Does the adapted sex offender treatment programme reduce cognitive distortions? A meta-analysis

Chris Patterson (Bangor University, Bangor, UK)

Journal of Intellectual Disabilities and Offending Behaviour

ISSN: 2050-8824

Publication date: 12 March 2018



For the 3.8 per cent of people with intellectual disabilities (IDs) who have offended sexually, the main form of treatment is the group-based, cognitive-behavioural, adapted sex offender treatment programme (ASOTP) that focusses on challenging cognitive distortions condoning sex offending. The purpose of this paper is to provide an evaluation of how effective the ASOTP is at reducing ID sex offenders’ cognitive distortions.


Three databases were searched systematically: PsycINFO, MEDLINE and Web of Science. Six studies met the inclusion criteria, yielding 118 participants. Using a random-effects model, effect sizes were calculated using pre- and post-treatment scores on a measure of cognitive distortions. The standardised mean difference (SMD) was 1.77 (95 per cent CI: 1.06, 2.46), which was statistically significant (p<0.001) and “large”. Sensitivity analysis demonstrated that this SMD was robust, and a check for publication bias revealed that it was unlikely that the “file drawer problem” confounded the meta-analysis.


These results indicated that the ASOTP can significantly reduce ID sex offenders’ cognitive distortions, regardless of treatment length, IQ level, language abilities, or offence type. Consistent with earlier reports, longer treatment resulted in the greatest reductions: the optimum treatment length was 24 months.

Research limitations/implications

The ASOTP’s current evidence is comprised wholly of case and quasi-experimental studies, none of which employed control groups. This paper highlights how there is a dire need for high-quality experimental evaluation of the ASOTP.

Practical implications

Clinicians are advised to continue using the ASOTP as the main treatment for ID sex offenders until the effectiveness of the ASOTP is further examined using randomised controlled trials.


This is the first meta-analytic review of the effectiveness of the ASOTP.



Patterson, C. (2018), "Does the adapted sex offender treatment programme reduce cognitive distortions? A meta-analysis", Journal of Intellectual Disabilities and Offending Behaviour, Vol. 9 No. 1, pp. 9-21.

Download as .RIS



Emerald Publishing Limited

Copyright © 2018, Emerald Publishing Limited


In the early 1990s, the sex offender treatment programme (SOTP) – group-based cognitive-behavioural treatment for sex offenders – was piloted across a number of prisons in the UK. Prior to this, sex offender treatment was dominated by behavioural interventions, such as aversion therapy (to decrease inappropriate sexual arousal) and masturbatory reconditioning (to increase appropriate sexual arousal) (Brown, 2010), which were believed to be unethical (Hall, 1995; Quinsey and Earls, 1990; Laws and Marshall, 2003) and unsupported by evidence (Wakeling et al., 2005). Evidence initially suggested that the SOTP reduced recidivism rates significantly better than earlier behavioural interventions (Hanson et al., 2002). The Ministry of Justice have recently evaluated the impact of their prison-based SOTP, and reported that 10 per cent of treatment completers had committed a further sexual offence after eight years. This was significantly higher than the 8 per cent of matched controls that committed a further sexual offence in the same follow-up period (Mews et al., 2017).

Notably, the SOTP’s inclusion/exclusion criteria stipulated that its recipients must have IQs of 80 or above, as it was believed that individuals with IQs of 79 or below could not benefit from CBT as they lacked the necessary cognitive abilities (Hastings et al., 2013). This criterion, therefore, precluded people with intellectual disabilities (IDs) from accessing the SOTP. In the late 1990s, however, evidence emerged indicating that people with ID could benefit from CBT as long as necessary adjustments were made to accommodate for their cognitive deficits (Bruce et al., 2010). In line with this, the SOTP was adapted for use with sex offenders with ID (e.g. Lindsay et al., 1998). The evidence-base for the adapted sex offender treatment programme (ASOTP) consists mainly of studies that have small samples and are uncontrolled. Nevertheless, the majority of the studies that have evaluated the ASOTP have reported promising results (Jones and Chaplin, 2017).

Definition of ID

Criteria for a diagnosis of ID are: deficits in intellectual and; adaptive behaviour functioning; beginning before the age of 18 years (DSM-5; American Psychiatric Association, 2013). Impaired intellectual functioning is usually characterised by an IQ score of below 70 on a Wechsler Adult Intelligence Scale (WAIS).

Prevalence of sex offending in the ID population

Due to numerous methodological issues, there is little agreement on the proportion of people with ID who have committed a sexual offence (Lindsay, 2009). In spite of this, the most widely cited prevalence estimate suggests that 3.7 per cent of the ID population have offended sexually (Hayes, 1991). This is slightly less than the proportion of the non-ID population who have committed a sexual offence: 4 per cent (Swanson and Garwick, 1990). Moreover, Lindsay (2011) has reported that 2-11 per cent of sex offenders have an ID. These statistics go some way to disprove the notion that sexual offending is overrepresented in the ID population (Daskalou, 2001).

Typically, when people with ID are sexually abusive, they are infrequently reported to the police (Green et al., 2002). In the rare cases where they are reported, the police tend not to proceed and the individuals continue to live in the community (McCarthy and Thompson, 1997). When the police choose to proceed, there are distinct outcomes for those found to be guilty of a sexual offence (Hayes, 1991): if deemed fit, ID sex offenders can serve a prison sentence; if deemed unfit to serve a prison sentence, ID sex offenders can be diverted to hospital under the Mental Health Act or treated in specialist community services (Murphy et al., 2007).

Theories of sex offending in the ID population

An array of theories has attempted to explain why people with ID commit sexual offences. However, only Finkelhor’s (1984) precondition model of child sex abuse will be expounded in this section, as it indicates a specific factor/variable as being central to the process of sex offending that has been tested empirically. Finkelhor (1984) theorised that there are four preconditions to child sex abuse that must be satisfied in order for the abuse to occur: motivation, overcoming internal inhibitors, overcoming external inhibitors, and overcoming the resistance of the child. Only the second precondition will be elaborated upon here, as the concept of overcoming internal inhibitors has sparked considerable interest in sex offender treatment research and programme development.

Lindsay (2009) identified that cognitive distortions are central to the process of sex offenders overcoming internal inhibitors. In this context, the term cognitive distortion refers to a belief that condones sexually inappropriate behaviour. Lindsay (2009) suggested that there are myriad cognitive distortions that sex offenders might hold, which permit them to overcome their internal inhibitors and to subsequently commit a sexually offensive act (i.e. in the case that they are able to overcome external inhibitors and the resistance of the child as well). For example, a person who is motivated to rape a child (motivation) is likely to be aware that this act is illegal (internal inhibitor); however, they manage to overcome their internal inhibitor with the beliefs (cognitive distortions) that the act will not hurt the child, and the child will “get over it”. Then, if they overcome any external inhibitors and the resistance of the child, the abuse is very likely to occur.

Considering the growing consensus that cognitive distortions are central to the process of sexual offending (Broxholme and Lindsay, 2003; Goodman et al., 2008), ID sex offender programme developers (i.e. Lindsay, 2009) have designed treatment manuals that place great importance upon the process of challenging and restructuring cognitive distortions.


Various clinicians/researchers in the field of ID sex offending have created treatment manuals (e.g. Lindsay, 2009). As highlighted by Lindsay et al. (2011), however, there is little difference in the treatment procedures described in these manuals, as the majority of the cognitive-behavioural approaches to ID sex offender treatment have stemmed from the SOTP approach. As mentioned above, the main modules in the ASOTP include challenging and restructuring cognitive distortions and increasing sexual knowledge and socio-sexual skills.

All treatment manuals recommend conducting treatment in a group format (six to ten members), as this has been shown to be the most effective means of working with sex offenders with ID (Lindsay, 2009). It is widely believed that the main mechanism of change in the ASOTP is the process of group members listening to, challenging and advising each other (Lindsay, 2009). Treatment typically lasts for a minimum of 12 months. Research indicates that there is a linear relationship between treatment effects and treatment length: the outcomes of treatment are increasingly more positive as the length of treatment increases. Sessions are held on a weekly basis and last for approximately two hours, with a half an hour break in the middle of the session. Facilitators are typically highly trained professionals (i.e. forensic and clinical psychologists, clinical nurse specialists) who have extensive experience of working with sex offenders and people with ID. Moreover, facilitators sometimes undergo some preparatory training before commencing the treatment.

Evidence base

Although CBT has been used with ID sex offenders since the late 1990s, there is yet to be a randomised controlled trial (RCT) examining its efficacy. Instead, research investigating the effectiveness of CBT for this population has been limited to studies using non-experimental (e.g. case studies) or quasi-experimental designs. This means that any treatment effects reported hitherto cannot be considered cause-and-effect, unlike if they utilised the gold-standard experimental RCT design. Nevertheless, as captured by a recent systematic review, a number of promising findings have emerged from case studies and non-randomised, non-controlled studies (Jones and Chaplin, 2017). Cognitive behaviour therapy has been shown to “significantly” reduce cognitive distortions and increase both sexual knowledge and victim empathy in ID sex offender populations. As pointed out by Jones and Chaplin (2017), however, CBT’s impact on reoffending, which is arguably the most important indicator of treatment efficacy (McGuire, 2002), has been unclear. This is due to studies using different criteria for reoffending, not statistically analysing reoffending data, and using unreliable records of reoffending.

Due to the lack of high-quality experimental evidence, formal guidance on how to treat sex offenders with ID is currently limited. At present, clinicians are advised to base their treatment of this population on their experience of what works best, and findings from mainstream sex offender treatment literature (Ashman and Duggan, 2008).

At present, there is a paucity of qualitative research exploring the understandings and experiences of individuals participating in ASOTPs. However, Goodman et al. (2008) conducted qualitative research examining the ways in which ID sex offenders viewed their sexual offences pre- and post-treatment. Goodman et al. (2008) reported improvements in denial, victim blaming and responsibility taking. In addition, they reported that participants found an aspect of the treatment highly acceptable: the process of graded offence disclosure. For example, participants were quoted saying “I’m glad I’ve got it off my chest”, “It felt like climbing a mountain”, and “I’m glad I’ve told someone”. Goodman et al.’s (2008) findings support the above-mentioned observation that the main process that facilitates change in the ASOTP is the group listening and offering different perspectives to each other.

Aims and objectives

This meta-analysis aimed to determine how effective the ASOTP (i.e. group-based CBT) is when implemented with sex offenders with ID. The outcome of interest was pre-, post-intervention changes on scales measuring attitudes supporting sex offending (i.e. cognitive distortions). This outcome measure was selected because, as mentioned, existing literature strongly indicates that cognitive distortions are central to the process of sex offending.


This meta-analysis was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.

Literature search

The PsycINFO, MEDLINE and Web of Science databases were searched systematically for articles reporting on the effectiveness of group-based CBT for sex offenders with ID in reducing cognitive distortions. The search terms used were as follows: sex* offend* and cognitive behav* or treatment or therapy and intellectual disabil* or learning disabil* or special needs or mental retardation. At this initial identification stage, exclusion criteria were that articles must have been written in English and published (in a peer-reviewed journal) in the January 1990 to May 2017 period. These searches yielded 749 studies. Three additional studies were identified as a result of searching through the reference lists of key relevant studies. The titles of these 752 studies were screened. Figure 1 exhibits the flow of these studies through this meta-analysis. All studies with irrelevant titles were excluded, as were any duplicates, leaving 621 articles.

Next, the abstracts of the remaining articles were viewed and assessed for relevance. The 589 studies deemed as irrelevant at this stage were excluded. In all, 29 full papers were read, key information was recorded using a data extraction form and methodological quality was evaluated using Downs and Black’s (1998) methodological quality checklist (MQC) adapted by Cahill et al. (2010). The MQC has repeatedly been cited as the most appropriate psychotherapy research quality assessment (Liebherz et al., 2016).

Table I illustrates how each included study fared in relation to the MQCs criteria in the form of percentages. The mean overall quality appraisal criteria percentage was 63.0 per cent. Levels of reporting (83.3 per cent) and internal reliability (80 per cent) were found to be the highest, whereas, internal validity (selection bias) was shown to be the lowest (16.7 per cent). This latter result was unsurprising given that none of the studies included in this meta-analysis utilised a true control group: an issue that is endemic in ID sex offender research. The quality of individual studies ranged from 50.0 per cent (Rose et al., 2002) to 71.9 per cent (Murphy et al., 2010).

To be included in the meta-analysis, studies were required to have utilised group-based CBT, used an assessment of cognitive distortions as an outcome measure (e.g. QACSO), and provided the pre- and post-intervention means, standard deviations and ns for the outcome measure. Unfortunately, one study (i.e. Murphy et al., 2007) was excluded from the analysis at this stage because the authors did not report the standard deviations of their pre- and post-intervention outcome measures. Contact with the lead author of the Murphy et al.’s (2007) study revealed that this study’s participants were the same as those in the Murphy et al. (2010). The lead author of the Murphy et al.’s (2007) study stated that this study should not be included in this meta-analysis.

Analysis procedure

Prior to conducting the main analysis, it was decided that a random-effects model would be applied. This meta-analysis involved calculating the effect sizes and 95 per cent confidence intervals for each of the studies. An overall effect size was then generated with a 95 per cent confidence interval. Next, a sensitivity analysis was conducted to ascertain the robustness of the result; that is, whether the standardised mean difference (SMD) changed markedly when individual studies were removed from the analysis. The I2 heterogeneity statistic was then computed. Each study was assigned a weight, according to the precision of its estimated effect size. Precision refers to within-study error, which is highly associated with sample size: Studies with greater sample sizes are likely to have less error and will thus be given more weight. All statistical analyses were performed using the R statistical computing and graphics software environment (R Development Core Team, 2008).

Study population and characteristics

A total of 124 participants were obtained from 6 studies. Full data could only be obtained for 118 participants, however, due to participant attrition. Notably, Lindsay et al.’s (2011) study included two treatment groups, which were treated as separate studies in the meta-analysis. Sample sizes ranged from 6 (Rose et al., 2002) to 40 (Murphy et al., 2010). All participants were males aged between 19 and 61.

The majority of studies measured participants’ intellectual functioning pre-intervention, using a version of WAIS; mainly WAIS-III. Reported IQs ranged from 49 (Rose et al., 2012) to 83 (Murphy et al., 2010). Craig et al. (2012) and Keeling et al. (2006) reported mean IQs above 70 (73.0 and 71.8, respectively). Additionally, Murphy et al. (2010) stated that although 45 per cent (18/40) of their sample scored above 70 on the WAIS, 91 per cent of their sample had a formal ID diagnosis and all of their participants were involved with ID services.

Adaptive behaviour functioning, receptive language abilities and traits of autism spectrum disorder (ASD) were also assessed. Scores on the Vineland Adaptive Behaviour Scale ranged from 10.1 (Murphy et al., 2010) to 14.0 years (Craig et al., 2012) (age equivalent). British Vocabulary Picture Scale scores ranged from 10.9 (Murphy et al., 2010) to 14.1 years (Craig et al., 2012) (age equivalent). Craig et al. (2012) reported that 38 per cent of their sample met criteria for ASD, and Murphy et al. (2010) identified that nine of their participants had previously been diagnosed with ASD.

Nearly all of the studies included participants who had sexually offended against men, women and children in a variety of ways. However, the participants in Murphy et al.’s (2010) study had not been convicted of a sexual offence, but had displayed sexually abusive behaviour. Participants were all recruited from community (Craig et al., 2012; Lindsay et al., 2011; Murphy et al., 2010; Rose et al., 2002, 2012), secure hospital (Murphy et al., 2010) and prison settings (Keeling et al., 2006).


Group-based CBT was used in all of the studies. Interventions were carried out by consultant and trainee forensic psychologists (Craig et al., 2012), clinical psychologists (Murphy et al., 2010), and therapists (Rose et al., 2012), some of whom were supported by team members from other disciplines (Murphy et al., 2010). Standardised manuals guided treatment in all of the studies. Craig et al. (2012) and Murphy et al. (2010) used manuals based on the work of Sinclair et al. while Lindsay et al. (2011) and Rose et al. (2002, 2012) used manuals developed by Lindsay et al. (1998), Lindsay (2009). The targets for intervention included sex education, cognitive distortions, victim empathy and awareness, relapse prevention, and offence cycles. There was some variability in the length of treatment across studies. For instance, Rose et al.’s (2002) intervention lasted four months, while Lindsay et al.’s (2011) intervention lasted 36 months. In spite of this, the majority of studies were 10 to 14 months long. Apart from Keeling et al.’s (2006) intervention, which included four sessions per week, all other interventions were conducted on a weekly basis. Sessions were always two to two and a half hours long.


Six articles were included in this meta-analysis. As mentioned, Lindsay et al.’s (2011) study included two different intervention groups. As such, the study was treated as two separate interventions in this meta-analysis: “a” and “b” (Lindsay et al., 2011). Table II presents an overview of the selected articles. The effect sizes and accompanying 95 per cent confidence intervals of the included studies are exhibited in Figure 2. It is apparent that all of the CBT ID sex offender treatment groups reduced cognitive distortions, evidenced by their positive effect sizes. Using the random-effects model, the SMD was 1.76 with a 95 per cent confidence interval of 1.06 to 2.46. According to Cohen (1992), a SMD of this magnitude is “large” (>0.8). The null hypothesis, that the mean of these effects is 0, can be rejected, as Z (n=7) = 4.92, p<0.001.

Statistical tests of heterogeneity were calculated across the seven sample estimates (I2=78.0 per cent). This indicates that the included studies were highly heterogeneous, which was not an issue in this meta-analysis, as it was decided a priori to use a random-effects model. Borenstein (2009) recommend the random-effects model when heterogeneity is 75 per cent or above and the fixed-effects model when heterogeneity is below 75 per cent. The funnel plot in Figure 3 illustrates that publication bias is unlikely to significantly influence the results of this meta-analysis, as the majority of the included studies fall symmetrically into funnel shape. A sensitivity analysis was performed to determine how robust the results were. This was accomplished through running the meta-analysis seven times, each time with one of the studies dropped, to establish if the outcome changes markedly when one of the studies was dropped (Borenstein 2009). The 7 overall effect sizes ranged from 1.45 to 1.90, which is relatively robust.


This meta-analysis was the first of its kind in the field, as it pooled together the results from several studies that evaluated ID sex offenders’ scores on a cognitive distortion measure before and after participating in ASOTPs. The results from this meta-analysis were clear: ID sex offenders’ cognitive distortions reduced significantly after completing the ASOTP; the size of this effect was large; the overall effect size was relatively robust; and the possibility of publication bias influencing the results was low.

Comparison with previous studies

Extant literature in the ID sex offender field suggested that group-based CBT could significantly reduce cognitive distortions, increase sexual knowledge and increase victim empathy. Until now, however, the “significant” results/treatment effects reported in these studies had not been pooled together and analysed using meta-analysis. This is likely because, typically, meta-analyses pool together effect sizes from experimental studies (e.g. RCTs), which compare particular treatment methods against control groups. The ASOTP’s evidence-base, however, is comprised of findings from studies using non-experimental (i.e. case studies) and quasi-experimental designs (i.e. non-randomised, non-controlled studies). Therefore, the preferred method for synthesising findings has, thus far, been the systematic review (e.g. Ashman and Duggan, 2008; Jones and Chaplin, 2017).


Assessment of the methodological quality of the studies included in this meta-analysis showed that, on average, studies met 63 per cent of Downs and Black’s (1998) MQC criteria for psychotherapy outcome studies. While the reporting (83.3 per cent) and internal reliability (80.0 per cent) of the included studies was strong, the internal validity was very poor (16.7 per cent). The main reason for this low level of internal validity was the lack of control groups employed: none of the studies included in this analysis included a control group. Without control groups, it is difficult to establish whether treatment effects were a result of the treatment or whether they would have occurred in the absence of the treatment.

As explained by Brown (2010), numerous ethical and methodological issues have prevented researchers from conducting studies using the gold-standard experimental design, the RCT, to examine the effectiveness of the ASOTP. Principally, it has been deemed unethical to randomly allocate sex offenders to either a sex offender treatment group or a control group (Marshall and Marshall, 2007), due to the risk of those assigned to control groups reoffending as a consequence of not receiving potentially effective treatment (Brown, 2010). There are two main methodological issues that make it difficult to randomise ID sex offenders to experimental groups (Brown, 2010). First, the numbers of ID sex offenders requiring treatment are usually small. With small sample sizes it is difficult to ensure that experimental groups are matched on key variables. Second, the way in which, and the reasons why, ID sex offenders enter sex offender treatment makes randomisation to experimental groups unethical. For example, courts and/or tribunals often order ID sex offenders to complete sex offender treatment and make their movement to community settings contingent upon them successfully completing the treatment.

Previously, it has been suggested that the ASOTP becomes increasingly more effective as the length of treatment increases (Lindsay and Smith, 1998). As only six studies/seven interventions were sufficient to be included in this meta-analysis, it seemed meaningless to divide sample into subgroups. However, two subgroup analyses/meta-regressions could have been conducted. First, included studies could have been divided into three subgroups: interventions lasting less than 12 months; interventions lasting 12-24 months; and interventions lasting over 24 months. There was one study that could be included in subgroup 1, five that could be included in subgroup 2, and one that could be included in subgroup 3. Individual effect sizes for the studies in subgroup 1 and subgroup 2 could have been calculated, and compared with a standardised effect size for the five studies in subgroup 2. Second, the studies could have been split to determine whether treatment manual moderated the effect of treatment (i.e. Lindsay, 2009).

It is also important to consider the possibility that the reductions in cognitive distortions might not reflect changes in ID sex offenders’ attitudes that condone sex offending, but changes in their ability to provide socially desirable responses. All of the studies included in this meta-analysis used the QACSO as the main outcome measure for assessing cognitive distortions. The QACSO has an in-built scale for assessing whether respondents are completing the measure in such a way as to appear socially desirable. Future studies should report participants’ scores on this scale to help identify whether improvements in cognitive distortions reflect genuine improvements or not.

Implications for clinical practice

The results from this meta-analysis could have considerable clinical significance, as they demonstrate that the ASOTP can reliably reduce cognitive distortions in the ID sex offender population: positive effect sizes were reported in all of the included studies, regardless of the length of treatment, extent of ID, level of language difficulty, extent of ASD comorbidity, or sex offence type. For example, Craig et al. (2012) reported significant reductions in cognitive reductions after only 12 weeks of treatment, and Rose et al. (2002) reported a positive effect with a sample comprised of sex offenders with moderate ID.

Cognitive distortions have already been shown to be key in the process of sexual offending, and it is believed that risk of reoffending could be reduced as a product of reducing cognitive distortions (Lindsay et al., 1998). However, this link is only implied, as there is yet to be any substantial evidence indicating that reoffending can be reduced through changing cognitive distortions. This is likely because there are issues that complicate the assessment of reoffending after ASOTP participation. As highlighted by Murphy et al. (2010), sexually abusive acts committed by ID sex offenders often go unreported to the police. Yet, studies examining rates of reoffending typically rely upon the criminal justice system for data. Therefore, reoffending estimates put forward by such studies are likely to be inaccurate. As such, this study’s main result cannot be taken as direct evidence that the ASOTP reduces reoffending, which is arguably the most important measure of the ASOTP (McGuire, 2002).

Implications for research

As argued by Quinsey et al. (1993), researchers in this field have an “ethical duty” to determine whether the ASOTP is effective and the only way of accomplishing this is by conducting an RCT. Quinsey et al. (1993) argued that preventable sexual abuse could occur if sex offenders are allowed to re-enter the community after completing the ASOTP, when in fact their risk of reoffending remained high, as a consequence of the treatment being less effective than it was believed to be. While there is an ethical issue associated with randomly allocating ID sex offenders to control groups, the ASOTP needs to be evaluated in a scientifically rigorous way to establish whether it is truly effective: there are real risks associated with offering a treatment that is not known to reduce reoffending.

This meta-analysis did not examine whether improvements in cognitive distortions were maintained post-treatment. This was because, apart from Murphy et al.’s (2010) study, studies included in this meta-analysis did not include follow-up measurement in their designs. Murphy et al. (2010) reported that improvements in cognitive distortions were maintained at six-month follow-up. When designing RCTs to examine the effectiveness of the ASOTP in the future, researchers should measure cognitive distortions at follow-up to provide evidence as to whether improvements can be maintained long term.

Any future RCT should also include a cost-benefit analysis, as it is currently unclear as to how much the ASOTP costs or benefits society. Until then, it will be difficult to make a strong case for/against ASOTPs because, as with any long-term treatment group requiring specialised input from experienced clinicians, the creation and maintenance of these groups will be expensive. Shanahan and Donato (1999) are the only researchers to have conducted a cost-benefit analysis of the financial implications of sex offender treatment. They reported that prison-based SOTP, in Australia, resulted in substantial net economic benefits, ranging from 60,000 to 1,000,000 dollars per 100 successfully treated offenders.

Additionally, there is a need for qualitative research examining group members’ understandings and experiences of the treatment. Research of this kind would give an indicator of the acceptability of the ASOTP to its participants. When RCTs have been conducted, further meta-analyses can be performed. With necessary qualitative research and cost-benefit analyses, this evidence can be used to develop guidelines on the treatment of sex offenders with ID.


PRISMA flow of studies through the meta-analysis

Figure 1

PRISMA flow of studies through the meta-analysis

Forrest plot showing the individual effect sizes and 95 per cent CIs of included studies, as well as the meta-analysis’ standardised mean difference and 95 per cent CI

Figure 2

Forrest plot showing the individual effect sizes and 95 per cent CIs of included studies, as well as the meta-analysis’ standardised mean difference and 95 per cent CI

Funnel plot to illustrate potential publication bias

Figure 3

Funnel plot to illustrate potential publication bias

Percentages of Downs and Black’s criteria addressed by included studies

Quality criteria
Reporting (%) External validity (%) Internal reliability (%) Internal validity – confounding (selection bias) (%) Overall (%)
Craig et al. (2012) 90.1 54.5 80.0 20.0 65.6
Keeling et al. (2006) 90.1 54.5 80.0 0.0 62.5
Lindsay et al. (2011) 72.7 45.5 80.0 60.0 62.5
Murphy et al. (2010) 90.1 63.6 100.0 20.0 71.9
Rose et al. (2002) 63.6 54.5 60.0 0.0 50.0
Rose et al. (2012) 90.1 54.5 80.0 0.0 65.6
Mean 83.3 54.5 80.0 16.7 63.0

Characteristics of studies included in the meta-analysis

Authors and year Study characteristics Intervention Outcome measures Quality rating
Craig et al. (2012) Quasi-experimental
Single treatment group, pre- post-intervention design. Sample was followed up three and six months post-treatment. Sample recruited from UK community and probation services
Mean age: 35.0 years (range 19-65); Mean IQ: 73 (range 67-79); Mean VABS: 14.0 years; Mean BVPS: 14.1 years; 38 per cent of the sample met diagnostic criteria for ASD
Group-based CBT, following Sinclair et al.’s manual for treating sex offenders with ID. Sessions were conducted on a weekly basis for 14 months QACSO
n=14; M=45.21; SD=21.21
n=12; M=19.25; SD=11.56
21 (65.6%)
Keeling et al. (2006) Quasi-experimental
Single treatment group, pre- post-intervention design. Sample recruited from prison services in Australia
Mean age: 35.2 years (range 25-46); Mean IQ: 71.8 (SD=6.8)
Group-based CBT four times a week for 12 months. The treatment approach used was adapted from Feelgood and Cortoni and Kow and Mamoni (2002) QACSO
n=18; M=26.64; SD=15.09
n=11; M=8.27, SD=11.70
20 (62.5%)
Lindsay et al. (2011) Quasi-experimental
Two separate treatment groups: (A) offenders against women, and (B) offenders against children. These two groups were treated as two separate studies in the meta-analysis. Pre- post-intervention design. Sample recruited from community forensic services
Group A: Mean age: 32.7 years; Mean IQ: 65.4
Group B: Mean age: 36.4 years; Mean IQ: 63.2
Group-based CBT, following Lindsay’s (2009) sex offender treatment manual. Sessions lasted two hours and were conducted on a weekly basis. Treatment lasted 36 months QACSO
Group A
n=15; M=33.3; SD=6.5
n=15; M=12.5; SD=4.2
Group B
n=15; M=34.8; SD=8.2
n=15; M=14.1; SD=6.3
20 (62.5%)
Murphy et al. (2010) Quasi-experimental
Single treatment group, pre- post-intervention design. There was also a six-month follow-up. Sample was recruited from 13 sites, community and secure provision, to form two groups
Mean age: 35.3 years; Mean IQ: 68.0 (SD=7.6); Mean VABS: 10.1 years; Mean BVPS: 10.9 years; nine had ASD diagnosis
Group-based CBT, following Sinclair et al.’s approach to sex offender treatment. Sessions were two hours long and conducted on a weekly basis for 12 months QACSO
n=40; M=51.4 SD=20.7
n=38; M=28.0; SD=20.6
23 (71.9%)
Rose et al. (2002) Quasi-experimental. Multiple case study design. Single treatment group, pre- post-intervention design. Sample recruited from specialist community services
Mean age: 32 years (range 17-43)
Mean IQ: 63.3 (range 54-71)
Group-based CBT, following Lindsay’s (2009) approach to sex offender treatment. Sessions were two hours long and conducted on a weekly basis for four months QACSO
n=4; M=73.0; SD=17.3
n=4; M=55.8; SD=15.1
16 (50%)
Rose et al. (2012) Quasi-experimental
Single treatment group, pre- post-intervention design. Sample recruited from the community
Mean age: 39.5 years (range 20-65)
Mean IQ: 58 (49-70)
Group-based CBT, following Lindsay’s (2009) manual for treating sex offenders. Sessions were two hours long and were conducted on a weekly basis. The treatment lasted ten months QACSO
n=16; M=37.3; SD=10.8
n=12; M=23.5; SD=7.9
21 (65.6%)


American Psychiatric Association (2013), Diagnostic and Statistical Manual of Mental Disorders (DSM-5®), American Psychiatric Pub, Washington, DC.

Ashman, L.L. and Duggan, L. (2008), “Interventions for learning disabled sex offenders”, The Cochrane Library, doi:10.1002/14651858.CD003682.pub2.

Borenstein, M. (2009), “Effect sizes for continuous data”, in Cooper, H., Hedges, L. and Valentine, J.C. (Eds), The Handbook of Research Synthesis and Meta-Analysis, Russell Sage Foundation, New York, NY, pp. 221-35.

Brown, S.J. (2010), “An introduction to sex offender treatment programmes and their risk reduction efficacy”, in Harrison, K. (Ed.), Managing High Risk Sex Offenders in the Community, Willan, Cullompton, Devon, available at:

Broxholme, S.L. and Lindsay, W.R. (2003), “Development and preliminary evaluation of a questionnaire on cognitions related to sex offending for use with individuals who have mild intellectual disabilities”, Journal of Intellectual Disability Research, Vol. 47 No. 6, pp. 472-82, doi: 10.1046/j.1365-2788.2003.00510.x.

Bruce, M., Collins, S., Langdon, P., Powlitch, S. and Reynolds, S. (2010), “Does training improve understanding of core concepts in cognitive behaviour therapy by people with intellectual disabilities? A randomized experiment”, British Journal of Clinical Psychology, Vol. 49 No. 1, pp. 1-13, doi: 10.1348/014466509X416149.

Cahill, J., Barkham, M. and Stiles, W.B. (2010), “Systematic review of practice – based research on psychological therapies in routine clinic settings”, British Journal of Clinical Psychology, Vol. 49 No. 4, pp. 421-53, doi: 10.1348/014466509X470789.

Cohen, J. (1992), “A power primer”, Psychological Bulletin, Vol. 112 No. 1, pp. 155-9, doi: 10.1037/0033-2909.112.1.155.

Craig, L.A., Stringer, I. and Sanders, C.E. (2012), “Treating sexual offenders with intellectual limitations in the community”, The British Journal of Forensic Practice, Vol. 14 No. 1, pp. 5-20, doi: 10.1108/14636641211204423.

Daskalou, M. (2001), “Life in prisons for sex offenders with learning disabilities”, unpublished manuscript, University College, Northampton.

Downs, S.H. and Black, N. (1998), “The feasibility of creating a checklist for the assessment of the methodological quality both of randomised and non-randomised studies of health care interventions”, Journal of Epidemiology and Community Health, Vol. 52 No. 6, pp. 377-84, available at:

Finkelhor, D. (1984), “Four preconditions-a model”, in Finkelhor, D. (Ed.), Child Sexual Abuse: New Theory and Research, The Free Press, New York, NY, pp. 53-68.

Goodman, W., Leggett, J., Weston, C., Phillips, S. and Steward, J. (2008), “Group treatment for men with learning disabilities who are at risk of sexually offending: themes arising from the four – stage model to offending”, British Journal of Learning Disabilities, Vol. 36 No. 4, pp. 249-55, doi: 10.1111/j.1468-3156.2008.00497.x.

Green, G., Gray, N.S. and Willner, P. (2002), “Factors associated with criminal convictions for sexually inappropriate behaviour in men with learning disabilities”, The Journal of Forensic Psychiatry, Vol. 13 No. 3, pp. 578-607, doi: 10.1080/0958518021000019407.

Hall, G.C.N. (1995), “Sexual offender recidivism revisited: a meta-analysis of recent treatment studies”, Journal of Consulting and Clinical Psychology, Vol. 63 No. 5, p. 802, doi: org/10.1037/0022-006X.63.5.802.

Hanson, R.K., Gordon, A., Harris, A.J., Marques, J.K., Murphy, W., Quinsey, V.L. and Seto, M.C. (2002), “First report of the collaborative outcome data project on the effectiveness of psychological treatment for sex offenders”, Sexual Abuse, Vol. 14 No. 2, pp. 169-94, doi: 10.1177/107906320201400207.

Hastings, R.P., Hatton, C., Lindsay, W.R. and Taylor, J.L. (2013), Psychological Therapies for Adults with Intellectual Disabilities, John Wiley & Sons, Ltd, Chichester.

Hayes, S. (1991), “Sex offenders”, Australian and New Zealand Journal of Developmental Disabilities, Vol. 17, pp. 221-7.

Jones, E. and Chaplin, E. (2017), “A systematic review of the effectiveness of psychological approaches in the treatment of sex offenders with intellectual disabilities”, Journal of Applied Research in Intellectual Disabilities, Special issue, pp. 1-22, doi: 10.1111/jar.12345.

Keeling, J.A., Rose, J.L. and Beech, A.R. (2006), “An investigation into the effectiveness of a custody-based cognitive-behavioural treatment for special needs sexual offenders”, The Journal of Forensic Psychiatry & Psychology, Vol. 17 No. 3, pp. 372-92, doi: 10.1080/14789940600658293.

Laws, D.R. and Marshall, W.L. (2003), “A brief history of behavioral and cognitive behavioral approaches to sexual offenders: part 1. Early developments”, Sexual Abuse, Vol. 15 No. 2, pp. 75-92, doi: 10.1177/107906320301500201.

Liebherz, S., Schmidt, N. and Rabung, S. (2016), “How to assess the quality of psychotherapy outcome studies: a systematic review of quality assessment criteria”, Psychotherapy Research, Vol. 26 No. 5, pp. 573-89, doi: 10.1080/10503307.2015.1044763.

Lindsay, W.R. (2009), The Treatment of Sex Offenders with Developmental Disabilities: A Practice Workbook, John Wiley & Sons, West Sussex.

Lindsay, W.R. (2011), “People with intellectual disability who offend or are involved with the criminal justice system”, Current Opinion in Psychiatry, Vol. 24 No. 5, pp. 377-81, doi: 10.1097/YCO.0b013e3283479dc9.

Lindsay, W.R. and Smith, A.H. (1998), “Responses to treatment for sex offenders with intellectual disability: a comparison of men with 1 – and 2 – year probation sentences”, Journal of Intellectual Disability Research, Vol. 42 No. 5, pp. 346-53, doi: 10.1046/j.1365-2788.1998.00147.x.

Lindsay, W.R., Neilson, C.Q., Morrison, F. and Smith, A.H. (1998), “The treatment of six men with a learning disability convicted of sex offences with children”, British Journal of Clinical Psychology, Vol. 37 No. 1, pp. 83-98, doi: 10.1111/j.2044-8260.1998.tb01281.x.

Lindsay, W.R., Michie, A.M., Steptoe, L., Moore, F. and Haut, F. (2011), “Comparing offenders against women and offenders against children on treatment outcome in offenders with intellectual disability”, Journal of Applied Research in Intellectual Disabilities, Vol. 24 No. 4, pp. 361-9, doi: 10.1111/j.1468-3148.2010.00615.x.

McCarthy, M. and Thompson, D. (1997), “A prevalence study of sexual abuse of adults with intellectual disabilities referred for sex education”, Journal of Applied Research in Intellectual Disabilities, Vol. 10 No. 2, pp. 105-24, doi: 10.1111/j.1468-3148.1997.tb00012.x.

McGuire, J. (2002), “Integrating findings from research reviews”, Offender Rehabilitation and Treatment: Effective Programmes and Policies to Reduce Re-Offending, pp. 3-38.

Marshall, W.L. and Marshall, L.E. (2007), “The utility of the random controlled trial for evaluating sexual offender treatment: the gold standard or an inappropriate strategy?”, Sexual Abuse: A Journal of Research and Treatment, Vol. 19 No. 2, pp. 175-91, doi: 10.1007/s11194-007-9046-3.

Mews, A., Di Bella, L. and Purver, M. (2017), Impact Evaluation of the Prison-Based Core Sex Offender Treatment Programme, Government, Ministry of Justice, available at: (accessed 15 October 2017).

Murphy, G., Powell, S., Guzman, A.M. and Hays, S.J. (2007), “Cognitive – behavioural treatment for men with intellectual disabilities and sexually abusive behaviour: a pilot study”, Journal of Intellectual Disability Research, Vol. 51 No. 11, pp. 902-12, doi: 10.1111/j.1365-2788.2007.00990.x.

Murphy, G.H., Sinclair, N. and Hays, S.J. (2010), “SOTSEC-ID: effectiveness of group cognitive behavioural treatment for men with intellectual disabilities at risk of sexual offending”, Journal of Applied Research in Intellectual Disabilities, Vol. 23 No. 6, pp. 537-51, doi: 10.1111/j.1468-3148.2010.00560.x.

Quinsey, V.L. and Earls, C.M. (1990), “The modification of sexual preferences”, in Marshall, W.L., Laws, D.R. and Barbaree, H.E. (Eds), Handbook of Sexual Assault: Issues, Theories, and Treatment of the Offender, Plenum, New York, NY, pp. 279-95.

Quinsey, V.L., Harris, G.T., Rice, M.E. and Lalumiere, M.L. (1993), “Assessing treatment efficacy in outcome studies of sex offenders”, Journal of Interpersonal Violence, Vol. 8 No. 4, pp. 512-23, doi: 10.1177/088626093008004006.

R Development Core Team (2008), R: A Language and Environment for Statistical Computing, R Foundation for Statistical Computing, Vienna, ISBN 3-900051-07-0, available at:

Rose, J., Rose, D., Hawkins, C. and Anderson, C. (2012), “A sex offender treatment group for men with intellectual disabilities in a community setting”, The British Journal of Forensic Practice, Vol. 14 No. 1, pp. 21-28, doi: 10.1108/14636641211204432.

Rose, J., Jenkins, R., O’Connor, C., Jones, C. and Felce, D. (2002), “A group treatment for men with intellectual disabilities who sexually offend or abuse”, Journal of Applied Research in Intellectual Disabilities, Vol. 15 No. 2, pp. 138-50, doi: 10.1046/j.1468-3148.2002.00110.x.

Shanahan, M. and Donato, R. (1999), The Economics of Implementing Intensive In-Prison Sex-Offender Treatment Programs, Australian Institute of Criminology, Canberra, available at: (accessed 1 June 2017).

Swanson, C.K. and Garwick, G.B. (1990), “Treatment for low-functioning sex offenders: group therapy and interagency coordination”, Mental Retardation, Vol. 28 No. 3, pp. 155-61, available at:

Wakeling, H.C., Webster, S.D. and Mann, R.E. (2005), “Sexual offenders’ treatment experience: a qualitative and quantitative investigation”, Journal of Sexual Aggression, Vol. 11 No. 2, pp. 171-86, doi: 10.1080/13552600412331321323.

Further reading

Lambrick, F. and Glaser, W. (2004), “Sex offenders with an intellectual disability”, Sexual Abuse, Vol. 16 No. 4, pp. 381-92, doi: 10.1177/107906320401600409.

Langdon, P.E. and Talbot, T.J. (2006), “Locus of control and sex offenders with an intellectual disability”, International Journal of Offender Therapy and Comparative Criminology, Vol. 50 No. 4, pp. 391-401, doi: 10.1177/0306624X06287105.

Michie, A.M., Lindsay, W.R., Martin, V. and Grieve, A. (2006), “A test of counterfeit deviance: a comparison of sexual knowledge in groups of sex offenders with intellectual disability and controls”, Sexual Abuse: A Journal of Research and Treatment, Vol. 18 No. 3, pp. 271-8, doi: 10.1007/s11194-006-9019-y.

Corresponding author

Chris Patterson can be contacted at:

About the author

Chris Patterson is a Trainee Clinical Psychologist at Bangor University, Bangor, UK.