Background – the first wave

Custody Diversion Teams for Mentally Disordered Offenders began proliferating across England and Wales at the beginning of the 1990s following publication of Home Office Circular 66/90 (Home Office, 1990a), the Reed Report (Department of Health and Home Office, 1992), and Home Office Circular 12/95 (Home Office, 1995). The broad aim of these services was to divert mentally disordered offenders away from the criminal justice system and custody to care and treatment by health and social services because of concerns about failures in the policy of community care and the growing prevalence of psychiatric disorder in prison populations. Diversion was a humanitarian response to the care and treatment needs of offenders with mental health problems; it was an attempt to improve public protection and risk management by reducing the likelihood of re-offending; and it was a support mechanism providing expert knowledge and input to enable the smooth operation of the criminal justice pathway (National Association for the Care and Resettlement of Offenders, 1993).

The National Schizophrenia Fellowship (1999) found that CJLDS generally shared a number of objectives, including diverting mentally disordered offenders from prosecution by assessing them in police custody, and reducing the number serving a custodial sentence by providing information to the Crown Prosecution Service, Probation Service and courts. However, no national guidelines were published and the Social Services Inspectorate (SSI, 1997) discovered local variations had developed according to perceptions of need and the availability of dedicated resources.

Challenges for the first wave

This first wave of diversion services did produce pockets of good practice. For instance, one multi-agency CJLDS established in the north east of England (NE CJLDS) in 1995 had adopted a broad definition of their client group (in order not to restrict access to the service they could provide) and offered a wide-ranging service from arrest to sentence. This meant that referrals did not necessarily fit neatly with the original diversion policy to divert individuals away from the criminal justice system and custody, and into a psychiatric hospital. Many of those referred to the NE CJLDS team did not have a severe mental disorder – a significant proportion were ‘misusing drugs and/or alcohol’ or had a non-severe ‘mental health problem’ – and did not require admission to hospital. Despite the fact that many were committing significant offences (violence against the person, burglary, theft, etc.), they were not necessarily at risk of a custodial sentence. A discontinuation of criminal proceedings was also not appropriate for most were not ill enough and had committed fairly serious offences. In other words, there was not one type of person referred to the NE CJLDS team but instead many different types of people with different psychiatric and criminal histories. There was no one single aim but instead many different aims, including diverting people to health and social care whilst criminal charges were processed. There was not one type of outcome but instead many variations on outcomes, including admission to hospital and for some a prison sentence. In short, the NE CJLDS team ‘tailored’ their service to the needs of individual clients (Dyer, Reference Dyer2001).

However, despite some evidence of good practice, development of this first wave of diversion services faced four difficulties: national variability between services; geographical coverage and the lack of services provided in some areas; services which changed over time, losing their original focus; and the difficulties of demonstrating the effectiveness of these complex services.

Lack of government guidelines, competing professional interests and tensions inherent in multi-agency working led to national differences in the development of CJLDS models (National Association for the Care and Resettlement of Offenders, 2004; Brooker and Ullman, Reference Brooker and Ullmann2009; Pakes and Winston, Reference Pakes and Winstone2009; James, Reference James2010), including: single practitioners versus multi-agency schemes; part-time or on-call services compared with full-time dedicated teams; panel assessment schemes compared with front-line proactive teams; reactive compared with proactive screening; and variations in the services offered by a CJLDS which might include some or all of the following: mental health assessment in police stations, reports to court, providing recommendations on sentence and management, managing and sharing information with probation and prisons, short-term treatment and access to inpatient beds.

More recently, local and national surveys (Dyer, Reference Dyer2009a; National Association for the Care and Resettlement of Offenders, 2011) have again confirmed huge variations in coverage, size, composition, governance, funding arrangements and quality of services provided. According to the annually updated NACRO directory of criminal justice mental health liaison and diversion schemes in England and Wales, there are approximately one hundred liaison and diversion schemes operating at court, police station or both. At present there are 335 magistrates’ courts in England, which means that at best only one third of magistrates’ courts currently have liaison and diversion services. In addition, many police custody suites have no provision whatsoever.

These variations exist even within a single commissioning area. For example, a review of the NE CJLDS (Dyer, Reference Dyer2009a) discovered five schemes in operation; however, two large geographical areas (one rural, one urban) had no access to any CJLDS, either currently or historically. Only one of these schemes proactively screened those detained in police custody, the remainder relied on referrals or other methods of indirect screening. The benefits of proactive screening include higher levels of service user identification and engagement, whereas reactive or referral-based methods rely on police custody officers whose priorities often lie elsewhere. One of the schemes operated part-time only, with others providing a service on Monday to Friday, 9a.m. to 5p.m. Two of the schemes consisted of a single community psychiatric nurse (CPN) compared with another which offered three CPNs and an approved mental health practitioner (AMHP).

In addition to variations between services in relation to model of service provision and geographical coverage, changes have also occurred over time within existing CJLDS. For example, a review of the NE team described above, fourteen years after it was originally established (Dyer, Reference Dyer2009b), uncovered a very different service. Over the years, the team had been subject to varying levels of interest and input from managers, as well as decisions about service operation and location and decreasing investment in terms of staff and other resources, all of which had impacted on service direction and delivery. The NE CJLDS was no longer operating as a ‘diversion’ service in the strictest sense, either in line with their original operational policy, which stated the aim of the service was to ‘divert mentally disordered offenders away from the criminal justice system except when public interest required prosecution’ (Morrison, Reference Morrison1995), or with previous government publications (e.g. Home Officer, 1990; Department of Health and Home Office (1992). The service no longer focused on people at the start of the criminal justice pathway (that is, in police stations or magistrates courts) when diversion from the criminal justice system or to an appropriate sentence remained an option. Instead the NE CJLDS were now providing a treatment service for clients of the local probation service who were already serving a community sentence. The service, while appreciated by local probation officers, focused at the end of the criminal justice pathway when having an impact on decisions to ‘divert’ was no longer an option.

As well as a lack of national direction, an additional issue which has threatened the development of CJLDS has been ‘how could we evaluate the overall impact of such very diverse services?’ For instance, in terms of the potential client group, Peay (Reference Peay, Maguire, Morgan and Reiner2007) argues ‘mentally disordered offenders . . . are not a single, easily identifiable group’ – mental illness comes and goes. A diagnosis of mental illness does not necessarily mean a lifetime of symptoms. A potential client group might include people who have a serious mental illness subject to detention and treatment under the Mental Health Act (2007) and those with more vague mental health and emotional problems, substance misuse and learning difficulties. Issues surrounding age, gender and ethnicity also complicate any attempt at evaluation. In terms of service provision, what worked for one individual might not work for others. There was no suggestion that there was a single ‘good treatment’. Instead it was possible there would be a variety of ways of arriving at a good as opposed to bad outcome, but the problem was how to distinguish good pathways from bad? The aim of diversion was not always clear or straightforward, that is did it mean diversion from prosecution, or diversion to care and treatment by health and social services while prosecution continues? Unresolved tensions inherent across both the continuum of ordered−disordered behaviour and that of law abiding−law breaking behaviour acted to confuse the aims of CJLDS (Peay, Reference Peay, Maguire, Morgan and Reiner2007; Prins, Reference Prins2010). Notions of care and treatment are usually reserved for those who are ‘disordered’, while demands for protection and punishment exist for those who offend. Just as the severity of mental health problems can vary as described above, so can the nature of the offence, from the minor to the very serious. Tensions between ‘care or control’, ‘treat or punish’ are confounded where disorder and offending exist side-by-side in one individual. As described earlier with the NE CJLDS, services of this type deal with many different types of people with different psychiatric and criminal histories, and as such they must contend with many different aims and many variations on outcomes. These variations and tensions must then have an impact on evaluation: what outcomes could or should be expected and how ‘success’ might be defined and measured.

There have been attempts to evaluate services, but results have varied according to the geographical location and model of scheme in operation (e.g. Holloway and Shaw, Reference Holloway and Shaw1992; Joseph and Potter, Reference Joseph and Potter1993; Greenhalgh et al., Reference Greenhalgh, Wylie, Rix and Tamlyn1996; Purchase et al., Reference Purchase, McCallum and Kennedy1996; James, Reference James2000; Riordan et al., Reference Riordan, Wix and Kenny-Herbert2000; McGilloway and Donnelly, Reference McGilloway and Donnelly2004). Because there was no ‘definition of success’, it was therefore difficult to determine if this early wave of diversion services were working or not.

Criminal justice liaison and diversion – the second wave

The start of the second wave of interest in CJLDS began formally with the publication of the Bradley Report (Bradley, Reference Bradley2009). This was an independent review to determine to what extent offenders with mental health problems or learning disabilities could be diverted from prison to other services and what were the barriers to such diversion. A number of issues prompted the review, including continued concerns that the numbers of prisoners with mental health problems remain high and that prison can itself have a detrimental impact on mental health (Singleton et al., Reference Singleton, Meltzer and Gatward1998; Birmingham, Reference Birmingham2003; Rickford and Kimmett, Reference Rickford and Kimmett2005; HM Inspectorate of Prisons, 2007; Loucks, Reference Loucks2007; Prison Reform Trust, 2009). There were also arguments that public protection and reducing re-offending might be better served by addressing the multiple problems that many of the most persistent offenders face, such as poor health (Social Exclusion Unit, 2002). The Centre for Mental Health, Rethink and the Royal College of Psychiatrists (2011) argue that increasing evidence from international experience and from local schemes in this country suggests that well-designed interventions can reduce re-offending by 30 per cent or more. Growing concerns around the increasing cost of imprisonment also meant that alternatives which offered better ‘value for money’ became increasingly attractive. Two reports published subsequent to the Bradley Report have aimed to evidence the cost/benefit of CJLDS. Rethink and the Sainsbury Centre for Mental Health (2010) argue that even on fairly conservative assumptions the diversion of an offender with mental health problems from a prison sentence towards effective treatment in the community could result in savings to society of over £20,000 because of reductions in future offending, and the Centre for Mental Health, Rethink and the Royal College of Psychiatrists (2011) argue that while national coverage of CJLDS will cost £50 million per year, this will be offset by the increased use of cheaper community orders and reductions in re-offending. For example, it costs £4,200 for an intensive two-year community order, involving twice-weekly contact with a probation officer, eighty hours of unpaid work and mandatory completion of accredited anti-offending programmes, compared with £5,000 to keep someone in prison for six weeks.

Whether motivated by an ideological wish to drive down costs or to impact on the revolving door of re-offending, the current UK Coalition Government's ‘Rehabilitation Revolution’ aims to reduce the current prison population by 3,000 by 2014, part of which will be achieved by diverting mentally disordered offenders away from prison to more appropriate forms of treatment and punishment. Breaking the Cycle (Ministry of Justice, 2010), the government's Green Paper, along with the cross-government strategy No Health Without Mental Health (HM Government, Reference Government2011) describe the intention to continue the renewed focus on CJLDS. The importance of integrated care pathways and continuity of care are recognised throughout the mental health report. Perhaps in an attempt to resolve some of the challenges faced by the first wave of CJLDS, both publications describe plans to identify ‘best practice, quantify the benefits and develop appropriate quality standards’. The government aims to achieve this by identifying a number of CJLDS ‘pathfinder’ pilot sites which will be subject to ongoing monitoring, research and evaluation. Roll out of a national CJLDS implementation programme will follow an assessment of the success of these pathfinder projects.

Current practice and challenges for the second wave

The current development of diversion services faces three key challenges: (1) CJLDS tend to be uncoordinated and to focus on discrete key stages of the offender pathway (pre-arrest, police custody, courts, prison and on release from prison) rather than the pathway as a whole; (2) despite (or more likely because of) the focus on stages, the services provided at each stage are problematic; and, (3) specific groups are receiving a particularly poor service.

A concern emerging from current literature (Winstone and Pakes, Reference Winstone and Pakes2010), and based on observation of the reasons why earlier CJLDS struggled, involves the continued lack of overall coordination and planning across the offender health pathway and the role CJLDS have to offer in terms of managing patient-centred care. CJLDS tend to focus on ‘key stages’ in the offender pathway (e.g. police station or court), providing actions to meet the needs of the services at these discrete stages rather than adopting a patient-centred approach which recognises the impact of action on the longitudinal institutional careers (criminal justice, health and social care) of their clients. For instance, early intervention and prevention strategies and services are key in stopping the offender pathway developing further but are a much neglected part of service coordination and development. Reports suggest that the level of contact between the police and the mentally ill has increased substantially in recent years (Patch and Arrigo, Reference Patch and Arrigo1999; Lamb and Weinberger, Reference Lamb and Weinberger2005; Price, Reference Price2005; Wells and Schafer, Reference Wells and Shafer2006; NPIA, 2008). A study carried out in Scotland to investigate police officers’ views on their roles in dealing with people with mental health problems, and with mental health services (Mclean and Marshall, Reference Mclean and Marshall2010) reported a number of recurrent themes, including the problems and failures of collaborative working with health services. Pre-arrest diversion was recognised as an option by police officers – there may be no indication for arrest, and they did not want to arrest individuals unnecessarily – but gaps in health services or failures in collaborative working could result in inappropriate detention in police cells, an echo of the so-called ‘mercy booking’ reported in the USA (Watson et al., Reference Watson, Morabito, Draine and Ottati2008). Many early intervention and prevention services are already in place across the UK but appear piecemeal (Dyer, Reference Dyer, Williams and Vogt2011a). There is no overall local or national knowledge or coordination of services at this stage of the offender health pathway. Consequently, sharing good practice is limited – what is being provided to whom with what effect? – and therefore it is not clear what or where the service or support gaps are.

In terms of the identification of mentally disordered offenders in police custody, Mckinnon and Grubin (Reference McKinnon and Grubin2010) argue that a significant amount of health morbidity is present among detainees; however, current police screening procedures detect only a proportion of this. One solution would be to use various different screening strategies and tools to support the identification of those who might otherwise go undetected, including women, people with learning difficulties and those with dual-diagnosis (Jacobson, Reference Jacobson2008; Scott et al., Reference Scott, Mcgilloway and Donnelly2009; Winstone and Pakes, Reference Winstone and Pakes2010). However, again a lack of coordination and planning continues to cause problems in terms of standardising support and training provided to police officers, and determining what action should be taken following a positive screen.

CJLDS provision of support and information to the Probation Service and the Courts should prevent unnecessary requests for psychiatric reports and resulting additional costs and delays (Rethink and the Sainsbury Centre for Mental Health, 2010). Unfortunately, the provision of psychiatric reports continues to offer challenges and highlights tensions in the relationship between the criminal justice and healthcare systems. There are already a number of alternatives to custody, for example Community Orders with Mental Health Treatment Requirements (MHTR). However, according to a report by the Sainsbury Centre for Mental Health (Khanom et al., Reference Khanom, Samele and Rutherford2009) few people were given a MHTR, despite two-fifths of people on community sentences having mental health problems. The report found that the purpose of the MHTR and the group of people to whom it can be given are not clear to sentencers, probation staff or health professionals. Long delays in the production of court psychiatric reports were a major barrier to the use of the MHTR, and without an offer of treatment from local services the courts could not make an MHTR. There was also widespread confusion among health, probation and court staff about how an MHTR can be breached and what should be the consequence. The authors argue:

For those mentally disordered offenders who do receive a custodial sentence, there are standardised prison reception practices and screening tools available (Carson et al., Reference Carson, Grubin and Parsons2003). However, it appears that they do not meet need, particularly in relation to the identification of those with learning difficulties (Durcan and Knowles, Reference Durcan and Knowles2006). Individual prisons have made their own amendments or improvements to screening tools (Shaw et al., Reference Shaw, Senior, Hassan, Weston, King and Mwasambilli2009), but these lack overall coordination, planning and evidence of ‘good practice’.

At the end of the pathway, as with the start, prison release and resettlement suffer from serious management neglect (Dyer et al., Reference Dyer, Biddle and Irving2010). As with early intervention, services are already in place across the UK, often provided by non-statutory, third-sector organisations. However, service availability appears piecemeal and with no overall knowledge or management, information sharing or sharing of good practice it is not clear what or where the service and support gaps are.

In addition to problems with lack of management coordination and planning, another problem which faces the current development of CJLDS is the failure to meet the needs of specific groups of mentally disordered offenders. For example, people with learning difficulties have a reduced chance of being identified at all key stages of the offender pathway because of the lack of ‘expert’ input from a forensic learning disability nurse, and screening and assessment tools which do not include items to support their identification (Jacobson, Reference Jacobson2008; Shaw et al., Reference Shaw, Senior, Hassan, Weston, King and Mwasambilli2008; Beebee, Reference Beebee2010; Jones and Talbot, Reference Jones and Talbot2010; Department of Health, 2011). Issues surrounding dual-diagnosis (substance misuse plus a mental health problem) appear to be mirroring those which surrounded all mentally disordered offenders at the beginning of the 1990s, that is the patients no-one owns (Harris and Web, Reference Harris and Webb1999). The literature suggests that the issue is not identification/screening or assessment, rather it is the fact that this group often falls between services – Substance Misuse and Mental Health (Scott et al., Reference Scott, Mcgilloway and Donnelly2009; Winstone and Pakes, Reference Winstone and Pakes2010). In order for diversion and liaison services to work, there must be somewhere to divert to. The literature also confirms a concern over the needs of women. Women, it is argued, have needs which are more than or sufficiently different from those of men so that they require dedicated service provision to increase chances of identification and engagement with assessment, support and treatment services (Corston, Reference Corston2007; Hunter et al., Reference Hunter, Boyce and Penfold2007; Scott et al., Reference Scott, Mcgilloway and Donnelly2009; Staddon, Reference Staddon2009; Hean et al., Reference Hean, Heaslip, Warr, Bell and Staddon2010). However, such services are rarely provided.

Good practice suggestions for the second wave

A review of the literature does provide numerous suggestions for good practice which could alleviate some of the problems which beset the first wave of diversion services. For instance, multi-agency commissioning and governance arrangements that manage and monitor a diversion team which has its own identity, a minimum of three practitioners and which provides a proactive, holistic service across the whole offender pathway (Winstone and Pakes, Reference Winstone and Pakes2010).

A need for ‘holism’ or a holistic approach is a useful theme emerging from current literature. There is a recognition that mentally disordered offenders present with complex needs, which means that assessment, management and the provision of support and service input should not focus on any one element, such as mental health problems, or any one discrete or isolated stage, such as police station or court. Successful outcomes will be based on the identification of physical, mental and social needs, and the provision of an individualised support package which aims to improve overall emotional health and wellbeing. This means that the assessment process must be holistic, and access to services to divert people sufficiently varied but coordinated across the whole offender pathway (National Association for the Care and Resettlement of Offenders, 2005; Confederation of British Industry, 2009; Revolving Doors Agency, 2010; Winstone and Pakes, Reference Winstone and Pakes2010). The Bradley Report (Bradley, Reference Bradley2009) suggests that the one service which could be common across the offender pathway are CJLDS. Improvements to the structure, function and coverage of these services therefore represents the single biggest opportunity for progress in this area. Bradley envisioned a new model of liaison and diversion services which he called ‘Criminal Justice Mental Health Teams’. These teams would be responsible for managing continuity of care across the whole offender pathway (including community, police custody, courts, prison/community sentence and resettlement).

Finally, in terms of identifying success, perhaps one solution to the complex nature of the CJLDS process might be to identify and ‘map’ offender pathways (Williams and Dyer, Reference Williams, Dyer, Byrne and Ragin2009; Dyer, Reference Dyer, Williams and Vogt2011b). Rather than focusing on discrete stages in the criminal justice process or activities undertaken by CJLDS, this method would adopt a case-centred approach and track individual journeys (including for those groups failed by current service provision) through the various stages in an attempt to uncover common types of pathways which lead to particular outcomes.

CJLDS ‘process’ different people in different ways. Therefore, there may not be only one type of outcome but many variations in outcomes for different people who have been processed in different ways. What worked for some might not work for others. This, Watson (Reference Watson, Watson and Grounds1993) argued, renders inadequate any simple focus on clearly delineated and significant episodes. Instead, he suggested that what is needed is an attempt to understand the complex ways in which some individuals become channelled through particular institutional and extra-institutional careers. In a review of the international literature on the epidemiology of mentally disordered offenders, Badger et al. (Reference Badger, Nursten, Williams and Woodward1999) noted that long-term studies which track people or individual cases are rare. They therefore recommend that:

This method has been trialled by a study using longitudinal data supplied by the NE CJLDS described earlier (Dyer, Reference Dyer2006). The research developed and applied an approach to data analysis based on time-ordered cluster analysis which made it possible to map out the pathways followed by individual mentally disordered offenders, and identify common patterns within those pathways. The aim was to identify and map the different institutional careers experienced by people referred to the team and the different paths their careers took as a consequence of the team's actions. Five different types of career were identified (see Figure 1).

Figure 1.

Career typology of individuals referred to the NE CJLDS (n = 1,011)

Careers 1 and 2 describe experiences of offenders – violent offenders with no psychiatric history who were referred, assessed by the community psychiatric nurses (Career 1), or the probation officer or social worker (Career 2) and diagnosed but had no health or social care needs identified and were not referred again. Careers 3 and 4 describe experiences of mentally disordered offenders – violent offenders with a psychiatric history half of whom (Career 3) were referred, assessed and diagnosed, had health or social care needs identified and were not referred again; the remainder (Career 4) were not assessed, nor did they have needs identified (possibly because they had committed less serious offences and were already in receipt of care and support) and consequently all were re-referred repeatedly. Career 5 represents neither – individuals referred for information and for whom little else is known.

A number of important and positive outcomes arose from this approach which had practical significance for the development of the service. For example, Career 4 represented a ‘revolving door’ outcome, where everyone was re-referred repeatedly to the diversion team. Future developments attempted to avoid this outcome by determining that everyone with a psychiatric and criminal history be assessed and their needs identified and met, regardless of the nature of their offending or the existence of current care packages. That said, acquiring a robust and reliable psychiatric and criminal history alongside an assessment of needs was not always easy or straightforward and was impacted by a number of issues including multi-agency working and information sharing across agencies, training and resourcing.

Defining outcomes was also problematic as previously discussed. The NE CJLDS activity involved screening, assessment, referral and signposting. ‘Diversion activity’ (for example, fewer people with mental health problems entering custody inappropriately, improving mental health, increasing social stability and reducing reoffending) happened and was recorded elsewhere. Consequently, the nearest outcome measure available to this longitudinal study to denote successful/unsuccessful pathways was re-referral/no re-referral (indicating diversion occurred post-discharge from the NE CJLDS or did not and resulted in a re-referral to the team).

These issues were problematic across the first wave of CJLDS and continue to cause concern for the success of the second wave. In particular, if diversion and liaison is to demonstrate that it ‘works’ and therefore justifies further investment, that will mean demonstrating some impact on expectations, including fewer people with mental health problems entering custody inappropriately, reducing reoffending and demonstrating potential savings. Defining and accessing reliable outcome data will be an imperative part of any attempt to map pathways for this second wave of CJLDS.

Mapping pathways within UK health and social care services is not a new concept. Integrated Care Pathway (ICPs) are clinician-led and driven, focussing on service users and best practice which aimed to have the right people doing the right thing, in the right order, at the right time, in the right place, to the right standard and with the right outcome. Emphasis is given to the importance of identifying and measuring ‘critical indicators’ – outcomes from interventions that make the biggest difference to ‘recovery’. Far from being linear, ICPs expect complexity in the form of variations and change as people move along the pathway. The causes of variations can be recorded and monitored over time, allowing the ICP to be altered to include or manage some of the most common reasons or risk factors. These events or actions can then be changed or removed. Variations should always lead to some kind of action. They are now used extensively across the UK (Kings Fund, 2007), including an Offender Mental Health Care Pathway (Department of Health and National Institute of Mental Health in England, 2005). While not without their problems (Rees et al., Reference Rees, Huby, McDade and McKechnie2004; Croucher, Reference Croucher2005; McDonald et al., Reference McDonald, Whittle, Dunn and Luc2006; Williams, Reference Williams2009), the concept or ideas behind ICPs could also offer some solution to the tensions inherent in the provision of CJLDS, as well as ideas for more appropriate evaluation methodologies.