Abstract
The aim of this Campbell systematic review is to assess the efficacy of interventions with learning disabled sex offenders.
Sex offending is of increasing public concern with calls for longer terms of imprisonment and closer supervision of such offenders in the community. Currently a variety of treatment approaches are used including medication and talking therapies, though little is known about their success rates.
The small group of sex offenders with learning disabilities pose a particular challenge as talking therapies need to be modified to account for the offender's limited understanding. We could find no randomised controlled trial evidence to guide us in the treatment of learning disabled sex offenders.
Synopsis
Sex offending is of increasing public concern with calls for longer terms of imprisonment and closer supervision of such offenders in the community. Currently a variety of treatment approaches are used including medication and talking therapies, though little is known about their success rates. The small group of sex offenders with learning disabilities pose a particular challenge as talking therapies need to be modified to account for the offender's limited understanding. We could find no randomised controlled trial evidence to guide us in the treatment of learning disabled sex offenders.
Abstract
Background
The management of sex offenders is a major public concern. Behavioural and pharmacological interventions have been used for many years and more recently cognitive behavioural based interventions have become popular around the world. Programmes designed for the general population have been modified for those sex offenders with learning disability, to address their cognitive deficits. The efficacy of these modified programmes is unclear.
Objectives
To determine the efficacy of interventions with learning disabled sex offenders.
The reviewers searched electronically EMBASE, PsycINFO, Medline, Cinahl, Cochrane Library, SPECTR, National Clearinghouse on Child Abuse and Neglect Information and National Criminal Justice Reference Service and Biological Abstracts.
Selection criteria
All randomised controlled trials comparing an intervention for learning disabled sex offenders to any other, or no intervention.
Data collection & analysis
Data were independently extracted.
Main results
No randomised controlled trial was identified.
Reviewers’ conclusions
Using the methods described the reviewers found no randomised controlled trial evidence to guide the use of interventions for learning disabled sex offenders. Until the urgent need for randomised controlled trials is met, clinical practice will continue to be guided by either extrapolation of evidence from randomised controlled trials involving sex offenders without learning disability or non-randomised trial evidence of interventions for the learning disabled sex offender.
Background
For the purpose of this review we use the term ‘learning disability’ to describe those people with a significant impairment of intelligence and social functioning with onset in the first eighteen years of life. This corresponds to ‘mental retardation’ as described in the major taxonomies of DSM IV (APA 1994) and ICD 10 (WHO 1992), other terms include ‘mental handicap and ‘intellectual disabililities’.
Few studies examine the prevalence of sexual offences in learning disabled offenders. In the United Kingdom, (Day 1994) carried out a survey of 47 learning disabled male sex offenders. They had committed a total of 191 sexual offences, 55% were heterosexual offences, 26% were indecent exposure and 12% were homosexual. In addition he reported higher recidivism rates with learning disabled sex offenders. (Hawk 1993) found that the point prevalence rate of sex offence charges were nearly twice as high amongst learning disabled defendants than amongst defendants without a learning disabilty. Cooper argued that people with learning disability were over-represented amongst sexual offenders. He stated that the prevalence of learning disability (including borderline intellectual functioning) in the general population is 9% whereas but that individuals with a learning disability commit 10 to 15% of all sex offences (Cooper 1995).
Further, much of this research on prevalence has relied on data from prison populations. These figures do not take into account individuals diverted from the criminal justice system such as those admitted to hospital with charges not pursued, those found unfit to plead and those whose offending behaviour is never reported to the police. Lyall, Holland and Collins (Lyall 1995) investigated the attitudes of staff and the policies of the services to offending behaviour by learning disabled adults in a community setting. They found that tolerance levels towards offending behaviour were extremely high and that theft and criminal damage were hardly ever reported. Staff in only three of the thirty establishments visited stated a sexual assault or indecent exposure would always be reported if it was to occur. Therefore, it is likely that the point prevalence of sex offences in the learning disabled population is higher than that reported in the literature.
The learning-disabled sex offender is more likely to commit offences against both males and females and is less likely to know their victim than non-learning disabled offenders (Murrey 1992). Sexual naivety, a lack of relationship skills and poor impulse control are prominent features in learning disabled sex offenders (Sellings 1939, Gebhard 1965).
The management of sex offenders has been the subject of much public debate in the Western world with recent examples in Belgium (Guardian 2000b), Megan's Law in the United States of America (New Jersey 1994) and tabloid headlines in the United Kingdom (News of the World, Guardian 2000a). An evidence base for the efficacy of sex offender treatment programmes is beginning to emerge (McKenzie 1999, White 2001). Often these programmes specifically exclude those of below average intelligence.
Interventions can be classified as follows: Pharmacological These treatments include antilibidinals (cyproterone acetate in the United Kingdom and medroxyprogesterone acetate in the United States), antipsychotic mediction (such as benperidol) and more recently introduced Selective Serotonin Reuptake Inhibitors (SSRIs). They are used to reduce sexual drive. Psychological Behavioural interventions are based on the modification of frequency, intensity and salience of deviant behaviours using a variety of methods including counter-conditioning and overcorrection to modify deviant sexual behaviours or preferences. Cognitive behavioural therapy (CBT) aims to teach sex offenders ways of controlling their inappropriate sexual behaviour by systematically identifying and challenging key distorted and permissive thinking patterns known to support sexually aggressive behaviour for example minimisation, justification and normalisation
Psychological treatment programmes of non-learning disabled sex offenders are usually cognitive behaviour therapies employing cognitive restructuring techniques to challenge the distorted cognitions the offender may have about their behaviour (Bremble 1999). Pharmacological and behavioural interventions are most often used with the learning disabled population.
Cognitive behavioural treatment of learning disabled sex offenders is a recent development and is not widely available or standardised. Individuals with a learning disability are likely to have limited reasoning and poor adaptive and verbal skills. They may also have poor concentration skills and low levels of understanding of abstract concepts and inappropriate behaviours (Allam 1997). Most cognitive behavioural treatment programmes have to be modified to compensate for these deficiencies. This includes a breaking down of tasks, more regular repetition of key points, less use of metaphor and greater involvement of key staff in the development of relapse prevention strategies.
Little is known about the effectiveness of these interventions with learning disabled sex offenders. This review examines whether interventions with learning disabled sex offenders reduces the likelihood of future offending.
Objectives
To evaluate the effectiveness of pharmacological (including antilibidinal and psychotropic preparations) and psychological treatments in reducing the target sexual acts, urges and thoughts of learning disabled sex offenders.
Criteria for considering studies for this review
Types of studies
Relevant randomised controlled trials.
Types of participants
Males or females with learning disability (defined as IQ <70) and borderline learning disability, (defined as IQ between 71 and 80), either convicted of a sexual offence or with sexually offensive behaviour. Aged 18 years and older and treated within the community, hospital or prison.
Types of interventions
Cognitive behavioural therapy - in this context we shall define it as an approach, either group or individual, that focuses on teaching skills for the offender to control their sexual behaviours.
Behaviour therapy - in this context means modifying deviant sexual behaviour by behavioural means (and hence sexual offending) such as covert counter-conditioning.
Pharmacological treatment - these reflect the theory that sex offending is result of hormonal drives and thus can be managed by reducing testosterone levels with antilibidinals. Whereas the rationale of using SSRIs is that the behaviour has a compulsive quality. One of the adverse effects of antipsychotic medication is a reduction in libido and for this reason they are prescribed to reduce sexual interest.
All interventions were compared to placebo or ‘standard care’.
Types of outcome measures
The primary outcome measures were: recidivism people lost to follow up. psychometric scores (that measures deviant arousal and pro-offending cognitions)
Other outcomes examined were: death (suicide, all causes); other non-sexual offence; adverse effects; level of security of placement
It was hoped to present outcomes in the short-term (< 1 year), medium term (1–10 years) and long term (> 10 years)
Many rating scales are available to measure outcomes in mental health and criminological trials. These scales vary in quality and many are poorly validated. It is generally accepted that measuring instruments should have the properties of reliability (the extent to which a test effectively measures anything at all) and validity (the extent to which a test measures that which it is supposed to measure). Before publication of an instrument, most scientific journals insist that both reliability and validity be demonstrated to the satisfaction of referees. As a minimum standard, data were excluded from unpublished rating scales. In addition, the rating scale should be either: (i) a self report; or (ii) completed by an independent rater or relative. More stringent standards for instruments may be set in future editions of this review.
Search strategy for identification of studies
Electronic searches Searching references from relevant articles. Contact with pharmaceutical companies marketing antilibidinal medication in the United Kingdom (Schering and Pharmacia). Contact with authors of relevant studies.
Methods of the review
Search strategy for identification of studies
A. Electronic searches
1. Biological Abstracts (Jan 1980 - September 2000) was searched using Silver Platter and the following terms:
[(clin∗ near trial∗) or (singl∗ or doubl∗ or trebl∗ or tripl∗) near (blind∗ or mask∗) or ((randomi∗ or random∗) near (allocat∗ or assign∗) or crossover)]
And
[(mental∗ or intell∗ or learning∗ or cognitive∗) near2 (handi∗ or retard∗ or impair∗ or difficult∗ or disab∗) or (subnormal) or (oligophreni∗) or (phenylketonuria) or (fragile∗ or ((down or down's) near1 syndrome)]
And
[(sex offen∗) or (sex∗ devia∗) or fetish∗ or exhibition∗ or masturbat∗ or voyeur∗ or paedophil∗ or pedophil∗ or child∗ molest∗ or (child∗ sex∗ abuse∗) or pederast∗ or masoch∗ or bondag∗ or sadis∗ or necrophil∗ or or frotteur∗ or necrophil∗]
2. The Cochrane Library (2001 issue 1) was searched using the phrase:
(sex∗ offen∗) or (sex∗ abus∗) or (sex∗ devian∗) or paedophil∗ or pedohil∗ or pederast∗ or rap∗ or fetish∗ or exhibition∗ or necrophil∗ or masturbat∗ or (child∗ molest∗) or (child∗ sex∗ abus∗)
3. EMBASE (1980 - December 2000) was searched using Ovid online and the following terms:
[exp clinical trial/or ((singl? or doubl? or trebl? or tripl?) adj1 (blind? or mask?)).or (random? adj1 (allocat? or assign?)). or (“random?” adj1 (“allocation” or allocated or assign?)). or (exp crossover procedure/)1 (exp randomized controlled trial/) or (exp controlled study/) or (exp double blind procedure/) or (exp controlled study/) or (exp single blind procedure/) or (exp randomization/)]
And
[((mental? or Learning or cognitive?) adj2 (retard? or handicap? or disab? or difficult? or impair?)) or (“subnormal?”) or (exp Learning disorder/) or (exp Mental deficiency/) or (exp intellectual impairment/) or (intellectual impairment)]
And
[exp sexual abuse/) or (exp sexual deviation/) or (sex? adj5 offen?) or (sex? adj5 offen?) or (“paraphilia?”) or (exp sexual crime/) or (“sex offenses”) or (sex? adj5 (crime? or assault? or molest?)) or (sex? adj5 abuse?) or (sex? adj5 inappropriate) or (incest?) or (exhibitionism) or (fetish?) or (masochis?) or (pedophil? or paedophil?) or (sadis?) or (pederast?) or (bondage?) or (frotteur?) or (necrophil?) or (voyeur?) or (masturbat?) or (pornograph?) or (child? adj5 molest?)]
4. MEDLINE (1966 - Dec 2000) was searched using Ovid online and the following terms:
[(randomized controlled trial in pt) or (explode randomized-controlled-trials / all subheadings) or (explode random-allocation / all subheadings) or (explode double-blind-method / all subheadings) or (explode single-blind-method / all subheadings) or (explode clinical trial in pt or clinical-trials / all subheadings) or (clin?adj2 trial?) or ((singl? or doubl? or trebl? or tripl?) adj2 (blind? or mask?)) or ((random? or randomi?) adj2 (allocate? or assign?)) or (controlled clinical trial in pt) or (crossover)]
And
[(explode learning-disorders / all subheadings) or (explode mental-retardation / all subheadings) or (explode developmental disabilities / all subheadings) or ((mental∗ or learning or cognitive?) adj2 (retard? or handicap? or disab? or difficult? or impair?)) or (oligophreni?) or (subnormal?) or ((fragile or down?) adj2 syndrome) or (phenylketonuri?)]
And
[(explode sex offenses/all subheadings) or (“sex offenses “) or (explode paraphilias/ all subheadings) or (explode child abuse, sexual/ allsubheadings) or (explode pedophilia/ all subheadings) or (sex adj2 offen?) or (sex adj2 devia?) or (fetish?) or (exhibition?) or (masturbat?) or (voyeur?) or (paedoph?) or (pedoph?) or (child? adj2 molest?) or (pederast?) or (sadis?) or (masoch?) or (bondag?) or (frotteur?) or (necrophil?) or (pornograph?)]
5. PsycINFO (1887 −2001(02)) was searched on Silver Platter using the following terms:
[randomi∗ or (singl∗ or doubl∗ or trebl∗ or tripl∗ near (blind∗ or mask∗) or (clin∗ near trial∗) or placebo∗ or (placebo- in de) or crossover or (treatment-effectiveness-evaluation in de) or (mental-health-program-evaluation in de) or random∗ near (assign∗ or allocate∗)]
And
[(explode learning-disorders in de) or (explode mental-retardation in de) or (explode “developmental-disabilities”) or oligophreni∗ or (mental∗ or learning or developmental or cognitiv∗) near2 (handi∗ or disab∗ or difficult∗ or disorder∗ or impair∗) or down∗ near2 syndrome or fragile near2 syndrome or phenylketonuri∗]
And
[(explode “sex-offenses”) or (explode “paraphilias”) or (sex near2 offen∗) or (sex near2 devia∗) or fetish∗ or exhibition∗ or masturbat∗ or voyeur∗ or paedoph∗ or pedophil∗ or child∗ near2 molest∗ or (child∗ near2 sex∗ near2 abus∗) or pederast∗ or sadis∗ or masoch∗ or bondag∗ or frotteur∗ or necrophil∗]
6. Cinahl (1982 - December 2000) was searched with Silver Platter using the following terms: [(clin∗ near trial∗) or ((singl∗ or doubl∗ or tripl∗ or trebl∗) and (mask∗ or blind∗)) or ((random∗ near (allocate∗ or assign∗)) or (“random-assignment”/ all topical subheadings / all age subheadings) or (explode “clinical-trials”/ all topical subheadings / all age subheadings) or (explode “meta-analysis”/ all topical subheadings / all age subheadings)]
And
[(explode learning-disorders in / all topical subheadings / all age subheadings in de) or (explode mental-retardation / all topical subheadings / all age subheadings in de) or (explode “developmental-disabilities”/ all topical subheadings / all age subheadings) or ((mental∗ or learning or developmental or cognitiv∗) near2 (handi∗ or disab∗ or difficult∗ or disorder∗ or impair∗)) or oligophreni∗ or (down∗ near2 syndrome) or (fragile near2 syndrome) or phenylketonuri∗]
And
[(explode paraphilias/ all topical subheadings / all age subheadings) or (explode “sex-offenders”/ all topical subheadings / all age subheadings(or (explode “sexual-abuse”/ all topical subheadings / all age subheadings) or (sex∗ near2 offen∗) or (sex∗ near2 devia∗) or (sex∗ near2 crim∗) or (sex near2 (perver∗ or inappropriate∗)) or (child∗ near2 molest∗) or incest∗ or exhibition∗ or fetish∗ or masochis∗ or pedophil∗ or paedophil∗ or masturbat∗ or bondag∗ or frotteur∗ or necrophil∗]
7. Other databases
Database compiled by Adams and Cure (Adams 2000) and SPECTR were searched using the following terms:
((mental∗ or learning or developmental or cognitiv∗) near2 (handi∗ or disab∗ or difficult∗ or disorder∗ or impair∗)) and (sex∗)
Two American databases the National Clearinghouse on Child Abuse and Neglect Information and National Criminal Justice Reference Service were searched online using the term mental retardation.
B. Reference lists
All references of articles selected were searched for further relevant trials.
C. Authors of studies
The reviewers would have contacted authors of studies when necessary to clarify data, and request information on possible additional studies.
D. Pharmaceutical companies
The reviewers contacted Schering HC and Pharmacia Ltd, the pharmaceutical companies that market cyproterone acetate and medroxyprogesterone respectively, in the United Kingdom, to request unpublished data and unpublished trials.
Methods of the review
1. Selection of studies
Two reviewers (LD, LA) independently inspected all reports of identified studies. It was usually possible to resolve any disagreement by consensus; however, where doubt remained the full article was acquired. The reviewers independently decided whether these met the review criteria. No blinding to the names of authors, institutions and journal of publication took place.
2. Assessment of methodological quality
Reviewers (LD, LA) allocated trials to three quality categories, A - adequate concealment, B - concealment unclear and C -inadequate concealment, as described in the Cochrane Collaboration Handbook (Clarke 2001). A decision was taken at the protocol stage to include only trials in category A and B.
3. Addressing publication bias
Data from all selected trials would have been entered into a funnel graph (trial effect against trial size) in an attempt to investigate the likelihood of overt publication bias (Egger 1997).
4. Data extraction
We intended to independently extract data and resolve any disagreement by discussion. If this was not possible we would have sought further information from trial authors.
5. Data synthesis
It was decided that if, for a given outcome, more than 50% of the total numbers randomised were not accounted for, results should not be presented, as such data are impossible to interpret with authority. If however, more than 50% of those in one arm of a study are lost, but the total loss is less than 50%, data will be presented, marked with an asterisk ‘∗’ to indicate the result may be prone to bias.
5.1 Intention to treat analysis
The reviewers intended to analyse data on an intention-to-treat basis where possible and assumed that those who had not been accounted for had the less positive outcome. We tested this assumption with a sensitivity analysis. For continuous data it is impossible to manage the data in this way therefore ‘completer’ data were presented binary data - where possible, the reviewers converted continuous scores to dichotomous data.
5.2 Binary data
In this edition of the review we did not identify any useable data. If, however, data are identified in the future we will analyse binary data where appropriate by calculating the relative risk (RR) statistic with a 95% confidence interval (CI) and use a random effects model. In addition, as a measure of efficiency, we would estimate the number needed to treat (NNT) or the number needed to harm (NNH) from the pooled totals.
5.3 Continuous data
Continuous data may be presented from different scales, rating the same outcome. In this event, the reviewers would have presented all data without summation and inspected the general direction of effect.
Data on continuous outcomes are frequently skewed, the mean not being the centre of the distribution. The statistics for meta-analysis are thought to be able to cope with some skew, but were not formulated for non-parametric data. To avoid this potential pitfall, the following standards would have been applied to all data before inclusion: (i) standard deviations and means were reported or obtained from authors; and (ii) for data with finite limits, such as endpoint scale data, the standard deviation (SD), when multiplied by two, was less than the mean. Otherwise the mean is unlikely to be an appropriate measure of the centre of the distribution (Altman 1996). The reviewers would have reported data that did not meet the first or second standard in the ‘Other data’ tables.
For change data (endpoint minus baseline), the situation is even more problematic. In the absence of individual patient data it is impossible to know if data are skewed, though likely. After consulting the ALLSTAT electronic statistics mailing list, the reviewers would have presented change data in MetaView in order to summarise available information. In doing this, it is assumed either that data were not skewed or that the analyses could cope with the unknown degree of skewness. Without individual patient data it is impossible to test this assumption. If both change and endpoint data were available for the same outcome category we would only present endpoint data.
Where possible, reviewers would have entered data in such a way that the area to the left of the line of no effect indicated a favourable outcome for the treatment.
6. Test for heterogeneity
To test differences between results of trials the reviewers would have inspected the graphical display and used Chi-squared tests of heterogeneity (limit value, p>0.1).
Description of studies
Please see ‘Included and excluded studies’ tables.
Studies were excluded for a variety of reasons. Studies were often not randomised; participants were not diagnosed as having both learning disability and sexually offensive behaviour. Although we identified a single randomised controlled trial that appeared to meet the inclusion criteria Cooper 1992, we excluded it on the grounds that there was no control for the one participant with borderline intellectual functioning.
Methodological quality of included studies
We could find no randomised controlled trials that fulfilled the inclusion criteria.
Results
We could find no randomised controlled trials that fulfilled the inclusion criteria.
Discussion
There is no randomised controlled trial-based evidence for the effectiveness, or ineffectiveness, of any intervention for those sexual offenders with learning disability.
Reviewers’ conclusions
Implications for practice
Clinicians
Professionals in both criminal justice and mental health settings are expected (and often mandated) to offer treatments that reduce recidivism in learning disabled sex offenders. At the present time they cannot base their choice of intervention on randomised controlled trial evidence. Until better evidence is forthcoming, clinicians will have to continue to base practice on clinical experience and evidence from the non-learning disabled population. The courts, recipients of care and carers should be informed of the basis on which an intervention is given.
Criminal justice agencies, recipients of care or their carers
Currently criminal justice agencies, recipients of care or their carers should know that the use of these interventions is based on data relating to non-learning disabled population of sex offenders.
Implications for research
The lack of included trials was not felt to be a result of over-strict inclusion criteria but reflected a genuine dearth of useable material. This review has highlighted the lack of randomised controlled trials of interventions for learning disabled sex offenders.
Trials often exclude the learning disabled population. In the one study where a person with borderline intellectual functioning was included it was impossible to tell the outcome for that individual as the results were not analysed with reference to intellectual functioning.
There is an urgent need for randomised controlled trials of efficacy of interventions for learning disabled sex offenders.
Characteristics of excluded studies