To compare the effectiveness of psychological treatment for sex offenders, compared with those receiving no treatment or treatment judged to be inadequate or inappropriate.

PsycLIT and the NCJRS Abstracts Database were searched; the search terms were described in the paper. In addition, articles were sought from reference lists and those of review articles in the area. Established researchers in the field were contacted for unpublished or overlooked manuscripts.

Study designs of evaluations included in the review

Any comparative design. The studies were assigned to one of the following six design categories:

random assignment;

incidental treatment;

any treatment attendance (including drop-outs) versus refusers;

treatment completers versus drop-outs;

drop-outs versus refusers;

assignment based on need, e.g. treatment given to those assessed as requiring treatment.

Specific interventions included in the review

Any current treatments being offered or any different cognitive- behavioural treatment delivered since 1980. To be included, the programmes must have provided predominantly psychological treatment (e.g. group therapy, aversive conditioning). The specific interventions assessed were cognitive behavioural (29 studies), behavioural (2 studies), systemic (2 studies), other psychotherapeutic interventions (7 studies) or unknown (2 studies).

Participants included in the review

Studies that compared treatment by psychological interventions (rather than treatment with medication) for sex offenders versus untreated sex offenders were included in the review. Most of the studies focused on adult, male sex offenders, although four studies specifically examined adolescents. Only one study indicated that the sample contained female offenders (less than 5% of the study sample).

Outcomes assessed in the review

The outcomes included recidivism rates of either a general or sexual nature for both treated and untreated sex offenders.

How were decisions on the relevance of primary studies made?

The initial codings were reviewed by one of the committee members. Potentially controversial cases, even when agreement had been reached, were referred to the committee. In the few cases where consensus could not be reached, the final coding was based on the majority opinion of the committee.

One trained rater coded each study using a standardised coding manual. The authors of the studies were contacted for further clarification when needed. Five studies were coded by a committee. In addition, the authors of four of the reports coded all the data except the design classification, which was done by the committee. The authors do not state how many of the reviewers performed the validity assessment.

The authors do not state how the data were extracted for the review, or how many of the reviewers performed the data extraction. Data were extracted on the country of origin, type of treatment, sample size, follow-up period, study design and method of assigning the participants. It is notable that only examples of what data were extracted are given in the paper; there is no definitive listing of all the data extracted.

How were the studies combined?

The studies were combined in a meta analysis.

How were differences between studies investigated?

Variability across the studies was indexed by the Q statistic.

Forty-three studies were included in the review. The total number of participants was 9,454 (5,078 treatment; 4,376 comparison).

The 43 studies examined a total of 5,078 treated sex offenders and 4,376 untreated sex offenders. Averaged across all studies, the sexual offence recidivism rate was lower for the treatment groups (12.3%) than for the comparison groups (16.8%, 38 studies unweighted average). There was considerable variability across the studies, as indicated by a large Q statistic (145.02, d.f.=37, p<0.001). The treatment effect was much stronger in unpublished studies than in published studies; the odds ratios (ORs) were 0.65 (95% confidence interval, CI: 0.52, 0.81, k=17) and 0.95 (95% CI: 0.79, 1.15, k=21), respectively. For general recidivism, the rates were higher (treatment 27.9%, comparison 39.2%, in 30 studies). Across all studies (n=31) treated offenders had significantly lower general recidivism rates than untreated offenders (OR=0.56), but there was substantial variability across the studies (Q=120.08, d.f.=30, p<0.001). The 16 published studies and the 15 unpublished studies showed similar results. The recidivism rates were based on an average 46-month follow-up period. It appears that current treatment, which included cognitive-behavioural interventions (k=13) and/or systemic interventions (k=2), were associated with reductions in both sexual recidivism (17.4 to 9.9%) and general recidivism (51 to 32%). Older forms of treatment operating prior to 1980 appeared to have little effect.

The meta-analysis provided evidence about the overall effectiveness of treatment, but little direction on how to improve current practice. This review did find that the recidivism rates of treated sex offenders were lower than those of untreated sex offenders. The authors believe that the balance of available evidence suggests that current treatment reduces recidivism, but firm conclusions require further research.

The descriptions of the methods used to conduct the review werebrief. The search was limited to only two databases and whilst experts were contacted and handsearches conducted, there remains the possibility that some studies may have been missed. The authors did not clearly describe how or what data were extracted for the review. There was a limited amount of detail on the included studies. The analysis was appropriate and the authors adequately addressed differences in the results by recourse to study design. However, differences between the types of intervention were not explored. Despite the limitations in reporting, the authors did present a balanced view of the evidence.

Practice: The authors state that this review provided little direction on how to improve current practice. However, they also noted the need for clinicians to collect data that enables the systematic review of treatment regimes.

Research: The authors state that further research is needed to make reliable distinctions between types of treatment and types of offenders.

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.