A look through the PRISM

It was all a bit of a mystery – Scotland's 'most violent men' were not being violent. As a young clinical psychologist, your first author was confronted by this mystery while working in the Barlinnie Special Unit (Boyle, 1977; Cooke, 1989). The unit was a small experimental facility that contained violent prisoners who had proved themselves unmanageable in other prisons. The majority had killed, at least once: however, murder was only part of the pattern of persistent and serious violent acts they perpetrated. They had been brought up in a subculture in which violence was normalised; starting at an early age they had progressed through diverse and prolific criminal careers. They suffered from many forms of dysfunction; personality pathology, substance misuse and, in some cases, episodes of major mental illness. Despite this concentration of risk, over the Unit's 21-year history only two assaults took place. The prisoners as a whole remained dominant, forceful, challenging, impulsive and prone to anger, yet their violence was curbed.

Perhaps, in retrospect, this should not have been a surprise, over four decades ago Mischel (1968) told us that the situations in which individuals find themselves have as much, if not more, influence on their behaviour as individual risk factors. But as psychologists – by inclination and training – our focus is on the individual and not the situation. Reflecting on the Special Unit suggested that a mosaic of situational factors probably underpinned this change in behaviour; these factors range from improved staff selection, training and morale, through a focus on relational rather than structural security, to access to a rich variety of activities (Cooke, 1989).

To some the Barlinnie unit was a brave penal experiment; to others it was 'Porridge with Cream'. Whatever perspective is taken, the radical and rapid change in the violence of these high-risk prisoners demonstrated the putative power of situational factors.

The saliency of situations was driven home by two other vivid and formative experiences. In the late 1980s the Scottish prison system was wracked by rooftop incidents; prison officers were taken hostage and threatened with death; control was lost. Directly observing four of these incidents as part of the incident management team, it became apparent that there was something rotten in the state of Scottish prisons. Other commentators pointed up the role that ill treatment by staff, limited privileges, impoverished regimes, and inadequate visit facilities played in fomenting unrest (Coyle, 1987; Scratton et al., 1991). Perhaps more startling, at least in retrospect, is the fact that six months after the Special Air Service rescued a prison officer from the roof of Peterhead prison a conference was held there in camera. It was attended by senior civil servants, outside experts, and most critically of all, 10 prisoners who had participated in the riots. This three-day meeting resulted in a dramatic volte-face in the prevailing view concerning the roots of the violence in Scottish prisons; a switch from a focus on personal pathology to a focus on situational factors (Cooke et al., 2008). With the realisation that risk factors could be located in the functioning of prisons – for example, the absence of sophisticated frontline staff, shifts in parole policy, shoddy physical conditions, overcrowding, geographical remoteness, impoverished regimes and deterrent sentences – came a paradigm shift which can be characterised as a shift from doing things to prisoners to doing things with prisoners. How could these lessons be harnessed?

The rise and rise of risk assessment

Violence risk assessment is not a new or novel phenomenon and the last two decades have witnessed dramatic developments in the techniques and technology employed. Pioneers such as Stephan Hart, Randall Kropp and Chris Webster have blazed the trail in developing evidenced-based approaches to risk assessment and management. Using what they term structured professional judgement (SPJ) approaches they have developed a paradigm that is scientifically robust yet applicable – and ethical – and which allows the formulation of an understanding both of why an individual might be violent and, critically, of how identified risk might be managed.

Various guides have been developed for the whole gamut of violent acts – sexual, spousal, child abuse, elder abuse, stalking, etc. Understandably, the focus of these guides is primarily on individual risk factors – drug abuse, mental disorder, relationship problems, employment difficulties, and so on. The SPJ procedures require the assessor to consider a number of risk factors known to be associated with violence risk. Information about risk factors is gathered through interview, document review and formal testing. If risk factors are present the assessor has to determine whether they are relevant to future violent offending; relevant either because they are in some sense causally linked to future violence, or because they may adversely affect the risk management plan. The assessor is required to formulate an account of why the individual may be at risk of violence, consider what form that violence might take and, finally, describe risk-management strategies designed to counter the risk.

While the SPJ approach continues to evolve and flourish (Hart et al., 2011), we realised that focusing primarily on individual risk factors missed half of the equation; as Simon (1990) remarked, behaviour is shaped by a scissors, one blade being the characteristics of the actor, the other blade being the characteristics of their environment. Furthermore, whilst risk assessment has predominantly been the task of forensic practitioners, there is increasing recognition that broader clinical risks – suicide, self-harm, self-neglect, and other forms of challenging behaviour – require similarly robust methods of assessment (Webster et al., 2004). But how can we understand and measure the impact of the environment?

Mapping the terrain of situational risk factors

The statement that institutional factors affect violence is rather opaque. It is clear that you cannot measure what you cannot describe; you cannot intervene effectively with that which you cannot measure. We saw a clear need to complement the SPJ guides developed to consider individuals; we could see that the SPJ paradigm could apply to institutions as well as individuals. Our overarching aim was to develop evidence-based practice guidelines that could be implemented in forensic settings and help to reduce the incidence of violence. We called these guidelines PRISM (Promoting Risk Interventions by Situational Management; Johnstone & Cooke, 2008).

We developed PRISM in four steps guided by the principles of evidence-based practice. First, we performed a systematic review of the literature on situational risk factors for violence that occurs in prisons and forensic hospitals. We were struck by the paucity of systematic consideration of these risk factors (Gadon et al., 2006). Thus, in the second step we obtained evidence from prisoners and from hospital and prison staff through semistructured interviews. This was revealing. Not only did we gather evidence about what aspects of the institution should be considered, but also we struck a rich seam of explanations about why such factors might serve to promote violent incidents; factors that engender a sense of injustice (e.g. lack of facilities to meet basic needs such as hygiene and privacy; illegitimate or misuse of authority; inconsistency and inequities in access to resources), that entailed disrespectful treatment (verbal or physical abuse from staff), that promoted a sense of uncertainty or a sense of frustration (e.g. having tribunals or hearings cancelled, having transfers without warning) or fear (e.g. being in an environment that lacked basic safety and security procedures or had a high tolerance for violence), or conditions that could be regarded as deprivation (e.g. lack of sensory stimulation). All these processes serve to promote violence risk in some individuals.

On the third step we took the evidence derived from both research and practice and used rational criteria to formulate the PRISM guidelines. Finally, we field-tested the approach in a pilot study of five Scottish prisons (Johnstone & Cooke, 2010).

We identified 22 risk factors and decided that they could be usefully grouped into five distinct domains; History of Institutional Violence; Physical and Security Factors; Organisational Factors; Staff Features; and Case Management.

The History of Institutional Violence domain focuses on the nature, frequency and pattern of violence in the last two years; this allows us to set a benchmark against which to compare this institution with itself over time, but also, with other comparable institutions. Evidence from this domain provides invaluable information for the scenario planning process used to project what form future violence might take in that institution.

The Physical and Security Factors domain focuses on both the quality of the built environment (e.g. structural quality, cleanliness, noise, temperature, space) and security: Is the built environment fit for purpose and does it conform to health and safety and human rights concerns? What is the quality of supervision and control imposed; does
it match the level of risk imposed or not? Either too rigorous or too lax security can serve to promote violence, the first through a need to save face, the second because of the anxiety generated.

The Organisational Factors domain is a broad domain as it is concerned with the strengths and weaknesses of the institution being considered, but also it is concerned with the organisation in which the institution is embedded e.g. health authority or prison system. The focus is on the management of violence: Is there someone in charge of relevant policies, practices and procedures? Is the management focused on zero-tolerance of violent behaviour? Is conflict and change managed systematically and effectively?

The Staff Features domain is perhaps the most salient domain when it comes to violence management, it focuses on the strengths and weaknesses of the staff compliment: Are appropriate staff recruited and retained? Are appropriate numbers available? Is the skills/experience mix correct? Do staff receive appropriate training for the management of potentially violent individuals? Do staff engage and communicate appropriately with those in their care? Do they receive required levels of support to do their jobs?

The final domain, the Case Management domain, is focused on the services available to potentially violent inmates: Does the institution have a systematic approach to the evaluation of individual risks and needs? Are appropriate intervention programmes available? And more broadly, to what extent do inmates have access to positive experiences including education, recreation, contact with family?

PRISM in practice

Dreaming up an evaluation process is one thing, getting it into practice is another challenge. We designed PRISM to be action-orientated and collaborative; we strove to avoid the evaluation being viewed as 'an inspection'. The first stage in the process is to recruit and train a multidisciplinary team to acquire the relevant information, to evaluate that information, to assess its relevance for future violence, to speculate systematically about what might happen in the institution in the future and, most critically, to propose and implement changes designed to obviate perceived risks. This approach increases 'buy-in' across the institution, perhaps the most rewarding comments are of the form 'you haven't told us anything we don't know about the institution but you have allowed us to think about it systematically, and do something pro-active'. Buy-in increases the likelihood of change. It is critical that risk interventions should be both realistic and achievable. It is essential that proposed changes fit with the capacities and capabilities of the institution and are implemented within an appropriate time frame. Some changes can be implemented immediately within current resources; others require long-term planning and the acquisition of suitable resources. But does it work? Case studies can be informative.

The PRISM process has now been applied in secure hospitals and prisons in Scotland, England, New Zealand, Norway and Barbados (e.g. Cooke & Wozniak, 2010; Cregg & Payne, 2010; Johnstone & Cooke, 2010; Wilson & Tamatea, 2010). Two brief illustrations might demonstrate PRISM's utility.

Hell in paradise – prisons in transitions

In March 2005 a fight broke out amongst a small group of prisoners in Her Majesty's Prison Glendairy, Barbados. This prison was a remnant of empire, a large Victorian prison built in 1855 that by 2005 was in a dilapidated state. For three days this incident escalated, the authorities lost control, prisoners rioted, they set fires and engaged in so much destructive behaviour that the prison was no longer habitable. The only prison on the island was lost; a major crisis ensued. Glendairy had contained almost a thousand prisoners, adults and young offenders, males and females, those who had been convicted and those still on remand; a group of adult males were on death row. They all had to be housed.  The authorities acted swiftly; all metal fabrication on the island ceased and temporary accommodation of 'cages' made from reinforcing rods were created in a former naval base at Harrison's Point. Containment was the key priority.

We were invited by the government and prison service in Barbados to examine their system; we applied the PRISM process to provide a critical incident review of the riot at Glendairy and to provide an analysis of the problems inherent in the regime at Harrison's Point (Cooke & Wozniak, 2010), but more importantly, our aim was forward-focused. A new prison was being constructed and our primary concern was not about attributing blame for past problems, but rather, our task was ensuring that the dysfunctional penal culture that had evolved in Glendairy – and which had been hardened in the fire and its aftermath – would not be transported to the new prison.

We carried out staff and prison surveys, interviews and focus groups and implemented the PRISM process with a team from the Barbados prison service. The team made a large number of recommendations regarding more inclusive approaches to prison management, a programme for improving staff skill, staff morale and leadership, procedures for enhancing staff–prisoner relationships, methods for evaluating risk and needs of individual prisoners and the provision of offender-behaviour programmes. These recommendations, while drawing on best practice from an international perspective, were tailored to fit both the resources and cultures of Barbados: recommendations have to be implementable to be effective (Cooke & Wozniak, 2010).

Recommendations included the implementation of a coherent and comprehensive communication strategy designed to enhance communication within the staff group and facilitate decision making as far down the hierarchy as possible. Staff training was improved with a greater emphasis being given, for example, to effective interpersonal skills and human rights training. Frontline staff felt alienated and disempowered, leadership training designed to reduce the overly hierarchical and constrained management style was recommended. The implementation of appropriate information systems was recommended to counter the identified problem that little systematic information was held about prisoners and that this lack led to capricious decision making. The implementation of offending-behaviour programmes for selected prisoners was suggested. It is heartening to note that recently a group of international prison experts, led by Dr Frank Porporino, have engaged with the Barbados prison service to deliver these, and other, changes.

Eschewing the blame culture

The correctional service in New Zealand has the substantial challenge of managing members of organised criminal groups including members of Black Power and the Mongrel Mob. Nick Wilson and Armon Tamatea carried out a sophisticated study of three maximum-security units in Auckland prison. These units had witnessed serious, high-profile assaults. The problems of these units had a system-wide impact: prisoners in other prisons would be instructed by their gangs to perpetrate assaults in order to be transferred to the maximum-security units and thus maintain their gang's position of power in those units. In common with many penal systems under stress the approach was to lock down – to impose extreme levels of control. This type of approach may have short-term benefits but the long-term consequences are invariably negative.

Two observations were particularly heartening. First, it was clear that the process of carrying out the PRISM was an intervention in and of itself: unit staff having had initial training spontaneously injected greater flexibility and variety into the quality-of-life experiences of the prisoners, they improved staff mix and implemented an active management approach to challenging prisoners. PRISM provided evidence to support the development of training in de-escalation techniques so that staff gained the skills – and confidence – to deal with problems without resorting to force. Second, Wilson and Tamatea observed that staff bought into the PRISM process; they saw it as relevant to their needs, but perhaps more importantly, they regard the application of scenario planning as emblematic of a shift from a blame culture focused on past problems to a focus on proactive interventions.

PRISM is being applied in other settings and for a variety of purposes. Johnstone and Cooke (2010) carried out a multiple-case study comparing the violence risk of five Scottish prisons and Cregg and Payne (2010) used PRISM to evaluate the functioning of an institution for young people. The functioning of three prisons and two secure units in Norway is ongoing. As a whole, this work suggests that a situational approach to the management of violence in closed institutions is viable. The approach has a number of advantages. First, frequently there are cost benefits in intervening at the level of the institution rather than focusing all resources at the level of the individual; current assets being reorganised to provide more effective interventions (Wortley, 2002). Second, almost by definition many of the troublesome patients or prisoners are the least likely to engage in individualised interventions; however, they are not immune to situational interventions. Third, in our experience, situational interventions frequently generate positive changes in the quality
of life of the institution.

Looking towards the future

We have come a long way from the blunt observation that an individual's violent behaviour is shaped by not only who they are, but critically, where they are. Over the last 20 years SPJ guides have comprehensively mapped out the terrain of individual risk factors; PRISM provides a first step towards mapping the terra incognita of situational risk factors. Yet there is still lots to do.

There are a number of key applied questions and a number of theoretical questions. For example, will the situational approach work with other risks and with other populations? Violence to others is not the only risk posed by those in secure settings; does this approach assist in the understanding and management of other risks: self-harm, suicide, self-neglect, victimisation or absconding? We, and others, have applied the PRISM process in low-, medium- and high-security prisons and forensic hospitals, but would the broad approach assist in other settings where violence is prevalent: challenging behaviour units, residential childcare environments, accident and emergency units, care homes for older adults, wards for dementia sufferers? These are key applied questions.

While the development of PRISM goes some way toward answering the 'what?' question – What risk factors operate to increase the likelihood of violence? A more interesting question is 'Why?'. Our examination of the experiences in Barbados suggested multifarious processes ranging from disrespectful treatment, fear and uncertainty, through loss of agency, to frustration and deprivation play a role. An enhanced understanding of the mechanisms of risk should assist in the formulation of more targeted and effective situational interventions.

The cost of violence at the individual and societal level can be substantial: violence remains a major obstacle to the safe and effective running of any institution. Perhaps we should close by considering the views of the most famous prisoner of the last century – Nelson Mandela – they should give us hope. He remarked, 'Many who live with violence day in and day out assume that it is an intrinsic part of the human condition. But this is not so. Violence can be prevented.' (Mandela, 2002).

David J. Cooke is at Glasgow Caledonian University and the University of Bergen
[email protected]

Lorraine Johnstone is at NHS Greater Glasgow & Clyde & Glasgow Caledonian University
[email protected]

References

Boyle, J. (1977). A sense of freedom. London: Handbooks.
Cooke, D.J. (1989). Containing violent prisoners: An analysis of the Barlinnie Special Unit. British Journal of Criminology, 29, 129–143.
Cooke, D.J. & Wozniak, E. (2010). PRISM applied to a critical incident review. International Journal of Forensic Mental Health Services, 9(3), 159–172.
Cooke, D.J., Wozniak, E. & Johnstone, L. (2008). Casting light on prison violence. Criminal Justice and Behavior, 35(8), 1065–1078.
Coyle, A.G. (1987). The Scottish experience of small units. In A.E. Bottoms & R. Light (Eds.) Problems of long-term imprisonment. Aldershot: Gower.
Cregg, M. & Payne, E. (2010). PRISM with incarcerated young people. International Journal of Forensic Mental Health Services, 9, 173–179.
Gadon, L., Johnstone, L. & Cooke, D.J. (2006). Situational variables and institutional violence. Clinical Psychology Review, 26(5), 513–534.
Hart, S.D., Sturmey, P., Logan, C. & McMurran, M. (2011). Forensic case formulation. International Journal of Forensic Mental Health Services, 10, 118–126.
Johnstone, L. & Cooke, D.J. (2008). PRISM: Promoting Risk Intervention by Situational Management: Structured professional guidelines for assessing situational risk factors for violence in institutions. Glasgow: Northern Networking.
Johnstone, L. & Cooke, D.J. (2010). PRISM: A promising paradigm for assessing and managing institutional violence. International Journal of Forensic Mental Health Services, 9(3), 180–191.
Mischel, W. (1968). Personality and assessment. New York: Wiley.
Mandela, N. (2002). Foreword. In E.G. Krug, J.A. Dahlberg, J.A. Mercy et al. (Eds.) World report on violence and health. Geneva: WHO.
Scratton, P., Sim, J. & Skidmore, P. (1991). Prisons under protest. Milton Keynes: Open University Press.
Simon, H.A. (1990). Invariants of human behaviour. Annual Review of Psychology, 41, 1–19.
Webster, C.D., Martin, M., Brink, J. et al. (2004) Short-term assessment of risk and treatability (START). Hamilton: St Joseph's Healthcare.
Wilson, N.J. & Tamatea, A. (2010). Beyond punishment. International Journal of Forensic Mental Health Services, 9, 192–204.
Wortley, R. (2002). Situational prison control. Cambridge: Cambridge University Press.