Developing a theoretical model of the intervention (phase 1, workstream 1)

Developing solutions to complex problems within complex systems is challenging, as it requires an in-depth understanding of the nature of the intervention and the implementation context.31 Relatively little attention is paid to the developmental phase of the complex intervention cycle and it has been argued that realist-informed methods may become routine once sufficient empirical examples are available to guide practice.32,33

The aim of workstream 1 was to develop an empirical example of a theoretical model of the Engager intervention. We used realist methods34 as an overall framework with which to establish the mechanisms involved and the causal pathways and contextual conditions that might lead to positive outcomes. Behaviour change theory35 influenced our understanding of the mechanisms needed to bring about change. Throughout the development, the intervention was seen as a two-step behaviour change model, that is, how the system and research team supported practitioners and then how practitioners supported offenders. The former became the intervention delivery platform (IDP) and the latter was developed into a detailed set of practitioner behaviours that were designed to support a heterogeneous group of offenders towards better well-being and rehabilitation, and is expressed as the programme theory. Through testing this programme theory, we offer empirical examples to guide future practice for other researchers considering realist-informed approaches to developing and evaluating interventions with complex groups.

There were three stages to the theory development. An initial starting point model, articulated in the original funding bid, was based on the team’s clinical and research knowledge and previous patient and public involvement work. The first stage involved a realist review, focus groups and organisational case studies, which were synthesised (i.e. synthesis 1) into a prototype intervention that was then delivered in the pilot trial. The second stage involved a realist formative process evaluation within the pilot trial, a further organisational case study and focus group, and a realist review of implementation literature. The results of these were brought together (i.e. synthesis 2) to inform the intervention theory and IDP (e.g. training, manual, supervision), which were then operationalised in the main trial. Throughout this workstream, the peer researchers commented on the academic researchers’ findings and the intervention development, helping the academic researchers to identify gaps in understanding. The sections below detail the components of workstream 1 and illustrate how each component contributed to the development of the programme theory of the Engager intervention.

Realist review

The realist review has been published36 and, in accordance with RAMESES II (Realist And Meta-narrative Evidence Syntheses: Evolving Standards II) reporting standards,31 the realist review aimed to specify how an integrated person-centred system to improve the mental health of offenders with common mental health problems might work. Therefore, the focus was on in-depth theory-building as a key contribution to the development of the intervention. The review is methodologically novel in the way that it operationalises a realist approach [e.g. use of ‘if–then’ statements, structured questions for synthesis of context–mechanism–outcome (CMO) configurations] in parallel with the timelines and methods of a large research project. The following three-stage process was used:

1. We identified sources providing rich descriptions of interventions (and/or their delivery) providing care for offenders with common mental health problems through hand-searching a core set of 16 journals, citation-chasing, elicitation from experts (including men with lived experience of prison and release) and database searches. Searches were extended beyond offender health where insufficient sources were located. Screening of sources, therefore, included those relating to mental health care, and health and social inclusion service delivery in vulnerable groups.
2. We identified explanatory accounts (i.e. ideas about how a problem can best be addressed) (n = 347). ‘Working backwards’ from outcomes (including those from meetings with the Peer Researcher Group of men with lived experience), explanatory accounts were extracted and tabulated, and then expressed as ‘if–then’ statements, specifying (as far as the source allowed) CMO configurations.
3. We consolidated explanatory accounts (n = 75) and produced a conceptual platform (i.e. a core set of processes about how intended effects could be achieved). To enable feedback about expression and scope, consolidated explanatory accounts were shared throughout the process with our expert group of practitioners and academics.

An overview of the conceptual platform developed for the Engager intervention depicts the key agents and their relationships (Figure 2). The three columns in Figure 2 describe different key interactions relating to participating in a mental health improvement intervention both before and after release from prison. The contextualised interactions can be between practitioners and offenders, between the practitioner/offender and other practitioners, or between the offender and family members, peers or mentors. At the centre of the intervention are the core interactions between practitioners and offenders. It is within these interactions that the intervention had the potential to activate mechanisms that affect offenders’ thinking, emotions and behaviours. In between these core interactions, the interactions of both the practitioner and offender with other people (i.e. the central three circles in Figure 2) can further affect changes in thinking, emotions and behaviours, and how these interact with their contexts. These changes influence the subsequent interaction between the practitioner and offender and generate other potentially beneficial effects.

FIGURE 2 Conceptual platform developed for Engager. Reproduced with permission from Pearson et al.36 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: https://creativecommons.org/licenses/by/4.0/. The figure includes minor additions and formatting changes to the figure. (Figure not reproduced here; see the original article.)

This realist review contributed substantively to our understanding of how an integrated person-centred system might work. Findings suggested that supporting positive mechanisms for change involved optimising the local support available and minimising the potential for complex relationships with services to destabilise an individual. Box 1 outlines the core set of processes proposed in the conceptual platform.

1. Different systems, in particular health and criminal justice, and having goals that are consistent with one another.
2. Attaining consistency between strategic goals and the goals of practitioners.
3. Making referral pathways and links between organisations comprehensible to practitioners and providing opportunity for the development of constructive working relationships.
4. Practitioners being facilitated and enabled to balance factors that can be in tension, for example ‘knowing how’, as well as ‘knowing that’, analysing one’s own behaviour while remaining attentive to emotions, and working towards an individual’s goals.
5. Practitioners being facilitated and enabled to apply scientific and experiential knowledge judiciously in working with individual offenders and colleagues, and the systems in which services are delivered.
6. Practitioners having sufficient knowledge about mental health and how to develop supportive relationships with people with mental health issues.
7. Recognising the individuality of offenders throughout all interactions in the criminal justice, health and social care systems.
8. Aligning resources to enable offenders to achieve their collaboratively agreed goals.
9. Practitioners supporting reconnection with, and/or development of, networks of support outside prison.
10. Offenders having reasons to trust practitioners, services and systems.

For pragmatic resource reasons, we did not take on the task of reviewing psychological therapies, having made an early decision to use mentalisation-based approaches (MBAs)37 to support existing competencies because of high co-occurrence of traits, such as emotional reactivity linked to the diagnosis of a personality disorder. The realist review also provided a detailed methodological example of theory-driven intervention development. In particular, the conceptual platform was used to help us understand where we had gaps in our understanding and how the organisational case studies, focus groups and peer researchers could help fill these gaps. For example, the conceptual platform highlighted a gap in our understanding of how families and peer mentors may inform the intervention and, therefore, one of the focus groups was developed to look at the role of families and another organisational case study site involved analysis of the use of peer mentors.

Organisational case studies

The organisational case studies article is published.38

The organisational case studies aimed to augment our emerging core intervention theory by gaining learning from a range of promising services that provided support and/or treatment for people experiencing common mental health problems within the CJS. The organisational case studies also aimed to identify key elements of practice, for example what was and was not effective in engaging people and maintaining contact with the service. Organisational case studies were particularly useful in eliciting grounded knowledge that was unpublished and, therefore, unavailable for the realist review.

Four cases were selected. Organisational case studies 1 and 2 were undertaken prior to the formative evaluation and organisational case studies 3 and 4 were undertaken after the formative evaluation. Data included documents (e.g. service standard operating procedures), field notes and semistructured interviews. Realist-informed interview schedules (see Report Supplementary Material 1) were constructed using the CMO configurations in our developing theory. All documents were analysed using deductive content analysis. Interview data were interrogated through a qualitative framework approach.39

Seven main themes were identified: (1) collaboration, (2) client engagement, (3) client motivation, (4) supervision, (5) therapeutic approach, (6) peers and (7) preparations for ending. The case studies highlighted that engaging and motivating clients was dependent on the relationship built with the professional. Consistent, respectful, open and honest communication over time was needed to activate mechanisms that produced a sense of trust and rapport. Professionals were often unable to provide this level of support if they did not work in multiagency collaborations, supported by supervision.

Engager organisational case studies highlighted the clear need for an intervention to have in its toolbox a range of psychological therapies. In addition, the case studies suggested that the addition of a MBA would be beneficial to support the relationship between the practitioner and client, as well as to help deal with mood lability. Supporting individual offenders to mentalise effectively about their situations also aligned with the realist notion of activating generative reasoning mechanisms that enable change in thinking, emotions and behaviours.

The organisational case studies were also critical to our understanding of the importance of a ‘strong’ IDP (i.e. manual, training and supervision) that would support the practitioners in operationalising the Engager theory. The findings focused specifically on the role of supervision in ensuring ongoing support for practitioners as part of a team and on the additional need for external supervisors.

Furthermore, the organisational case studies highlighted the importance of preparatory work in prison to build trust, rapport and engagement. The case studies also prioritised the need to prepare clients for the end of the intervention as a positive step in a longer journey, including providing clients with a summary of their progress.

Finally, the initial prototype and the realist review highlighted the use of peers to support engagement and improve outcomes. The organisational case studies evidenced benefits for peers delivering support, but there was less evidence of benefit for the recipient. The Peer Researcher Group strongly asserted the importance of peer support. However, the research team reluctantly decided, following protracted discussions with the Peer Researcher Group, that this potential component was not logistically deliverable within the resources of this project. Instead, Engager practitioners would encourage and facilitate access to peer/mentoring services.

Focus groups

This work has been published.40

The published work focuses on the methodological development of the focus groups. The aim of the focus groups was to explore the views of subgroups of the target population whose experiences were not represented by previous published data, the realist review and organisational case studies.

The following four focus groups were convened: (1) homelessness on release from prison, (2) young men who had served prison sentences, (3) people with a ‘loved one’ in prison and (4) men recently released from prison. Each group was recruited through a community-based service that already knew the participants and continued to support them afterwards.

Data were coded and analysed deductively using our theorical framework. This analysis consisted of four columns representing time points (i.e. initial meeting with the participant, actions in prison, actions in the community and preparations for endings) and six rows representing the main functions of the intervention (i.e. shared understanding of individual goals, shared action plan to achieve individual goals, therapeutic approach, family and community capabilities, peer or other volunteer mentor support, and liaison by practitioner and participant with other professionals/teams).

Conversations within the focus group predominantly related to the following areas within our framework.

Developing the shared understanding by making links between a person’s emotions, thinking, behaviour and social outcome

Members of the focus groups talked about a reluctance to discuss emotions for fear of appearing weak. Often, emotions were being masked by other behaviour (e.g. using substances or ‘putting on a front’) and this highlighted the need to build trust and engagement before moving too quickly into talking deeply about emotions; however, periods of abstinence may be windows of opportunity for this work.

Developing the shared action plan: matching goals to resources

Men in the focus groups talked about a range of resources within prison that they could use to pass the time and/or could help if they were feeling low, including education, workshops, gym and chaplain. These resources would be key resources that the Engager practitioners could link with in the prison.

Developing the shared action plan: other professionals

The men talked about how, in their experience, a diagnosis of mental illness was often not helpful, as it could limit access to other services or professionals and it may cause professionals to treat them more negatively. This supported our initial thoughts that, although the focus of the Engager intervention is for those with common mental health problems, it would be a person- rather than diagnostic-centred intervention.

Developing the shared action plan: other people as a resource

The roles of other people, especially family members, were talked about at length in the focus groups. The men highlighted that often families receive little information. Family members are often reluctant to ask for information, help or support because of the perceived stigma of having a family member in prison. The men also talked about preparing both the soon-to-be released prisoner and family about expectations. Prisoners cannot always expect to return to situations as they left them. Proactive expectation management can facilitate communication, including creating awareness of how a family member or friend has had to do things differently to manage on their own and has had to make other changes over time. A partner may have grown in confidence and independence; this can be difficult for released prisoners to adjust to and also the family member may feel guilty about this. These discussions highlighted a gap within the developing intervention, that is, the Engager practitioner may also have to work closely with family members to provide information, to support contact with services and, importantly, to prepare both sides for what release would be like.

What to do in prison to prepare for community contact

The focus groups gave us many practical actions that the Engager practitioners could do to prepare released prisoners for contact in the community. The men talked about accommodation and money as key concerns during the initial release period and, therefore, the Engager practitioners could work alongside other services to try to secure accommodation and money on release. We recognised that this would not always be possible and that some people would be homeless upon release. The men suggested making sure that prisoners knew all of the locations a person may present at if they were homeless (e.g. drop-in centres, homeless shelters or soup kitchens).

Community (first meetings)

The men suggested that it can be hard to establish a routine in the first few weeks and this was a key factor in the likelihood of retuning to prison, alongside insecure accommodation and financial issues. This indicated to us that another role the Engager practitioner could have is working with the person to develop a new routine.

The data from the realist review, focus groups, organisational case studies and peer researcher meetings were synthesised into a prototype intervention (i.e. synthesis 1), which was then delivered and evaluated in the pilot trial and formative process evaluation. Figure 3 depicts the overarching structure of the Engager logic model theory. More detail of that synthesis process is detailed in Report Supplementary Material 2. During the synthesis, narratives were produced of the core components (see Report Supplementary Material 3 for details of the theory for shared understanding, a shared action plan, action in prison and action in the community).

FIGURE 3 Engager logic model. Reproduced with permission from Brand et al.41 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for non-commercial use, provided the original work is properly cited. See: https://creativecommons.org/licenses/by-nc/4.0/. The figure includes minor additions and formatting changes to the original figure. (Figure not reproduced here; see the original article.)

The peer researchers contributed to the discussions and challenged the academic researchers throughout the group meetings. The peer researchers required the academic researchers to prioritise the importance of ‘self-care’ and ‘family support’, even when the published data and financial capabilities suggested that these might not be achievable. Consequently, focus groups addressing these issues were held and a decision was made that Engager practitioners should work with participants’ families when appropriate.

Formative process evaluation

The formative process evaluation is published.41

The formative process evaluation continued the realist approach. Realist approaches addressed the need to build theory prior to outcome evaluation33 and the need for greater attention to be paid to context during delivery and to effective scale-up during later implementation.42

Face-to-face semistructured realist interviews were digitally audio-recorded and transcribed verbatim. Interviews were conducted with Engager practitioners, Engager supervisors, prison leavers and CJS, health and/or social care professionals. In addition, Engager practitioners used a pro forma to record notes about what practitioners had carried out in each session with, or on behalf of, participants, and any significant issues that had been discussed. Scanned copies of practitioner notes were examined using a structured pro forma to describe which components of the intervention were delivered, where they were delivered and over what time.

The findings presented in the published paper are both methodological and substantive. The paper provides a consolidated version of the core programme theory, with illustrative examples. A full overview of all the core programme theories examined and not included in the published paper are provided in Report Supplementary Material 4.

Key learning from the formative process evaluation led to direct modifications to the intervention and adjustments to IDP resources to support practitioner teams. The direct modifications included the following:

• Emphasising the importance of engagement and building a trusting relationship.
• Activating trust mechanisms (e.g. feeling cared for) as an important mechanism for undertaking meaningful goal-setting.
• A reinforcement of the importance of ‘quick wins’ for activating trust mechanisms (i.e. the practitioner delivering something for them that may seem small but is of significance to the participant and gives a concrete demonstration of help).
• The need for a more protocolised sharing of the written plan with the participant and other services at release to help galvanise resources to support personalised goals.
• Highlighting the difficulties of balancing the competing demands of immediate (often social, but sometimes emotional) crises and working towards longer-term goals.
• Practitioners needing to anticipate barriers and also respond flexibly when crises emerged.

Adjustments to IDP included the following:

• The addition of a ‘content of the manual’ session to the training to ensure consistent practice delivery.
• The need for clearer and more consistent risk assessment processes and support for practitioners in making risk assessments from team supervisors.
• The need to specify the expected content of the supervision process more clearly. The theory of supervision needed to focus on mechanisms that would develop practitioner competencies to deliver the intervention as intended.
• The need for a more robust team structure and management at each site, including re-emphasising the importance of joint decision-making and meeting together regularly to reflect on cases.
• The need to add a section to the manual about practitioners taking care of their own emotional well-being.

Intervention delivery platform: rapid realist review

In recognition of the importance of effective implementation and delivery of complex interventions in challenging contexts, an IDP was developed as a means of supporting practitioners to deliver the intervention as intended. The IDP incorporated training, a manual, supervision and organisational support. Having previously conducted the realist formative process evaluation41 with evidence from the organisational case studies, the research team acknowledged that they lacked adequate, relevant knowledge about intervention implementation. Rather than conducting a second realist review, we decided to use our existing realist review resource to deepen our understanding of relevant implementation issues and supportive strategies.

Studies were identified from the search conducted for the realist review. The flow of sources through the IDP review and information about the 29 included sources are shown in Report Supplementary Material 5. We used the Consolidated Framework for Implementation Research36 as a framework for structuring our identification and synthesis of evidence, and these findings are also shown in Report Supplementary Material 5.

This rapid realist review synthesised knowledge from a wider literature about complex implementation issues. This enabled us to go beyond simple descriptions of ‘practitioners’ experiences’ to provide a sounder, contextually refined foundation on which to base implementation activities. Although the review was not decisive on any one aspect of the implementation strategy, across the board it played a significant role in supporting emerging findings from other parts of the research programme. For example, working iteratively with the formative process evaluation, the review identified the importance of the manual catering for different learning styles and describing the intervention in ways that aligned with existing service design. The review also informed our understanding of how supervision should be delivered in a way that is theoretically consistent with the intervention. Some parts of the review simply played a supportive role, for example by alerting us to areas where there was scant existing research evidence or insufficient detail in the published evidence, and this was particularly the case for organisational agreements. The requirement for these organisational agreements became evident in discussions with practitioners and managers during set-up and delivery, and the agreements were necessary in order to put in place the formal clinical governance requirements that would enable teams to operate under research conditions but within NHS and voluntary sector services.

After the formative evaluation, the IDP was strengthened through acknowledging the need for top-up training and additional supervision for mentalisation. A decision was made to combine the role of Engager team leader and supervisor to create coherence and promote decisions in the local setting, taking into consideration the whole caseload and workload, as well as fidelity to the model for each individual (see Report Supplementary Material 3 for our developing theory of supervision).

The Engager model going into the trial

At this point, the results of the realist formative process evaluation and of the realist review of implementation literature, plus synthesis 1, were brought together in a second round of synthesis (i.e. synthesis 2). This synthesis was to inform the intervention theory (i.e. what practitioners do and why) and the IDP (i.e. training, manual and supervision), which would be then operationalised in the main trial.

Details of synthesis 2 are detailed in Report Supplementary Material 2.

There are several ways of gaining a description of the Engager intervention in addition to what is already presented:

• Box 2 describes the key elements and rationale behind the intervention theory going into the main trial.
• Report Supplementary Material 6 describes the details of the practitioners, training and supervision arrangements in place going into the trial.
• The Engager programme provides a link for requesting a copy of the trial version of the Engager manual (URL: www.plymouth.ac.uk/research/primarycare/criminal-justice/engager; accessed 9 June 2022).

The intervention was originally conceived to be an adaptation of collaborative care intervention proven to be effective for individuals with depression. The principles of collaborative care – a psychological intervention, joint working between professionals and case-based supervision – were extended on the basis of the findings of the realist review and case studies to incorporate flexibility of delivery and practitioner behaviours to overcome distrust; co-ordination across all services involved in an individual’s care and continuity with a single practitioner, if possible; engagement in prison; meeting on the day of release; and continued care for 3–5 months following release.

Specific evidence-informed mental health and psychological therapy components were included. Early on, a MBA was selected as the supportive psychological method most likely able to assist practitioners in dealing with emotional lability (an aspect of personality disorder) and to help them to understand different perspectives and support psychological thinking. A generic approach to psychological formulation was taken, and simplified, as a shared understanding and action plan to be comprehensible for support workers with limited training and offenders with limited capacities. The shared understanding aims to understand the links for the individual between emotions, thinking, behaviour and current situation. The aim was to identify the means to enhance positive behaviours and to reduce negative ones, such as substance use, self-harm and aggressive behaviour.

The concept of the plan was also extended beyond what services could deliver to bring together all the relevant available resources to address those goals individuals had selected as the most important. Resources included not just the therapeutic support of practitioners, but the potential support of other teams, friends and family and the individual’s own personal resources for self-care. Finally, the concept of ‘discharge’ was transformed into the possibility of next steps, either alone or supported by other services.