Types of studies

Randomised controlled trials (RCTs) evaluating pharmacological interventions to reduce, eliminate or prevent relapse and/or criminal activity among drug‐using offenders were included. Relapse was used to refer to individuals who may have returned to an incarcerated setting or been subsequently arrested, who relapsed back into drug misuse or who did both.

Types of participants

Illicit drug‐misusing offenders were included in the review regardless of gender, age, ethnicity or psychiatric illness. Drug use was deemed relevant to any studies that referred to individuals using occasional drugs, being dependent on drugs or known to abuse drugs. We excluded studies that referred to tobacco or alcohol use. Offenders were defined as individuals who were involved in the criminal justice system. Offenders were those in police custody, being processed by the court system, residing in secure establishments (e.g. special hospitals, prisons or jails) or based in the community (i.e. under the care of probation or parole services).

Types of interventions

Included interventions were designed wholly or in part to eliminate or prevent relapse to drug use, criminal activity or both among participants. Different types of interventions were included in the review:

Experimental interventions included in the review:

• Methadone.

• Naltrexone.

• Buprenorphine.

• Diamorphine.

• Combined pharmacological and counselling interventions.

• Pharmacological treatments for cocaine and amphetamine dependence.

Control interventions included in the review.

• No treatment.

• Minimal treatment.

• Waiting list.

• Treatment as usual.

• Other treatment (e.g., pharmacological or psychosocial)

Types of outcome measures

Electronic searches

Electronic searches

The update searches identified records from 2004 to March 2013.

• MEDLINE (1966 to March 2013).

• EMBASE (1980 to March 2013).

• CENTRAL (1980 to March 2013).

• PsycINFO (1978 to March 2013).

• Pascal (1973 to November 2004)^a.

• SciSearch (Science Citation Index) (1974 to March 2013).

• Social SciSearch (Social Science Citation Index) (1972 to March 2013).

• ASSIA (1987 to March 2013).

• Wilson Applied Science and Technology Abstracts (1983 to October 2004)^a.

• Inside Conferences (1993 to November 2004)^a.

• Dissertation Abstracts (1961 to October 2004)^a.

• NTIS (1964 to March 2013).

• Sociological Abstracts (1963 to March 2013).

• HMIC (to March 2013).

• PAIS (1972 to March 2013).

• SIGLE (1980 to June 2004)^b.

• Criminal Justice Abstracts (1968 to March 2013).

• LILACS (2004 to March 2013).

• National Research Register (March 2004)^c.

• Current Controlled Trials (December 2009).

• Drugscope (February 2004)－unable to access.

• SPECTR (March 2004)^d.

^aUnable to access further to 2004 search.
^bDatabase not updated since original 2004 search.
^cNo longer exists.
^dNow Campbell Collaboration searched on line.

To update the original review (Perry et al, 2006), the search strategy was restricted to studies that were published or unpublished from 2004 onwards. A number of original databases were not searched for this update (indicated by the key at the end of the database list). Pascal, ASSIA, Wilson Applied Science and Technology Abstracts, Inside Conferences and Dissertation Abstracts were not searched. These databases are available only via the fee‐charging DIALOG online host service. We did not have the resources to undertake these searches. The National Research Register no longer exists, and SIGLE has not been updated since 2005. Drugscope is available only to subscribing members. The original searches were undertaken by Drugscope staff.

Search strategies were developed for each database to exploit the search engine most effectively and to make use of any controlled vocabulary. Search strategies were designed to restrict the results to RCTs. No language restriction was placed on the search results. We included methodological search filters designed to identify trials. Whenever possible, filters retrieved from the InterTASC Information Specialists' Sub‐Group (ISSG) Search Filter Resource site (http://www.york.ac.uk/inst/crd/intertasc/) were used. If filters were unavailable from this site, search terms based on existing filters were used instead.

In addition to the electronic databases, a range of relevant Internet sites (Home Office, National Institute of Drug Abuse (NIDA) and European Association of Libraries and Information Services on Alcohol and Other Drugs (ELISAD)) were searched. Directory web sites, including OMNI (http://www.omni.ac.uk), were searched up until November 2011. The review did not place any language restrictions on identification and inclusion of studies in the review.

Details of the update search strategies and results and of the Internet sites searched are listed in Appendix 1; Appendix 2; Appendix 3; Appendix 4; Appendix 5; Appendix 6; Appendix 7; Appendix 8; Appendix 9; Appendix 10; Appendix 11; Appendix 12; Appendix 13.

Searching other resources

Reference checking

The reference lists of all retrieved articles were scrutinised for further references. Searches of the catalogues of relevant organisations and research founders were also undertaken.

Personal communication

Experts were contacted and were asked about their knowledge of other studies, published or unpublished, relevant to the review.

Selection of studies

Two independent review authors inspected the search hits by reading the titles and abstracts. Each potentially relevant study located in the search was obtained as a full article and was independently assessed for inclusion by two review authors. In the case of discordance, a third independent review author arbitrated. Translation of articles not written in the English language was undertaken by a single review author.

The screening process was divided into two key phases. Phase one used the initial eight key questions reported in the original review (Perry 2006). These included the following.

Prescreening criteria: phase one

• Is the document written in 2004 or later? [If "no," exclude document.]

• Is the document an empirical study? [If "no," exclude document.]

• Does the study evaluate an intervention, a component of which is designed to reduce, eliminate or prevent relapse among drug‐using offenders?

• Are the participants referred by the criminal justice system at baseline?

• Does the study report pre programme and post programme measures of drug use?

• Does the study report pre programme and post programme measures of criminal behaviour?

• Does the study include a comparison group?

• Do the outcome measures refer to the same length of follow‐up for two groups?

After relevant papers from phase one had been identified, phase two screening was performed to identify papers reporting on pharmacological interventions. Criteria included the following.

Prescreening: phase two

• Is the intervention a pharmacological intervention? [if "yes" include document]

Pharmacological interventions (excluding those focused only on alcohol outcomes, only on tobacco outcomes or on both) were implied if the programme was targeted at reducing drug use in a group of individuals or if their use could be ascertained from the background characteristics of the group.

Offenders included individuals residing in special hospitals, prisons or the community (i.e. under the care of the probation service) and offenders who were diverted from court or placed on arrest referral schemes for treatment.

The study setting could change throughout the process of the study. For example, offenders could begin the study in prison but then progress through a work release project into a community setting.

Finally, studies need not report both drug and criminal activity outcomes. If either one of these was reported, the study was included in the review.

Data extraction and management

Data extraction forms were used to standardise the reporting of data from all included studies as potentially relevant. Data were extracted by two independent review authors.

Assessment of risk of bias in included studies

Four independent review authors (AEP, JMG, MM‐SJ, MN) assessed risk of bias in all included studies using risk of bias assessment criteria recommended in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).

The risk of bias assessment for RCTs in this review was performed using the criteria recommended by the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). The recommended approach for assessing risk of bias in studies included in a Cochrane Review involves the use of a two‐part tool that addresses six specific domains, namely, sequence generation and allocation concealment (selection bias), blinding of participants and providers (performance bias), blinding of outcome assessor (detection bias), incomplete outcome data (attrition bias), selective outcome reporting (reporting bias) and other sources of bias. The first part of the tool involves describing what was reported to have happened in the study. The second part of the tool involves assigning a judgement related to the risk of bias for that entry in terms of low, high or unclear risk. To make these judgements, we used the criteria indicated by the Cochrane Handbook for Systematic Reviews of Interventions as adapted for the addiction field.

The domains of sequence generation and allocation concealment (avoidance of selection bias) were addressed in the tool by a single entry for each study.

Blinding of participants, personnel and outcome assessor (avoidance of performance bias and detection bias) was considered separately for objective outcomes (e.g. dropping out, using substance of abuse as measured by urinalysis, relapsing of participants at the end of follow=up, engaging of participants in further treatments) and subjective outcomes (e.g. duration and severity of signs and symptoms of withdrawal, participant self‐reported use of substance, side effects, social functioning as integration at school or at work, family relationships).

Incomplete outcome data (avoidance of attrition bias) were considered for all outcomes except dropping out of treatment, which very often is the primary outcome measure in trials on addiction. See Appendix 14 for details

For studies identified in the most recent search, the review authors attempted to contact study authors to establish whether a study protocol was available.

Measures of treatment effect

For continuous outcome measures a mean difference (MD) with 95% confidence intervals (CI) was used. Higher scores for continuous measures are representative of greater harm. Dichotomous outcomes were presented as risk ratios (RRs) with 95% CIs.

Dealing with missing data

When data were missing from the original publication, review authors attempted to contact the study author via email to obtain missing data.

Assessment of heterogeneity

Heterogenity was assessed using I² and Q statistics.

Data synthesis

The RevMan software package was used to perform a series of meta‐analyses for continuous and dichotomous outcome measures. A random‐effects model was used to account for the fact that participants did not come from a single underlying population. A narrative review were performed to address each of the key questions outlined in the objectives. The narrative tables included a presentation of study details (e.g. author, year of publication, and country of study origin), study methods (e.g. random assignment), participants (e.g. number in sample, age, gender, ethnicity, age, mental health status), interventions (e.g. description, duration, intensity, setting), outcomes (e.g. description, follow‐up period, reporting mechanism), resource and cost information and resource savings (e.g. number of staff, intervention delivery, estimated costs, estimated savings) and notes (e.g. methodological and quality assessment information). For outcomes of criminal activity, data were sufficient to allow the review authors to divide this activity into re‐arrest and re‐incarceration categories.

Subgroup analysis and investigation of heterogeneity

A subgroup analysis of setting (secure versus community) and pharmacological drug type (buprenorphine, methadone and naltrexone) was conducted.

Sensitivity analysis

When appropriate, sensitivity analyses were planned to assess the impact of studies with high risk of bias. Because of the overall high risk of bias of the included studies, this analysis was not conducted.