The impact of treatment on reconviction for drug-related offences
Effectiveness bank home page. Opens new windowResearch analysis

This entry is our analysis of a study considered particularly relevant to improving outcomes from drug or alcohol interventions in the UK. The original study was not published by Findings; click Title to order a copy. The summary conveys the findings and views expressed in the study. Below is a commentary from Drug and Alcohol Findings.

Links to other documents. Hover over for notes. Click to highlight passage referred to. Unfold extra text Unfold supplementary text
Copy title and link | Comment/query |

The impact of treatment on reconviction for drug-related offences.

National Treatment Agency for Substance Misuse.
[UK] National Treatment Agency for Substance Misuse, 2012.

England's National Treatment Agency for Substance Misuse finds known offenders substantially less likely to be reconvicted after (re)starting addiction treatment, especially when they stay in treatment or complete it after overcoming their dependence.

Summary This is one of several reports from the National Treatment Agency for Substance Misuse – a special health authority which aims to improve treatment for drug problems in England – presenting a picture of this treatment based largely on data from the National Drug Treatment Monitoring System to which services send information about the people they are treating and the treatments provided. The featured report draws on this and other data to estimate crime reductions arising from treatment in terms of the reduction in conviction rates from before to after entering treatment. The report limited itself to known recent offenders (a minority of the treatment caseload) and to offences which trigger a test for drug use (drug offences and forms of acquisitive crime) and prostitution, 'drug-related' crimes thought most likely to be committed to obtain money to pay for drugs.

Main findings

Of 53,851 adults (re)starting treatment in England in 2006/07 who were either absent from or could reliably be matched to police records, 19,570 had been convicted for one or more drug-related offences committed in the previous two years. Their convictions for offences committed in the next two years enabled an estimate of the reduction in the number of convictions, assumed to reflect (if imperfectly) a reduction in crime consequent on treatment entry.

Within this minority of known recent offenders, convictions for the crimes included in the analysis were on average 26% fewer Except for figures relating to individual types of offences, the reduction estimates were slightly adjusted to cater for the likelihood that some reduction in convictions would have happened due to random fluctuations – so-called 'regression to the mean'. after than before starting treatment. In numbers, the reduction was greatest for theft (other than of a vehicle), because this was by far the crime patients were most likely to have been convicted before (re)starting treatment. Of the different drugs, at 20%, patients in treatment for both their opiate and crack cocaine use registered the smallest reduction in convictions; opiate-only users were the next lowest at 25%. Whether the patient (re)entered treatment via the criminal justice system, by self-referral, or through some other route, made no difference to the proportionate reduction in convictions, though with a higher pre-treatment entry conviction rate, this proportion translated in to a greater drop in the number of offences for criminal justice entrants.

The conviction rate fell most steeply – by 47% – among patients who remained virtually continuously In practice, no more than a three-week break. in treatment over the full two years since (re)starting treatment. Those continuously in treatment for at least a year but less than the full two years had the next greatest reduction at 24%. The degree of reduction then waned only slightly until it bottomed at 17% among patients who had stayed continuously in treatment for under three months. These trends largely reflected trends among patients treated for opiate and/or crack use problems.

Fine tuning this analysis showed that the advantage gained by remaining in treatment was mainly due to the relatively poor record (just a 15% drop in convictions) of patients who left treatment prematurely, that is, before being judged by the service as free of dependence, of the need for addiction treatment, and/or of opiate and crack cocaine use. In contrast, 'successful' treatment leavers who left in a planned way after meeting these criteria reduced their convictions by only slightly less – 41% v. 47% – than patients continuously in treatment for two years. The gap altogether disappeared if the successful leavers had left after at least six months continuous treatment. Especially among problem users of opiates and/or crack cocaine, 'sucessfully' leaving treatment was considerably less successful in terms of conviction reductions if it happened after shorter periods of up to year.

Leaving treatment after successfully completing and not returning later (presumably after relapsing) is being seen as a good indicator that treatment has indeed been successful. In terms of known crime, the featured report supports this assumption. Patients who (as assessed by their treatment service) left free of dependence, of the need for addiction treatment, and/or of opiate or crack cocaine use – the criteria for successful completion – and did not return to treatment within the two years, reduced their convictions by 61% compare to the two years before they (re)started treatment, by far the best record. In contrast, leaving without successfully completing treatment was more likely to be followed by treatment re-entry and among those who did return, the conviction rate barely dropped at all, by just 5%.

The authors' conclusions

The clear trend across the board is a marked reduction in average conviction rates following treatment entry for those identified as recently convicted offenders. This was the case regardless of how individuals were referred into treatment, the substance(s) misused, or length of time spent in treatment, although all these factors were related to the degree of reduction.

Without knowing what would have happened without treatment, it is not possible to say what proportion of these improvements were due to treatment. Probably it played the major part in the process, particularly for patients retained for the whole period or who completed treatment successfully, but other changes in their lives or 'natural recovery' (change which would have happened even without intervention) could account for reduced crime among some patients. Nevertheless, these observations are consistent with earlier studies and reinforce some established conclusions. In particular, the longer a patient remained in treatment, the greater was the reduction in the conviction rate. Sizeable reductions also occurred when a relatively long period in treatment was combined with its ending in successful completion.

Though broadly true for patients as a whole, this pattern appears most applicable to dependent users of opiates (with or without crack cocaine), who represent the majority. Dependent users of other drugs are much less likely to have pre-treatment convictions, but those who do appear to need less time to reap the most benefit, and even seem to benefit significantly without leaving treatment successfully. The few patients who use only crack cocaine appear to benefit earlier than opiate users and to a greater degree. In general, however, patients who successful complete their treatment in the two years after (re)starting register double the reduction in convictions of those who drop out. If their successful completion occurs after at least six months continuous treatment, they appear to reap the same benefit as those retained in treatment for the whole two years.

In interpreting these results, it should be borne in mind that it is entirely possible that the various subgroups of patients were different in ways which affected the observed outcomes. This report is based on observations of treatment, not random allocation which ensures that (for example) patients allocated to a longer time in treatment are the same as those allocated to a shorter time. Discharge from treatment was based on individual clinical need, not random allocation. This means the findings, while not an argument for keeping people in treatment indefinitely, do argue for retaining patients for the optimal period demanded by individual clinical need in order to reap the greatest reduction in convictions.

Importantly whether patients seek treatment because of criminal sanctions, on their own initiatives, or for any other reason, appears to have little influence over the proportionate reduction in convictions.

Findings logo commentary As the report comments, these findings echo those of other studies, particularly DTORS which sampled patients starting treatment for drug problems in England in 2006–2007, and its predecessor NTORS, whose patients started treatment in 1995. In both, crime too fell substantially after treatment started, the main reason why calculations indicated that treatment saved more money for society than it cost. For example, in DTORS, during the four weeks before seeking treatment, 40% of the sample had committed an acquisitive offence (mainly relatively minor), itself probably a reduction on prior offending. Within three to five months this had halved to 21%, then fell by a year to 16%. As in the featured study, the reduction flattened out after about six months in treatment. Even if offending did not stop, on average there was a substantial decrease in its volume and/or the costs associated with it. In DTORS too, crime reductions were equally evident whether or not patients entered treatment via the criminal justice system.

Particularly for opiate/crack users (in practice, overwhelmingly dependent opiate users who may also use crack), the featured study implies that if they have not yet reached the point where they are considered ready to leave treatment, in crime reduction terms it is best to keep them in if possible. Keeping people in treatment for the full two years was generally the way patients and society were protected from continuing convictions and it is presumed crime; nearly twice as many patients were retained as successfully completed. Opiate-addicted patients who for whatever reason left before being considered ready to leave continued to be convicted after entering treatment at almost the same high rate as before.

It should not be assumed that the reduction in convictions seen in the featured study would apply across the treatment caseload. Restricting the analysis to recently convicted patients also confined it to patients whose records could get better. The conviction records of the excluded majority could only stay the same or get worse. Including them ( below) might have halved the average reduction. However, the implications for how treatment should be organised in order to maximise crime reduction would probably remain valid, since crimes and the potential for them to be reduced are concentrated among high-crime offenders (1 2), likely to be among those with a recent conviction included in the featured analysis.

Neither can it be assumed that what changed in the two years before versus after starting treatment in 2006/07, was that the patient was in treatment afterwards, but not before, and that therefore being in treatment was a possible explanation for crime reductions. The fact that around a third of patients who leave treatment re-enter within 12 months makes it inevitable that in any year many patients starting a new treatment journey will have been on just such a journey within the last two years. For these patients, any change in their conviction record after they re-entered treatment in 2006/07 could not be due to treatment entry as such. Below further or more detailed consideration of problems in attributing the observed crime reductions to treatment entry.

The fact that the featured study's findings concur with those of other studies reduces but does not eliminate concerns that limited sampling mean the figures were not representative of the drop in convictions across the treatment caseload. The report was only concerned with the minority of treatment starters who were known recent offenders. Among patients excluded from the analysis, any change in the conviction record could only be for the worse, because their past two-year record was clean when they started treatment. Even if they remained with a clean record, this mass of patients with no improvement in the record would have diluted the overall fall in the conviction rate. How much can be appreciated from another analysis which reflected only patients retained in treatment for two years after starting, but did not confine itself to recently convicted patients. Instead of 47% as in the featured study, the overall fall in convictions was 24%. Adding in patients who were not fully retained in treatment would reduce this figure further – in the featured study (so among only recently convicted patients), by nearly half.

The greater reduction in convictions among patients continuously in treatment could be due to treatment's protective impact and suggest that a break in treatment opens up a window during which crime escalates and conviction is more likely. But it could also reflect a reverse process, of conviction causing a break in treatment. In theory the same could be true of the finding that dropping out of treatment was associated with only a small reduction in convictions, but in this case we are told that 70% of the convictions were for offences committed after the patient had left the treatment system. This suggests that the main causal direction was from unplanned treatment exit to (relative to patients who remained continuously in treatment or who left in a planned way free of dependence) increased crime and with this an increased chance of being convicted. Interestingly, the 70% figure applied also to successful completers; if they are to be convicted, it tends to be for crimes committed after they left treatment. This suggests that even after the service has declared one free of dependence, leaving treatment opens up the opportunity for relapse and increased crime. Complicating this analysis is the fact that the conviction record was limited to certain crimes and in particular did not include offences which breached court orders while offenders were supervised by the probation service. Nearly half (45%) of the offenders in the sample had entered treatment via criminal justice routes and many of these must have been on court orders. Conceivably it was breaching these orders which led some to be forced to leave treatment.

A related issue the report was unable to cater for is the degree to which patients were free to commit and be convicted for crimes over the four years. In particular, some will have been in prison for some of that time. If they were imprisoned after starting treatment they would have been recorded as discharged from treatment, presumably generally as unsuccessful leavers. If this was the case then per year free to commit crimes, the post-treatment entry conviction rate among unsuccessful leavers would be even worse relative to retained patients and successful leavers.

A population selected to be atypical on a variable such as recent convictions will naturally tend to normalise their record over time, creating a reduction which might falsely be attributed to intervention. Additionally, over the four years convictions were tracked, the drug users in the study had aged by this same amount, moving closer on average to the stage in their lives when they will have had enough of the rigours of dependent use of illicit drugs involving for them high levels of crime. As the overall treatment population ages, this is likely to have become more of a factor in trends. In 2005/06, 73,217 were under 30 and 32,406 aged 40 or more. By 2010/11 the under-30s had shrunk to 60,578 but the 40+ groups had expanded to 58,617.

Without making any specific reservations about the featured report, it should also be borne in mind that analysts with an interest in the success of a programme they are evaluating tend to produce more positive analyses than independent analysts – in research parlance, the 'allegiance effect'. In several social research areas,1 programme developers and other researchers with an interest in the programme's success have been found to record more positive findings than fully independent researchers.

1. See articles at the following web addresses:
It is part of the remit of the National Treatment Agency for Substance Misuse to have an interest in the success of addiction treatment in England, to improve this, and to show this has been done by producing reports such as the featured report.

Thanks for their comments on this entry in draft to Tim McSweeney of the Institute for Criminal Policy Research at the University of London. Commentators bear no responsibility for the text including the interpretations and any remaining errors.

Last revised 28 June 2012. First uploaded

Open Effectiveness Bank home page

Top 10 most closely related documents on this site. For more try a subject or free text search

STUDY 2009 The Drug Treatment Outcomes Research Study (DTORS): final outcomes report

STUDY 2012 Estimating the crime reduction benefits of drug treatment and recovery

STUDY 2010 A long term study of the outcomes of drug users leaving treatment

DOCUMENT 2011 Drug treatment and recovery in 2010–11

STUDY 2013 Criminal justice responses to drug related crime in Scotland

STUDY 2013 Drug treatment in England 2012–13

STUDY 2014 Drug treatment in England 2013–14

STUDY 2018 Adult substance misuse statistics from the National Drug Treatment Monitoring System (NDTMS) 1 April 2016 to 31 March 2017

STUDY 2016 Adult substance misuse statistics from the National Drug Treatment Monitoring System (NDTMS) 1 April 2015 to 31 March 2016

REVIEW 2017 An evidence review of the outcomes that can be expected of drug misuse treatment in England