the preferred form of ST with more severe PDs (Young et al., 2003). Young defined 11 maladaptive schema modes, to which Bernstein et al. (2007) added 4 “forensic” modes: Angry Protector Mode, Predator Mode, Conning and Manipulative Mode, and Over-Controller Mode (Obsessive and Paranoid subtypes). Ideally, an individual also has a strong Healthy Adult Mode that is aware of the various maladaptive modes and can moderate and integrate them (Young et al., 2003). Schema Mode Work consists of a set of techniques to help the patient mitigate or eliminate his individual maladaptive Schema Modes, and to develop a stronger Healthy Adult Mode that can assist in meeting basic emotional needs in a more prosocial manner. Similar to DBT, ST has a strong focus on emotion recognition and regulation, but in addition, ST tries to link maladaptive modes (including the emotions that go with them) to failures of early caregivers to meet the child’s basic needs for warmth, guidance and limit setting. With the “limited re-parenting” technique, the ST therapist attempts to meet these thwarted developmental needs within the confines of the therapeutic relationship. Regular supervision and support for the therapists are needed to ensure the quality of ST delivery to forensic patients (Bernstein et al., 2007).

Research suggests that early maladaptive schemas, assessed with self-report, are common in patients with psychopathic traits (Chakhssi et al., 2014a). Keulen-de Vos et al. (2016) tested the underlying theory of forensic ST, which states that offending behavior can be understood as a sequence of maladaptive schema modes. The authors coded schema modes on the basis of descriptions of forensic patients’ (N = 95) offenses in their charts, which typically included statements by the patient as well as victims’ and witnesses’ statements. For the sample as a whole, vulnerable child modes, accompanied by feelings of abandonment or shame, were evident in the events leading up to the offenses, while over-compensatory modes, such as bully and attack and predator modes, had a stronger presence during the offenses. This finding concurs with forensic ST’s view that states involving aggression compensate for contrary emotional states, such as those involving feelings of weakness, fear, humiliation, or helplessness. The associations between schema modes and PCL-Rpsychopathy revealed a number of interesting findings. The bully and attack mode and the conning and manipulative mode were positively correlated with the interpersonal facet. The affective facet showed negative correlations with vulnerable child modes, during events leading up to the offense. The detached self-soother mode, i.e., alcohol or drug use, was positively related to the lifestyle facet, both before and during the offense (Keulen-de Vos et al., 2016).

Preliminary findings of a multicenter randomized clinical trial (RCT) using ST with forensic patients with PD suggests treatment reduces future violence risk and improves the ability to be open and vulnerable during treatment (Bernstein et al., 2012). This paper did not report interaction effects of psychopathy and treatment, however, because of limited sample size. A single case study documented the process of individual Schema Therapy (ST) in a Dutch forensic patient with psychopathic traits (Chakhssi et al., 2014b). The patient had been a victim of extreme physical and emotional abuse as a child and the therapist used different ST techniques in an attempt to alter the resultant maladaptive schema modes of the patient. After the ST treatment, the antisocial modes, such as the predator, bully and attack, and self-aggrandizer modes, were clearly less prevalent than at the beginning. There was also more room for healthy adult modes of being and for vulnerable emotions, although mistrust schemas could still be easily triggered. The case study also showed the patient’s PCL-R total score changed from 27 at baseline to 14 after four years of intensive ST; pre- and post PCL-R ratings were performed by two independent assessors who were not involved in the patient’s treatment. Remarkably, the Affective facet showed the largest change: from 7 to 1; the Interpersonal facet decreased from 4 to 1. This finding, although just an N = 1 result, challenges the notion that the affective and interpersonal ‘core’ of psychopathy is immutable (Olver, 2016).

14.7 The Future of Psychopathy Treatment

We started to work as scientist-practitioners in a forensic psychiatric hospital in 1995. The first author had just spent the first nine years of her career in general outpatient psychiatry, assessing and treating patients with anxiety and mood disorders. We were struck by the lack of evidence-based treatments for forensic psychiatric patients, including those with psychopathic PD, compared to the empirical knowledge base that informed psychiatric treatments for anxiety and depressive symptoms at the time. Considering the social cost of psychopathy to society, both in terms of human emotional and physical suffering, it would appear wise to invest in the development of effective treatments for this disorder. Now, twenty-five years later, science still cannot provide a clear answer to the question of what works for psychopathy. Obviously, conducting effectiveness research within criminal justice environments is a huge challenge; controlled research in these settings is difficult and requires cooperation at many levels (e.g., institutional leadership, treatment staff engagement and supervision, and patient/offender cooperativeness). RCTs are virtually impossible to conduct, because the legal system and ethical considerations make random allocation to treatment vs. no-treatment undesirable.

14.8 Lessons Learned

With these thoughts in mind, we would like to end our contribution with a set of “lessons learned” and “pointers to the future” concerning the treatment of psychopathy:

  • Lesson #1.Psychopathy is not untreatable. For treatment to be effective, it will have to be more tailored than most current “one-size-fits-all” offender (group) treatments and require a longer duration, which includes working through past traumatic experiences and a period of aftercare.

  • Lesson #2. Negatively toned misconceptions about psychopathy and psychopathic behaviors among professionals lead to diminished hope for change. As an example, it is rather widely assumed that high psychopathyoffenders seek treatment to manipulate others and reach desired outcomes, such as early release from detention. However, a recent study tested this idea in a sample of 217 jail inmates and failed to find an association between PCL:SV total, Factor 1 and Factor 2 scores and treatment seeking (including psycho-educational and support groups and substance use treatment while detained (Schrader et al., 2018). As a second example, many authors contend that psychopathy, and Factor 1 in particular, has a negative effect on the therapeutic alliance. However, several studies that examined relationships between psychopathy scores and working alliance scores, as reported by both therapists and clients, do not find significant associations (Polaschek & Ross, 2010; Walton et al., 2018).

  • Lesson #3. Psychopathic patients are able to engage in a therapeutic alliance. To achieve this, therapists need to be nonjudgmental and validate the clients’ thoughts, feelings and behaviors (Chakhssi et al., 2014b; Galietta & Rosenfeld, 2012). Obviously, violent and harmful behavior are never to be validated, but the underlying feelings and thoughts, such as “I felt belittled by that remark” may have validity. In the words of Gullhaugen and Nøttestad (2012): “Empathy may be taught through the process of considering the psychopathic offender’s needs, which herein lays the irony in treatment of psychopathy” (p. 648).

  • Lesson #4. Psychopathic patients are a heterogeneous group. Any type of treatment should start with a thorough assessment of the patient’s offending behavior through methods such as behavioral chain analysis or schema mode sequences. An interesting case study of a male violent offender with a PCL-R score of 38, demonstrated the utility of the Adult Attachment Interview and the Rorschach Performance Assessment System in identifying underlying unresolved loss and trauma, as well as strong denial of fear and vulnerability, counteracted by extreme outbursts of anger (Nørbech et al., 2013).

14.9 Pointers to the Future

  • Pointer #1. Positive, strengths-focused interventions, such as mindfulness meditation, yoga, and aerobic exercise could serve as a positive adjunctive treatment to present-day risk-reduction approaches (for a discussion and underlying rationale, see de Ruiter, 2018). Some of these are already part of therapeutic interventions, such as DBT.

  • Pointer #2. Because RCTs will remain an exception for effectiveness studies in offender treatment, we believe alternative, quasi-experimental designs, including case series analysis, can be helpful in moving the field of psychopathy treatment forward. Of note, not only treatment “successes”, but also treatment “failures” can be informative.

  • Pointer #3. A non-repressive, therapeutic climate is an essential component of any offender rehabilitation program. In forensic ST, it is made explicit that its effectiveness depends on an institutional environment that is sufficiently safe and supportive of the patient’s recovery (Bernstein et al., 2007). In the ST model, a harsh institutional environment would reinforce precisely the kinds of aggressive, maladaptive Schema Modes in forensic patients that ST is attempting to change. A recent, qualitative study of opiate maintenance treatment (OMT) in a Norwegian prison demonstrated that repressive and collective control measures clearly undermined the rehabilitative aims of the OMT (Mjåland, 2015). This will apply a fortiori to offenders high in psychopathy, who are particularly sensitive to feeling controlled, and will likely respond with higher than average levels of resistance and aggression.

14.10 Conclusion

Present-day psychopathy research can be traced back to Hervey Cleckley’s (1941/1988) The Mask of Sanity, and Robert Hare’s operationalization of the disorder in the Psychopathy Checklist and allied instruments. Cleckley’s view that psychopaths were unable to develop an emotional attachment needed for effective psychotherapy and therefore failed to benefit from treatment, lingers on until today. We believe it is time to leave this view behind, given present-day evidence to the contrary. We do not claim treating patients with psychopathic PD is easy, on the contrary, a high level of experience and theoretical sophistication on the part of treatment developers and implementers is needed. The efforts made by the scientist-practitioners in this field, as summarized in the present chapter, will hopefully inspire professionals involved in the treatment of this fascinating disorder.