This section discusses the evidence in relation to the original review questions and highlights key findings that cross all review questions, including the development of a typology of peer-based interventions and the impact on peer deliverers. Review findings are discussed in terms of relevance to health services, but particularly those services that are operating within and in partnership with prisons in England and Wales. Key issues for policy and practice identified through the expert symposium and listening exercises with serving prisoners are discussed in relation to the review findings.
Developing a typology of peer-based interventions in prison settings
This study has confirmed that there is considerable heterogeneity in the range of peer-based interventions in the prison setting, in terms of both the health issues addressed and the mode of delivery (see Chapter 4). To group studies to review and summarise evidence it was necessary to develop a new categorisation of peer-based interventions in prison settings. The original role categories for community-based lay health workers developed through the People in Public Health study97 did not provide a good fit for the interventions described in studies included in this review, although some of the roles reflected dimensions identified in that study. For example, peer advisors, who provide housing and resettlement advice, demonstrated some aspects of bridging roles in terms of helping prisoners to access welfare services.
A new classification for peer interventions was developed that provided a better fit with the data ( and see Appendix 7).
The typology includes a number of intervention models that are currently operating in the prison system in England and Wales, including the Listener scheme, which covers most prisons in England and Wales, insiders, peer advisors and health trainers. The typology broadly reflects the range of peer support schemes identified by Levenson and Farrant32 in 2002 and Edgar and colleagues25 in 2011. The range of intervention types undoubtedly reflects the review strategy, with the inclusion of an expert symposium and the emphasis on identifying UK grey literature. A limitation is that these models may not be transferable to other contexts.
The final typology is not necessarily comprehensive. The review found that there are few standard models and there is much variation in intervention design, peer roles, recruitment, training and implementation. Some of the included studies reported that peers may undertake additional roles outside of the intervention.38 Developing the classification inductively from reported definitions limits the extent to which we have been able to group interventions. For example, it can be argued that on a theoretical level peer mentoring is not a distinct intervention but is a form of peer support as it provides appraisal support.94 However, for the purposes of the review the typology provided a useful framework and offers a basis for further analysis of intervention modes.
Review question 1: what are the effects of peer-based interventions on prisoner health?
Peer education
Peer education in the prison setting involves prisoners receiving training and then acting as educators, communicating information and encouraging the uptake of healthy (or less risky) behaviours.24 The review confirmed that there is a body of literature reporting the effects of prison-based peer education interventions; most of these refer to the prevention of HIV/hepatitis C virus infection.24,35 Although a sizeable number of peer education studies were included (n = 18), they were of variable quality, with only a small number of studies having a strong design, which limits the conclusions that can be drawn. There was moderate evidence from quantitative studies that peer education interventions can result in changes in HIV/hepatitis C virus knowledge, but equivocal results for effects on behaviour change intentions and health beliefs. For health behaviours there was consistent evidence of peer education resulting in the reduction of risky behaviours, for example sharing needle equipment or not using a condom at first intercourse post release. Additionally, there was weak evidence indicating an association between the uptake of screening/HIV testing and peer health education programmes.125,144 These findings support rationales for peer education as a means of increasing social influence and positive social norms,24,95 but further research is needed to explore the relative importance of peer education as a factor in the uptake of these health services.
Although there was limited evidence on peer education from qualitative studies, the study by Scott and colleagues131 reported an interesting finding that peer education was diffused outside of the prison to family and friends. This is an area that would merit further exploration in future intervention studies and may have implications for cost-effectiveness. The development and empowerment of peer educators can be an important component in some peer education approaches,166 and there was moderate evidence from qualitative studies that peer educators benefited through rewarding experiences, the acquisition of skills, the development of supportive networks and improved mental health.
Overall, the findings on peer education support the conclusions of earlier reviews, including the systematic review on peer health promotion conducted by Wright and colleagues,35 which was based on a smaller number of included studies. Interpreting the findings within a harm reduction approach,104,105 the review provides evidence that peer education interventions are effective at reducing risky behaviours, which can be regarded as intermediate health outcomes.102 Despite the moderate evidence of effects for peer education, the review findings have low relevance for health services operating in prisons in England and Wales. Two descriptive reviews of volunteering in prison25,32 and the results from the expert symposium suggest that peer education programmes are not prominent in current practice. There are, however, some promising results for Toe by Toe, the peer-based literacy scheme widely implemented across the prison service in England and Wales. A high proportion of learners reported positive educational outcomes and had high levels of satisfaction with the programme. Education is a social determinant of health that may be associated with other positive outcomes for the prison population.3 As the review identified only one poor-quality cross-sectional study on this initiative, there is insufficient evidence to draw conclusions and further research is recommended on this specific scheme.
Peer support
Peer support in a prison setting involves prisoners providing practical help, social and emotional support and advice to other prisoners in a paid or voluntary capacity.22,25 The review has confirmed that the focus of the intervention and the role of the peer support worker vary considerably between different interventions. The review did not examine evidence about the value of informal peer support and natural social networks in the prison setting,94 which may be a considerable protective factor for prisoners’ mental health.103
Quantitative evidence on peer support was exclusively drawn from the Canadian PST model. All six included studies were based on a common model of peer support within women’s prisons, which allows some tentative conclusions to be drawn. The PST programme had no demonstrable effects on prisoners or the prison environment, but the programme was rated highly in terms of satisfaction across a number of variables including the usefulness of peer support sessions, the approachability of PST members, levels of trust and handling crisis interventions. The model, which is described as a women-centred approach aiming for greater prisoner empowerment,153,155 has some similar features to other schemes in operation in England and Wales, such as Insiders and Listeners, with its focus on befriending and emotional support to meet the needs of prisoners on an individual basis.
Across the range of peer support interventions, there was moderate qualitative evidence from 10 studies on the positive effects of peer support and this triangulated with some of the survey data on satisfaction from the evaluations of the Canadian PST model. The review found that peer support was beneficial in terms of both practical assistance and helping prisoners overcome mental health problems such as anxiety, loneliness, depression and self-injury. This supports rationales for peer support as a mechanism to support coping when faced with external stressors.94,239 The timing of the intervention may be a critical factor. Jacobsen and colleagues,140 reporting on the Insiders scheme, suggested that the provision of peer support in the early days of custody was particularly valuable. There was strong qualitative evidence of the positive effects for peer deliverers, including enhanced self-awareness and life perspective, increased knowledge and skills, increased sense of purpose and relief of boredom. Negative effects were related to the burden of care, which is discussed below in relation to listeners and in the section on review question 2.
On balance, and taking into account some of the triangulation of the results, there is moderate evidence that peer support services can provide an acceptable source of help within the prison environment and can have a positive effect on recipients and peer deliverers, but there is scope for more research to obtain definitive evidence of effectiveness in terms of mental health outcomes.
The Listener scheme
The Listener scheme is a specific type of peer support intervention focusing on the prevention of suicide and self-harm. Listeners are volunteers who provide confidential emotional support to fellow prisoners who are experiencing distress. The review team decided that the distinctive focus of the intervention, the specific role of the listener and the relatively standardised model whereby training and supervision are managed by a single organisation (the Samaritans)22,31 all meant that studies on listeners and similar interventions could most usefully be reviewed separately from studies on other peer support interventions.
The review found consistent evidence from three qualitative studies and one quantitative study which strongly suggests that contact with a listener (or similar role) at a time of need was helpful in reducing anxiety, depressive thoughts and intention to self-harm, improving emotional health and helping with adjustment to the institution. There was evidence that the Listener scheme was acceptable and accessible to prisoners, from the perspective of both users and non-users. For the impact on incidence of suicide and self-harm, there was only weak, mainly anecdotal evidence.
There was consistent qualitative evidence from six studies on the benefits for the peer deliverer of becoming a listener; this was seen across a number of areas of well-being including relationships with staff, other prisoners and their families; self-esteem, self-worth and confidence; changing attitudes; social skills; and knowledge and awareness of mental health issues. There was some evidence of negative effects because of the emotional burden of care.
Overall, there was a large and consistent body of qualitative evidence which suggested that the Listener scheme is an effective means of providing targeted emotional support for individual prisoners who identify need. There is weak evidence on the impact of the scheme on the incidence of suicide and self-harm. There are positive effects in terms of mental health and well-being for those who take on the listener role, although there can an associated emotional burden. The results from the review of listeners have high relevance to health services as Listener schemes are in operation across most prisons in England and Wales.
Peer-based interventions for behaviour change
Two intervention modes, health trainers and peer mentors, focused on changing behaviours. Peer mentoring interventions in prison settings are based on the development of an affirmative relationship between a mentor and a mentee, with the mentor offering support, education and encouragement based on his/her own life experience of being in custody. These types of interventions predominantly take place in the pre-release period and mentoring can continue outside the prison gate. The review found weak evidence that mentoring can result in positive effects in terms of health behaviours, treatment adherence, abstinence from drug taking and propensity to reoffend.
Health trainers are lay public health workers who use a client-centred approach to support individuals around health behaviour change. The model has been adapted for the prison setting and some other criminal justice settings in England and Wales. There was moderate qualitative evidence that the process of training and then becoming a health trainer had a positive effect on peer deliverers, with reported effects including increased knowledge about healthy lifestyles, attitudinal and behaviour change, increased self-esteem and development of transferable skills. There was a lack of evidence of effects on health trainer recipients; however, there was some limited evidence showing that health trainers discussed a range of lifestyle issues with clients and referred individuals to other services. The results have high relevance for health services as the health trainer is an established community model with standardised competencies with the potential for transferability into the prison setting. Given the high prevalence of long-term conditions and risky health behaviours in the prison population,3,57,58 the finding that health trainers are connecting with prisoners on a range of lifestyle issues is positive; however, there is insufficient evidence on impact and more research is needed to determine the outcomes for prisoners and health services.
Positive outcomes for peers
There was consistent evidence from a large number of studies, predominantly those reporting qualitative findings, that being a peer worker was associated with positive effects on mental health and the determinants of mental health and well-being (see Barry240). Reported effects included increased self-worth and self-esteem, enhanced self-awareness and understanding, increased knowledge, having a purposeful role, relief from boredom, development of social support networks, an enhanced sense of compassion and empathy, life enrichment and improved social and communication skills. The findings that taking up a peer role could lead to positive outcomes was consistent across a number of different models, including peer education, peer support, the Listener scheme, prison hospice volunteers, health trainers and peer advisers (housing). Skills development, including having transferable employment skills, was also identified in relation to peer advisors and health trainers. Although there were some negative effects in relation to experiencing a burden of care, particularly for those roles involving emotional support, the emphasis in the evidence was on positive effects. These findings support the findings of other research on the positive impact of the act of volunteering on mental health and well-being and individual capacity.241,242 Much of the evidence comes from interventions that are well established and feature across prisons in England and Wales; therefore, the results have high relevance for health services. The review findings were also reflected in the experiences of prisoners attending the listening exercises.
Review question 2: what are the positive and negative impacts on health services in prison settings of delivering peer-based interventions?
A number of factors influencing the delivery and maintenance of peer-based interventions were identified in the review. Most of this evidence came from qualitative or mixed-method studies. The issues can be grouped into process issues, which are internal to the delivery of the intervention, and contextual factors, which are external to the intervention.
Process issues
The review found that factors relating to the maintenance of security and the management of risk are often represented in selection criteria for the recruitment of peers. Other selection criteria included interpersonal skills, levels of knowledge and the time that a prisoner is likely to be staying in that institution. There was very little evidence about selection procedures and how the criteria were applied, the exception being for the Listener scheme. None of the included studies examined the relationship between recruitment and selection processes and any outcomes, even qualitatively. This finding is somewhat surprising given the theoretical basis of peer interventions95,96 and the questions raised in the expert symposium on the contextual nature of peer identity. Further research is needed to explore assumptions about the attributes of peers in relation to the effectiveness of peer-based interventions.
The results show that training processes vary between interventions in terms of content, duration and intensity. There were some examples of standardised models, for example the Canadian PST training. However, there was no evidence on the relative effectiveness of different training packages. There was only weak qualitative evidence suggesting that mental health topics should be covered in training and, in relation to health trainers, that training should be flexible. The qualitative evidence on the benefits to peer deliverers would suggest that there is a link between participation in training and individual benefits, such as the development of skills and confidence, but it is difficult to separate out training from other aspects of the peer experience. The added value of gaining accreditation was identified and this confirms one of the themes in the expert symposium.
There was strong and consistent qualitative evidence that retention of peer deliverers was an important process issue and that attrition because of prisoner movement between prisons was a negative factor. This finding was also reflected in the expert symposium. No studies examined the issue of incentives in depth and there is scope for more process evaluations on the factors that support retention.
The importance of role boundaries was a recurring theme and overall the review provides some insight into factors affecting relationships between peer deliverers, fellow prisoners and staff. Peers overlap two distinct cultures (prisoner culture and staff culture) and this has implications for issues such as confidentiality. There is moderate evidence that peer deliverers can recognise role boundaries and when to refer to staff or other professionals, but problems such as dependency may arise. In many studies this dynamic is rarely considered. For the peer deliverers ongoing supervision, in both one-to-one and group meetings, was found to be helpful. This was an issue that was specifically identified for the Listener scheme, in which supervision and support are provided by the Samaritans. This issue was also reflected in both the expert symposium and the listening exercises. Overall, there was a range of process issues that have high relevance for current practice in the prison system in England and Wales.
Contextual factors in the prison system
A range of factors that may influence whether or not prisoners choose to utilise peer-based interventions was identified in the review. These included a lack of awareness amongst prisoners and staff; personal need; concerns about confidentiality and breaches of trust; preference for support from other sources such as staff; language barriers; and fear of demonstrating weakness by using a peer service. Many of the included studies focused on the views of peer deliverers and staff and there was more limited evidence on the views of service recipients. More research is needed to examine issues of acceptability from the perspective of recipients and those who choose not to receive peer support.
There was strong and consistent evidence, mostly drawn from qualitative studies, on the importance of organisational support within the prison, including building acceptance and support amongst staff. Resistance from staff was identified as a negative factor inhibiting the implementation of peer-based interventions. This theme is likely to have relevance for the management of peer-based interventions. The problem of staff resistance and the need for support from prison governors and service managers was also highlighted in the expert symposium.
The review found that there are modifying factors within the prison system, such as organisational support, that influence the delivery of peer-based interventions and potentially impact on outcomes. At the same time, the results indicate that peer interventions can impact on the prison environment and service provision. There was equivocal evidence that peer interventions had a positive impact on prison culture and ethos, with the most positive effects being reported in relation to peer support, prison hospice volunteers and the Listener scheme. A number of studies reported that having a cadre of peer workers can increase service capacity and reduce demand from paid staff, but there was only limited evidence on the reported impact on the prison workforce or health services. The review identified that peer interventions may increase security risks through peers distributing drugs, tobacco and mobile phones as peers often have enhanced freedoms to move and associate with other prisoners. Potential abuses of trust by peer deliverers was a process issue that programme managers needed to be aware of. The expert symposium also highlighted that security concerns and risks require active management.
Overall, the review findings indicate that peer interventions cannot be considered ‘stand-alone’ interventions that are independent of the organisation and culture of the prison. Instead, there are multiple interactions between the intervention and different levels of the prison system, in line with understandings of complex interventions.46 This also relates to sociological understandings of prisons as ‘total institutions’.243
Review question 3: what is the effectiveness of peer delivery compared with professional delivery?
Overall, only a limited number of studies compared peer delivery with professional delivery and it was not possible to triangulate the quantitative and qualitative results as the qualitative evidence related only to prisoner preferences. However, there was consistent evidence across 10 qualitative studies that peer delivery was preferred to professional delivery, with cross-cutting themes including peer deliverers demonstrating empathy because of lived experiences, being non-judgemental, being trusted by prisoners and being able to offer more time than staff. Accessibility was also a theme, with prisoners feeling more at ease talking to peer deliverers. Results support the rationales advanced for lay involvement and peer support, which emphasise lay designation and the role of peers in connecting with the community of interest.94,244,245 The review findings were confirmed by prisoners attending the listening exercises.
Reported preferences for peers in some studies could not be linked to the four quantitative studies in which a direct comparison was made, as the intervention modes were different. There was consistent evidence from four quantitative studies that peer educators were as effective as (but not more effective than) professional educators in the prevention of HIV transmission for all of the outcomes measured. Although the peer observer intervention showed some positive effects for peers compared with professionals, this was only one study about a single intervention and there is therefore insufficient evidence to draw any conclusions.
Review question 4: what is the cost-effectiveness of peer-based interventions in prison settings?
There were two components to the economic analysis in this study, namely the systematic review of economic evaluations of peer-based interventions in prison settings and the development of an economic model. Overall, there was a dearth of robust evidence on the cost-effectiveness of peer-based interventions to improve and maintain the health of prisoners, with little economic evaluation even of schemes with evidence of effectiveness. The systematic review of economic evaluations identified only one study that assessed cost-effectiveness and here the focus of analysis was costs rather than health outcomes.171 Evidence from this study suggests that TC activities involving peers may help to reduce or control prison management costs; however, it is difficult to draw further conclusions based on a single study. The review points to the need for more and better-quality research to estimate the economic value of peer-based interventions in prison settings. Despite the limitations of current studies, economic analysis has high relevance to the prison system in England and Wales, and to the wider criminal justice system. The expert symposium highlighted a number of resource issues and the potential for cost savings from peer delivery.
The prison setting provides an ideal opportunity for education and prevention because of the high concentration of high-risk individuals.123 Using the review findings and an additional literature review on economic modelling of the prevention of HIV infection, it was possible to develop a limited economic model that estimated the total number of cases of HIV infection averted by a peer-led intervention compared with a professionally led intervention and a ‘do nothing’ scenario. The results, although based on data of variable quality and a number of assumptions, indicate that both peer-led and professionally led interventions prevent HIV infections and are cost saving for all parameter values implemented in the sensitivity analyses. The peer-led intervention is dominant compared with the professionally led intervention for all parameter values implemented in the sensitivity analyses. The results are most sensitive to changes in the lifetime cost of HIV treatment and to changes in QALY estimates. The model has limitations because of the assumptions made about health behaviours and costs; nonetheless, it represents an important contribution to the evidence base on the cost-effectiveness of peer education.
Revised logic model
A preliminary logic model was developed to guide the study implementation, including development of the search strategy and the inclusion/exclusion criteria.45 This linked the wider determinants of prison health, types of peer-based interventions, the mechanisms of change and likely outcomes, both intermediate and long term.102 The study did not use a single theoretical framework as peer interventions have been linked to a range of different theories. Instead, the logic model was an attempt to draw together theoretical perspectives based on a social model of prison health and understandings of peer support.
The review results have been used to revise the logic model to provide a better fit with the quantitative and qualitative findings and expert evidence on contextual matters. Specific areas to be incorporated were:
delivery to include voluntary and community sector organisations
peer intervention modes to be changed to reflect the new typology of intervention modes
outcomes identified through the review for peer deliverers cross-referenced to intervention modes
outcomes identified through the review for the target population cross-referenced to intervention modes
contextual relationships and the place of the intervention in the prison system – this should reflect the four overarching thematic categories derived from the qualitative synthesis: peer recruitment, training and support; prisoner relationships; organisational support; and prison life.
Two logic models were then developed based on the study findings. The first is the logic model for the effects of peer-based interventions on the prison population (). This uses the same programme logic as in the original model but with the additions as described above. It explains how peer-based interventions work for service recipients or prisoners in general and what outcomes result. Outcomes have been grouped into (1) harm reduction outcomes focused on health behaviour change; (2) mental health and well-being outcomes relating both to the alleviation of individual mental health needs and to the development of positive mental health, for example better coping; and (3) improvements in social determinants, for example education, skills, housing and access to services, that enable individuals to exercise healthy choices. Outcomes at an organisational level are also represented and these are grouped into (4) the uptake of services and (5) improvements in the culture and ethos of the prison.
The logic model shows how the interventions link to intermediate outcomes and the possible links with long-term health goals. represents the different mechanisms of change but these are not explicitly linked to types of intervention mode as the review showed that peer interventions may be based on more than one mechanism of change, for example insiders provide both social support and improved access to other services. The logic model, together with the typology, will help in future research as it provides the basis for the generation of specific hypotheses to test the effectiveness of peer-based interventions using the most appropriate measures and can be matched, when appropriate, by the behaviour change technique taxonomy of Michie and colleagues.246
The second logic model () was developed to reflect the strong evidence around the positive effect on peer deliverers and the need to account for the wider impacts on the prison system. The original logic model was based on the assumption that effects on peer deliverers were part of a linear intervention chain. The review results show that becoming a peer health worker is associated with a range of benefits to the individual, and broadly similar positive effects are reported across different intervention types. There was very little evidence making a direct link between the effects on peer deliverers and the effects on service recipients. The implications are that the development of peer deliverers needs to be considered as a distinct component of an intervention, requiring an additional, non-linear logic model. This logic model also needed to take account of the thematic categories derived from the qualitative synthesis and the finding that peer interventions could not be considered ‘stand-alone’ interventions as there were multiple interactions between peer interventions and aspects of the prison setting. The evidence indicated that features of the context could not be represented solely as modifying factors as peer-based interventions also impacted on the wider prison as an institution. The final model has been termed a health capacity logic model as it attempts to show the inter-relationships between developing individual capacity to act as a change agent and organisational capacity to create a supportive environment for that change. It also acknowledges that peer-based interventions are ultimately coconstructed with staff. The review findings are clear that both strands need to be considered in developing and implementing peer interventions in prison. The model can be used as a platform for developing research looking at the wider impact of prisoner involvement in prison settings, as well as for process evaluations.
