Evaluation of the Community Reintegration Project

2 The Implementation and Operation of the CRP

2.1 This chapter provides a descriptive overview of the implementation of the CRP within the four prisons and the two associated CJSW areas. It considers issues of 'fidelity' - the extent to which the basic process elements of the project were implemented as intended - but also whether the activities or outputs associated with the project were at the expected levels. It begins with a description of the CRP process and how it was implemented in each site, before looking at quantitative indicators of CRP activity across the different stages.

2.2 We will argue that the implementation within the four sites was broadly consistent with the original model and guidance, but that there were also significant variations in how the CRP was implemented across settings and that its operation evolved over the course of the project. We will also show that the number of offenders engaging with the project was relatively low - especially within the women's prisons - and that there was a high degree of attrition across the various stages of the project. That said, the evaluation also provides evidence that offender needs which might otherwise have been missed were identified and potentially addressed through referrals to and contact with community-based services.

Outline of the CRP process

2.3 In broad terms, the CRP involves the identification and assessment of offender needs in custody; referral to services and support to address those needs while in custody; referral to CJSW to allow a plan to be put in place to facilitate transition to the community; and access to continuing CJSW support following release. More specifically, however, the project is intended to be underpinned by a consistent and structured process with a number of key stages. These are outlined below, and a chart summarising the stages of the process is included at Annex B.

Sentencing of individuals

• Transfer of the CJSW report from the sentencing court to the receiving prison (CRP Stage 1).
• Immediate Needs Screening (CRP Stage 2): A CRP specific Core Screen is carried out in the first few days in custody. This incorporates the first night in custody checklist and screening process carried out for all offenders on arrival at prison in identifying needs (part 2a), but those eligible to take part in CRP are offered the opportunity to do so at this point and are taken through a fuller screening process (part 2b).
• Referral to CJSW: The relevant CJSW team is notified of those engaging with the scheme; a CJSW worker then arranges to meet the participant in custody and will arrange subsequent meetings as required.
• Comprehensive Needs Screening (CRP Stage 3): The Comprehensive Screen is carried out between the offender and his or her Personal Officer (PO) within 28 days of entering custody. The CJSW report forms the basis of the discussion and leads to fuller assessment of the offender's needs and referrals to appropriate services.
• Referral to services and support in custody: The participant is referred on to appropriate services in response to the identified needs.
• Standard Review (CRP Stage 4): Standard Reviews are held on a monthly basis in order to review needs and check on the progress of referrals.
• Summary Review (CRP Stage 5): The Summary Review is held 5 or 6 weeks prior to release and involves a meeting between the offender and their PO in order to review progress to date and update the Community Integration Plan (CIP)^[13] prior to the pre-release meeting.
• Pre-release meeting: This meeting is held around 4 weeks prior to liberation and brings together the offender, the CJSW representative, the PO and any other key personnel in order to agree a plan for release and reintegration into the community and make/confirm arrangements for referrals to community-based services.

Transition to the community

• CJSW case management: This is flexible but generally involves an initial appointment and follow-up supervision and appointments as required. Cases are reviewed after three months, and closed once a participant is stable in the community and no longer requires CJSW support or when the participant disengages from the process. When a participant disengages from the process they remain eligible to request voluntary throughcare for 12 months following their release from prison.

2.4 The prison-based stages of the process use specifically designed forms, allowing staff to record identified needs, indicators of engagement, and a brief narrative account of the meeting. Completed forms are then uploaded to the SPS case management system, PR2.

Implementation of the process in the project sites

2.5 The research found that the CRP process implemented in each of the sites was broadly in line with the staged process as set out in the original guidance. However, while the CRP process was recognisable at each site there were notable differences in the way in which it was implemented and delivered in terms of the management and supervisory arrangements, types of staff involved, and the links with CJSW teams. A brief description of each site is presented below, followed by a discussion of some of the key aspects of variation identified.

HMP Perth/Dundee CJSW

2.6 HMP Perth is a male-only establishment with most offenders coming from the surrounding area (of Perth and Kinross, Dundee, Fife and Angus). For the purposes of the project, the prison-CJSW link was between HMP Perth and Dundee CJSW. The introduction of the CRP at the prison in spring 2012 was supported by training sessions offered to residential staff with additional follow-up sessions in mid-2013. The Link Centre manager was also available to provide advice and assistance to staff on an ad hoc and ongoing basis.

2.7 Key features of the CRP within HMP Perth included a central role for the Link Centre, supervisory input provided at residential manager level, a focus on specific residential halls and a CJSW team with the use of office space in the Link Centre and access to PR2. The CRP process operated as follows:

• Accommodation of CRP participants: Offenders from the designated Dundee postcode areas were housed within two residential halls.
• Who was involved within the prison: The process was overseen by Link Centre staff. They oversaw the identification of eligible participants, carried out Stage 2b and maintained the CRP database. All POs in the relevant residential areas (potentially) carried out the process with offenders (representing a change from the original approach of a small number of designated officers). Administrative support was put in place from late 2013.
• Who was involved within CJSW: The CRP project was staffed by two main CJSW support workers, and one additional support worker. A social worker team leader oversaw the work and provided case supervision. There was administrative support for maintaining records.
• Liaison between the prison and CJSW: The CJSW team had office space in the Link Centre and spent (at least) one day in prison each week. There was regular communication and liaison between the Link Centre and CJSW.
• Recording and sharing information: Information was recorded on the CRP forms and then uploaded to PR2 by the relevant Link Centre staff and POs. The CJSW team had access to PR2 in prison for background information - and completed the Stage 5 report on PR2 - but did not have access from their own office in the community.
• Prison activity: The Link Centre manager had responsibility for determining eligibility and Link Centre staff carried out stage 2b, made initial referrals and notified CJSW of eligible cases. Cases were then added to the CRP database and hall managers had responsibility for allocating cases to individual POs. Hall managers supervised the process, following up non-completion of stages and disengagement. POs carried out Stage 3 and 4 meetings and made referrals - some directly on PR2, and others as paper referrals - for the Link Centre to pick up. An officer attended the pre-release meeting, although this was not always the relevant PO.
• Referrals to CJSW: The Link Centre sent notification of new CRP participants to CJSW on a regular basis.
• CJSW activity in prison: CJSW picked up CRP referrals and drew up a schedule of meetings (new cases and follow-up appointments) for their forthcoming prison visits. The Link Centre then liaised with residential areas about notifying individuals and bringing them over for meetings. On the day of meetings, Link Centre staff took a proactive approach to following up non-attendance of those listed for meetings. Following an initial meeting and needs assessment, the CJSW team saw individuals as many times as required (generally on between two and six occasions) to ensure that support and services were put in place for their release.
• CJSW activity in the community: Participants were given a community appointment to attend on release though most tended to disengage at this point. Those who did not attend were offered a second appointment and a further follow-up letter was then sent before the case was closed. Post-release activity included occasional gate pick-ups and intensive support in community.
• Referrals to other agencies: POs recorded referrals on paper forms and occasionally directly onto PR2 to be picked up by staff/agencies in the Link Centre. CJSW workers were able to see referrals made on PR2 and to make referrals and negotiate with external agencies on behalf of clients.
• Pre-release meeting: This was organised by the Link Centre and typically attended by the participant, CJSW and a residential officer (although this was often not the relevant PO). Attendance by other agencies was not common. A subsequent meeting with CJSW was held prior to release to confirm arrangements.

HMP Cornton Vale/Lanarkshire CJSW (and Dundee CJSW)

2.8 HMP Cornton Vale is a women-only national facility taking all categories of short and long-term female offenders from across the country. It is the initial receiving establishment for all convicted female offenders in Scotland, although some may subsequently be transferred to other more local establishments. The prison regime involves a progression through different blocks based on an assessment of risk. The implementation of the CRP at Cornton Vale involved a main link between HMP Cornton Vale and North Lanarkshire CJSW. Training sessions were provided at the outset of the project targeting CRP 'champions' in residential areas, and the Offender Outcomes Manager had offered follow-up awareness sessions and operated an open door policy in offering advice and support to staff.

2.9 Key features of the process at Cornton Vale included the role of the Integrated Case Management (ICM) Unit and the use of administrative support to assist with adherence to the process. The process operated as follows:

• Accommodation of CRP participants: Offenders from the designated postcode areas were housed throughout the prison and progressed through the estate in line with standard practice.
• Who was involved within the prison: The CRP process was overseen by the ICM Unit, where administrative staff maintained a database and issued paperwork and reminders for staged meetings. Reception staff carried out the Immediate Needs Screen (parts 2a and 2b). Any PO across the establishment could potentially be required to carry out CRP duties if allocated an eligible offender.
• Who was involved within CJSW: Two CJSW support workers carried out CRP duties alongside a mixed criminal justice caseload. The social worker team leader provided ad hoc support and regular case supervision.
• Liaison between prison and CJSW: Contact was on a case-by-case basis, and sometimes included liaison between CJSW and the Vulnerable Prisoner Officer. CJSW meetings with clients generally took place in the agents' room at the prison gate, meaning that there was limited contact with prison staff.
• Recording and sharing information: The ICM administrator was responsible for maintaining records in prison - information was recorded on CRP forms and then typed up and uploaded to PR2 by the administrator. CJSW did not have direct access to PR2, but were provided with a copy of the SPS CIP in preparation for the pre-release meeting.
• Prison activity: The ICM Unit notified residential managers of CRP participants who allocated cases to POs. The ICM administrator emailed the residential manager when each CRP stage was due, prepared paperwork for collection and then recorded the outcome on PR2 once paperwork was returned; paperwork was signed for on uplift and signed back in on return. POs confirmed offenders' consent to participate and carried out the Comprehensive Screen and subsequent stages.
• Referrals to CJSW: Referrals were made on an individual basis using the existing voluntary throughcare request form.
• CJSW activity in prison: CJSW aimed to see offenders three or four times if possible during their time in custody, with activity involving an initial needs assessment and appropriate follow-up to ensure support and services were put in place for release.
• CJSW activity in the community: Participants were given a community appointment for the day of release (or the day after), with most taking up the offer or responding to a reminder; a second appointment was offered and a follow-up letter sent before cases were closed. The support provided included gate pick-ups and ongoing engagement in the community - such as regular meetings, visits, phone calls, attending appointments, acting as advocate in dealing with other agencies and services - though some engagement was relatively short-term.
• Referrals to other agencies: Referrals were recorded on PR2 by the ICM administrator and picked up by staff/agencies in the Link Centre.
• Pre-release meeting: This was organised by the ICM administrator in advance and typically attended by the participant, CJSW support worker and an officer (though not necessarily the participant's PO); other attendees sometimes included the Vulnerable Prisoner Officer and representatives of other support agencies.

HMP Edinburgh/Dundee CJSW

2.10 HMP Edinburgh is a predominantly male prison but also accommodates some female offenders from the east of Scotland in a single residential hall. Women offenders may be transferred to HMP Edinburgh following initial admission to Cornton Vale. The CRP link was between HMP Edinburgh and Dundee CJSW. No formal CRP training was provided to staff in the prison; the staff involved had relied on the CRP documentation for information and guidance, and the hall manager provided ongoing support to frontline staff. Key features of the CRP process at Edinburgh were the hall-based focus and the allocation of all CRP cases to a single designated PO (and deputy). The HMP Edinburgh process is outlined below.

• Accommodation of CRP participants: Offenders from the designated Dundee area were housed within the single women's hall in the prison.
• Who was involved within the prison: Admission staff carried out the first stages of the process, with offenders consenting to the CRP allocated to a designated PO by the hall manager. The PO (or deputy) then carried out subsequent stages.
• Who was involved within CJSW: Two CJSW support workers carried out CRP duties alongside a general women's criminal justice caseload. The social work team leader provided ad hoc support and regular case supervision.
• Liaison between prison and CJSW: Contact was on a case-by-case basis. CJSW visited the residential hall for meetings, which allowed for liaison; administrative arrangements for organising meetings and booking rooms were dealt with via the hall manager and PO.
• Recording and sharing information: Information was recorded on the CRP forms and then uploaded to PR2 by the PO. The prison shared the CIP and minutes of pre-release meeting with CJSW. Although CJSW did not have access to PR2, other agencies did.
• Prison activity: Admission staff carried out the initial stages with CRP participants who were then allocated to the designated PO to carry out subsequent stages. There was some flexibility in relation to timing as offenders were transferred to Edinburgh at different stages in their sentence and in the CRP process. The PO reconfirmed consent and carried out the Comprehensive Screen, encouraging participants to make self-referrals which were sent to the Link Centre and then put on PR2 for agencies to pick up. A narrative approach was adopted to the paperwork associated with Stage 4.
• Referrals to CJSW: Referrals came from the prison on a case-by-case basis, but also came from other agencies (e.g. Phoenix) or as self-referrals. CJSW then made contact with the offender outlining the service and made arrangements to visit.
• CJSW activity in prison: CJSW workers focused on engagement in the six months leading up to release. Support workers visited the residential hall for meetings and were able to approach people informally in that setting.
• CJSW activity in the community: Participants were given a community appointment for release; a second appointment was offered to those not attending, and a follow-up letter was issued before cases were closed. The team offered gate pick-ups, and community engagement sometimes involved frequent initial contact, reducing in response to progress but lasting as long as required. Intensive support involving multi-agency working was initiated when required.
• Referrals to other agencies: Participants were encouraged to make self-referrals within the prison; CJSW also made referrals and negotiated with external agencies on behalf of clients.
• Pre-release meeting: This was organised by the PO and ideally accommodated within shift patterns. It was generally held two to three weeks before release and attended by the participant, CJSW and PO, with other agencies also invited to attend.

HMP Greenock/Lanarkshire CJSW

2.11 HMP Greenock is a predominantly male prison which also accommodates female offenders from the west of Scotland in a single residential hall. Women offenders may be transferred to Greenock following initial admission to Cornton Vale. The CRP link is with North Lanarkshire CJSW. A training session for staff was held at the prison at the beginning of the project.

2.12 Key features of the CRP process at Greenock were the hall-based focus and the integration of the CRP within a standard PR2-based review system for all short-term offenders. Alongside the CRP process, Greenock was also piloting a Throughcare Support Officer (TSO) scheme offering assistance to both male and female offenders six weeks before and six weeks after release. The CRP process at Greenock included the following:

• Accommodation of CRP participants: Offenders from the designated Lanarkshire postcode areas were housed within the single women's hall in the prison.
• Who was involved within the prison: Initial screening of transfers-in was carried out in the Link Centre. The hall manager identified eligible individuals and residential staff carried out subsequent stages.
• Who was involved within CJSW: Two CJSW support workers carried out CRP duties alongside a mixed criminal justice caseload. The social worker team leader provided ad hoc support and regular case supervision.
• Liaison between prison and CJSW: Contact was on a case-by-case basis. CJSW developed links with relevant staff and made contact following referrals to arrange meetings.
• Recording and sharing information: Greenock did not use the CRP forms; instead information was recorded directly in narrative form onto PR2 by the PO or other relevant staff. CJSW did not have access to PR2 but a printout of information was provided to CJSW for meetings.
• Prison activity: Following initial screening in the Link Centre, the Comprehensive Screen was carried out by a residential officer in the following few days and prior to the allocation of a PO. The hall manager then identified CRP-eligible individuals and alerted the allocated PO to the need for a CJSW referral. The prison then followed a PR2-based system of regular reviews carried out by POs for all offenders, with the referral to CJSW being the defining feature of CRP cases.
• CJSW activity in prison: CJSW arranged to meet people following referral, with meetings held in the hall. Activity involved an initial needs assessment and appropriate follow-up to ensure support and services were put in place for release. Meetings were arranged by CJSW.
• CJSW activity in the community: Participants were given a community appointment for the day of release (or the day after) with most taking up the offer or responding to a reminder; a second appointment was offered and a follow-up letter sent before cases were closed. Support included gate pick-ups and ongoing engagement in community - regular meetings, visits, phone calls, attending appointments, acting as advocate in dealing with other agencies and services.
• Referrals to CJSW: Referrals were made on an individual basis using the existing voluntary throughcare request form.
• Referrals to other agencies: Referrals were recorded on PR2 by the PO and picked up by staff/agencies in the Link Centre. CJSW also made referrals and negotiated with external agencies on behalf of clients.
• Pre-release meeting: A pre-release meeting was held by the prison six weeks prior to release - typically attended by the participant and SPS officer (though not necessarily the participant's PO) although TSOs and other agencies (most frequently, addictions services) were sometimes involved. CJSW organised a meeting with the offender closer to the release date.

Variation between the CRP sites

2.13 As the above accounts illustrate, the same staged CRP process with related CJSW referral was broadly followed at each site, but there were also some important variations in how the project was implemented. These are summarised below.

2.14 Training, management and supervisory arrangements: The level of CRP training provided across the sites varied. While on-site training was offered at all but one of the prisons (HMP Edinburgh) at the outset of the project, follow-up familiarisation sessions provided by local staff featured at two of the sites only (HMP Perth and Cornton Vale). Otherwise, less formal on-the-job training was the main approach used to equip staff for CRP duties, although this was supplemented at Perth and Cornton Vale by support from the Link Centre/ICM Unit.

2.15 In terms of day-to-day oversight, the Greenock and Edinburgh hall-based systems relied on the residential managers in the relevant halls to supervise the process, and to liaise with staff, although no explicit supervisory pathway was identified. At Cornton Vale, the ICM Unit oversaw the process: administrative staff prepared the paperwork and prompted individual staff and hall managers to undertake the various stages of the process; and the Offender Outcomes Manager operated an open door policy in terms of offering assistance to individual members of staff. At Perth, the Link Centre took a lead role in overseeing the process and worked with hall managers to improve delivery: the Link Centre manager delivered a series of familiarisation sessions for staff; disengagement and non-completion of stages was monitored and followed up by hall managers.

2.16 Types of staff involved: At each of the four prisons, POs - as envisaged - were key to the delivery of the CRP. At Perth, Cornton Vale and Greenock all POs were potentially involved in CRP cases if allocated an offender eligible for the scheme (albeit in Perth and Greenock all such offenders were accommodated in one or two halls). In Edinburgh the scheme was delivered by a single designated PO (and a deputy), with all eligible offenders allocated to this officer. In practice, at Cornton Vale and Greenock, where there were relatively few eligible offenders, most members of staff had only limited experience of the CRP.

2.17 Link Centre staff at Perth conducted the key CRP Stage 2b meeting, offering offenders the opportunity to participate in CRP and dealing with the initial needs assessment and referrals - a change from the original arrangement. Link Centre staff there also took a role in facilitating the attendance of offenders at meetings. Link Centre staff at all sites were involved in picking up referrals. Only Cornton Vale had dedicated CRP administrative support for the project throughout the pilot, although Perth had administrative support for the latter part of the pilot period; elsewhere individual staff were responsible for completing forms and maintaining records on PR2.

2.18 Links between prisons and CJSW teams: Perth had ongoing links with the Dundee CJSW team, with the support workers having use of an office in the Link Centre and spending one or two days there each week, and also having access to PR2. On the women's side, contact was less regular and on a case-by-case basis, with personal links between individual officers and social workers often being cited as key to the process. Cornton Vale's administrator-led process, however, incorporated a systematic initial referral to CJSW and the administrator also took the lead in organising pre-release meetings. While three of the prisons experienced the CRP as a one-to-one link with a single CJSW team, Cornton Vale dealt with North Lanarkshire and (to a lesser extent) Dundee CJSW. Correspondingly, North Lanarkshire CJSW dealt with two prisons (Cornton Vale and Greenock) while each of the Dundee CJSW teams dealt with one.

2.19 Information recording and sharing: Practices varied here. At Cornton Vale, the project administrator had a role in typing up meetings and uploading CRP information onto the system; elsewhere, this was the responsibility of individual POs. Staff at Greenock reported that they had subsumed the CRP within a PR2-based prison-wide case management system for short-term offenders rather than using the CRP paperwork. Perth was the only establishment where the CJSW team had direct access to PR2, although Greenock provided a printout from PR2 of all activity as background for meetings with participants, and other sites provided copies of relevant CRP reports and CIPs.

2.20 These different arrangements appeared to both influence and be influenced by the delivery of the project. On the women's side, the small number of eligible offenders spread across three institutions (see below) meant there was less exposure for individual staff members and less frequent contact between CJSW and the prison staff, with the result that regular visits and provision of prison accommodation would therefore be harder to justify. In contrast, the number of eligible offenders at Perth provided a critical mass that merited the CJSW-embedded approach and allowed closer working relationships to develop. Similarly, the low throughput of CRP-eligible offenders at Edinburgh led to the approach involving a designated PO for all cases. The impact of these and other arrangements - e.g. the roles of Link Centre and administrative staff - are considered in later sections of the report.

An evolving process

2.21 It is important to note that the CRP process evolved during the implementation period and that there were a number of examples of modifications or adaptations at a project-wide or local level.

• Following feedback from the CRP Operational Group that the 10 day target for the Comprehensive Screen was too short, it was agreed by the Steering Group that this should be extended to 28 days (with appropriate monitoring arrangements in place) to allow offenders longer to settle down into prison routine.
• At three of the four sites, the formal pre-release meeting or an additional subsequent meeting was held two to three weeks prior to liberation. This meant that account could be taken of changes in offender circumstances which often occur in the critical period immediately before release.
• HMP Edinburgh took the decision to move to a system based on a designated PO (and deputy) - this was a response to the low number of cases at the prison and was intended to allow the designated staff to build up a level of expertise in dealing with the CRP.
• HMP Perth moved Stage 2b of Immediate Needs Screening to the Link Centre and introduced follow-up by a residential manager whenever a participant disengaged from the CRP process.

2.22 All these refinements reflect staff input in developing the process to meet local needs and the needs of the short-term prison population - themes which we return to at various points during the later chapters of the report.

Data on levels and patterns of CRP activity

2.23 Having provided a brief account of what the CRP process was intended to look like and how it was implemented in each setting, we turn now to some quantitative indicators of the activity associated with the project. These are drawn from two main sources: unit-level data relating to the CRP process within the participating prisons during the period covered by the CRP as a whole; and aggregate-level statistics compiled by each participating prison and by partners in CJSW for the last six months of the project. We also present some information provided directly by the CJSW teams. Information collected from Scottish Court Service (SCS) relating to the transfer of CJSW reports is covered in Chapter 3.

2.24 The unit-level records collated by the prisons relate to the period from May 2012 to March 2014 and are based not on individual offenders per se, but on individual admissions, as some individuals were admitted on more than one occasion during that period. Although it was possible to conduct some analysis using these unit-level data, there were several important limitations:

• Some of the participating prisons customised the original standard data-collection sheets, greatly reducing the scope for comparison across sites and the creation of a single overarching dataset.
• There was some duplication of records - e.g. as a result of offender transfers.
• There were gaps in individual records and it was often not clear whether a field in the sheet was blank because the relevant stage/meeting had not taken place, or simply because such activity had not been recorded.

2.25 In response to some of these limitations, participating prisons and other key agencies were asked by the Scottish Government to provide additional, aggregate-level data, which they did for the period between October 2013 and March 2014 (inclusive). Although these data do not cover the whole period in which the CRP was running, for some purposes they arguably represent a more reliable indicator of how the CRP was operating during its 'mature' phase once any initial problems had been identified and addressed.

2.26 However, neither of these two datasets (offering unit-level or aggregate information) should be regarded as offering a definitive source of information about levels and patterns of CRP activity. The two do not map neatly onto each other and, in the unit-level data in particular, it is often impossible to determine whether the absence of data indicates that a particular activity (e.g. a Stage 3 meeting) had not happened or simply has not been recorded. As such, the following estimates should generally be regarded as worst case estimates of the proportion of cases reaching each stage, since we have counted only those cases where there is clear evidence that a meeting has taken place or that the relevant paperwork has been uploaded.

2.27 These failings in the systems for monitoring relevant activity represent a significant problem in the implementation of the CRP and a lesson for future initiatives of this type (and the potential roll-out of the CRP itself). Nevertheless, with the above caveats about data quality and completeness in mind, it is possible to build a rough picture of how extensive the activity has been around the CRP, in each of the participating areas and across the project as a whole.

Numbers of offenders identified as eligible to participate in the CRP

2.28 The unit-level data suggest that, between May 2012 and the end of March 2014, around 436 offenders (or individual admissions) were identified as eligible to participate within HMPs Perth, Cornton Vale and Edinburgh. (No comparable data were available for HMP Greenock, but it appears from other sources that the numbers would have been very small). Of these, the majority (around three-quarters) were male admissions at HMP Perth (293). Of the remainder (all of which were female admissions), the largest single group of offenders identified as eligible for the CRP was at HMP Cornton Vale (125), followed by HMP Edinburgh (18).

2.29 In terms of patterns over time, the early phases of the project (prior to March 2013) were marked by relatively low numbers of offenders being identified as eligible to participate in the project. Beyond March 2013, at which point the project was given additional resource within the Scottish Government, numbers increased significantly, driven specifically by a sharp increase in the identification of eligible offenders within Perth. Figure 2.1 shows the best estimate of the number of eligible offenders by establishment by quarter.

Figure 2.1 Number of offenders identified as eligible to participate in the CRP by date of admission, by site

Source: Unit-level monitoring data (unique admissions), n=399

Note: Does not include data for HMP Greenock or cases where date of admission not available

2.30 This variation over time suggests that the number of offenders identified as eligible for the CRP should not be regarded as a straightforward indicator of the size of the potential target population for the intervention, and that the identification of eligible participants is partly a product of SPS staff activity. In other words, the greater number of offenders identified as eligible in the period after March 2013 is likely to reflect changes in staff awareness of and attention to the CRP rather than an underlying shift in the proportion of inmates from relevant home postcodes. This suggests that any intervention based on eligibility criteria (such as postcode) needs to be actively promoted and monitored to ensure that it is actually offered to the relevant group.

Numbers of offenders identified as engaging with the CRP within prison

2.31 It is also difficult to establish definitively the proportion of those offered the CRP who actually engaged with the project. In principle, offenders might be said to have engaged at the point at which they initially agreed to take part - i.e. when offered the opportunity to do so as part of the Immediate Needs Screen (at Stage 2). Evidence of such agreement is not, however, recorded systematically within the monitoring data. Moreover, a proportion of those who initially agreed to participate will have effectively disengaged before any meaningful CRP activity occurred.

Participation in Stage 3 meeting

2.32 A more useful indicator of engagement with the project is, therefore, participation in the Comprehensive Screen (Stage 3) - a process that lies at the heart of the CRP approach and which is not available, as a matter of course, to other short-term offenders. The unit-level dataset suggests that approximately 223 Stage 3 meetings took place during the period of the project as a whole representing 51% of those admissions identified as eligible for the CRP.

2.33 Not surprisingly, in absolute terms, many more such meetings were held with male than female offenders (and in Perth than in other sites, where there was evidence of 182 such meetings having been held, compared with 44 at Cornton Vale and 8 at Edinburgh). Rates of participation in Stage 3 meetings, however, were higher among female offenders (61% compared with 46%) - see Figure 2.2.

2.34 Although there is little further that can be said about offender characteristics, rates of engagement at Stage 3 appear to have been slightly higher among those serving longer sentences. For example, among those whose earliest release date was within 90 days of admission, 39% had evidence of a Stage 3 meeting, compared with a figure of 53% among those whose earliest release date was more than 91 days after their date of admission.^[14] The fact that those with relatively little time to serve upon admission were less likely to have a Stage 3 meeting may reflect the differing needs of this group; but it may also result from differences in how they are dealt with by SPS staff or in the opportunities for constructive engagement.

Participation in Stage 4 and 5 meetings

2.35 The unit-level data also provide an estimate of the number and proportion of CRP-eligible offenders for whom there was evidence that each subsequent stage of the process had been completed. This suggests that female offenders were somewhat more likely to have evidence of a Stage 4 meeting (39%, compared with 31% of males) and to slightly less likely to have evidence of a Stage 5 meeting (19%, compared with 25% of males).

2.36 Overall, the attrition rate is striking, even after offenders have engaged meaningfully with the project - for example, of all CRP-eligible offenders who took part in a Stage 3 meeting, only 35% took part at Stage 5. However, it should not be assumed that the process of attrition is entirely linear: for example, 10% of those for whom there was no evidence of a meeting at Stage 4 did appear to have taken part in a Stage 5 meeting. While this may be partly about incomplete records for individuals involved in the project, it may also suggest that some of those who disengage (formally or informally) during the course of their sentence may re-engage as they approach release and develop a stronger awareness of their needs or motivation to address them.

Figure 2.2 Percentage of eligible admissions with evidence of meetings at Stages 3, 4 and 5

Source: Unit-level monitoring data (based on unique admissions between May 2012 and March 2013)

Note: Does not include data for HMP Greenock

Additional evidence from the aggregate-level data for the most recent period

2.37 The aggregate data, covering the period October 2013 to March 2014 (inclusive), offer an alternative perspective on the issue of engagement. These show, within that period, the total number of offenders eligible for the CRP and the number who were due to have, and actually had, a meeting at each stage (see Table 2.1). It should be noted that there are some inconsistencies in these returns and that no direct comparison is possible with the unit-level data (as these do not represent a single cohort of offenders moving through the process but the level of each type of activity within a specific time frame, and with a changing cast of offenders). Overall, however, these figures reinforce the suggestion from the unit-level data that there is significant attrition at each stage and across the process as a whole.

Table 2.1 Number of offenders engaged at each stage of the CRP, by gender (October 2013-March 2014 inclusive)

	Males	Females	All
Qualify for the CRP	122	57	179
Initially engaged in the CRP	116	46	162
Stage 2a
Meetings required	118	49	167
Undertaken	59	31	90
Stage 2b
Meetings required	117	50	167
Undertaken	96	40	136
Stage 3
Meetings required	95	32	127
Undertaken	20	23	43
Stage 4
Meetings required	339	45	384
Undertaken	64	34	98
Stage 5
Meetings required	56	14	70
Undertaken	37	13	50

Source: Aggregate-level monitoring data (based on activity by stage within period Oct 2013-March 2014)

2.38 For example, Table 2.1 shows that in total, 127 Stage 3 meetings were scheduled for this period and, of these, only 43 (or 34%) actually took place. (It is worth noting that this represents a lower level of activity than was suggested by the unit-level data, and it is not clear why the number of Stage 3 meetings was so low during this period.) The number of Stage 4 and Stage 5 reviews required also significantly outnumbered those conducted. At Stage 4, for example, 384 reviews were required during this period but only 98 (26%) were recorded as having taken place. During this period, the proportion of Stage 4 reviews going ahead was much higher for female than male offenders (76% compared with 19%). At Stage 5, some 70 reviews were required, 50 of which (or 71%) took place. Again, the proportion going ahead was higher for female than male offenders (93% compared with 66%).

2.39 There are two main possible explanations for this gap between meetings 'required' and proceeding: first, that there are failings on the staff side to progress, facilitate or attend such meetings; second, that offenders are disengaging, formally or informally, in advance of meetings being held. Both possibilities are explored through the qualitative data in subsequent chapters.

Indicators of engagement with CJSW and other CRP-related activity 'beyond the prison gate'

2.40 Although the main focus of the evaluation was on the operation of the CRP within the four prison settings, aggregate data provided by CJSW also shed some light on CRP-related activity within partner organisations. In this section, we review evidence about the extent of offender engagement with CJSW and other agencies in the community.

Engagement between CJSW and offenders

2.41 The aggregate-level monitoring data suggest that, between October 2013 and March 2014, CJSW conducted 145 CRP interviews with CRP participants within prison settings. Not surprisingly, given the total number of male offenders involved in the project, the vast majority of these involved Dundee CJSW (139). Even allowing for overall levels of engagement, however, the number of CRP meetings recorded by Lanarkshire CJSW during this period (just 6) was especially low and does not appear to be entirely consistent with soft indicators of CJSW activity from the qualitative interviews. The reason for this apparent discrepancy is not clear.

2.42 Supplementary information provided to the evaluation team by CJSW in the two areas also allows a basic comparison of the number of requests for voluntary throughcare in the period 2011 to 2014 . The figures for the more recent period include those offenders who were referred to CJSW through the CRP, in addition to those who might otherwise have requested support (via a prison social worker or once released into the community) under the existing voluntary throughcare provision. This provides some evidence - particularly in relation to male offenders - of a sharp increase in such requests in comparison to the period prior to the introduction of CRP.

Table 2.2 Requests for voluntary throughcare in Dundee and Lanarkshire CJSW areas (2011-2014)

	Dundee (Male offenders)	Lanarkshire (Female offenders)
2011-12	63	20
2012-13	200	22
2013-14	255	37

Source: Supplementary statistics provided by CJSW in relevant areas

2.43 Unfortunately, neither the unit-level nor aggregate-level monitoring data provide any means of gauging the number or proportion of offenders who engaged directly with CJSW in the community upon release (for example, by attending an initial meeting). Some indication of the extent of community engagement is, however, provided by the qualitative data in Chapter 5.

Referrals made by CJSW to external service providers

2.44 Aggregate-level data from CJSW for the period October 2013 to March 2014 do, however, provide an indication of the level of referral to external agencies. As the following table shows, during this time, a total of 589 referrals were made (it should be noted that multiple referrals may have been made for any one individual). Not surprisingly, given the numbers associated with the CRP at HMP Perth, the vast majority of these were in Dundee. Across the CRP as a whole, needs were most likely to be identified in relation to housing, addictions and benefits. The overall volume and the pattern of referrals provides some indication of where additional resource may be required as a result of a more systematic approach to voluntary throughcare - an issue we return to in the concluding chapter.

Table 2.3 Referrals to external partner agencies by CJSW in Dundee and Lanarkshire (October 2013-March 2014 inclusive)

	Referral made to external partner
Identified needs	Dundee	Lanarkshire	All
Housing	181	8	189
Addictions	183	13	196
Health Care	9	8	17
Benefits	177	8	185
Education	2	0	2
Total	552	37	589

Source: Aggregate-level monitoring data (based on activity by stage within period Oct 2013-March 2014)

2.45 From these data, it is not possible to determine the difference that the CRP made - in other words, the extent to which the project resulted in referrals that would not otherwise have been made.

Summary

2.46 All four sites operated a process that was recognisably part of the CRP, in that it involved structured engagement with eligible offenders through five broad stages; referral to CJSW; and an element of coordinated pre- and post-release support.

2.47 But there were significant variations in how each establishment organised itself to deliver these activities and, to a lesser extent, in the character of the activities themselves. There were also some modifications to the process during the period in which the CRP was running, both locally and at the level of the project as a whole.

2.48 The total number of offenders eligible for and actually participating in the CRP was relatively small in all prisons except HMP Perth, raising questions about the scope to generalise from a pilot phase in which it was not necessarily possible to achieve the 'critical mass' needed for all relevant staff to be fully aware of (and trained in) the project and for the project to have a sufficiently high profile to 'compete' with other services and initiatives already operating within the prisons.

2.49 Not all eligible offenders were necessarily identified. The sharp increase in numbers identified as eligible after March 2013 suggests that this was not happening systematically in the period before then.

2.50 An understanding of the potential and actual engagement with the CRP requires a more comprehensive and accurate set of monitoring data than was generated in the course of the project to date. Although it has been possible to piece together a broad picture of the operation of the CRP, this needs to be treated cautiously because of problems with the data. Any future provision will require better information.

2.51 What might be considered the minimal level of meaningful engagement with the CRP - participation in a Comprehensive Screen (Stage 3) - appears to have happened in relation to around half of eligible admissions.

2.52 The unit-level data indicate that this level of minimum engagement with the CRP was slightly higher (proportionately) among female than male offenders, and among those serving slightly longer sentences.

2.53 Both sets of monitoring data suggest a fairly high degree of attrition beyond Stage 3. The aggregate data, in particular, indicate that many of the meetings due to take place (between October 2013 and March 2014) did not actually proceed. The question of why engagement with the later stages of the CRP process within prisons tailed off in this way is explored in the following chapters.

2.54 Monitoring data from CJSW - especially in Dundee - provide evidence that offenders are being referred to other services in the community. What is less clear is the extent to which such referrals would have been generated in other ways in the absence of the project or the extent to which offenders are interacting directly with CJSW.