Since the early 1990s British policy for offenders with mental disorders has been to divert them, wherever possible, from the criminal justice system to psychiatric services. The policy is intended to be in effect at all levels of the criminal justice system, applying both to the transfer of highly dangerous offenders from prison to maximum-security forensic hospitals, known in the United Kingdom as special hospitals, and to the diversion of petty offenders from police stations or magistrates' courts to local outpatient and community psychiatric services. In addition to encouraging diversion, British policy aims to ensure that offenders with mental disorders who need incarceration are detained in levels of security as low as is commensurate with public safety and as close to their home community as possible.
Having such clear goals has brought welcome consistency to the nation's approach to planning services for this group. However, implementation of the principles has not been easy, partly because of the need to achieve agreement and establish close working relations among the various public agencies that share responsibility for offenders with mental disorders. The High Security Psychiatric Services Commissioning Board, part of the executive branch of the National Health Service, has responsibility for mentally ill offenders who require the highest level of security. In England, highly secure care is provided by the special hospitals of Broadmoor, Ashworth, and Rampton. Carstairs Hospital in Scotland provides a similar service for Scotland and Northern Ireland and offers medium-security care as well. These four hospitals hold about 2,000 patients, or about three to four people per 100,000 population in the United Kingdom (1). In addition, about 40 regional secure units offer medium-security care both as a step-down facility and as a secure treatment base for patients who do not require a higher level of security. Unlike the special hospitals, the regional secure units are designed for only shorter-term care, for example, stays of 18 to 24 months.
Patients who need medium- or low- security services are the responsibility of local health authorities, which serve populations that vary considerably in size. The size of the population served has implications for service provision to a small diverse group such as offenders with mental disorders.
The third major public body involved with these offenders is the Prison Service, which comes under the jurisdiction of the Home Office but includes some privately operated prisons. Provision of mental health services to offenders with mental disorders in the Prison Service is problematic because the Prison Health Service is separate from the National Health Service, and health records are not transferred between the two services. This separation hinders planning for the release from prison of psychiatrically vulnerable people, particularly those who are homeless and are probably not registered with a community physician. In cases where continuity of treatment is important or a risk of self-harm exists, communication between the two services is dependent on the initiative of individual prison doctors and doctors in the community.
Coordination of services is a particular concern, given the relatively high proportion of the population in jails and prisons with mental health problems. Swyer and Lart (2), who evaluated a plan to improve discharge arrangements for prisoners with mental health problems leaving Winchester Prison, found that about 25 percent of sentenced men reported a history of mental health problems. In a study of the British prison system by Gunn and colleagues (3), about 37 percent of sentenced male prisoners had a mental disorder of some kind. The researchers estimated that 3 percent of all male sentenced prisoners needed transfer to a psychiatric hospital.
In addition, the behavior of a number of persons in the general population of psychiatric patients raises issues of security and safety of staff and fellow patients. These patients may be in a hospital or living in the community, but they are unlikely to be represented in statistics relating to criminal activity, especially if they have committed minor crimes and may not even have been charged with an offense.
The attitudes and responses of care staff and of the general public may contribute to underreporting. For example, a study of adults with learning disabilities (mental retardation) in Cambridge found very low rates of contact with the criminal justice system and of prosecution, partly because care staff were reluctant to report even serious assaults to the police (4). Yet a small number of people with mental retardation do present considerable management and security problems and require secure care.
In most parts of the United Kingdom, all elements of the mental health system are under severe pressure to meet the demand from referral agents. Acute care units and services that provide greater security for patients with higher levels of dependence are particularly in demand.
Acute care psychiatric units are serving an increasing number of more severely disturbed patients detained under sections of the Mental Health Act of 1983 that call for periods of intensive care due to risk to themselves or others. Patients whose needs cannot be met within local acute care units often have to wait several weeks before transfer to a more secure environment. Access to appropriate care and treatment may require recourse to expensive and often private out-of-area placements many miles from the patient's home and family. Other patients may become "stuck" in high-security placements that were suitable at the time of referral but have become inappropriate for their current needs. Their move to a less restrictive placement may be delayed, often for a considerable time, due to the scarcity of suitable accommodations at lower levels of security.
The category of offenders with mental disorders covers a wide range of persons with criminal behavior, from those who have committed violent crimes to those who are petty offenders. Although program planning is made more difficult by this broad definition, a wide scope encompassing all mentally ill offenders is necessary if the policy of diversion to psychiatric services is to be fully realized and if local services are to provide appropriate levels of security. The High Security Psychiatric Services Commissioning Board has responded to the challenge of program planning for this diverse population by launching several new research initiatives. As a first step the board commissioned several systematic literature reviews to inform the next round of funded research. One of these reviews—on the epidemiology of criminal offenses by persons with mental disorder—was carried out by three authors of this paper (DB, PW, and MW).
Although epidemiological findings would constitute a valuable guide for service planning, the literature on offenders with mental illness in the United Kingdom is quite limited, and no general population study of this group has been done. Thus the United Kingdom must use results from studies in other countries to plan for services. Such transfer of findings is seldom straightforward, even when the research reports clearly identify the subjects' diagnoses. When the broadly defined phenomena of mental disorder and criminality are considered, international comparisons become even more problematic. Data from studies in other countries are probably not useful for predicting prevalence or incidence of offenses by persons with mental disorders in the United Kingdom, although such studies can identify risk factors that may apply cross-nationally.
For example, the Swedish Metropolitan Project, a follow-up study of more than 15,000 persons born in Stockholm in 1953, showed that mental disorder is positively associated with crime of all types, more so among women than among men, and that major mental disorders are more strongly associated with violent crime (5). The study also found that offenders with major mental disorders tended to commit more offenses than offenders with no mental disorder. These findings are useful and generalizable. However, the study's findings that the lifetime prevalence of offenses committed by persons with mental disorder was 2 to 8 percent for men and 1 to 4 percent for women cannot be directly applied in another country, due to differences in criminal laws and in the types and levels of security of available psychiatric services. Woodward and associates (6) have provided a fuller discussion of these results and issues.
The literature on the prevalence of offenses committed by persons with mental disorders and on the characteristics of the offenders suggests that regional variation exists regardless of whether the data were gathered in the criminal justice system, general psychiatric services, or secure psychiatric services. This variation is to be expected, because both criminal behavior and the incidence of mental disorder are related to demographic characteristics that vary in different populations. Thus national statistics from the study of specific service agencies with large or undefined catchment areas are of limited use for regional or local planning.
These findings can be applied at the regional level only if the original study included detailed demographic data on the study population and if information is available to link the demographic data with data on the prevalence of offenses by people with mental disorders and the characteristics of the offenders. These data are rarely available from national or agencywide studies. Thus there is no substitute for regional and local surveys that collect data on the local population for whom services are being planned.
Creating a regional consortium of agencies is one approach to strategic planning and commissioning of services for offenders with mental illness. A notable example in the United Kingdom is the Wessex consortium, which consists of five local health authorities and one social services department in southern England. The consortium's catchment area has a population of 2.5 million, which is large enough to support the range of resources needed for viable planning. The consortium was established in February 1997.
To determine the consortium's service requirements, a project team was established to quantify and qualify the needs of mentally ill offenders who required secure care and specialist mental health services. The team consisted of a project manager and two project workers drawn from community psychiatric nursing and psychiatric social work backgrounds. The team's task was to identify the number of offenders who were detained in special hospitals, regional secure units, out-of-area placements, and the prison system and carry out a needs assessment for each offender to determine the individual's requirements for longer-term placement. The aim was to achieve appropriate placements in the least restrictive environment and provide a guide for developing local secure services.
The study used a "rates-under-treatment" approach to identify 110 individuals from the consortium area who were known to be receiving treatment in secure and specialist mental health settings. The project team decided to give priority to 58 individuals with more immediate and unresolved placement needs. These needs assessments were targeted to three groups of offenders: those who were located in special hospitals and were due to be transferred or discharged within two years; those who were located in the regional secure unit and were detained for more than 12 months with a supposed maximum length of stay of up to two years, by which time plans should be made for their future placements; and those who were in out-of-area placements.
Clinicians in the special hospitals and regional secure unit were asked to complete a questionnaire detailing each patient's clinical background and history of offenses and indicating the patient's future needs for secure care. A member of the project team visited the out-of-area placements to make similar assessments jointly with the local clinicians. At the same time, a survey of the area covered by the consortium was undertaken to identify residential facilities that offered high levels of security and care. A full account of the methodology is given elsewhere (7). The needs assessments were completed in February 1998. The survey of the residential facilities was completed in March 1998.
Of the 58 individuals with unresolved placement needs, 34 could be matched with appropriate local facilities or with appropriate placements outside the consortium area. However, appropriate placements were not available for the other 24, who remained in placements with higher levels of security than were needed given their current condition. Most needed to move to long-stay low- or medium-security settings. The needs assessments done by the project team for offenders with mental disorders in the prison system revealed a constant pressure to transfer seriously disturbed offenders to medium-security psychiatric facilities, adding to the already significant demand for medium-security beds. As a result, many seriously mentally disturbed prisoners continued to be held within the prison system.
The needs assessments produced useful information on the social and psychiatric characteristics of the consortium area's offenders with mental illness and their patterns of offenses. They were primarily male and had histories of repeated and long-term inpatient and prison care. Their previous poor response to treatment and rehabilitation, together with enduring problems of mental illness, mental retardation, or personality disorder, marked them as a highly institutionalized group. Two-thirds had been convicted of serious violence against others, including homicide, and half had been convicted of sexual offenses. It was clear that they required containment in conditions of security for a protracted period.
The individuals requiring long-stay care were fairly evenly distributed across the areas served by the five health authorities in the consortium. Individually, each health authority would not have had enough mentally ill offenders with similar needs to justify the development of resources for specialist care. Collectively, however, the consortium agencies had enough offenders to warrant the development of two long-stay medium-security units—one for offenders with mental retardation and personality disorder and one for those with mental illness, personality disorder, or both. The survey of residential facilities had revealed a complete lack of this type of service. A program is currently in place to develop these units locally.
Finally, detailed analysis of the needs assessments of the 58 offenders with unresolved placement issues allowed the project team to develop profiles of the types of patients found in different secure settings. The profiles were used to create a model for service requirements for each group of patients, thus reducing some of the uncertainties in describing different levels and types of secure care.
For example, offenders who required medium-security care typically had committed serious index offenses against persons. They were determined absconders who would continue to pose an immediate risk to others if they were at large. Offenders who needed low-security care were more likely to have committed less serious index offenses and be at greater risk to themselves. They were less likely to make serious attempts to abscond, and, if they did, they posed less of an immediate threat to the public.
These differences have implications for the levels of security required for each group. Medium-security settings would need to be secure enough to maintain public safety and the safety of their occupants. Their aim should be to provide a range of security conditions that can be used flexibly to meet the needs of individuals at different stages of their stay and rehabilitation. A range of security levels can be achieved by a combination of effective staff deployment; procedural guidelines; internal planning of the building; appropriate design of windows, doors, and walls; and use of locks, electronic security systems, and perimeter security measures.
Offenders who require a low level of secure care should encounter a less intrusive level of security than that found in a medium-security setting. In low-security settings, safety and security are achieved primarily through environmental and relational methods rather than by using high-tech equipment. Security measures should be designed to impede rather than completely prevent patients from absconding, and more emphasis should be placed on staff members' roles than on physical security measures.
Further distinctions were found between offenders who needed long-term care of more than two years and those who needed shorter-term care for up to two years. The former group had enduring mental health and mental retardation problems and had experienced repeated or long-term institutional care. They responded poorly to various interventions over a prolonged period and had weak community and family links. In contrast, those who required shorter-term care were more acutely disturbed and had avoided overt institutionalization. Moreover, they retained recent and more active community and family links.
Differences in the length of stay patients require influence the types of facilities and programs of therapy and rehabilitation they need. Those who are detained for a protracted period of time should reside in a facility with adequate space and access to natural surroundings. Ample opportunities for leisure, work, and recreation should be available on site. A full range of medical, psychological, and social interventions should be available to counter the harmful aspects of institutionalization and provide ongoing maintenance and a gradual rehabilitation process.
Those who require short-term care may need a similar range of services from a team of treatment professionals, but their treatment should focus on more rapid treatment and rehabilitation with the aim of moving them to less restrictive settings as soon as possible.
The detailed information obtained from the assessments allowed the consortium to develop accurate specifications for services to guide the commissioning of the new facilities (8).
The system of care for offenders with mental disorders in the United Kingdom has developed piecemeal over the years, and responsibilities have been divided between the criminal justice system, the National Health Service, and local health authorities and social services departments. The specialized secure facilities have also evolved in an ad hoc way throughout the country, without standardized models of service delivery.
Policies such as diversion to treatment and containment at the lowest level of security commensurate with public safety have highlighted the need for better information about the characteristics and service requirements of offenders with mental disorders and the need for agreed-on definitions to identify such offenders. Such information can best be obtained by large-scale epidemiological studies that have so far not been conducted in the United Kingdom. However, the additional policy objective that care should be provided in the patient's home area wherever possible has drawn attention to the value of local surveys examining the needs of offenders with mental illness and the resources available to them. The results of such surveys are useful for service planning, provided that the population of the region studied is large enough to support the resources needed for specialized care.
This work was supported by the United Kingdom High Security Psychiatric Services Commissioning Board and the Wessex consortium. The authors thank Jean Nursten, M.S.W.
Mr. Badger, Mr. Vaughan, and Mr. Williams are with the department of professional education in community studies and Dr. Woodward is with the department of applied statistics at the University of Reading in Earley, Reading, England. Address correspondence to Mr. Badger, University of Reading, Bulmershe Court, Earley, Reading RG6 1HY, United Kingdom (e-mail, firstname.lastname@example.org). This paper is part of a special section on mentally ill offenders.