6.2.1.1. Parent training for parent-child attachment for women with sub-threshold symptoms of depression

Two RCTs (N = 308) met the eligibility criteria for this review. Sleed et al. (2013) evaluated Better Start, a Better Start, a manualized intervention including group parent training sessions and home visits for women visits for women following release from specialized mother and baby units within prisons in England and Wales. England and Wales. Menting et al. (2014) was a Dutch study evaluating New Beginnings, a manualised manualised attachment-based group intervention developed specifically for mothers and babies in prison. An babies in prison. An overview of the trials included in the analysis can be found in Table 47. Further information about both included and excluded studies can be found in Appendix L. Summary of findings can be found in N= total number of participants

1 Number randomised

NR= Not reported

Table 48. The full GRADE evidence profiles and associated forest plots can be found in Appendices N and O, respectively.

Table 47

Study information for trials included in the analysis of parent training for parent-child attachment for women with sub-threshold symptoms of depression.

Table 48

Summary of findings table for parent training versus treatment as usual for parent-child attachment for women with sub-threshold symptoms.

6.2.1.2. Yoga for promoting mental health and wellbeing

One RCT (N = 167) met the eligibility criteria for this review. Bilderbeck et al. (2013) was a study of the impact of study of the impact of a ten-week yoga course on the psychological wellbeing of prisoners. Prisoners diagnosed Prisoners diagnosed with psychiatric illness were excluded. An overview of the trial included in the analysis can in the analysis can be found in Table 49. Further information about both included and excluded studies can be found in Appendix L. Summary of findings can be found in N= total number of participants

1 Number randomised

NR= Not reported

Table 50. The full GRADE evidence profiles and associated forest plots can be found in Appendices N and O, respectively.

Table 49

Study information for studies included in the analysis of Yoga versus Waitlist Control.

Table 50

Summary of findings table for yoga versus waitlist control for promoting mental health and wellbeing.

6.2.1.3. Meditation for promoting mental health and well-being

One RCT (N = 33) met the eligibility criteria for this review. Sumter et al. (2009) was a US study of meditation in a study of meditation in a residential detention facility for nonviolent female probationers. An overview of the trial overview of the trial included in the meta-analysis can be found in Table 51. Further information about both included and excluded studies can be found in Appendix L. Summary of findings can be found in N= total number of participants; 1Number randomised.

Table 52. The full GRADE evidence profiles and associated forest plots can be found in Appendices N and O, respectively.

Table 51

Study information table for studies included in the meta-analysis of meditation.

Table 52

Summary of findings table for meditation for promoting mental health and well-being.

6.2.1.4. Physical exercise programmes versus exercise as usual for promoting mental health and well-being

One RCT (N =75) met the eligibility criteria for this review. Battaglia et al. (2015) was a three arm trial comparing two different physical exercise programs with exercise as usual in an Italian prison. An overview of the trial can be found in

Table 53. Further information about both included and excluded studies can be found in Appendix L. Summary of findings can be found in N= total number of participants

1 Number randomised

Table 54. The full GRADE evidence profiles and associated forest plots can be found in Appendices N and O, respectively.

Table 53

Study information table for studies included in the analysis of physical exercise programmes versus exercise as usual.

Table 54

Summary of findings table for physical exercise programmes versus exercise as usual.

6.2.3.1. Parent training for parent-child attachment for women with sub-threshold symptoms

Low quality evidence from a single study (N=115) found no clinically important effects of parent training on depression as measured by mean scores on the Center for Epidemiological Studies Depression Scale or the number of participants with symptoms of depression for women with sub-threshold symptoms of depression resident in specialized mother and baby units in prison.

Low quality evidence from single study data (N=88-109) also showed no clinically important effects of parent training on:

• measures of mother-child attachment
• mother-child interaction
• or maternal perceptions of their child

A single study (N=102-3) found low quality inconsistent evidence for effects of parent training on maternal perceptions of parenting as measured by the Alabama Parenting Questionnaire (APQ) for women recently released from prison.

• For the Inconsistent discipline sub-scale, a clinically important effect was found in favour of parent training.
• However, for the Positive parenting subscale there was a clinically important effect in favour of the treatment as usual control.
• While for the Involvement, Poor monitoring/supervision and Corporal Punishment subscales no clinically significant effects were found.

Low quality evidence from two studies (N=308) did not indicate a clinically significant difference in drop-out between the parent training and treatment as usual conditions for women in prison or recently released from prison.

6.2.3.2. Yoga for promoting mental health and wellbeing

A single study (N=100) found very low quality evidence for clinically significant effects of yoga on increasing positive affect, decreasing negative affect and reducing perceived stress and for reducing psychological distress for participants in prison.

However, this study (N=167) also found that there was a clinically significant number of participants who dropped out of the study in the yoga condition.

6.2.3.3. Meditation for promoting mental health and well-being

A single study (N=33) found very low quality evidence for a clinically important effect of meditation for non-violent female probationers in a residential detention facility on reducing the desire to throw things or hit people and on being bothered by nail biting as measured by a study-specific scale

However, confidence in these effect estimates was very low due to serious risk of bias (unclear randomisation method and allocation concealment, lack of blinding and lack of a valid and reliable outcome measure) and very serious imprecision).

This same study found no evidence for clinically significant effects of meditation on feelings of guilt or hopelessness or being bothered by sleeping difficulties.

6.2.3.4. Physical exercise programmes for promoting mental health and well-being

Low quality evidence from one randomised controlled trial (N=75) indicated that a 9 month physical exercise programme had a clinically important beneficial effect on psychological wellbeing as measured by the Symptom Checklist-90-Revised (SCL-90-R) when compared to exercise as usual.

6.4.1.1. Substance misuse

6.4.1.1.1. Psychological interventions

There were 23 studies that met the criteria for this review: Alemi 2010 (Alemi et al., 2010), Annis 1979 (Annis, 1979), Binswanger 2015 (Binswanger et al., 2015), Brown 1980 (Brown, 1980), Carroll 2006 (Carroll et al., 2006), Carroll 2012 (Carroll et al., 2012), Crane 2015b (Crane et al., 2015b), Easton 2000 (Easton et al., 2000), Easton 2007c (Easton et al., 2007c), Forsberg 2011 (Forsberg et al., 2011B), Gordon 2008 (Gordon et al., 2008), Gordon 2014 (Gordon et al., 2014), Kinlock 2007 (Kinlock et al., 2007), Kinlock 2009 (Kinlock et al., 2009), McKenzie 2012 (McKenzie et al., 2012), Miller 1975 (Miller, 1975) M, Proctor 2012 (Proctor et al., 2012), Sinha 2003 (Sinha et al., 2003), Stuart 2013 (Stuart et al., 2013), Villagara-Lanza 2013 (Villagra Lanza & Menendez, 2013), Villagara-Lanza 2014 (Lanza et al., 2014), Witkiewitz 2014 (Witkiewitz et al., 2014) and Zlotnick 2009 (Zlotnick et al., 2009).

The interventions studied included acceptance and commitment therapy (ACT), cognitive behavioural therapy (CBT), other cognitive and behavioural therapies, mindfulness-based approaches, counselling, motivational interviewing techniques, self-help and psychoeducation.

Cognitive behavioural therapy versus active intervention for substance misuse

There were 3 RCTs (N=254) that met the eligibility criteria for this review: Carroll 2012, Easton 2007c and Zlotnick 2009 (Carroll et al., 2012; Easton et al., 2007c; Zlotnick et al., 2009).

An overview of the trials can be found in Table 56. Further information about both included and excluded studies can be found in Appendix N. Summary of findings can be found in N= total number of participants

NA= Not applicable

1 Number randomised

Table 57. The full evidence profiles and associated forest plots can be found in Appendices O and P.

The Zlotnick 2009 and Easton 2007c studies both describe 2-arm trials. The Zlotnick 2009 trial compared a variation of CBT tailored specifically to substance misuse (seeking safety) with treatment based upon the 12-step recovery model, whilst the Easton 2007c trial compared a variation of CBT designed for co-occurring substance misuse and domestic violence with the 12-step model. Finally, the Carroll 2012 trial was a 4-armed trial in which service users were allocated to one of the following conditions; standard CBT alone, CBT plus contingency management for adherence, contingency management for abstinence or CBT plus contingency management for abstinence. Only the CBT alone and CBT plus contingency management for adherence arms were included within this sub-review (CBT versus active intervention). The contingency interventions in Carroll 2012 were looked at under different sub-reviews (Contingency management versus Active intervention). The Carroll 2012 and Easton 2007c studies were both conducted in the community whilst the Zlotnick 2009 trial was conducted in a residential facility. The Easton 2007c trial intervention was delivered in a group setting whilst treatment in the other 2 studies was delivered individually (Carroll 2012) or a mixture of the two (Zlotnick 2009).

The evidence for this review was low to very low quality. No data was available for the outcomes of service utilisation, adaptive functioning or rates of self-injury.

Table 56

Study information table for trials included in the analysis of CBT versus active intervention for substance misuse.

Table 57

Summary of findings table for the analysis of CBT versus active intervention for substance misuse.

Cognitive behavioural therapy versus wait-list control for substance misuse

One RCT (N=27) met the eligibility criteria for this review: Villagara-Lanza 2014 (Lanza et al., 2014).

An overview of the trial can be found in Table 58. Further information about both included and excluded studies can be found in Appendix N. Summary of findings can be found in N= total number of participants;

1 Number randomised.

Table 59. The full evidence profiles and associated forest plots can be found in Appendices O and P.

The Villagara-Lanza 2014 study was a 3-arm trial, with groups receiving CBT, ACT or no treatment (waitlist control). The comparison of CBT and waitlist control group is described here. Treatment was delivered in a group format within a prison setting.

The evidence for this review was low to very low quality. No data was available for the outcomes of service utilisation, adaptive functioning or rates of self-injury.

Table 58

Study information table for trials included in the analysis of CBT versus waitlist control.

Table 59

Summary of findings table for the analysis of CBT versus wait-list control for substance misuse.

Acceptance and commitment therapy versus cognitive behavioural therapy for substance misuse

One RCT (N=30) met the eligibility criteria for this review: Villagara-Lanza 2014 (Lanza et al., 2014).

An overview of the trial can be found in Table 60. Further information about both included and excluded studies can be found in Appendix L. Summary of findings can be found in N= total number of participants

1 Number randomised

Table 61. The full evidence profiles and associated forest plots can be found in Appendices N and O. respectively.

Villagara-Lanza 2014 was a 3-armed trial comparing Acceptance and Commitment Therapy (ACT), CBT and control. The comparisons of ACT versus control and CBT versus control are detailed elsewhere in this chapter. Treatment was delivered in groups in prison.

The evidence for this review was low to very low quality. No evidence was available for the outcomes of offending and reoffending, service utilisation, adaptive functioning or rates of self-injury.

Table 60

Study information table for trials included in the analysis of ACT versus CBT for substance misuse in adults within the criminal justice system.

Table 61

Summary of findings for the analysis of ACT versus CBT for substance misuse in adults within the criminal justice system.

Acceptance and commitment therapy versus waitlist for substance misuse

Two RCTs (N=61) met the eligibility criteria for this review: Villagara-Lanza 2013 (Villagra Lanza & Menendez, 2013) and Villagara-Lanza 2014 (Lanza et al., 2014).

An overview of the trials can be found in Table 62. Further information about both included and excluded studies can be found in Appendix L. Summary of findings can be found in N= total number of participants;

1 Number randomised

Table 63. The full evidence profiles and associated forest plots can be found in Appendices N and O, respectively.

The Villagara-Lanza 2013 study was a 2-armed trial comparing ACT and a waitlist control group. The Villagara-Lanza 2014 study was a 3-armed trial comparing ACT, CBT and waitlist control groups. The comparisons of ACT versus CBT and CBT versus control are included elsewhere within the chapter. Both trials delivered interventions in groups within a prison setting.

The evidence for this review was low to very low quality. No data were available for the outcomes of offending and reoffending, service utilisation, adaptive functioning or rates of self-injury.

Table 62

Study characteristics for the analysis of ACT versus waitlist control for substance misuse in adults within the criminal justice system.

Table 63

Summary of findings for the analysis of ACT versus waitlist control for substance misuse in adults within the criminal justice system.

Mindfulness-based relapse prevention (MBRP) versus Cognitive Behavioural Therapy (CBT) for substance misuse

One RCT (N=105) met the eligibility criteria for this review: Witkiewitz 2014 (Witkiewitz et al., 2014).

An overview of the trial can be found in Table 64. Further information about both included and excluded studies can be found in Appendix L. Summary of findings can be found in N = total number of participants

NR=Not reported

RCT=randomised controlled trial

1 Number randomised

Table 65. The full evidence profiles and associated forest plots can be found in Appendices N and O, respectively.

This was a 2-armed trial comparing mindfulness-based CBT with a CBT-based relapse prevention intervention. The authors hypothesised that the addition of a mindfulness-based component would help service users to identify when they were on ‘automatic pilot’ and accordingly assist them in identifying triggers for cravings and help prevent relapse. The interventions were provided in a group format in a residential detention setting.

The evidence for this review was very low quality. No data were available for the outcomes of offending and reoffending, service utilisation, adaptive functioning or rates of self-injury.

Table 64

Study characteristics for the analysis of mindfulness-based relapse prevention versus active intervention for substance misuse in adults within the criminal justice system.

Table 65

Summary of findings table for the analysis of MBRP versus CBT.

Contingency management versus active intervention for substance misuse

Four RCTs (N=461) met the eligibility criteria for this review: Carroll 2006 (Carroll et al., 2006), Carroll 2012 (Carroll et al., 2012), Prendergast 2015 (Prendergast et al., 2015) and Sinha 2003 (Sinha et al., 2003).

An overview of the trials can be found in

Table 66. Further information about both included and excluded studies can be found in Appendix L. Summary of findings can be found in N= total number of participants;

1 Number randomised

Table 67. The full evidence profiles and associated forest plots can be found in Appendices N and O, respectively.

The Carroll 2006 study was a 4-armed trial comparing CBT plus motivational enhancement and contingency management, motivational enhancement plus CBT only, drug counselling plus contingency management and drug counselling alone. Only the contingency management plus drug counselling and drug counselling alone arms are included within this review. The Carroll 2012 trial was a 4-armed trial in which service users were allocated to one of the following conditions; standard CBT alone, CBT plus contingency management for adherence, contingency management for abstinence or CBT plus contingency management for abstinence. Here the 3 contingency management arms are combined and compared with the CBT arm. By combing the 3 contingency management arms to create a single pair-wise comparison with CBT alone randomisation was preserved, the interventions being similar enough to not downgrade this evidence for indirectness. The Sinha 2003 study was a 2-armed trial, where participants received either contingency management plus motivational enhancement therapy, or motivational enhancement therapy alone. The Prendergast 2015 study was also a 2-armed trial comparing contingency management with a psychoeducational intervention called ‘attendance education group’.

The evidence for this review was low to very low quality. No data was available for the outcomes of adaptive functioning, offending and reoffending or rates of self-injury.

Table 66

Study information table for trials included in the analysis of contingency management versus active intervention for substance misuse.

Table 67

Summary of findings table for the analysis of contingency management versus active intervention for substance misuse.

Contingency management versus Treatment as usual (TAU) for substance misuse

One RCT (N=20) met the eligibility criteria for this review: Miller 1975 (Miller, 1975).

An overview of the trial can be found in

Table 68. Further information about both included and excluded studies can be found in Appendix L. Summary of findings can be found in N= total number of participants; TAU = treatment as usual

1 Number randomised

Table 69. The full evidence profiles and associated forest plots can be found in Appendices N and O, respectively.

The Miller 1975 study was a 2-arm trial, with groups receiving contingency management or no treatment. Contingency management consisted of the provision of goods and services in exchange for sobriety. This trial was conducted in the community with treatment provided on an individual basis.

The evidence for this review was low quality. No data was available for the outcomes of mental health, service utilisation, adaptive functioning or rates of self-injury.

Table 68

Study information table for trials included in the analysis of contingency management versus treatment as usual substance misuse.

Table 69

Summary of findings table for the analysis of contingency management versus TAU for substance misuse.

Motivational enhancement therapy versus active intervention for substance misuse

Three RCTs (N=362) met the eligibility criteria for this review: Carroll 2006 (Carroll et al., 2006), Easton 2000 (Easton et al., 2000) and Stuart 2013 (Stuart et al., 2013).

An overview of the trial can be found in

Table 70. Further information about both included and excluded studies can be found in Appendix L. Summary of findings can be found in Table 71. The full evidence profiles and associated forest plots can be found in Appendices N and O, respectively.

Easton 2000 and Stuart 2013 were both 2-armed trials comparing motivational enhancement with psychoeducation. Carroll 2006 was a 4-armed trial with participants randomly allocated to receive one of the following: motivational enhancement therapy plus CBT and contingency management, motivational enhancement therapy plus CBT without contingency management, drug counselling plus contingency management or drug counselling alone. The drug counselling comparisons are described elsewhere within this chapter. The interventions in the Carroll 2006 trial were delivered individually.

The available data for this review were of very low quality. No data were available for the outcomes of offending and reoffending, adaptive functioning, rates of self-injury or service utilisation.

Table 70

Study characteristics table for the comparison of motivational enhancement therapy versus active intervention.

Table 71

Summary of findings for the analysis of motivational enhancement versus active intervention.

Motivational interviewing (MI) versus control or treatment as usual (TAU) for substance misuse

Four RCTs (N=492) met the eligibility criteria for this review: Alemi 2010 (Alemi et al., 2010), Crane 2015b (Crane et al., 2015b), Davis 2003 (Davis et al., 2003) and Forsberg 2011 (Forsberg et al., 2011B).

Three trials were 2-armed (Alemi 2010, Crane 2015b and Davis 2003) and compared motivational interviewing with control or no treatment. Forsberg 2011 was a 3-armed trial that compared two different forms of motivational interviewing with treatment as usual. These two forms (with workshop training only or with peer group supervision in addition) have been combined here. All trials delivered the intervention of interest individually. Crane 2015b conducted their trial in the community whilst Davis 2003 and Forsberg 2011 conducted trials in prison settings.

An overview of the trials can be found in Table 72. Further information about both included and excluded studies can be found in Appendix L. Summary of findings can be found in N= total number of participants; NR=Not reported

1 Number randomised

Table 73. The full evidence profiles and associated forest plots can be found in Appendices N and O, respectively.

The evidence for this review was low to very low quality. No data were available for the outcomes of adaptive functioning or rates of self-injury.

Table 72

Study characteristics table for the analysis of motivational interviewing or motivational feedback compared with control or treatment as usual.

Table 73

Summary of findings for the analysis of motivational interviewing or motivational feedback versus control or treatment as usual.

Group counselling versus treatment as usual for substance misuse

1 RCT (N=150) met the eligibility criteria for this review: Annis 1979 (Annis, 1979).

An overview of the trial can be found in Table 74. Further information about both included and excluded studies can be found in Appendix L. Summary of findings can be found in N= total number of participants

TAU=treatment as usual

1 Number randomised

Table 75. The full evidence profiles and associated forest plots can be found in Appendices N and O, respectively.

This was a 3-armed trial with service users being allocated to group counselling either with or without video-feedback, or treatment as usual. Outcomes with and without video feed-back are combined here. In the analysis the data for the two group counselling arms were pooled together and compared to the treatment as usual arms, preserving randomisation. When appraising quality it was considered that the two counselling arms used very similar interventions - differing only in the use of video feedback.

The available data for this review were of very low quality. No data were available for the outcomes of adaptive functioning, rates of self-injury or service utilisation.

Table 74

Study information table for trials included in the analysis of group counselling versus treatment as usual for substance misuse.

Table 75

Summary of findings for the analysis of group counselling versus treatment as usual.

Self-help versus control for substance misuse

1 RCT (N=183) met the eligibility criteria for this review: Proctor 2012 (Proctor et al., 2012).

An overview of the trial can be found in Table 76. Further information about both included and excluded studies can be found in Appendix L. Summary of findings can be found in N= total number of participants;

1 Number randomised

Table 77. The full evidence profiles and associated forest plots can be found in Appendices N and O, respectively.

The RCT had 2 arms, with service users randomly allocated either to either complete a self-help journal or to receive no intervention. The purpose of the journal was to assist service users to make a connection between their substance misuse and criminal activity and was based upon the trans-theoretical model of change.

The data for this review was of a low quality. No data were available for the outcomes of mental health, service utilisation, adaptive functioning or rates of self-injury.

Table 76

Study information table for trials included in the analysis of self-help versus control for substance misuse.

Table 77

Summary of findings for the analysis of self-help versus control for substance misuse.

6.4.1.1.2. Pharmacological interventions

Opioid antagonists

These drugs bind to opioid receptors without activating them, preventing the body from responding to opioids and endorphins in the same way as they would otherwise. Naloxone is also used as an antidote drug in instances of opioid overdose, whilst Naltrexone can help reverse the long-term neurochemical after-effects of opioid misuse, which is hypothesised to help prevent relapse.

Five RCTs (N=394) met the eligibility criteria for this review: Cornish 1997 (Cornish et al., 1997), Coviello 2010 (Coviello et al., 2010), Hanlon 1977 (Hanlon et al., 1977), (Lee et al., 2016; Lee et al., 2015) and Lobmaier 2010 (Lobmaier et al., 2010).

Naloxone versus placebo

One RCT (N=154) met the eligibility criteria for this review: Hanlon 1977

An overview of the trial can be found in Table 78. Further information about both included and excluded studies can be found in Appendix L. Summary of findings can be found in N= total number of participants

1 Number randomised

Table 79. The full evidence profiles and associated forest plots can be found in Appendices N and O, respectively.

The RCT by Hanlon 1977 had 2 arms and was conducted in a community setting.

The data for this review was of very low quality. No data were available for the outcomes of offending and reoffending, service utilisation, adaptive functioning and rates of self-injury.

Table 78

Study information table for trials included in the analysis of Naloxone versus placebo for drug misuse.

Table 79

Summary of findings table for the analysis of Naloxone versus placebo for drug misuse.

Naltrexone versus active intervention

Four RCTs (N=514) met the eligibility criteria for this review: Cornish 1997 (Cornish et al., 1997), Coviello 2010 (Coviello et al., 2010), Lee 2016 (Lee et al., 2016; Lee et al., 2015) and Lobmaier 2010 (Lobmaier et al., 2010).

An overview of the trials can be found in

Table 80. Further information about both included and excluded studies can be found in Appendix L.

Summary of findings can be found in N= total number of participants;

1 Number randomised

Table 81. The full evidence profiles and associated forest plots can be found in Appendices N and O, respectively.

These were all 2-armed trials with Naltrexone treatment in one arm and either a psychosocial intervention (3 trials) or Methadone (1 trial) in the other.

The data for this comparison was of very low quality. No data were available for the outcomes of quality of life and adaptive functioning.

Table 80

Study information table for trials included in the analysis of naltrexone versus active intervention for drug misuse.

Table 81

Summary of findings table for the analysis of naltrexone versus active intervention for drug misuse.

Opioid maintenance treatment

Opioid maintenance treatment aims to minimise the harms associated with opioid use, such as blood-borne illnesses associated with needle sharing.

There were 8 RCTs (N=1,565) that met the eligibility criteria for this review: Cropsey 2011 (Cropsey et al., 2011), Dolan 2003/²⁰⁰⁵ (Dolan et al., 2003; Dolan et al., 2005), Howells 2002 (Howells et al., 2002), Magura 2009 (Magura et al., 2009), Rich 2015 (Rich et al., 2015), Sheard 2009 (Sheard et al., 2009), Shearer 2006 (Shearer et al., 2006) and Wright 2011 (Wright et al., 2011).

Methadone maintenance versus waiting list control

Four papers from 3 separate RCTs (N=1,047) met the eligibility criteria for this review: Dolan 2003 and Dolan 2005 (Dolan et al., 2003; Dolan et al., 2005), Rich 2015 (Rich et al., 2015) and Shearer 2006 (Shearer et al., 2006).

An overview of the trials can be found in

Table 82. Further information about both included and excluded studies can be found in Appendix L.

Summary of findings can be found in N= total number of participants

TAU=treatment as usual

1 Number randomised

Table 83. The full evidence profiles and associated forest plots can be found in Appendices N and O, respectively.

These were 2-armed trials with service users randomly allocated to either a Methadone treatment arm or waiting list control or forced withdrawal arm.

The data for this comparison were of low to very low quality. No data were available for the outcomes of adaptive functioning and quality of life.

Table 82

Study information table for trials included in the analysis of methadone maintenance versus waiting list control for drug misuse.

Table 83

Summary of findings table for the analysis of methadone versus waiting list control for drug misuse.

Alpha-adrenergic agonists versus opioid maintenance for substance misuse

One RCT (N=68) met the eligibility criteria for this review: Howells 2002 (Howells et al., 2002).

An overview of the trial can be found in

Table 84. Further information about both included and excluded studies can be found in Appendix L.

Summary of findings can be found in N= total number of participants;

1 Number randomised

Table 85. The full evidence profiles and associated forest plots can be found in Appendices N and O, respectively.

This was a 2-armed trial comparing Lofexidine, an alpha-adrenergic agonist, with methadone. Lofexidine is typically managed in these settings to minimise symptoms of opiate withdrawal. This trial was conducted within a prison setting.

The quality of evidence for this review was low. No evidence was available for the outcomes of offending and reoffending, service utilisation, adaptive functioning or rates of self-injury.

Table 84

Study information table for trials included in the analysis of alpha-adrenergic agonists versus opioid maintenance for substance misuse.

Table 85

Summary of findings table for the comparison of alpha-adrenergic agonists versus opioid maintenance for substance misuse.

Opioid substitution therapy versus active intervention or placebo

Four RCTs (N=450) met the eligibility criteria for this review: Cropsey 2011 (Cropsey et al., 2011), Magura 2009 (Magura et al., 2009), Sheard 2009 (Sheard et al., 2009) and Wright 2011 (Wright et al., 2011).

An overview of the trials can be found in

Table 86. Further information about both included and excluded studies can be found in Appendix L.

Summary of findings can be found in N= total number of participants;

1 Number randomised

Table 87. The full evidence profiles and associated forest plots can be found in Appendices N and O, respectively.

Each study had 2 arms with buprenorphine in one arm and an alternative opioid substitute or placebo in the other. Three studies were conducted within a prison setting whilst one was conducted in the community.

The data were low to very low quality. No data were available for the outcomes of quality of life or adaptive functioning.

Table 86

Study information table for trials included in the analysis of opioid substitution versus active intervention for substance misuse.

Table 87

Summary of findings table for the analysis of opioid substitution versus active intervention or placebo for substance misuse.

6.4.1.2. Combined psychological and pharmacological interventions

Antidepressants plus psychological therapy versus psychological therapy alone for substance misuse

One RCT (N=60) met the eligibility criteria for this review: George 2011 (George et al., 2011).

An overview of the trial can be found in

Table 88. Further information about both included and excluded studies can be found in Appendix L.

Summary of findings can be found in N= total number of participants;

1 Number randomised

Table 89. The full evidence profiles and associated forest plots can be found in Appendices N and O, respectively.

The trial had 2 arms, with service users being randomly allocated to either receive fluoxetine, a selective serotonin reuptake inhibitor (SSRI), in addition to CBT and motivational therapy or just to receive CBT and motivational therapy. The authors report that fluoxetine was chosen for this study as SSRIs are hypothesised to modulate the processing of environmental stimuli to increase orbital frontal cortex function and accordingly reduce impulsive aggression. This trial was conducted in the community.

The available data for this review was of low quality. No data were available for the outcomes of offending and reoffending, adaptive functioning or rates of self-injury.

Table 88

Study information table for trials included in the analysis of antidepressants plus psychological therapy versus psychological therapy alone for substance misuse.

Table 89

Summary of findings table for antidepressants plus psychological therapy versus psychological therapy alone for substance misuse.

6.4.1.2.1. Support and educational interventions

Psychoeducation versus control or treatment as usual (TAU)

One RCT (N=60) met the eligibility criteria for this review: Brown 1980 (Brown, 1980).

An overview of the trial can be found in

Table 90. Further information about both included and excluded studies can be found in Appendix L.

Summary of findings can be found in N= total number of participants;

1 Number randomised

Table 91. The full evidence profiles and associated forest plots can be found in Appendices N and O, respectively.

The RCT had 3 arms with service users being allocated to either psychoeducation, educational drinking (where participants learned to control their drinking behaviour in an experimental bar facility) or treatment as usual. Only the psychoeducation and treatment as usual arms are included here. The psychoeducational intervention consisted of 3-hour sessions comprising a 30 minute talk, a 30 minute film and then a chaired group discussion. The types of topic covered included drinking and driving, effects of alcohol on physical health, effects upon family and how to modify drinking habits. Treatment as usual consisted of assigned tasks at the periodic detention centre.

The available data for this review was of very low quality. No data were available for the outcomes of offending or reoffending, adaptive functioning, service utilisation or rates of self-injury.

Table 90

Study information table for trials included in the analysis of psychoeducation versus control or treatment as usual for drug misuse.

Table 91

Summary of findings table for psychoeducation versus control or treatment as usual for drug misuse.

Employment workshop versus control or treatment as usual for substance misuse

Two RCTs (N=555) met the eligibility criteria for this review: Hall 1981 (Hall et al., 1981) and Webster 2014 (Webster et al., 2014).

An overview of the trials can be found in

Table 92. Further information about both included and excluded studies can be found in Appendix L.

Summary of findings can be found in N= total number of participants;

1 Number randomised

Table 93. The full evidence profiles and associated forest plots can be found in Appendices N and O, respectively.

Both studies were 2-armed trials conducted in the community. The experimental arm of both trials (employment workshop) consisted of a mixture of individual and group sessions designed to provide information, support and opportunities to practice skills needed to find and maintain employment and seek promotion. The control arm of the Hall 1981 study consisted of a 3-hour sign-posting meeting. The control arm of the Webster 2014 study consisted of treatment as usual. Although “treatment as usual” was not specified, it was assumed that the control group in Webster 2014, are likely to have had access to substance misuse support or treatment at least as effective as the control group’s 3-hour meeting in the Hall 1981 study. Thus, the outcomes were combined in the analyses.

The evidence for this review was of low to very low quality. No data were available for the outcomes of mental health, offending and reoffending, service utilisation or rates of self-injury.

Table 92

Study information table for trials included in the analysis of employment workshop versus control or treatment as usual for substance misuse.

Table 93

Summary of findings table for employment workshop versus control or treatment as usual for substance misuse.

6.4.1.2.2. Physical interventions

Acupuncture versus active intervention

Two RCTs (N=726) met the eligibility criteria for this review: Berman 2004 (Berman et al., 2004) and Konefal 1995 (Konefal et al., 1995).

An overview of the trials can be found in

Table 94. Further information about both included and excluded studies can be found in Appendix L.

Summary of findings can be found in N= total number of participants;

1 Number randomised

Table 95. The full evidence profiles and associated forest plots can be found in Appendices N and O, respectively.

Both studies were 2-armed trials. The Berman 2004 study compared two different forms of acupuncture, the NADA (National Acupuncture Detoxification Association) and the Helix protocols. This study was conducted within a prison setting. In the Konefal 1995 study service users in one arm received acupuncture in addition to frequent urine testing and in the other frequent urine testing only. This study was conducted in the community. Both studies used the NADA protocol for acupuncture in the intervention arm. The NADA protocol consists of 5 points chosen for their ability to assist with detoxification; Shen-Men, sympathetic, kidney, liver and lung. The Helix protocol involved acupuncture to the ear using five points on the helix of the ear

The evidence for this review was low to very low quality. No data were available for the outcomes of offending or reoffending, adaptive functioning or rates of self-injury.

Table 94

Study information table for trials included in the analysis of acupuncture versus active intervention for substance misuse.

Table 95

Summary of findings table for acupuncture versus active intervention for substance misuse.

6.4.1.3. Depression

Three RCTs (N = 206) met the eligibility criteria for this review: Gussak 2009, Johnson 2012 and Wilson 1990 (Gussak, 2008; Johnson & Zlotnick, 2012; Wilson, 1990). Gussak 2009 was arts-based psychotherapy (examples included construction of three-dimensional forms with few supplies); Johnson 2012 used interpersonal psychotherapy (IPT) intervention compared to psychoeducation whereas Wilson 1990 compared group cognitive treatment with individual supportive therapy. Due to the differences in the psychotherapy interventions data were not combined and separate analysis was done and presented for each study. In Johnson 2012, IPT was based on Wilfrey 2000 psychotherapy model while in Wilson 1990, group therapy was based on Hollon and Shaw 1979 cognitive treatment model.

An overview of the trials included in the meta-analysis can be found in

Table 96. Further information about both included and excluded studies can be found in Appendix L.

Summary of findings can be found in N= total number of participants

1 Number randomised

Table 97. The full GRADE evidence profiles and associated forest plots can be found in Appendices N and O, respectively.

No data were available for the outcomes of offending and reoffending, service utilization, adaptive functioning and rates of self-injury.

Table 96

Study information table for trials included in the meta-analysis of psychotherapy for depression.

Table 97

Summary of findings table of psychological intervention versus active intervention or no treatment for depression.

6.4.1.4. Individuals with suicidal risk

One RCT (N = 46) met the eligibility criteria for this review: Biggam 2002 (Biggam & Power, 2002). Principal training techniques in social problem-solving group therapy included instruction, active discussion, reflective listening and group exercises to practice the targeted skills. It was delivered in small group format (4-6 individuals/group). The participants in control did not receive principal training techniques.

An overview of the trials included in the meta-analysis can be found in

Table 98. Further information about both included and excluded studies can be found in Appendix L.

Summary of findings can be found in N= total number of participants; NR=Not reported

1 Number randomised

Table 99. The full GRADE evidence profiles and associated forest plots can be found in Appendices N and O, respectively.

No data were available for the outcomes of offending and reoffending, service utilization, adaptive functioning and rates of self-injury.

Table 98

Study information table for trials included in the meta-analysis of social problem-solving group therapy for vulnerable personality with suicidal risks.

Table 99

Summary of findings table of social problem-solving group therapy versus no treatment control for vulnerable personality with suicidal risks.

6.4.1.5. Anxiety Disorder

One RCT (N = 38) met the eligibility criteria for this review: Maunder 2009 (Maunder et al., 2009). The therapy was based on CBT-principles. The intervention group was provided with a booklet with a list of instructions and exercises and completed the time diary and thought about their personal reactions to the booklet. The participants in control group did not receive self-help booklets.

An overview of the trials included in the meta-analysis can be found in

Table 100. Further information about both included and excluded studies can be found in Appendix L.

Summary of findings can be found in N= total number of participants; NR=Not reported

1 Number randomised

Table 101. The full GRADE evidence profiles and associated forest plots can be found in Appendices N and O, respectively.

No data were available for the outcomes of offending and reoffending, service utilization, adaptive functioning and rates of self-injury.

Table 100

Study information table for trials included in the meta-analysis of self-help materials versus wait-list control for anxiety disorders.

Table 101

Summary of findings table of self-help materials versus wait-list control for anxiety disorders.

6.4.1.6. PTSD

Four RCTs (N =290) met the eligibility criteria for this review: Bradley 2003, Ford 2013, Cole 2007 and Valentine 2001 (Bradley & Follingstad, 2003; Cole et al., 2007; Ford et al., 2013; Valentine & Smith, 2001). Bradley 2003, Ford 2013 and Cole 2007 studies used group therapy method whereas Valentine 2001 applied traumatic incident reduction psychotherapy model. Bradley 2003 and Cole 2007 compared with no-contact and wait-list control respectively. Thus, outcomes were combined in analysis if they were measured by the same measurement tool. Ford 2013 study compared Trauma affect regulation: Guide for Education and Therapy (TARGET) with small group therapy (SGT). Moreover, Valentine 2001 study evaluated trauma incident reduction compared to wait-list controls. Data from these studies were analysed separately.

An overview of the trials included in the meta-analysis can be found in

Table 102. Further information about both included and excluded studies can be found in Appendix L.

Summary of findings can be found in N= total number of participants

TIR=Traumatic incident reduction

TARGET=Trauma Affect Regulation: Guide for Education and Therapy

PTSD=Post-traumatic stress disorders

1 Number randomised

Table 103. The full GRADE evidence profiles and associated forest plots can be found in Appendices N and O, respectively.

No data were available for the outcomes of offending and reoffending, service utilization, adaptive functioning and rates of self-injury.

Table 102

Study information table for trials included in the meta-analysis of psychotherapy versus no treatment/wait-list control/active treatment for post-traumatic stress disorders.

Table 103

Summary of findings table of psychotherapy vs wait-list control/No treatment/Active treatment for PTSD.

6.4.1.7. ADHD

Two studies (N = 84) met the eligibility criteria for this review: Ginsberg 2012 and Konstenius 2013 (Ginsberg et al., 2012b; Ginsberg & Lindefors, 2012a; Konstenius et al., 2013). The placebo and the methylphenidate capsules and packaging were identical in appearance. Data were combined by meta-analysis as the population, type of intervention and placebo were the same.

An overview of the trials included in the meta-analysis can be found in

Table 104. Further information about both included and excluded studies can be found in Appendix L.

Summary of findings can be found in N= total number of participants;

1 Number randomised

Table 105. The full GRADE evidence profiles and associated forest plots can be found in Appendices N and O, respectively.

No data were available for the outcomes of offending and reoffending, service utilization, adaptive functioning and rates of self-injury.

Table 104

Study information table for trials included in the meta-analysis of methylphenidate versus Placebo for Attention Deficit Hyperactivity Disorder (ADHD).

Table 105

Summary of findings table for Methylphenidate versus Placebo for Attention Deficit Hyperactivity Disorder.

6.4.1.8. Antisocial personality disorders

One RCT (N = 12) met the eligibility criteria for this review: Gowin 2012 (Gowin et al., 2012). This study took a total of 6 week, with no treatment in week 1, placebo to both groups in week 2, either tiagabine or placebo in week 3, 4 and 5 and placebo to both groups in week 6. Tiagabine was given orally in an escalating dose of 4, 8, 12 mg bd over 3 weeks. Outcomes were measured immediately post-treatment and during a period of follow-up

An overview of the trials included in the meta-analysis can be found in

Table 106. Further information about both included and excluded studies can be found in Appendix L. Summary of findings can be found in N= total number of participants;

1 Number randomised

Table 107. The full GRADE evidence profiles and associated forest plots can be found Appendices N and O, respectively.

No data were available for the outcomes of offending and reoffending, service utilization, adaptive functioning and rates of self-injury.

Table 106

Study information table for trials included in the meta-analysis of Tiagabine versus placebo for the antisocial personality disorder.

Table 107

Summary of findings table for Tiagabine versus placebo for the antisocial personality disorder.

6.4.1.9. Severe mental illness

6.4.1.9.1. Pharmacological interventions

One RCT (N=450) met the eligibility criteria for the review: Alphs 2015a (Alphs et al., 2015). In this study, Paliperidone was given intramuscularly with 234 mg on day 1 and 158 mg on day 8 with monthly maintenance range of 78-238 mg thereafter from day 38. Patients on oral regime received aripiprazole 33 (15.1%), haloperidol 15 (6.9%), olanzapine 36 (16.5%), paliperidone 48 (22%), perphenazine 20 (9.2%), quetiapine 29 (13.3%) or risperidone 37 (17%).

An overview of the trials included in the meta-analysis can be found in

Table 108. Further information about both included and excluded studies can be found in Appendix L.

Summary of findings can be found in N= total number of participants;

1 Number randomised

Table 109. The full GRADE evidence profiles and associated forest plots can be found Appendices N and O, respectively.

No data were available for the outcomes of offending and reoffending, service utilization, adaptive functioning and rates of self-injury.

Table 108

Study information table for trials included in the meta-analysis of IM pahpendone versus oral antipsychotics for schizophrenia.

Table 109

Summary of findings table for palip Schizophrenia.

6.4.1.9.2. Psychological intervention

Two RCTs (N = 204) met the eligibility criteria for this review: Bond 2015 and Clayton 2013 (Bond et al., 2015; Clayton et al., 2013). The Bond 2015 study recruited participants with no competitive job placement in previous three months. The study used Individual Placement and Support (IPS) model, supported by employment specialist and the aim was to help identify and prepare for job search. On the other hand, the Clayton 2013 study included participants with severe mental illness who had a criminal charge in the 2 years prior to enrolment for the Citizenship project. The project consisted of three integrated components: individual peer mentor support (8 hours/week), an 8-week citizenship class and an 8-week valued role component. The separate analysis was performed for different intervention.

An overview of the trials included in the analysis can be found in Table 110. Further information about both included and excluded studies can be found in Appendix L.

Summary of findings can be found in Table 111 and Table 112. The full GRADE evidence profiles and associated forest plots can be found in Appendices N and O, respectively.

No data were available for the outcomes of service utilization and rates of self-injury.

Table 110

Study information table for trials included in the analysis of psychosocial intervention versus treatment as usual for severe mental illness.

Table 111

Summary of findings table for the Citizenship Project versus TAU for Severe Mental Health Disorders.

Table 112

Summary of findings table for Individual Placement and Support (IPS) vs Work choice models for severe mental illness.

6.4.1.10. Interventions for uncategorized mental health disorders

6.4.1.10.1. Parenting from the inside

One RCT (N=176) met the eligibility criteria for this review: Loper 2011 (Loper, 2011). The intervention was based on cognitive behavioural strategy and the sessions were focused on connecting with one’s own children emotionally and guiding as a parent, while they were in prison.

An overview of the trials included in the meta-analysis can be found in

Table 113. Further information about both included and excluded studies can be found in Appendix L. N= total number of participants;

1 Number randomised

Table 114. The full GRADE evidence profiles and associated forest plots can be found in Appendices N and O, respectively.

No data were available for the outcomes of offending and reoffending, service utilization, adaptive functioning and rates of self-injury.

Table 113

Study information table for trials included in the analysis of parenting from inside versus wait-list control for mental health disorders.

Table 114

Summary of findings table for parenting from inside vs Wait-list control for mental health disorders.

6.4.1.10.2. Music Therapy

Two RCTs (N = 215) met the eligibility criteria for this review: Chen 2015 and Hakvoort 2013 (Chen et al., 2015; Hakvoort et al., 2013). The two studies used standard care and wait-list control as comparison respectively and the underlying mental health disorders and the reported mental outcome measures were also different. Thus, separate analysis was performed for each study.

An overview of the trials included in the meta-analysis can be found in Table 115. Further information about both included and excluded studies can be found in Appendix L.

Summary of findings can be found in N= total number of participants;

1 Number randomised

Table 116 and 1 Chen 2015 - Appropriate randomization with proper concealment; blinding of care administrators, but not participants; ITT analysis; appropriate outcome report

Table 117. The full GRADE evidence profiles and associated forest plots can be found in Appendices N and O, respectively.

No data were available for the outcomes of offending and reoffending, service utilization, adaptive functioning and rates of self-injury.

Table 115

Study information table for trials included in the analysis of music therapy versus standard care or wait-list control for mental health disorders.

Table 116

Summary of findings table for music therapy vs standard care for depression and anxiety disorders.

Table 117

Summary of findings table for music therapy vs wait-list control for antisocial personality disorders.

6.4.2.1. Systematic literature review

The systematic search of the literature identified 3 studies that assessed the costs and benefits of interventions for adults with substance misuse disorders who are in contact with the criminal justice system.

Of these:

• One study examined the costs and benefits associated with a psychosocial intervention in the US (Daley et al., 2004)
• Two studies examined the costs and benefits associated pharmacological interventions in Australia (Gisev et al., 2015; Warren et al., 2006)

No studies assessing the cost effectiveness of psychological, social, pharmacological or physical interventions for other disorders recommended in existing NICE guidance for adults who are in contact with the criminal justice system, were identified by the systematic search of the economic literature undertaken for this guideline. Details on the methods used for the systematic review of the economic literature are described in Chapter 3; full references and evidence tables for all economic evaluations included in the systematic literature review are provided in Appendix S. Completed methodology checklists of the studies are provided in Appendix R. Economic evidence profiles of studies considered during guideline development (that is, studies that fully or partly met the applicability and quality criteria) are presented in Appendix T.

6.4.2.2. Psychosocial interventions

6.4.2.3. Pharmacological interventions

6.4.3.1. Substance misuse

6.4.3.2. Depression

Very low quality evidence from one RCT (N=38) indicated interpersonal psychotherapy had a clinically important effect on depression by HRSD scales at post-treatment when compared to a psychoeducation intervention, but this effect was uncertain at 13-weeks follow-up.

Very low quality evidence from one RCT (N=10) indicated uncertainty about the effectiveness of group cognitive treatment compared to individual support therapy in terms of depression by BDI scales, Hopeless scales, MMPI D scales and Multiple affect adjective checklist D scales at post-treatment and in depression symptoms by MMPI D scales at 39-weeks follow-up.

Very low quality evidence from one randomized study (N=158) showed that arts-based therapy had a clinically significant effect on Adult Nowicki-Strickland Locus (ANS) of control scale and depression scales by BDI in male, female and combined groups in comparison with no treatment control. Low quality evidence from this trial indicated arts-based therapy had a clinically significant effect on two of the formal elements of the arts therapy scale rating guide (FEATS), prominence of colour and colour fit when compared to no treatment control.

6.4.3.3. Vulnerable inmates with suicidal risk

One randomized study (N=46) provided very low quality evidence of a clinically significant beneficial effect of a social problem solving group on depression symptoms by either HADS or Beck Hopeless scales at post-treatment and 13-weeks follow-up compared to no treatment control. Similarly, it is clinically significant that the social group therapy decreased anxiety symptoms by HADS scales at post-treatment and 13-weeks follow-up. Low quality evidence from this trial indicated decision making ability as measured by SPSI:R scales was improved by a clinically significant amount in the social problem solving group therapy compared to the no treatment control group.

6.4.3.4. Anxiety Disorders

Very low quality evidence from one randomized study (N=33) indicated uncertainty about the effectiveness of psychological therapy with self-help compared to wait-list control in terms of anxiety symptoms by HADS scales at post-treatment. This trial, however, reported a clinically significant reduction in anxiety with the self-help materials compared to wait-list control after 4-weeks of follow-up.

6.4.3.5. PTSD

Very low quality evidence from two randomized studies (N=40) indicated uncertainty about the effectiveness of psychological therapy (group method) compared to wait-list or no-contact control in terms of trauma symptoms by TSI scales.

Very low quality evidence from one RCT (N=123) reported a clinically significant decrease in depression symptoms by BDI scale and PTSD symptoms by PSS scales at either post-treatment or 13-weeks follow-up, increase in post-treatment generalised expectancy for success scales as well as increase in 13-weeks follow-up clinical anxiety scales with TIR intervention relative to wait-list control. The clinical effects on generalised expectancy for success scales at 13-weeks follow-up and clinical anxiety scales at post-treatment were uncertain and of very low quality.

One RCT provided moderate quality evidence of uncertainty about the effectiveness of TARGET intervention compared to SGT intervention in terms of post-treatment PTSD symptoms by CAPS scales (N=72) and Heartland forgiveness scales (N=32; low quality evidence).

Very low to low quality evidence from one randomized study (N=9) suggested a clinically important beneficial effect of focused group therapy on global severity index, positive symptom distress index and total positive symptom index by symptom checklist-90R compared to wait-list control at post-treatment.

Very low quality evidence from one RCT (N=31) indicated uncertainty about the effectiveness of group therapy compared to no contact control in terms of IIP-32 scales.

6.4.3.6. ADHD

Very low quality evidence from two RCTs (N=84) suggested a clinically important effect of methylphenidate on reduction in ADHD symptoms by CAARS:OSV scales at post-treatment as compared to placebo. Low quality evidence from one RCT (N=20) indicated a clinically significant reduction in ADHD symptoms at 3-years follow-up with methylphenidate compared to placebo.

Very low quality evidence from one RCT (N=54) indicated uncertainty about the effectiveness of methylphenidate compared to placebo in terms of the number of participants with drug negative urine at post-treatment.

6.4.3.7. Antisocial personality disorder

Very low quality evidence from one randomized study (N=12) indicated a clinically significant decrease in aggressive response with tiagabine compared to placebo at post-treatment.

However, there was uncertainty about difference in the number of adverse effects in the two groups.

6.4.3.8. Severe mental illness

6.6.1.1. Pharmacological interventions

Three RCTs (N = 84) met the eligibility criteria for this review. These trials were identified in a systematic review (Khan et al., 2015) of seven RCTs, four of which were excluded from this review because they involved psychiatric inpatients or were cross-over trials where the first phase data could not be extracted.

The included trials involved medroxyprogesterone acetate (MPA) a synthetic progesterone proposed to supress sexual desire by countering the libidinal effects of testosterone. Hucker et al. (1988) compared MPA with placebo for paedophilia in the outpatient setting. Langevin et al. (1979) and McConaghy et al. (1988) examined the addition of MPA to outpatient psychological interventions for exhibitionism or varied paraphilic disorders respectively.

An overview of the trials included in the meta-analysis can be found in

Table 119. Further information about both included and excluded studies can be found in Appendix L.

Summary of findings can be found in N= total number of participants

MPA = Medroxyprogesterone acetate

Psych= Psychosocial

mg/day = milligrams per day

CJS = Criminal justice system

1 Number randomised

Table 120 and 1 Downgraded for inconsistency

2. Confidence interval of the effect estimate includes appreciable benefit, harm and no effect

3. High risk of selection and performance bias

4. High risk of selection and performance bias.

Table 121. The full GRADE evidence profiles and associated forest plots can be found in Appendices N and O, respectively.

Table 119

Study information table for trials included in the meta-analysis of pharmacological interventions for paraphilia.

Table 120

Summary of findings table for medroxyprogesterone + psychological intervention compared to psychological intervention only for paraphilic disorders.

Table 121

Summary of findings table for medroxyprogesterone compared to placebo for paraphilic disorders.

6.6.1.2. Psychoeducational interventions

Three RCTs (N=779) and 23 non-randomised controlled trials (N=12317) met the eligibility criteria for this review. The randomised trials were identified in a systematic review (Dennis et al., 2012): Anderson Varney (1991) and Hopkins (1991) compared CBT based psychoeducational interventions to treatment as usual or waiting list control. Marques et al. (1994a) examined the California Sex Offender Treatment and Evaluation Project. No trials published after this systematic review were found.

The non-randomised controlled trials (Abracen et al., 2011; Aytes et al., 2001; Craissati & McClurg, 1997; Craissati et al., 2009; Di Fazio et al., 2001; Duwe & Goldman, 2009; Friendship et al., 2003; Hanson et al., 2004; Looman et al., 2000; Lowden et al., 2003; Marshall et al., 2008; McGrath et al., 2003; McGrath et al., 1998; McGuire, 2000; O’Reilly et al., 2010; Olver et al., 2013a; Redondo Illescas & Garrido Genoves, 2008; Ruddijs & Timmerman, 2000; Song & Lieb, 1995; Stalans et al., 2001; Turner et al., 2000) involved primarily group CBT based psychoeducation (including SOTP). Content of the psychoeducation included: offence disclosure, accepting responsibility, cognitive distortions/cognitive restructuring, victim empathy, offending cycle, individual risk factors and recognition cues, relapse prevention and social skills. Methods included group discussion, exposure to video or audio accounts presented by victims, positive modelling, role-play, skills practice and decision matrices. The control groups received either no treatment, treatment as usual (which was not specified) or were waitlist controls.

An overview of the trials included can be found in

Table 122. Further information about both included and excluded studies can be found in Appendix L.

Summary of findings can be found in CBT= cognitive behavioural therapy

SOTP= sex offender treatment programme

RCT= randomised controlled trial

Table 123 and 1 Anderson-Varney 1991 - unclear risk of selection bias; no blinding; low risk of attrition bias; low risk of selective outcome bias; low risk of other bias

2 The MID calculated from SD of control was +/-5.41.

3 The MID calculated from SD of control was +/-6.01.

4 Hopkins 1991 - Unclear selection bias; No blinding; low risk of attrition bias; low risk of selective outcome bias; low risk of other bias.

5 Hopkins 1991 - Participants involved roughly equal numbers of incarcerated paedophile and rapists

6 Indirectness – inpatient setting

7 Imprecision – the CI for the effect spans no effect and both MID thresholds

Table 124. The full GRADE evidence profiles and associated forest plots can be found in Appendices N and O, respectively.

Significant heterogeneity was noted in some of the outcomes and subgroup analysis was done according to country (which reduced heterogeneity) and reported in the summary of findings tables. No other sources of heterogeneity were identified: the non-randomized controlled trials were typically not adjusted for confounders.

Table 122

Study information table for trials included in the analysis of psychoeducational interventions for paraphilia.

Table 123

Summary of findings table (RCTs) for psychoeducational interventions, principally CBT-informed psychoeducation (including SOTP) versus treatment as usual, no treatment or waitlist control for paraphilic disorders.

Table 124

Summary of findings table (observational studies) for psychoeducational interventions, principally CBT-informed psychoeducation (including SOTP) versus treatment as usual, no treatment or waitlist control for paraphilic disorders.

6.6.1.3. Good Lives Model (GLM) versus Relapse Prevention (RP)

Two non-randomised controlled studies (N=1278) met the eligibility criteria for this review. Barnett et al. (2014) and Harkins et al. (2012) compared the Good Lives Model or the revised Better Lives Model with standard relapse prevention in UK sexual offenders against children.

An overview of the trials included can be found in

Table 125. Further information about both included and excluded studies can be found in Appendix L.

A summary of findings can be found in Table 126. The full GRADE evidence profiles and associated forest plots can be found in Appendices N and O, respectively.

Table 125

Study information table for trials included in the analysis of Good Lives Model versus Relapse Prevention for paraphilia.

Table 126

Summary of findings table for Good Lives Model (GLM) versus Relapse Prevention (RP) for paraphilic disorders.

6.6.1.4. Reintegration programmes

One RCT (Duwe, 2013)(N=62) and three non-randomised controlled trials (Bates et al., 2014; Wilson et al., 2009; Wilson et al., 2007b) (N=350) met the eligibility criteria for this review. All of the studies involved the Circles of Support and Accountability (COSA) intervention. The COSA inner circle consists of the core member (the sex offender) and up to six volunteers from the community. The COSA outer circle consists of, supervision agents, law enforcement personnel and treatment professionals. Volunteers are recruited and trained in preparation for their role, with topics covered such as typology, manipulation, personal boundaries and managing risk. The goal for each circle is to provide the core member with support during their reintegration into the community.

An overview of the trials included can be found in

Table 127. Further information about both included and excluded studies can be found in Appendix L.

Summary of findings can be found in Table 128 and Table 129. The full GRADE evidence profiles and associated forest plots can be found in Appendices N and O, respectively.

Table 127

Study information table for trials included in the analysis of reintegration programmes versus treatment as usual for paraphilia.

Table 128

Summary of findings table (RCTs) for reintegration programmes versus treatment as usual for paraphilic disorders.

Table 129

Summary of findings table (observational studies) for reintegration programmes versus treatment as usual for paraphilic disorders.

6.6.1.5. Therapeutic communities

One non-randomised controlled trial (N=1217) met the eligibility criteria for this review. Lowden et al. (2003) involved the Sex Offender Treatment and Monitoring Programme (SOTMP) phase 1 and 2, a modified sex offender therapeutic community housing inmates together in a therapeutic milieu where individuals work and live with others who are working on similar treatment issues.

An overview of the trial can be found in Table 130. Further information about both included and excluded studies can be found in Appendix L.

Summary of findings can be found in Table 131. The full GRADE evidence profiles and associated forest plots can be found in Appendices N and O, respectively.

Table 130

Study information table for trials included in the analysis of therapeutic communities versus no treatment for paraphilia.

Table 131

Summary of findings table for Therapeutic communities versus no treatment for paraphilic disorders.

6.6.1.6. Cognitive behavioural therapy

One non-randomised controlled trial (N=61) met the eligibility criteria for this review. Marshall et al. (1991) involved a treatment programme that conceptualized exhibitionism in cognitive and social terms, rather than simply sexual motivation. Treatment was aimed at teaching skills to deal with all sources of stress and intervention content included: assertiveness training; stress management; cognitive restructuring; training in relationship skills.

An overview of the trial can be found in Table 132. Further information about both included and excluded studies can be found in Appendix L.

Summary findings can be found in Table 133. The full GRADE evidence profiles and associated forest plots can be found in Appendices N and O, respectively.

Table 132

Study information table for trials included in the analysis of cognitive behavioural therapy versus treatment as usual for paraphilia.

Table 133

Summary of findings table for cognitive behavioural therapy (CBT) versus treatment as usual for paraphilic disorders.

6.6.1.7. Behavioural therapy

Two randomised controlled trials (McConaghy et al., 1985; McConaghy et al., 1988) and two non-randomised controlled trials (Marshall & Barbaree, 1988b; Marshall et al., 1991) (N=187) met the eligibility criteria for this review.

McConaghy et al. (1985) compared imaginal desensitization to covert sensitization for varied paraphilic disorders. McConaghy et al. (1988) examined the addition of imaginal desensitization to MPA for varied paraphilic disorders. Marshall & Barbaree (1988b) and Marshall et al., (1991) compared behavioural treatment programs to treatment as usual in paedophiles and exhibitionists respectively

An overview of the trials can be found in Table 134. Further information about both included and excluded studies can be found in Appendix L.

Summary of findings can be found in N= total number of participants; NR=Not reported; MPA = medroxyprogesterone

1 Number randomised

Table 135, 1 Marshall 1988a/^b/¹⁹⁹¹ - Non-RCT with 4 and 9-year follow-up; No baseline risk differences; No blinding; unclear attrition risk of bias; low risk of selective outcome bias; low risk of other bias

2 The 95% CI considered for imprecision was 0.80 to 1.25.

Table 136 and 1 McConaghy 1988 - unclear risk of selection bias, no blinding, low risk of attrition bias, high risk of selective outcome bias, low risk of other bias.

2 Unclear what percentage are currently in contact with the criminal justice system

3 The 95% CI considered for imprecision was 0.80 to 1.25.

Table 137. The full GRADE evidence profiles and associated forest plots can be found in Appendices N and O, respectively.

Table 134

Study information table for trials included in the analysis of behavioural therapy for paraphilia.

Table 135

Summary of findings table for behavioural therapies versus treatment as usual for paraphilic disorders.

Table 136

Summary of findings table behavioural therapy plus medroxyprogesterone versus other medroxyprogesterone alone.

Table 137

Summary of findings table for imaginal desensitization versus covert sensitization for paraphilic disorders.

6.6.1.8. Aversive conditioning training and milieu therapy

One non-randomised controlled trial (N=197) met the eligibility criteria for this review. Hanson et al. (1993) examined a specialised treatment programme provided in a separate minimum security setting. The programme aimed to increase the social competence of offenders through individual and group counselling and by creating a therapeutic milieu that encouraged the men to recognise and correct social and sexual adjustment problems. The offenders also received aversive conditioning training, involving pairing shocks to stimulus sets tailored for each participant on the basis of their offence history.

An overview of the trial can be found in Table 138. Further information about both included and excluded studies can be found in Appendix L.

Summary of findings can be found in N= total number of participants; NR=Not reported

1 Number randomised

Table 139. The full GRADE evidence profiles and associated forest plots can be found in Appendices N and O, respectively.

Table 138

Study information table for trials included in the analysis of aversive conditioning training and milieu therapy for paraphilia.

Table 139

Summary of findings table for aversive conditioning and milieu therapy versus treatment as usual for paraphilic disorders.

6.6.1.9. Individual and Group Psychotherapy

Two non-randomised controlled trials (Craissati 2009, Peters 1968; N=440) met the eligibility criteria for this review. Peters et al. (1968) involved a group psychotherapy programme for sex offenders. Craissati et al. (2009) involved individual supportive psychotherapy for sex offenders deemed inappropriate for more structured offence-focused treatment due to disruptive or antagonistic behaviour or denial of the offence

An overview of the trials can be found in Table 140. Further information about both included and excluded studies can be found in Appendix L.

Summary of findings can be found in N= total number of participants; NR=Not reported

1 Number randomised

Table 141. The full GRADE evidence profiles and associated forest plots can be found in Appendices N and O, respectively.

Table 140

Study information table for trials included in the analysis of psychotherapy versus no treatment or treatment as usual for paraphilia.

Table 141

Summary of findings table for psychotherapy versus no treatment or treatment as usual for paraphilic disorders.

6.6.1.10. Polygraph testing

One controlled non randomised study (N=208) met the eligibility criteria for this review. McGrath et al. (2007) examined the effectiveness of periodic polygraph compliance exams as a condition of probation or parole in a group of primarily sex offenders against children.

An overview of the trial can be found in Table 142. Further information about both included and excluded studies can be found in Appendix L.

Summary of findings can be found in N= total number of participants; NR=Not reported

1 Number randomised

Table 143. The full GRADE evidence profiles and associated forest plots can be found in Appendices N and O, respectively.

Table 142

Study information table for trials included in the analysis of polygraph testing versus treatment as usual for paraphilia.

Table 143

Summary of findings table for polygraph testing versus treatment as usual for paraphilic disorders.

6.6.3.1. Pharmacological interventions

Very low quality evidence from two randomised controlled trials (N=66) indicated uncertainty about the benefit of adding MPA to psychosocial interventions for paraphilia in terms of anomalous desires or behaviour.

Low quality evidence from one randomised controlled trial (N = 32) indicated that adding MPA to a psychosocial intervention for paraphilia increased the risk of study dropout by a clinically important amount compared to the psychological intervention alone.

Very low quality evidence from one randomised controlled trial (N=20) indicated uncertainty about whether MPA alone was more effective than imaginal desensitization alone in terms of anomalous desires or behaviour.

6.6.3.2. Psychoeducational interventions

Moderate quality evidence from one randomised controlled trial (N=60) indicated that a psychoeducational CBT intervention reduced cognitive distortions by a clinically important amount when compared to no treatment. This trial provided low quality evidence of uncertainty about the effects of psychoeducation on acceptance of accountability, sexual anxiety and anxiety.

The only evidence about reconviction from randomised studies was from a single trial in the inpatient setting (N =480), which provided low quality evidence of uncertainty about the benefit of psychoeducation in terms of sexual or violent reconviction rates.

Very low quality evidence from nine non-randomised controlled trials (N=2796) indicated that psychoeducational interventions led to a clinically important reduction in reconviction rates when compared to treatment as usual, no treatment or waitlist control. This was also the case when restricting the analysis to UK studies (1 study; N=338).

There was very low quality evidence from 11 non-randomised controlled trials (N=5261) that psychoeducational interventions were associated with a clinically important reduction in reconviction rates for sexual offenses when compared to treatment as usual, no treatment or waitlist control. There was very low quality evidence of uncertainty about the effect of psychoeducation on reconviction for sexual offenses when restricting the analysis to UK studies (three studies; N=2885).

There was very low quality evidence from 6 non-randomised controlled trials (N=2181) that psychoeducational interventions were associated with a clinically important reduction in reconviction rates for violent offenses when compared to treatment as usual, no treatment or waitlist control. There was very low quality evidence of uncertainty about the effect of psychoeducation on reconviction for violent offenses when restricting the analysis to UK studies (1 study; N=240).

There was very low quality evidence from 6 non-randomised controlled trials (N=2181) of uncertainty about the effect of psychoeducational interventions on revocation rates when compared to treatment as usual, no treatment or waitlist control. There was very low quality evidence of a clinically important reduction in revocation rates with psychoeducation when restricting the analysis to UK studies (1 study; N=240).

6.6.3.3. Good Lives Model (GLM) versus Relapse Prevention (RP)

Very low quality evidence from one non-randomised controlled trial (N=501) suggested that the Good Lives Model reduced cognitive distortions and emotional congruence with children by a clinically important amount when compared with relapse prevention.

Very low quality evidence from one non-randomised controlled trial (N=501) suggested no clinically important difference in the effectiveness of the Good Lives Model and relapse prevention in terms of victim empathy distortions.

Very low quality evidence from one non-randomised controlled trial (N=2698) suggested uncertainty about the relative treatment dropout rates associated with the Good Lives Model and relapse prevention.

Low quality evidence from one RCT (N=587) indicated no clinically important difference in the effectiveness of Good Lives Model and relapse prevention for reducing pro-offending attitudes.

There was no evidence about the relative effectiveness of the Good Lives Model and relapse prevention in terms of offending or reoffending.

6.6.3.4. Reintegration programmes

One randomised trial (N = 62) provided very low quality evidence that a support group, Circles of Support and Accountability, reduced rates of re-arrest at two years of follow-up by a clinically important amount compared to treatment as usual. From this trial there was very low quality evidence of uncertainty about the relative effectiveness of the support group compared to treatment as usual in terms of: sex offence re-arrest, reconviction, resentence and re-incarceration at two years follow up.

Very low quality evidence from three non-randomised controlled trials (N=350) indicated reintegration programmes were associated with clinically important reductions in reconviction rates (including for sexual offenses) when compared to treatment as usual. Restricting the analysis to UK only studies there was uncertainty whether reintegration programmes were more effective than treatment as usual.

6.6.3.5. Therapeutic communities

Very low quality evidence from one non-randomised controlled trials (N=1217) indicated therapeutic communities reduced rates of re-arrest, incarceration and revocation by a clinically important amount compared with no treatment. There was however uncertainty about the effectiveness of the therapeutic community intervention in terms of re-arrest or incarceration when looking at specific sexual or violent offenses.

6.6.3.6. Cognitive behavioural therapy

Very low quality evidence from one non-randomised controlled trial (N=38) indicated uncertainty about whether CBT reduces the rate of sexual reconviction when compared with treatment as usual.

6.6.3.7. Behavioural therapy

Very low quality evidence from a small non-randomised controlled trial (N=44) suggested a behavioural treatment programme had a clinically important effect on sexual reconviction rates at 4 years of follow up in sex offenders against children but there was uncertainty about its effectiveness for sexual reconviction rates at 9 years in exhibitionists.

Very low quality evidence from a randomised trial (N=20) indicated uncertainty about the relative effectiveness of imaginal desensitization plus MPA versus MPA alone in terms of anomalous desires and behaviours.

Very low quality evidence from a randomised trial (N=20) indicates uncertainty about the relative effectiveness of imaginal desensitization versus covert sensitisation alone in terms of anomalous desires and behaviours.

6.6.3.8. Aversive conditioning training and milieu therapy

Very low quality evidence from one non-randomised controlled trial (N=197) indicated uncertainty about whether aversive conditioning training and milieu therapy is more or less effective than treatment as usual in terms of sexual or violent reconvictions.

6.6.3.9. Individual and Group Psychotherapy

Low quality evidence from one non-randomised controlled trial (N=167) indicated a clinically important reduction in re-arrest rates following psychotherapy for paraphilic disorders when compared to treatment as usual.

Low quality evidence from two non-randomised controlled trials (N=335) indicated uncertainty about the effectiveness of psychotherapy for paraphilic disorders when compared to no treatment or treatment as usual in terms of sex-offence re-arrest or reconviction, violent reconviction and breaches of the sex offender register.

6.6.3.10. Polygraph testing

Very low quality evidence from a non-randomised controlled trial (N=208) indicated uncertainty about the effectiveness of periodic polygraph compliance exams when compared to treatment as usual in terms of sexual reconviction, incarceration or violation of supervision conditions. Violent reconviction, however, was reduced by a clinically important amount in the polygraph testing group.