Send email for updates
Palmer E., Hatcher R., McGuire J. et al.
Substance Use and Misuse: 2011, 46, p. 1072–1080.
Unable to obtain a copy by clicking title? Try asking the author for a reprint by adapting this prepared e-mail or by writing to Dr Palmer at email@example.com.
From the early 2000s cognitive-behavioural group therapy programmes have been relied on to improve the anti-offending record of UK probation services. Now the first independent evaluation of the main programme for substance users has found no impact on reconviction even among offenders who completed the 20 sessions.
Summary Criminal justice strategy in Britain aims to provide evidence-based interventions to convicted offenders to reduce their substance use and offending. The Addressing Substance-Related Offending (ASRO) programme is one such intervention, nationally accredited By the Correctional Services Accreditation Panel in England and Wales. for use with offenders with a major problem of substance use linked to their offending who are being supervised by probation in the community. It adopts a structured, cognitive-behavioural approach to reducing crime by targeting substance use as one underlying factor. Over 20 two-hour group sessions, the programme aims to enhance motivation to change, strengthen self-control, develop strategies to avoid relapse to problem substance use, and encourage lifestyle change to reduce the risk of a return to substance use and offending. Offenders accepted on to the programme should be at medium to high risk of reoffending and their substance use (though now stabilised) should have been related to their offending or likely to increase the risk that they will reoffend.
ASRO has been widely implemented by probation services in Britain, yet has never been evaluated by research teams independent of the criminal justice system. The featured study aimed to remedy this lack by examining the programme's impact on the chances that an offender would later be reconvicted.
The study sampled 319 male offenders serving community sentences supervised by English and Welsh probation services. Of these offenders, 141 had been required by the court as part of their sentence to attend an ASRO programme, of whom 41 (29%) completed it. Another 178 offenders formed a comparison group with substance use problems and similar sentences, but without such a requirement. Typically the offenders were in their late 20s and early 30s and had several previous convictions. Acquisitive crime and motoring offences were most commonly the reasons for their current sentence.
For those ordered on to an ASRO programme, official reconviction records were obtained for the year from the last session they attended. For the comparison group, the year started from the date they started their sentence. It was difficult to match individuals in the ASRO group with comparison offenders, so instead it was decided to adjust reconviction outcomes for the ways the groups differed which were associated with their risk of being reconvicted.
Over the follow-up year 164 – just over half – of all the offenders had been reconvicted, again most commonly for acquisitive crime or motoring offences. At 34%, least likely to be reconvicted were the ASRO completers, most likely (72%) ASRO non-completers; in between at 44% was the comparison group. [Editor's note: across all offenders sentenced to ASROs 61% had been reconvicted versus 44% not sentenced to ASROs.]
However, the three groups of offenders differed in various Offenders who completed their ASRO programmes were significantly older than the comparison group. They also scored significantly lower on a set (known as OGRS2) of demographic and criminal history variables related to risk of reconviction than either offenders who did not complete their ASROs or the comparison group. Compared to non-completers, they had significantly fewer previous convictions. The groups also differed in the types of offences they had been sentenced for. The comparison group were, for example, more likely to be motoring offenders and less likely to have been sentenced for theft or handling stolen property, while ASRO completers were more likely to have been sentenced for a violent offence. ways, so these raw results were adjusted for the offender's age, OGRS2 A set of demographic and criminal history variables related to risk of reconviction. score, number of previous convictions, and the type of offence they had been convicted for. After taking these in to account, whether offenders had been sentenced to an ASRO and, if they had, whether they had completed it, was still significantly associated with risk of reconviction. In summary, ASRO non-completers fared substantially and significantly worse than people who completed the programme or had never been referred to an ASRO. In detail, ASRO completers were significantly and substantially less likely to be reconvicted than non-completers but not significantly less likely than offenders in the comparison group. Offenders sentenced to ASROs but who did not complete their programmes were significantly and substantially more likely to be reconvicted than offenders in the comparison group. Similar results were found for the number of days offenders took to be reconvicted.
In this study 29% of offenders sentenced to an ASRO completed it, similar to completion rates for other such programmes at the time of the study. Compared to offenders not sentenced to an ASRO, offenders who had been and who had completed the programme were no less likely to be reconvicted within the following year. Offenders sentenced to an ASRO but who did not complete it were more likely than either to be reconvicted.
That the non-completers had particularly poor reconviction outcomes is in line with previous evaluations, and with research showing that factors related to non-completion of a therapeutic programme are similar to those related to a return to crime. It seems possible that their deeper involvement with substance use (11% had been sentenced for a drug offence compared to 2% of ASRO completers) detracted from their motivation to address this issue and their engagement with the ASRO programme.
However, the fact that programme completers were not reconvicted significantly less often than comparison offenders contrasts with previous research.
Substance using offenders vary in the nature of their substance use and its link with their offending. It could be that the ASRO programme has been insufficiently targeted and that some types of offenders will benefit more from other programmes. The effectiveness of the programme may also be weakened by the fact that offenders are required to attend as part of their court order and can be returned to court if they do not comply. A recent synthesis of relevant studies of offenders in custody and the community found that interventions which the offender volunteered to participate in had significantly greater effects than those the offender had been required to undergo. In particular, measured in terms of effect size, A standard way of expressing the magnitude of a difference (eg, between outcomes in control and intervention groups) applicable to most quantitative data. Enables different measures taken in different studies to be compared or (in meta-analyses) combined. Based on expressing the difference in the average outcomes between control and experimental groups as a proportion of how much the outcome varies across both groups. The most common statistic used to quantify this difference is called Cohen's d. Conventionally this is considered to indicate a small effect when no greater than 0.2, a medium effect when around 0.5, and a large effect when at least 0.8. in the community voluntary treatment (effect size 0.22) had twice the impact on recidivism as required treatment.
commentary In this the first study to compare the subsequent offending of substance using offenders referred versus not referred to an ASRO programme, the most surprising – and for the programme, most damaging – finding is that even offenders who completed the 20 group sessions were reconvicted over the following year no less often (after other factors had been taken in to account) than comparison offenders. In this comparison there can be no 'excuse' that incomplete (as opposed to possibly inadequate) implementation undermined the programme's impact. The complete programme failed to significantly better a sentence which did not include the programme at all, even though it presumably benefited from the likelihood that offenders who completed were relatively stable and committed to staying out of trouble – a so-called 'selection' effect.For several reasons (primarily to with the validity of the contrast with the comparison group; details below), the study is not a definitive verdict on ASRO, but remains a disappointing failure to find expected positive outcomes among those exposed to the full 20 sessions. Other similar programmes also have an unconvincing research record. Unlike the featured study, some studies have found that completers reoffend less often than comparison offenders. However, the acid test of whether crime is reduced across all offenders ordered in to these programmes has yet to be convincingly passed; details below.
While there is no disguising the seriousness of this challenge to ASRO's anti-crime credentials, the absence even of a selection effect casts doubt over the validity of the comparison between ASRO and non-ASRO offenders. Assuming that the differences between ASRO and non-ASRO offenders truly had been statistically ironed out, even if the programme was totally ineffective, we would have expected the select band of completers to have been reconvicted less often than comparison offenders. That this was not found suggests that despite adjustments made for other factors, the comparison group was not really comparable to ASRO-ordered offenders. It also seems conceivable that the non-completers were more often dependent users of drugs like heroin and crack who might have responded better to fully-fledged addiction treatment required by the court as part of a drug rehabilitation order more on comparability issues below.
About the comparison group we are told they had "substance use problems and ... similar sentences" to those required to attend an ASRO programme. How they were selected and the basis on which it was decided that they shared these similarities with the ASRO offenders is not detailed. If they were from the same probation areas, the question arises what caused them not to be sentenced to an ASRO programme when similar offenders were. Perhaps, for example, their substance use problems were relatively mild or they seemed likely to respond to a basic sentence. If they were from areas which did not offer ASRO programmes, comparability is undermined by differences between areas. It is not specified whether the non-ASRO offenders had been referred to an alternative substance use or anti-offending programme which perhaps happened to perform as well as ASRO, or whether they were sentenced to basic probation only. In the former case, the conclusion might be that ASRO is no less effective than other programmes, in the latter, that it is no better than basic probation supervision. Another source of non-comparability is the follow-up window. Over just a year, it might have made a big difference that the clock started for ASRO completers after they had finished 20 sessions, but for comparison offenders the moment they started their sentences. This presumably meant that the follow-up window was relatively delayed for ASRO completers. It seems likely that had the clock for them also started when their sentence started – meaning the follow-up period would have included the time they were on the programme – their reconviction rate would have been less and possibly significantly bettered the comparison group.
One reason why the ASRO completers were able to finish the programme could have been that they avoided re-arrest, and possibly some non-completers failed to finish because they had been arrested. This would constitute another selection effect favouring the completers. It would also entail a reverse causation – crime causing non-completion, rather than non-completion causing crime. Also, it is not specified whether some of their reconvictions were a consequence of non-completion because they arose from breach of probation conditions rather than new offences. Another possibility alluded to by the authors is that the non-completers were particularly deeply enmeshed in dependent substance use; compared to both other sets of offenders, more had been sentenced for drug offences and more (42%) for theft and handling stolen property, the type of offence typically committed by dependent users in order to fund their drug use.
The ASRO programme is one of several The alternative for higher risk or dependent users is the Offender Substance Abuse Programme (OSAP). accredited to be imposed by courts to address criminogenic substance misuse among offenders being supervised in the community. It is one of a family particularly of cognitive-behavioural programmes which from year 2000 were seen as key elements in the work of the emerging national probation service. Surveyed early in 2008, 16 of the 41 probation authorities in England and Wales said they offered the programme. ASRO seems most often used for offenders whose offending is related to use of drugs like heroin and cocaine, but is also considered suitable for higher risk or dependent drinkers.
Though in theory the panel which (among other programmes) accredited ASRO required evaluation evidence, in practice this was rarely available within the time scale required to meet government implementation targets. Instead it usually accredited programmes on the basis that they embodied the general principles of 'what works', which largely from North American evidence meant cognitive-behavioural therapeutic methods. Evidence on programmes as implemented in the UK derived largely from studies not capable of determining impacts on offending.
Moreover the panel lacked the means and except for a brief period the remit to ensure high fidelity implementation, leaving open the possibility that if programmes failed to dent reoffending, this was not due to their inherent ineffectiveness, but to poor implementation. A study of probation work in 2008 with drink-related offenders found that probation alcohol leads nationally seemed largely unaware of the level of accredited competence of staff involved in alcohol interventions, and in at least two of the six case study areas staff stressed their lack of alcohol-related training. In West Yorkshire, of all the accredited programmes they were involved with, probation staff were most negative about how well the ASRO programme was delivered in terms of adherence to intended structure, concepts and values.
In its 2008–2011 national drug strategy for offenders, the National Offender Management Service referred to research showing that re-offending rates fall by almost 7% for offenders placed on ASRO-type anti-offending programmes. This may refer to an unpublished Home Office evaluation Hollis V. Reconviction analysis of programme data using Interim Accredited Programmes Software (IAPS). London: RDS NOMS Research and Evaluation, 2007. not specific to the ASRO programme and which lacked a comparison group. Instead it compared predicted reconviction rates for offenders referred to programmes like ASRO with their actual convictions. The results appeared generally positive. Compared to a predicted rate of 61%, just 55% of all offenders were reconvicted within two years, while the reconviction rate for those completing a programme was 38% compared with a predicted rate of 51%. Though in the 'right' direction, several features of the design of the research mean the results cannot be relied on as indicating that the programmes reduced offending.
When from year 2000 ASRO-type cognitive programmes for offenders were being rolled out in Britain, an evaluation found no reduction in reconviction rates compared to offenders not placed on these programmes after other variables which might have influenced the findings had been taken in to account. There was however the familiar low level of reconviction among the minority of offenders who had completed the programmes, an effect which might have been due to a number of factors including their motivation to change, ability to do so and their stability, which might have improved their prospects, regardless of the programme they had been allocated to. Among these programmes was the prototype ASRO, trialled on 62 offenders of whom 21% had completed it. Results from the ASRO paralleled those of cognitive programmes in general.
Even when in a controlled study a cognitive programme been found effective, this has not necessarily been maintained in a larger scale roll-out, as found in British prisons. Interventions for offenders are, it has been argued, highly context-specific; what works in one culture at one time may well be ineffective in other settings and at other times.
A review of studies which had randomly allocated offenders to anti-offending programmes found two which had evaluated examples of the family of cognitive skills approaches of which ASRO is an example. These approaches created no statistically significant advantages on measures indicative of drug use or crime.
Thanks for their comments on this entry in draft to Russell Webster, an independent consultant on drugs and crime based in London, and Tim McSweeney of the Institute for Criminal Policy Research at Birkbeck College in London. Commentators bear no responsibility for the text including the interpretations and any remaining errors.
Last revised 12 April 2012. First uploaded
Comment/query to editor
Give us your feedback on the site (one-minute survey)
Open Effectiveness Bank home page
Add your name to the mailing list to be alerted to new studies and other site updates