Drug and Alcohol Findings home page in a new window Background notes

Review of the Glasgow & Fife drug courts

About the featured study

British drug court studies have been criticised for not implementing what the authors argued would have been a feasible model for a randomised trial. Such a study could recruit a sample of offenders suitable for and willing to be adjudicated by a drug court and randomly select for this option the fraction able to be accommodated by the courts, allocating the remainder to normal proceedings. Without randomisation, differences between the samples will normally hamper interpretation of whether the outcomes were due to these or due to the different adjudication procedures, leaving policymakers unclear whether the increased spending on drug courts was worthwhile.

It is possible to estimate Over the financial years 2005/6 to 2007/8 1677 orders were made across Scotland of which an estimated 23.6% were made by the drug courts amounting to 396 orders by the drug courts and 1281 by other courts. If 47% of the drug court orders were successfully completed this amounts to 186 successes, leaving 393 successes among the 1281 orders not made by drug courts. that if about a quarter of drug treatment and testing orders were made by the two drug courts, then the success rates of those not made by these courts was about 31%. In other words, for every three offenders who completed their orders without the enhancements a drug court, nearly five did so with these enhancements. However, this improved completion rate did not translate in to a lower proportion becoming reconvicted.

The report suggests that the lower reconviction rate among successful completers means that offenders with the resolve to complete their orders also go on to commit fewer crimes, an explanation which highlights the pre-existing motivation (and perhaps also recovery resources) of an offender as the key factor. This fits with the finding that though the courts seem to have been better at 'holding' offenders on the order, this did not result in fewer convictions than ordinary drug treatment and testing orders. The implication is that the experience of being sentenced and supervised by a drug court was less important than what the offenders themselves brought to the process, and the treatment and monitoring imposed by drug treatment and testing orders.

Costs accounted for included only the direct costs of the orders and of running the courts. The adverse cost balance could be reversed if drug courts helped avoid imprisonment. Since imprisonment is the most common response to an offender breaching a drug treatment and testing order, the fact that drug courts reduced breaches and revocations could have meant fewer spells in prison due to this cause. Also, at least one well controlled US study found consistent impacts on arrests and to a lesser degree on charges but not on convictions, the yardstick used by the featured study. It could be that this in some circumstances crime reduction impacts are real but are not reflected in conviction statistics.

Drug courts in England

Conducted in 2006 and 2007, a study

The report identified no critical fault lines in the implementation of the courts. However, these were particularly promising sites: the Leeds court built on a pre-existing system and in London court staff were enthusiastic about the proposal and had already been working towards creating a drug court. Nevertheless, offender throughput was lower than expected. Over the 17 months of the evaluation, the London court sentenced just 60 new offenders while in Leeds the total was 276. Low throughput raised costs per offender. Compared to a standard 12-month drug rehabilitation requirement Successor to the drug treatment and testing order. order implemented through normal adjudication, supervising the order through the drug courts cost £4633 more per offender. This is just the extra cost of the drug court elements. The report gave no indication of how much success would be needed to match the total costs incurred by the criminal justice system in implementing all the elements of a drug court-supervised drug rehabilitation requirement order.

Drug courts internationally

Reporting in 2005, the US Government Accountability Office assessed studies of US drug courts which featured a comparison group matched to the drug court participants group as closely as possible, or which statistically adjusted for known differences – an attempt to overcome the case-mix differences which hampered other studies. Based on these studies, they concluded that compared to other criminal justice options, drug courts led to fewer arrests and convictions both during and for up to a year after the sentence, though impacts on substance use were unclear. Cost-benefit analyses found net benefits to society due primarily to reduced recidivism and avoided costs to potential victims.

Around the same time a review of a similar set of studies went further by combining their recidivism findings in a meta-analysis. A study which uses recognised procedures to summarise quantitative results from several studies of the same or similar interventions to arrive at composite outcome scores. Usually undertaken to allow the intervention's effectiveness to be assessed with greater confidence than on the basis of the studies taken individually. Compared to normal traditional adjudication, overwhelmingly the studies found various measures of crime (generally arrests) were lower after offenders had been allocated to a drug court. When the findings were amalgamated, drug courts had a modest but statistically significant advantage in respect of all offences and drug offences in particular.

However, most of the 55 comparisons were vulnerable to case-mix differences or other influences which might have accounted for the results rather than any benefit of the drug courts. Methodologically, the strongest comparisons were the five based on random allocation of offenders. Of these, two were compromised by high drop-out. Among the remaining three, drug courts had a small but not statistically significant advantage, meaning that it could not safely be concluded that they were preferable to normal adjudication. In these studies the impact of the courts was appreciably smaller than in studies which did not randomise participants, a possible indication that these other studies did indeed suffer from bias. It was this which led the analysts to conclude that though the findings were consistently positive and persuasive, methodological shortcomings meant they could not be considered convincing evidence of the general effectiveness of drug courts.

The analysis went on to probe what might have led to the substantial variation in outcomes across the studies. For all offences (but not drug offences in particular), courts which predictably levied sanctions for non-compliance by waving prosecution or waving the imposition of a suspended or deferred sentence were significantly more effective than courts without such a clear and predictable model. There was also suggestive evidence that courts which referred to a single treatment provider had better outcomes, possibly because such providers were more likely to use cognitive-behavioural therapies, but also possibly due to a closer relationship between court and provider.

These reports are among several recent analyses amalgamating data on the crime reduction impact of drug courts which found extra reductions relative to alternative procedures. Set against this are concerns that the existence of this sentencing option might lead police to ramp up arrests for less serious drug offences, counter-productively drawing essentially law-abiding people in to the criminal justice system. Another concern is that non-compliance partly due to the intractability of addiction can lead to offenders being imprisoned for longer than would otherwise have been case – ethically questionable, economically costly, and contrary to hopes that drug courts would divert suitable offenders from prison.

More broadly, international experience and research relating to drug courts suggests it is important for courts to emphasise rewards as well as punishments, see offenders frequently enough to apply these swiftly in response to progress, deploy a range of rewards and sanctions short of revocation which are consistently applied, have a strong and sure ultimate sanction when the programme fails, make these consequences absolutely clear to offenders, have rapid access to a range of treatment options, maintain continuity in the judge dealing with the case, and attend to the range of the individual's needs. Willingness to continue despite some initial offending makes the structure imposed by stringent requirements and monitoring a positive feature rather than one which leads most offenders to fail. Consistent judicial supervision, the fact that this forces addicts (back) in to treatment, and drug testing which provides a shared measure of how treatment is progressing, probably all play their part.

The case-mix too is important. High-risk offenders who have more severe anti-social predispositions or a history of not having responded to standard community-based treatment services benefit from the intensive supervision, sanctions and regular hearings typical of drug courts. However, the lives of low-risk offenders who are likely to do well regardless of the intensity of the programme are unnecessarily and ineffectively intruded on by onerous – and expensive – requirements.

Best evidence

The most convincing demonstration that drug courts can exert a lasting anti-crime impact is the trial of the Baltimore City Drug Treatment Court which randomly assigned 235 eligible clients to the court or to normal adjudication. The Baltimore court is broadly typical of US models, providing an alternative to incarceration for drug involved, non-violent offenders. However, most of the offenders it sees are black heroin users, the programme entails intensive probation supervision, and extensive screening conducted by probation eliminates unsuitable candidates such as those with psychopathic tendencies. During the period when subjects were recruited to the study, the court had a manageable caseload and no waiting lists.

The study began in 1997. Until the third year, for outcomes it relied entirely on criminal justice records. Offenders were referred to the drug court by other courts and processed through to the point where they would normally have been offered the option of their case being heard by the drug court. Instead they were asked to agree to randomisation which gave them at least a 50-50 chance of attending the court. Those who agreed were included in the study. In other words, the study tests what happens when offenders fully eligible for the drug court, and who would it seems have accepted that option, are instead (with a few exceptions imposed by the court) processed through normal channels. 139 offenders were randomised to the drug court, 96 to the control condition. Typically they were older (average age 35), black unemployed, single heroin addicts with extensive criminal records. On average they had been arrested a dozen times before and convicted five times. The impression is of a sample with a high risk of re-offending. Drug court offenders all pleaded guilty as a condition of the programme. Judges tended to impose harsher sentences than were imposed on controls, but suspended these in part or whole conditional on compliance with the drug court regimen, providing an incentive to comply.

Key elements of the drug court programme were implemented but not uniformly, reducing the chances for the programme to differentiate itself from normal procedures. Over the year after entering the programme, drug court clients were far more likely than controls to be tested for drugs (78% v 14%) and to be reviewed by the judge (78% v 7%), though this was on average for just under eight times in the year. Far more of the drug court offenders entered addiction treatment programmes (48% v 14%) but still most did not and very few (5% and 7%) entered methadone or residential programmes which can be expected to be the most effective options for this client group. The main way they were distinguished from controls was in entering outpatient treatment, either routine (24%) or intensive (19%). Additional to addiction treatment, each of the eight providers used by the court offered educational opportunities, job and lifeskills training, and housing assistance. Offenders were assigned to the programme which best suited their treatment needs. On average treatment lasted about four months.

Impacts were evident within the first year. During the 12 months following randomisation into the study, 48% of drug court clients and 64% of controls were arrested for new offences and drug court assignees were also arrested or charged less often. The difference was most evident and significant with respect to charges for violent or sex offences; 3% of drug court assignees were charged compare to nearly 12% of controls. However, conviction rates were unaffected – about a third of both groups were reconvicted an average 0.5 times. The drug court's advantage was concentrated among the more serious cases heard in the circuit court; 32% of these drug court clients versus 57% of controls were rearrested. These figures did not take into account the fact that controls were more often sentenced to immediate imprisonment for the original offence and for longer, reducing their opportunity to re-offend. When these differences were taken into account, per free day controls were nearly three times more likely to be re-arrested than drug court assignees.

Over a further year, two years in total, 66% of drug court offenders and 81% of those not assigned to the court had been arrested for new offence. The number of new arrests (1.6 v. 2.3) and new charges (3.1 v. 4.6) were also significantly lower for drug court offenders, and remained so after taking into account the time not free to commit offences due to imprisonment. Results favoured the drug court for the number of charges for every type of crime examined, but the differences were statistically significant only for drug-related crime.

Some benefits remained discernable over the three year span after randomisation. By this time all but a few of the offenders sentenced by the drug court were no longer under its supervision; about 60% had been terminated for non-compliance, but only after on average nearly a year and a half on the programme. Across the three years the average numbers of new arrests (2.3 v. 3.4) and new charges (4.4 v. 6.1) were significantly fewer among drug court offenders, and arrests remained significantly fewer when adjusted for time in prison. The proportion of offenders charged with a drug offence at some time during the three years was also significantly lower (56% v. 68%). All but one of the other measures favoured the drug court, but none to a statistically significant degree and differences were in some cases minor. For example, 78% of drug court offenders had been arrested at some time during the three years versus 88%, and the figures for convictions were 58% and 64%. In respect of numbers of arrests, reductions attributable to the drug court were seen both while offenders were subject to its supervision and after this ended. A sophisticated attempt to disentangle what is was about the drug court which led to fewer arrests found that arrests diminished as offenders were increasingly subject to a package of interventions consisting of addiction treatment, drug testing, and court reviews of progress.

None of the above reports adjusted for the risk that some of the many differences tested for might have seemed significant purely by chance. Neither was a single primary outcome specified in advance. This makes the findings vulnerable to wrongly assessing chance variations as significant impacts. The fact that the outcomes were logically and probably also empirically related makes this less of an issue. Also the consistent direction of the impacts and the low risk of some of these occurring by chance were indicative of real anti-crime benefits. Puzzlingly, in all the reports, despite impacts on arrests and charges, the average number of new convictions was almost identical in drug court and normally adjudicated offenders. Finally, we do not know how many potential participants in the trial refused to join it and whether this left a selected set of participants not typical of all offenders eligible for the drug court.

All the above reports relied entirely on criminal justice records. At the three-year point, for the first time participants were interviewed about their crime and drug use and general welfare. Of the 235 who started the study, 157 were interviewed. Over the past year those assigned to the drug court were significantly less likely to have said they committed the most serious offences and other crime indicators (including likelihood of having been arrested) also favoured the drug courts. Similarly, substance use and related problems were lower, but not significantly so for some of the measures, and only when the sampling took in to account the fact that some offenders randomly allocated to the drug court were actually sentenced by other courts and vice versa. As with criminal record outcomes, impacts were greatest among the more serious cases heard in the circuit court. This study adopted an unusually lenient criterion That the difference would have occurred less than 1 in 10 times by chance and that any negative impacts could be dismissed as a meaningless fluke, which roughly doubles the chance of a significant finding. for accepting a difference as statistically significant but did (as many studies do not) adjust for the risk that some of the many comparisons it made might have seemed significant purely by chance. Just two thirds of the starting sample were reinterviewed but 71% of those still alive.

These three-year interviews were also analysed to discover what might have accounted for the extra impact of the drug court. The results suggested that the combination of drug testing and drug treatment restrained the frequency of multiple-drug use, while judicial hearings restrained both the variety of crimes committed by the offender and the variety of drugs they had used over the three years since joining the study and drug use. These elements had both direct effects (drug testing and hearings attended) and indirect effects (drug treatment). The indirect effects worked by increasing motivation to abstain due to perceived social and criminal justice pressures or influences (which reduces multiple-drug use), and by strengthening the offender's feeling that had been dealt with fairly by the court and other elements of the programme. This study used unusual measures of crime and drug use severity to do with the variety of drugs use or crimes committed or the number of times several drugs were used together. It is unclear what relationship these bear to more usual measured such as severity of dependence, of related problems, or intensity of substance use. Again this study report adopted an unusually lenient criterion That the difference would have occurred less than 1 in 10 times by chance and (it is presumed but not stipulated) that any negative impacts could be dismissed as a meaningless fluke, which roughly doubles the chance of a significant finding. for accepting a difference as statistically significant.

Across the three years drug court and non-drug court offenders had spent on average the same amount of time in prison, but this was due to the drug court which saw less serious cases resorting more often to imprisonment as a sanction for non-compliance. More serious offenders sentenced by the circuit drug court spent on average 54 days less in prison than their normally adjudicated counterparts. Since beneficial impacts were concentrated among these offenders, the implication is that modest use of the prison sanction is optimal. More extensive use of the prison sanction by the lower court almost certainly meant that across both courts there were no savings in costs to the criminal justice system and (because the drug court programme itself is more expensive than normal sentencing) costs were probably higher than normal adjudication.

Do drug courts work outside the USA?

Certainly those involved believe that they work, though reliable research evidence is scarce. A report drafted by (among others) international organisations based in South America has summarised views from 11 countries with drug courts as well as the USA. Not including the USA, the courts had all been established since 2001 and had enrolled over 3800 participants, of whom over 500 had successfully completed their respective programmes. Respondents said courts cut both recidivism and criminal justice costs and contributed to a healthier and safer society. An accompanying volume summarised such evaluation findings as there were. In the first report greatest reliance was placed on findings from Ireland and Australia. The Irish findings came from a preliminary study of short-term outcomes involving just 37 offenders whose offending and substance use declined after entering the programme, data which the evaluators saw as insufficient to determine the court's impacts. The larger and methodologically superior Australian trial is detailed below.

The randomised trial of a drug court which opened in 1999 in Sydney, Australia, was the first such trial outside the USA. It confirmed that these courts also can reduce reoffending compared to normal adjudication in a country whose addiction treatment and criminal justice systems are closer to that of the UK. In the court specialist judges determined and monitored the treatment of drug-related offenders and imposed sanctions and rewards, typically as an alternative to prison. The emphasis was on cooperation between all parties to secure the common objectives of rehabilitation and avoiding prison. Successful programme completion led to discharge of the original offence.

Treatment started with detoxification in prison. When beds were short, applicants were selected at random, producing 309 drug court participants, plus 191 controls who were returned to the referring court for normal adjudication. Most offenders had multiple convictions and had previously been imprisoned. Their re-offending (theft and drug prosecutions) was tracked for between three and 20 months. By the end, just 12 (4%) of the drug court sample had completed the programme, half were still in it, and 4 in 10 had been removed, disproportionately offenders with short (under six months) sentences hanging over them.

Despite having more opportunity (they spent less of the follow-up in custody), drug court participants did not offend more often than controls. When not in prison they were significantly less likely to re-offend, especially with respect to the main targeted drug (heroin) and crime (shoplifting). After 300 'free' days, 3% had been prosecuted for opiate use compared to 10% of controls. For shoplifting, over 250 days the corresponding proportions were 9% and 20%. However, for theft in general the differences were small. Fewer offences meant that the drug court programme cost £1752 less to achieve a day free of shoplifting prosecutions and £6778 less per day free of opiate use prosecutions. The gains were usually seen only in participants who had remained on or completed the programme. They had also substantially cut their spending, presumably due to reduced illegal drug use. Removed participants did no better (and often worse) than controls.

Unfortunately the study was driven to use a per day cost when the cost of a typical programme was the more appropriate measure. Prosecution as the key indicator of re-offending leaves the study vulnerable to bias introduced by decisions whether or not to prosecute. Nevertheless the findings echoed the mainly US evidence base and extended it to more serious offenders than those typically eligible for US drug courts. One complication was the unusual prison detoxification start, casting doubt on whether the findings would apply to programmes which conduct treatment entirely outside prison and not necessarily with a detoxification phase.

How effective was the DTTO comparator?

In the featured study, drug treatment and testing orders (DTTOs) imposed by a drug court were compared with those imposed by normal courts. Equivalent reconviction outcomes does not mean the drug court orders were ineffective, just that they were no more effective than normal DTTOs. In turn this raises the issue of the effectiveness of DTTOs themselves relative to less intensive orders such as probation or relative to treatment not ordered by a court.

DTTOs were first imposed at the three English pilot schemes set up in October 1998 and then in pilots in Fife and Glasgow in the year 2000. Methodological issues such as the delay between offence and conviction complicated attempts to assess impacts on crime rates. Figures for the year 12–24 months before the DTTOs started and the corresponding period after they started were less contaminated by these problems. In Scotland, the average number of convictions per offender fell from 2.4 to 1.3 and in England from roughly 4.4 to 3.1. In England the pre to post reduction in the frequency of offending was much greater among offenders sentenced to DTTOs than those sentenced to its predecessor, which coupled probation with a treatment requirement. Though suggestive of improved effectiveness, case-mix differences meant the findings were not conclusive.

Between the pilot areas and between England and Scotland, reconviction rates tended to vary with the rate the orders were successfully completed, suggesting that how a DTTO was implemented was a factor in both. Among the success factors were it seemed less onerous orders offering fewer opportunities for failure, and appropriate access to methadone maintenance rather than an immediate and blanket insistence on abstinence-oriented treatment.

Randomised trials of interventions for problem drug users in criminal justice systems should overcome the case-mix differences which hamper the interpretation of non-randomised trials. Internationally, these offer little support for the monitoring and treatment elements of a DTTO, and some studies find that reimprisonment rates increase due to more intensive supervision. The general lack of clear positive findings may reflect the inadequacy of the research rather than the inadequacy of the interventions.

Whether treatment entered in to as part of a drug treatment and testing order is more or less effective than other routes to treatment was the subject of a study of ten treatment sites in London and Kent in 2003 and 2004. At these sites 89 people started DTTO-ordered treatment and 68 started treatment without it being required by a court. Once case-mix differences had been taken in to account, both sets of clients made substantial and roughly equivalent reductions in their self-reported offending and substance use over the following 18 months. Rather than legal status, age, severity of drug problems at intake, and motivation to overcome these, were influential factors, again suggesting that given the same sort of treatment, the characteristics of the patient are more important than whether they have opted for this treatment by agreeing to a court order or in some other way. This study did however suffer from the fact that at each follow-up point about a third of patients could not be followed up, and even more so from the nearly half who could not be interviewed in the first place. Over two fifths Alex Stevens, Neil Hunt, Tim McSweeney, Paul Turnbull. QCT Europe: Research Bulletin 1, England. December 2004. of offenders sentenced to DTTOs did not enter treatment or did not stay long enough to be interviewed compared to just a quarter of the comparison sample. How these people fared could have tipped the balance of outcomes either way.

A similar study in Southampton and Portsmouth compared 25 offenders starting DTTOs with 27 new Started treatment within the past month. drug treatment clients who had a history of acquisitive crime but were not on DTTOs, though they may have been under other forms of criminal justice supervision. Researchers surveyed them at the start of the study and a year later, when substance use had diminished significantly more in the DTTO group but not crime and not other indicators of the extent of drug-related or other problems. This result was reached after adjusting for any differences between the samples on various characteristics, Age, sex, centre treated at (non-DTTO sample only), duration of drug use, duration of current episode of drug use, service contact and number of attempts to give up substance misuse. adjustments which eliminated most of the apparently significant differences in outcomes. It seems possible that further adjustments would have eliminated all the outcome differences, especially since two of the three factors found influential in the previous study – severity of drug problems and motivation – were not adjusted for. Despite the adjustments that were made, the possibility remains that case-mix differences accounted for some of the findings. Unlike the previous study there was no basis for thinking that the treatments the two groups had were similar, and the study conflated comparisons between DTTOs and patients under either no or other forms of criminal justice supervision. Differences and similarities between these groups may have been due not to their different legal status but to different treatments.

In the two studies referred to above similar proportions of court-ordered and other patients had previously been in treatment and (in the second of the studies) their service histories were similar. It seems possible that many of the patients diverted in to treatment via the criminal justice system would have entered treatment in any event.

Comment on these background notes Return to main entry
L10 Web Stats Reporter 3.15 LevelTen Hit Counter - Free PHP Web Analytics Script
LevelTen dallas web development firm - website design, flash, graphics & marketing