Intervention Review

You have free access to this content

Interventions for drug-using offenders with co-occurring mental illness

  1. Amanda E Perry1,*,
  2. Matthew Neilson1,
  3. Marrissa Martyn-St James2,
  4. Julie M Glanville3,
  5. Rachael McCool3,
  6. Steven Duffy4,
  7. Christine Godfrey1,
  8. Catherine Hewitt1

Editorial Group: Cochrane Drugs and Alcohol Group

Published Online: 2 JAN 2014

Assessed as up-to-date: 10 OCT 2013

DOI: 10.1002/14651858.CD010901


How to Cite

Perry AE, Neilson M, Martyn-St James M, Glanville JM, McCool R, Duffy S, Godfrey C, Hewitt C. Interventions for drug-using offenders with co-occurring mental illness. Cochrane Database of Systematic Reviews 2014, Issue 1. Art. No.: CD010901. DOI: 10.1002/14651858.CD010901.

Author Information

  1. 1

    University of York, Department of Health Sciences, York, UK

  2. 2

    University of Sheffield, School of Health and Related Research (ScHARR), Sheffield, South Yorkshire, UK

  3. 3

    York Health Economics Consortium, York, UK

  4. 4

    University of York, NHS Centre for Reviews and Dissemination, York, UK

*Amanda E Perry, Department of Health Sciences, University of York, Heslington, York, YO105DD, UK. amanda.perry@york.ac.uk.

Publication History

  1. Publication Status: New
  2. Published Online: 2 JAN 2014

SEARCH

 
Characteristics of included studies [ordered by study ID]
Cosden 2003

MethodsAllocation: random assignment
Randomisation method: adequate/low risk
Similar on drug use: unknown/unclear risk
Similar on criminal activity: yes
Blinding methodology: high risk
Loss to follow-up: partial/high risk


Participants235 adults
Age not reported
50.2% male
70.6% European American
Drug use not reported
Alcohol use not reported
100% psychiatric history
Eligibility criteria: adults charged with a crime or misdemeanour who were booked into county jail, had at least 1 prior booking and were diagnosed with a serious and pervasive mental illness and were residents of the county involved. Pre-plea participants were required to have no previous offences involving violence; post-adjudication participants with prior violence were eligible if they were considered to no longer pose a threat


InterventionsCourt-based sentencing and case management intervention vs. treatment as usual
(I) MHTC and ACT case management (n = 137) vs. (C) treatment as usual (n = 98)

(I) received weekly or bi-weekly court supervision and frequent contact with case managers, duration 18 months, followed by treatment as usual if required

C received traditional court proceedings and county mental health services as usual for at least 18 months, which was less intensive than (I)


OutcomesDrug use (Addiction Severity Index, self report) during the last 1 month at 12 months' follow-up


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskRandom number table

Allocation concealment (selection bias)Unclear riskNot reported

Blinding (performance bias and detection bias)
All outcomes
High riskThe interviewer was not blind to the condition of the
client

Blinding of participants and personnel (performance bias)
subjective outcomes
High riskThe interviewer was not blind to the condition of the
client

Blinding of participants and personnel (performance bias)
objective outcomes
High riskThe interviewer was not blind to the condition of the
client

Blinding of outcome assessment (detection bias)
subjective outcomes
Unclear riskNo information available

Blinding of outcome assessment (detection bias)
objective outcomes
Unclear riskNo information available

Incomplete outcome data (attrition bias)
All outcomes
High risk25% of the initial population could not be relocated at the end of 12 months

Selective reporting (reporting bias)Unclear riskNot reported

Other biasHigh riskThe relatively small number of clients in each group resulted in chance variation on some of the intake measures. Generalisability issues and concerns about self report measures and validity

Prendergast 2003

MethodsAllocation: random assignment
Randomisation method: unclear risk
Similar on drug use: yes
Similar on criminal activity: yes
Blinding methodology: unknown/unclear
Loss to follow-up: partial/high risk


Participants715 adults
Mean age 30.9 years
100% male
37.8% white
100% drug-using
100% psychiatric history
Eligibility criteria: male inmates


InterventionsPrison/secure establishment-based TC, counselling and aftercare vs. treatment as usual


OutcomesArrest for any offence (self report), arrest for a drug offence (self report), incarceration for any offence (official records) 12-months post release


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear risk"Random assignment and stratified for equal ethnic proportions". Method not stated

Allocation concealment (selection bias)Unclear riskConcealment method not stated

Blinding (performance bias and detection bias)
All outcomes
Unclear riskNo information reported

Blinding of participants and personnel (performance bias)
subjective outcomes
Unclear riskNo information reported

Blinding of participants and personnel (performance bias)
objective outcomes
Unclear riskNo information reported

Blinding of outcome assessment (detection bias)
subjective outcomes
Unclear riskNo information reported

Blinding of outcome assessment (detection bias)
objective outcomes
Unclear riskNo information reported

Incomplete outcome data (attrition bias)
All outcomes
High risk74% of the original sample were included in the follow-up

Selective reporting (reporting bias)Unclear riskNo information reported

Other biasHigh riskOnly the prison phase was randomised. Aftercare was voluntary and participants self selected. Concerns about prison drop-outs and completers p.87

Prendergast 2004

MethodsAllocation: random assignment
Randomisation method: unclear risk
Similar on drug use: yes
Similar on criminal activity: yes
Blinding methodology: unknown/unclear risk
Loss to follow-up: partial/high risk


Participants715 adults
Mean age 30.9 years (SD 7.4)
100% male
37.8% white
100% drug-using
Alcohol use not reported
100% psychiatric history
Eligibility criteria: offenders with a drug problem who were 9-14 months from parole. Offenders convicted of arson or sexual crimes to minors were not eligible


InterventionsSecure establishment-based TC vs. no treatment.
(I) Amity TC and voluntary residential aftercare vs. (C) waiting-list control.

(I) TC included a 2- to 3-month orientation phase, a 5- to 6-month treatment stage, and a 1- to 3-month re-entry phase; total duration 12 months. (I) included need assessment, education, group work, counselling and prison industry jobs. (I) aftercare duration up to 12 months

(C) duration not applicable


OutcomesIncarceration (official records) during the last 60 months at 60 months' follow-up
Drug use (self report) during the last 60 months at 60 months' follow-up


Notes5-year follow-up to Wexler 1999a and Wexler 1999b


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear risk"Random assignment and stratified for equal ethnic proportions". Method not stated

Allocation concealment (selection bias)Unclear riskConcealment method not stated

Blinding (performance bias and detection bias)
All outcomes
Unclear riskNo information reported

Blinding of participants and personnel (performance bias)
subjective outcomes
Unclear riskNo information reported

Blinding of participants and personnel (performance bias)
objective outcomes
Unclear riskNo information reported

Blinding of outcome assessment (detection bias)
subjective outcomes
Unclear riskNo information reported

Blinding of outcome assessment (detection bias)
objective outcomes
Unclear riskNo information reported

Incomplete outcome data (attrition bias)
All outcomes
High risk74% of the original sample were included in the follow-up

Selective reporting (reporting bias)Unclear riskNo information reported

Other biasHigh riskOnly the prison phase was randomised. Aftercare was voluntary and participants self selected. See comments above for Prendergast 2003

Sacks 2004a TC

MethodsAllocation: random assignment
Randomisation method: unclear
Similar on drug use: no
Similar on criminal activity: yes
Blinding methodology: unknown/unclear risk
Loss to follow-up: adequate/low risk


Participants236 adults
Mean age 34.3 years (SD 8.8)
100% male
49% white
100% drug-using
32% alcohol-using
100% psychiatric history
Eligibility criteria: prisoners who had both a serious mental disorder and a substance-use disorder.


InterventionsSecure establishment-based TC vs. treatment as usual
(I) Personal reflections TC and voluntary residential aftercare (n = 142) vs. (C) mental health programme (n = 94)
(I) TTC included psychoeducational classes, cognitive behavioural methods, medication and group therapy. Activities were attended 5 days per week for 4-5 hours per day with the rest of the day spent working in the prison; duration 12 months. (I) Aftercare included mental health counselling, medication and psychiatric services and basic skills. Activities were attended 3-7 days per week for 3-5 hours per day; duration 6 months

(C) Programme included intensive psychiatric services with medication, weekly individual therapy and counselling and specialised groups of cognitive behavioural work, anger management, therapy and education, domestic violence, parenting and weekly drug/alcohol therapy with a 72-hour course on substance abuse education and relapse prevention; duration 12 months


OutcomesCriminal activity regarding a new offence (official records) during the last 12 months at 12 months follow-up
Incarceration for a new offence (official records) during the last 12 months at 12 months follow-up


NotesDrug use activity at 12-months reported in Sullivan 2007


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskNoted as "randomly assigned" on p. 3, no other information provided

Allocation concealment (selection bias)Unclear riskNo information provided

Blinding (performance bias and detection bias)
All outcomes
Unclear riskNo information about blinding provided

Blinding of participants and personnel (performance bias)
subjective outcomes
Unclear riskNo information about blinding provided

Blinding of participants and personnel (performance bias)
objective outcomes
Unclear riskNo information about blinding provided

Blinding of outcome assessment (detection bias)
subjective outcomes
Unclear riskNo information about blinding provided

Blinding of outcome assessment (detection bias)
objective outcomes
Unclear riskNo information about blinding provided

Incomplete outcome data (attrition bias)
All outcomes
Low riskITT analysis performed. Missing data imputation reported

Selective reporting (reporting bias)Unclear riskNo information reported

Other biasUnclear riskThere was a large amount of crossover between the treatment and intervention groups. The cross-overs were removed from the analysis leaving unequal sample sizes p. 482. Difficult to assess to what extent this has caused a threat to validity within the study

Sacks 2004b TC + AC

MethodsAllocation: random assignment
Randomisation method: unclear
Similar on drug use: no
Similar on criminal activity: yes
Blinding methodology: unknown/unclear risk
Loss to follow-up: adequate/low risk


Participants236 adults
Mean age 34.3 years (SD 8.8)
100% male
49% white
100% drug-using
32% alcohol-using
100% psychiatric history
Eligibility criteria: prisoners who have both a serious mental disorder and a substance-use disorder


InterventionsSecure establishment-based TC vs. treatment as usual
(I) Personal reflections TC and voluntary residential aftercare (n = 142) vs. (C) mental health programme (n = 94)
(I) TTC included psychoeducational classes, cognitive behavioural methods, medication and group therapy. Activities were attended 5 days per week for 4-5 hours per day with the rest of the day spent working in the prison; duration 12 months. (I) aftercare included mental health counselling, medication and psychiatric services and basic skills. Activities were attended 3-7 days per week for 3-5 hours per day; duration 6 months

(C) Programme included intensive psychiatric services with medication, weekly individual therapy and counselling, and specialised groups of cognitive behavioural work, anger management, therapy and education, domestic violence, parenting and weekly drug/alcohol therapy with a 72-hour course on substance abuse education and relapse prevention; duration 12 months


OutcomesCriminal activity regarding a new offence (official records) during the last 12 months at 12 months' follow-up
Incarceration for a new offence (official records) during the last 12 months at 12 months' follow-up


NotesDrug use activity at 12-months reported in Sullivan 2007


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskNoted as "randomly assigned" on p. 3, no other information provided

Allocation concealment (selection bias)Unclear riskNo information provided

Blinding (performance bias and detection bias)
All outcomes
Unclear riskNo information about blinding provided

Blinding of participants and personnel (performance bias)
subjective outcomes
Unclear riskNo information about blinding provided

Blinding of participants and personnel (performance bias)
objective outcomes
Unclear riskNo information about blinding provided

Blinding of outcome assessment (detection bias)
subjective outcomes
Unclear riskNo information about blinding provided

Blinding of outcome assessment (detection bias)
objective outcomes
Unclear riskNo information about blinding provided

Incomplete outcome data (attrition bias)
All outcomes
Low riskITT analysis performed. Missing data imputation reported

Selective reporting (reporting bias)Unclear riskNo information provided

Other biasUnclear riskThere was a large amount of cross-over between the treatment and intervention groups. The cross-overs were removed from the analysis leaving unequal sample sizes p. 482. Difficult to assess to what extent this has caused a threat to validity within the study

Sacks 2011

MethodsAllocation: random assignment
Randomisation method: random number list
Similar on drug use: yes
Similar on criminal activity: yes
Blinding methodology: open label - no blinding
Loss to follow-up: unclear risk


Participants127 adults

Mean age 38.2 years (SD 9.9)

100% male

56% white

100% co-occurring substance use and mental illness

Alcohol use unknown

61.8% with clinical level of psychological distress as measured by Global Severity Index

Eligibility criteria: male, diagnosed with co-occurring mental and substance-use disorders, had participated in 1 of 2 prison substance-abuse treatment programmes, were approved for placement in a community corrections facility, were accepted by the provider agency for placement in a community corrections facility


InterventionsSecure establishment based TC vs. parole supervision case management

(I) re-entry modified TC (n = 71) vs. (C) parole supervision case management (n = 56)

(I) was a residential programme of 6 months' duration. Formal programme activities 3-7 days per week, 3-5 hours each day. Participants had progressively increasing independence, eventually being responsible for providing counsel, guidance and coaching for new members. Participants also worked in the community and saved money for independent living. There were weekly group psycho-educational classes to address the inter-relationship between mental disorders and substance abuse, as well as various other group and individual counselling sessions. Medication monitoring and psychiatric services were on site. Participants were given assistance with housing and encouragement for employment

(C) participants were released to a community corrections facility, and left the facility during the day to go to work, have treatment and report to parole officers. The intervention consisted of outreach and engagement activities, brokering community-based services, and direct provision of support and counselling services. There was a weekly relapse prevention group and daily medication monitoring. Psychiatric and substance-abuse services were provided by outside agencies (client helped to choose by community parole officers). Unlike in the intervention, criminal thinking and behaviour were not specifically addressed. The average participant attended 1 group per week and had monthly psychiatric assessment


OutcomesRate of re-incarceration, number of days until re-incarceration, involvement in self reported criminal activity, number of days until self reported criminal activity. Alcohol and drug offences (self reported) %. Other offences (self reported) %. All at 12 months post prison release


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskRandom number list

Allocation concealment (selection bias)Unclear riskNot reported

Blinding (performance bias and detection bias)
All outcomes
Unclear risk"Open-label trial", no blinding possible

Blinding of participants and personnel (performance bias)
subjective outcomes
Unclear risk"Open-label trial", no blinding possible

Blinding of participants and personnel (performance bias)
objective outcomes
Unclear risk"Open label trial", no blinding possible

Blinding of outcome assessment (detection bias)
subjective outcomes
Unclear riskNo information about blinding presented

Blinding of outcome assessment (detection bias)
objective outcomes
Unclear riskNo information about blinding presented

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskSome partial loss to follow-up

Selective reporting (reporting bias)Unclear riskNot reported

Other biasLow riskNo other obvious concerns with the study

Stein 2011

MethodsAllocation: random assignment
Randomisation method: random numbers table
Similar on drug use: yes
Similar on criminal activity: unknown/unclear risk
Blinding methodology: unclear/unknown
Loss to follow-up: partial/unclear risk


Participants189 adolescents

Mean age 17.12 years (SD 1.10). Range 14-19 years

85.7% male

32.8% white

88.9% marijuana use

63% alcohol use

68.5% had significant depressive symptomatology during past week at baseline (CES-D)

Eligibility criteria: 14-19 years old, sentenced to juvenile correctional facility for 4-12 months, engaged in at least monthly marijuana use or binge-drinking in the year before incarceration, used any alcohol or marijuana in the month prior to incarceration (or prior to the offence leading to incarceration)


InterventionsSecure establishment based MI intervention vs. relaxation treatment

(I) MI (n = 96) vs. (C) relaxation training (n = 85)

(I) was MI intervention designed specifically to reduce substance use and its associated risks and consequences. Consisted of 90-minute baseline intervention and 60-minute booster intervention within 2 weeks of release

(C) consisted of 90-minute baseline and 60-minute booster intervention, and involved relaxation techniques as well as advice on risky behaviours associated with substance use


OutcomesMean number of joints per day and mean percentage of days used marijuana at 3 months


NotesResults presented for both high and low depressive symptom scores. Results used in this review are for those identified as having high depressive symptomatology.

Appears to be same study population as Stein 2011b


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low risk"Random assignment using a random number table"

Allocation concealment (selection bias)Unclear risk"Random number was placed in an envelope and opened by research staff after the baseline assessment"

Blinding (performance bias and detection bias)
All outcomes
Unclear riskResearchers were blind until after the baseline assessment. Participants were not blinded.

Blinding of participants and personnel (performance bias)
subjective outcomes
Unclear riskResearchers were blind until after the baseline assessment. Participants were not blinded.

Blinding of participants and personnel (performance bias)
objective outcomes
Unclear riskResearchers were blind until after the baseline assessment. Participants were not blinded.

Blinding of outcome assessment (detection bias)
subjective outcomes
Unclear riskFollow-up assessments at 3 months were completed blind by the researchers but not at any other time point

Blinding of outcome assessment (detection bias)
objective outcomes
Unclear riskFollow-up assessments at 3 months were completed blind by the researchers but not at any other time point

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskSome attrition lost particularly for those individuals with more severe depression

Selective reporting (reporting bias)Unclear riskLimited information

Other biasHigh riskUnderpowered and small sample size, with a brief follow-up period and concerns about self report measures

Sullivan 2007

MethodsAllocation: random assignment
Randomisation method: unclear
Similar on drug use: no
Similar on criminal activity: yes
Blinding methodology: unknown/unclear risk
Loss to follow-up: adequate/low risk


Participants236 adults
Mean age 34.3 years (SD 8.8)
100% male
49% white
100% drug-using
32% alcohol-using
100% psychiatric history
Eligibility criteria: prisoners who had both a serious mental disorder and a substance-use disorder


InterventionsSecure establishment-based TC vs. treatment as usual
(I) Personal reflections TC and voluntary residential aftercare (n = 142) vs. (C) mental health programme (n = 94)
(I) TTC included psycho-educational classes, cognitive behavioural methods, medication and group therapy. Activities were attended 5 days per week for 4-5 hours per day with the rest of the day spent working in the prison; duration 12 months. (I) Aftercare included mental health counselling, medication and psychiatric services and basic skills. Activities were attended 3-7 days per week for 3-5 hours per day; duration 6 months

(C) Programme included intensive psychiatric services with medication, weekly individual therapy and counselling and specialised groups of cognitive behavioural work, anger management, therapy and education, domestic violence, parenting and weekly drug/alcohol therapy with a 72-hour course on substance-abuse education and relapse prevention; duration 12 months


OutcomesDrug use (self report) at 12 months


NotesCriminal activity outcomes at 12 months reported in Sacks 2004


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskParticipants were randomly assigned p. 824

Allocation concealment (selection bias)Unclear riskMethod of concealment not reported

Blinding (performance bias and detection bias)
All outcomes
Unclear riskNo information on blinding reported

Blinding of participants and personnel (performance bias)
subjective outcomes
Unclear riskNo information on blinding reported

Blinding of participants and personnel (performance bias)
objective outcomes
Unclear riskNo information on blinding reported

Blinding of outcome assessment (detection bias)
subjective outcomes
Unclear riskNo information on blinding reported

Blinding of outcome assessment (detection bias)
objective outcomes
Unclear riskNo information on blinding reported

Incomplete outcome data (attrition bias)
All outcomes
Low riskSome difference between the groups. At follow-up, 82% for the (I) group and 69% for the (C). ITT was performed and missing data computated

Selective reporting (reporting bias)Unclear riskLimited information

Other biasUnclear riskNo other obvious concerns with the study but difficult to assess

Wexler 1999a

MethodsAllocation: random assignment
Randomisation method: unclear/unknown
Similar on drug use: yes
Similar on criminal activity: yes
Blinding methodology: unknown/unclear risk
Loss to follow-up: adequate/low risk


Participants715 adults
Mean age 30.9 years (SD 7.4)
100% male
37.8% white
100% drug-using
Alcohol use not reported
100% psychiatric history
Eligibility criteria: offenders with a drug problem who were 9-14 months from parole. Offenders convicted of arson or sexual crimes to minors were not eligible


InterventionsSecure establishment-based TC vs. no treatment
(I) Amity TC and voluntary residential aftercare (n = 247) vs. (C) waiting-list control (n = 290)
(I) TC included a 2- to 3-month orientation phase, a 5- to 6-month treatment stage, and a 1- to 3-month re-entry phase; total duration 12 months. (I) included need assessment, education, group work, counselling and prison industry jobs. (I) aftercare duration up to 12 months

(C) Duration not applicable


OutcomesIncarceration (official records) during the last 12 months at 12 months' follow-up
Incarceration (official records) during the last 24 months at 24 months' follow-up


Notes36-month follow-up outcome in Wexler 1999b and 5-year outcomes in Prendergast 2004


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskIt was noted that the participants were 'randomly' assigned and stratified by ethnic makeup. Randomisation only occurred to the TC and not to aftercare

Allocation concealment (selection bias)Unclear riskNo information provided

Blinding (performance bias and detection bias)
All outcomes
Unclear riskNo blinding was conducted

Blinding of participants and personnel (performance bias)
subjective outcomes
Unclear riskNo information on blinding was provided

Blinding of participants and personnel (performance bias)
objective outcomes
Unclear riskNo information on blinding was provided

Blinding of outcome assessment (detection bias)
subjective outcomes
Unclear riskNo information on blinding was provided

Blinding of outcome assessment (detection bias)
objective outcomes
Unclear riskNo information on blinding was provided

Incomplete outcome data (attrition bias)
All outcomes
Low riskOutcomes for the first 12 months post-release were obtained for all 715 participants, and 12 months were obtained for 263 participants at risk for 24 months at time of record review

Selective reporting (reporting bias)Unclear riskInformation not reported

Other biasHigh riskOnly the prison phase was randomised. Aftercare was voluntary and participants self selected. Concerns about bias in self selection processes p.164-165

Wexler 1999b

MethodsAllocation: random assignment
Randomisation method: unclear/unknown risk
Similar on drug use: yes
Similar on criminal activity: yes
Blinding methodology: unknown/unclear risk
Loss to follow-up: adequate/low risk


Participants715 adults
Mean age 30.9 years (SD 7.4)
100% male
37.8% white
100% drug-using
Alcohol use not reported
100% psychiatric history
Eligibility criteria: offenders with a drug problem who were 9-14 months from parole. Offenders convicted of arson or sexual crimes to minors were not eligible


InterventionsSecure establishment-based TC vs. no treatment
(I) Amity TC and voluntary residential aftercare (n = 247) vs. (C) waiting-list control (n = 290)
(I) TC included a 2- to 3-month orientation phase, a 5- to 6-month treatment stage, and a 1- to 3-month re-entry phase; total duration 12 months. (I) Included need assessment, education, group work, counselling and prison industry jobs. (I) aftercare duration up to 12 months

(C) Duration not applicable


OutcomesIncarceration (official records) during the last 36 months at 36 months' follow-up


Notes12- and 24-month follow-up outcome in Wexler 1999a and 5-year outcomes in Prendergast 2004


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskIt was noted that the participants were 'randomly' assigned and stratified by ethnic makeup. Randomisation only occurred to the TC and not to aftercare

Allocation concealment (selection bias)Unclear riskNo information provided

Blinding (performance bias and detection bias)
All outcomes
Unclear riskNo blinding was conducted

Blinding of participants and personnel (performance bias)
subjective outcomes
Unclear riskNo information on blinding was provided

Blinding of participants and personnel (performance bias)
objective outcomes
Unclear riskNo information on blinding was provided

Blinding of outcome assessment (detection bias)
subjective outcomes
Unclear riskNo information on blinding was provided

Blinding of outcome assessment (detection bias)
objective outcomes
Unclear riskNo information on blinding was provided

Incomplete outcome data (attrition bias)
All outcomes
Low riskOutcomes for the first 12 months post-release were obtained for all 715 participants, and 12 months were obtained for 263 participants at risk for 24 months at time of record review

Selective reporting (reporting bias)Unclear riskInformation not reported

Other biasHigh riskThe study design did not randomly assign inmates to the aftercare. Concerns about bias in self selection processes, pp.164-5

 
Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion

Alemi 2010Study population did not have co-occurring mental illness

Alessi 2011Not original RCT. Data are from previous, older studies

Anglin 1999The study did not report relevant drug or crime outcome (or both) measures at both the pre- and post-intervention periods

Awgu 2010No relevant outcomes reported

Baldus 2011Study protocol only

Banks 2004Study population did not have co-occurring mental illness

Berman 2004The intervention was not appropriate for inclusion

Brady 2010Not RCT

Braithwaite 2005The study did not report relevant drug or crime outcome (or both) measures at both the pre- and post-intervention periods

Britt 1992aStudy population did not have co-occurring mental illness

Britt 1992bStudy population did not have co-occurring mental illness

Britt 1992cStudy population did not have co-occurring mental illness

Britt 1992dStudy population did not have co-occurring mental illness

Brown 20013-arm study in which only 2 arms were randomised - 1 treatment arm and control arm. Results presented as both treatment arms combined vs. control

Carr 2008The population of the study was not 100% drug-using offenders that were specifically referred by the criminal justice system to the intervention

Carroll 2006Study population did not have co-occurring mental illness

Carroll 2011Not offender population

Chandler 2006The study did not report relevant drug or crime outcome (or both) measures at both the pre- and post-intervention periods

Cogswell 2011Paper not available

Cornish 1997Study population did not have co-occurring mental illness

Cosden 2005The study did not report relevant drug or crime outcome (or both) measures at both the pre- and post-intervention periods

Coviello 2010Study population did not have co-occurring mental illness

Coviello 2010bStudy population did not have co-occurring mental illness

Cropsey 2011Study population did not have co-occurring mental illness

Cusack 2010Not a drug use intervention

Dakof 2010Study population is mothers of offenders, not offenders themselves

Dembo 2000The study did not report relevant drug or crime outcome (or both) measures at both the pre- and post-intervention periods
The follow-up periods reported for the different groups were not equivalent

Deschenes 1994Study population did not have co-occurring mental illness

Di Nitto 2002The follow-up periods reported for the different groups were not equivalent

Diamond 2006The study did not report relevant drug or crime outcome (or both) measures at both the pre- and post-intervention periods

Dolan 2003Study population did not have co-occurring mental illness

Dugan 1998The study did not report relevant drug or crime outcome (or both) measures at both the pre- and post-intervention periods

Forsberg 2011Study population did not have co-occurring mental illness

Freudenberg 2010Study population did not have co-occurring mental illness

Gagnon 2010Not offender population

Gil 2004The study did not report relevant drug or crime outcome (or both) measures at both the pre- and post-intervention periods

Gordon 2008Study population did not have co-occurring mental illness

Gottfredson 2002Study population did not have co-occurring mental illness

Gottfredson 2003Study population did not have co-occurring mental illness

Gottfredson 2005Study population did not have co-occurring mental illness

Gottfredson 2006Study population did not have co-occurring mental illness

Grohman 2002The study did not report relevant drug or crime outcome (or both) measures at both the pre- and post-intervention periods

Guydish 2011Study population did not have co-occurring mental illness

Haapanen 2002Study population did not have co-occurring mental illness

Haasen 2010Not offender population

Hall 2009a ST vs. DTStudy population did not have co-occurring mental illness

Hall 2009b ST vs. TPStudy population did not have co-occurring mental illness

Hanlon 1999Study population did not have co-occurring mental illness

Harrell 2001The study did not report relevant drug or crime outcome (or both) measures at both the pre- and post-intervention periods

Henderson 2010No relevant outcomes reported

Henggeler 1991The study did not report relevant drug or crime outcome (or both) measures at both the pre- and post-intervention periods

Henggeler 1999Study population did not have co-occurring mental illness

Henggeler 2002The study did not report relevant drug or crime outcome (or both) measures at both the pre- and post-intervention periods

Henggeler 2006a FC vs. DCStudy population did not have co-occurring mental illness

Henggeler 2006b FCvDCMSTStudy population did not have co-occurring mental illness

Hser 2011Unclear if study looks at offender population

Inciardi 2004Some participants were not randomly selected into the treatment groups

Jain 2011Paper not available and not clear from abstract if looks at offender population

Johnson 2011Study population did not have co-occurring mental illness

Katz 2007The population of the study was not 100% drug-using offenders that were specifically referred by the criminal justice system to the intervention

Kinlock 2005Study population did not have co-occurring mental illness

Kinlock 2007Study population did not have co-occurring mental illness

Kinlock 2008Study population did not have co-occurring mental illness

Kinlock 2009Conference proceedings only

Kinlock 2009a CO vs. CTStudy population did not have co-occurring mental illness

Kinlock 2009b CO vs. CMStudy population did not have co-occurring mental illness

Liddle 2011No relevant outcomes reported

Lobmaier 2010Study population did not have co-occurring mental illness

Lobmann 2007Study population did not have co-occurring mental illness

Magura 2009Study population did not have co-occurring mental illness

Marlowe 2003The study did not report relevant drug or crime outcome (or both) measures at both the pre- and post-intervention periods

Marlowe 2005Study population did not have co-occurring mental illness

Marlowe 2007Participants randomised to receive treatment were not randomised into the different treatment subgroups (selected by level of risk)

Marlowe 2008Study population did not have co-occurring mental illness

Martin 1993Study population did not have co-occurring mental illness

Mbilinyi 2011Participants not recruited through criminal justice system

McCollister 2009aStudy population did not have co-occurring mental illness

McCollister 2009bStudy population did not have co-occurring mental illness

McCollister 2009cStudy population did not have co-occurring mental illness

McKendrick 2007The study did not report relevant drug or crime outcome (or both) measures at both the pre- and post-intervention periods

Messina 2000The population of the study was not 100% drug-using offenders that were specifically referred by the criminal justice system to the intervention
The study did not report relevant drug or crime outcome (or both) measures at both the pre- and post-intervention periods

Milloy 2011Relevant results from original RCT not reported here

Needels 2005The population of the study was not 100% drug- using offenders that were specifically referred by the criminal justice system to the intervention

Nemes 1998The population of the study was not 100% drug-using offenders that were specifically referred by the criminal justice system to the intervention
The study did not report relevant drug or crime outcome (or both) measures at both the pre- and post-intervention periods

Nemes 1999The population of the study was not 100% drug-using offenders that were specifically referred by the criminal justice system to the intervention
The study did not report relevant drug or crime outcome (or both) measures at both the pre- and post-intervention periods

Nielsen 1996Study population did not have co-occurring mental illness

Nosyk 2010Not offender population

Petersilia 1992aStudy population did not have co-occurring mental illness

Petersilia 1992bStudy population did not have co-occurring mental illness

Petersilia 1992cStudy population did not have co-occurring mental illness

Petersilia 1992dStudy population did not have co-occurring mental illness

Petersilia 1992eStudy population did not have co-occurring mental illness

Petersilia 1992fStudy population did not have co-occurring mental illness

Petersilia 1992gStudy population did not have co-occurring mental illness

Petry 2011Not offender population

Polsky 2010Not offender population

Prendergast 2008a STVSDTStudy population did not have co-occurring mental illness

Prendergast 2008b STVSTPStudy population did not have co-occurring mental illness

Prendergast 2009No relevant outcomes reported

Prendergast 2011Study population did not have co-occurring mental illness

Proctor 2012Study population did not have co-occurring mental illness

Reimer 2011Not offender population

Robertson 2006The population of the study was not 100% drug- using offenders that were specifically referred by the criminal justice system to the intervention

Rosengard 2008No relevant outcomes reported

Rossman 1999Study population did not have co-occurring mental illness

Rowan-Szal 2005Paper not available and not clear from abstract if looks at offenders

Rowan-Szal 2009Not RCT

Rowe 2007The population of the study was not 100% drug- using offenders that were specifically referred by the criminal justice system to the intervention

Sacks 2008Study population did not have co-occurring mental illness

Sanchez-Hervas 2010Paper not available and not clear from abstract if looks at offenders

Schwartz 2006Not offender population

Sheard 2009The study did not report relevant drug or crime outcome (or both) measures at both the pre- and post-intervention periods

Siegal 1999Not RCT

Smith 2010Study population did not have co-occurring mental illness

Stanger 2009The population of the study was not 100% drug-using offenders that were specifically referred by the criminal justice system to the intervention

Staton-Tindall 2009No control group

Stein 2010Not offender population

Stevens 1998The study did not include an appropriate comparison group
The population of the study was not 100% drug- using offenders that were specifically referred by the criminal justice system to the intervention

Svikis 2011Not clear if offender population

Taxman 2006Study population did not have co-occurring mental illness

Thanner 2003Study population did not have co-occurring mental illness

Wang 2010Participants not in criminal justice system

White 2006Randomisation broken as 40% of control arm were allowed to receive treatment (acupuncture) outside of the intervention

Williams 2011Not RCT

Winstanley 2011Not clear if offender population

Witkiewitz 2010Not clear if offender population

Zlotnick 2009Study population did not have co-occurring mental illness

 
Characteristics of studies awaiting assessment [ordered by study ID]
Burdon 2013

MethodsAllocation: random assignment
Randomisation method: not yet assessed
Similar on drug use: not yet assessed
Similar on criminal activity: not yet assessed
Blinding methodology: not yet assessed
Loss to follow-up: not yet assessed

Participants143 female inmates
0% male (all female)
% white
% drug-using
Alcohol use not yet assessed
Eligibility criteria: not yet assessed

Interventions12-week prison-based intensive outpatient drug treatment or standard treatment plus positive behavioural reinforcement

OutcomesDepression

Criminal thinking

Treatment participation and progress

Treatment satisfaction

Psychosocial functioning

Notes

Lee 2011

MethodsAllocation: random assignment
Randomisation method: not yet assessed
Similar on drug use: not yet assessed
Similar on criminal activity: not yet assessed
Blinding methodology: not yet assessed
Loss to follow-up: not yet assessed

Participants24 inmates
100% male
% white
% drug-using
Alcohol use not yet assessed
Eligibility criteria: incarcerated individuals with a history of substance abuse

InterventionsMindfulness-based relapse prevention vs. treatment as usual

OutcomesDrug-use identification disorders

Drug avoidance self efficacy scale

Positive and negative outcome expectancies

Beck Depression Inventory

Notes

Wolff 2012

MethodsAllocation: random assignment - open trial
Randomisation method: not yet assessed
Similar on drug use: not yet assessed
Similar on criminal activity: not yet assessed
Blinding methodology: not yet assessed
Loss to follow-up: not yet assessed

Participants74 female inmates
0% male (all female)
% white
% drug-using
Alcohol use not yet assessed
Eligibility criteria: incarcerated women with axis I mental health disorder of post-traumatic stress disorder and substance-use disorder who self referred for speciality trauma treatment

InterventionsCognitive behavioural therapy - Seeking Safety vs. standard treatment

OutcomesNot yet assessed

Notes

 
Comparison 1. Self report drug use

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Self report drug use dichotomous2715Risk Ratio (M-H, Random, 95% CI)0.82 [0.44, 1.55]

 
Comparison 2. Criminal activity

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Criminal activity dichotomous4Risk Ratio (M-H, Random, 95% CI)Subtotals only

    1.1 Re-arrests
2663Risk Ratio (M-H, Random, 95% CI)1.00 [0.90, 1.12]

    1.2 Re-incarceration
2266Risk Ratio (M-H, Random, 95% CI)0.40 [0.24, 0.67]

 2 Criminal activity continuous2361Mean Difference (IV, Random, 95% CI)28.72 [5.89, 51.54]

 
Comparison 3. Therapeutic community

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Criminal activity3Odds Ratio (M-H, Random, 95% CI)Subtotals only

    1.1 Re-arrests
1428Odds Ratio (M-H, Random, 95% CI)0.90 [0.61, 1.33]

    1.2 Re-incarceration
2266Odds Ratio (M-H, Random, 95% CI)0.29 [0.16, 0.54]

 
Comparison 4. Mental health court

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Self report dichotomous criminal activity1235Risk Ratio (M-H, Random, 95% CI)1.05 [0.90, 1.22]

 
Comparison 5. Motivational interviewing and cognitive skills

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Self report drug use1162Mean Difference (IV, Random, 95% CI)-7.42 [-20.12, 5.28]

 
Table 1. Mental health diagnoses

Study, yearCriteria used for diagnosesDescription of mental health problem

Stein 2011CES-D ScaleScores > 16 indicate presence of significant depression. 69.8% had

significant depressive symptoms

Sacks 2004a TC; Sacks 2004b TC + AC; Sullivan 2007DISDiagnoses of lifetime Axis I or Axis II mental disorder

Antisocial personality disorder

Cosden 2003Determined by a psychiatrist/psychologist on the basis of a clinical interview and observationsMood disorder

Schizophrenia

Bipolar disorder

Other

Dual diagnosis

Prendergast 2003; Prendergast 2004; Wexler 1999a; Wexler 1999bNot specifiedAntisocial personality disorder

Phobias

PTSD

Depression

Dysthymia

Attention deficit

hyperactivity disorder

Sacks 2011DSM-IV diagnostic criteria

Beck Depression Inventory

Post Traumatic Stress Disorder Symptom Scale

Brief Symptom Inventory

Global Severity Index
Depression

PTSD

Psychological distress

 CED-D: Centre for Epidemiological Studies - Depression; DSM-IV: Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition; DIS: Diagnostic Interview Schedule; PTSD: post-traumatic stress disorder.
 
Table 2. Summary research evidence for the meta-analyses

Paper, yearinterventionComparisonFollow-upOutcome typeMeasurementActual outcome

Cosden 2003Sentencing and case management (mental health treatment court and assertive community treatment case management)Treatment as usual6 months

12 months
Criminal activity dichotomous

Self report drug use continuous
% and total

mean and SD
% arrested and spent some time in jail

% convicted of a new crime

mean Addiction Severity Index (drug) composite score

Prendergast 2003

Prendergast 2004

Wexler 1999a; Wexler 1999b
Therapeutic community, counselling and aftercareTreatment as usual and waiting list control12, 24, 36 months

up to 5 years
Biological drug use dichotomous

Criminal activity continuous

Criminal activity dichotomous

Self report drug use dichotomous
% and total

mean and SD
% testing positive for illicit drugs at 12 months' follow-up

Mean months incarcerated in the year following release

% any arrest

% arrested for drug crime

% arrested for property crime

% arrested for violent crime

% arrested for other crime

% used drugs heavily in past year at 5 years

Mean days until re-incarceration

% re-incarcerated

Mean days on parole to first return to custody

% returned to prison within 3 years post parole

Sacks 2004a TC therapeutic community only

Sacks 2004b TC + AC therapeutic community plus aftercare

Sullivan 2007
Modified therapeutic community (personal reflections therapeutic community and voluntary residential aftercare)Intensive psychiatric services12 monthsCriminal activity continuous

Criminal activity dichotomous

Self report drug use dichotomous
Mean and SD

% and total
Mean number of days until incarceration

Mean number of days until first crime

% re-incarceration

% criminal activity

% alcohol/drug offence

% other (non-alcohol/drug) offence

% illegal drug use

Sacks 2011Therapeutic community (re-entry modified)Parole supervision case management12 monthsCriminal activity dichotomous% with total% re-incarcerated

% self reported criminal activity

Stein 2011 (high depression score)Motivational interviewingRelaxation training3 monthsSelf report drug use continuousMean and SDMean joins per day

Mean % days used marijuana

 SD: standard deviation.