37 Suicide in prisons and remand centres: Screening and prevention

Early identification of prisoners with suicide risk, and the prevention of suicide, are of great importance given the high suicide rates in correctional facilities. General suicide screening instruments are not very useful in prisons and remand centres. A new screening instrument has been developed and tested consisting of eight questions. The cut-off score, retrospectively, allowed the identification of 82 per cent of all Dutch prisoner suicides. By screening new inmates, this set of questions classified 18 per cent of all prisoners in a high-risk group. All of these inmates require immediate and further assessment by a medical or mental health care officer. However, even without screening, these prisoners will visit medical or mental health services sometime later during their incarceration, because of their level of psychopathology. Regular follow-up interviews are necessary to assess the suicide risk. Suicidal prisoners, most of whom have severe psychiatric disorders, should receive standard treatment in the same way as ordinary people would receive treatment beyond prison walls.

Suicide rates in correctional facilities are several times higher than in the larger community (Backett 1987; Hayes 1989; Dooley 1990; Kerkhof and Bernasco 1990; Liebling 1992; Davis and Muscat 1993; Blaauw and Kerkhof 1999; Shaw et al. 2004; Blaauw et al. 2005; Fazel et al. 2005). Many people are affected by prison suicides: prison officers, the psycho-medical staff, fellow prisoners, relatives and partners. Prison officers can develop feelings of guilt for not having interpreted warning signs correctly. A suicide confronts fellow prisoners with a model of applying a drastic change to their lives. For relatives and partners, coming to terms with a suicide is difficult in itself, but a suicide in a penal institution may be painful as it often leaves even more questions unanswered. A suicide can cause disturbance in the institution and be the cause of negative publicity, because it seems to indicate that the authorities failed in their responsibilities regarding the safety of their prisoners. In view of all this, both the early identification of prisoners having a high suicide risk and the prevention of suicide are of great importance.

General suicide screening instruments are unlikely to be very useful in a population that is characterized by suicide vulnerability, in a situation, in which even mentally strong individuals are being tested to the limits of their coping resources.

The high prevalence of suicidal behaviours in prisons is the result of a complex interaction between the highly demanding prison environment and the vulnerability of large numbers of prisoners (Liebling 1992; Blaauw et al. 2005). Many prisoners enter prison with alcohol or drug dependence (Walker 1983; Gibbs 1987), poor coping skills (Liebling 1992; Toch 1992), histories of suicidal behaviour (Anno 1985; Hatty and Walker 1986; Marcus and Alcabes 1993) and current mental disorders (for a meta-analysis see Fazel and Danesh 2002), such as schizophrenia (4–6 per cent) and major depression (10–12 per cent). In addition, many prisoners struggle with severe personality disorders. A meta-analysis on almost 23,000 prisoners revealed that of men, 65 per cent had a personality disorder, including 47 per cent with antisocial personality disorder (Fazel and Danesh 2002). Of women, 42 per cent had a personality disorder, including 21 per cent with antisocial personality disorder. Furthermore, there is no debate that imprisonment is stressful (Sykes 1966; Toch 1992). Prisoners are deprived of their liberty, decisional authority, contacts with their loved ones, work activities, etc. They are also forced to live in a situation in which they are more likely to be confronted with individuals with addiction problems, mental disorders and/or with increased needs than individuals in the community. Therefore, it is not surprising that the vast majority of prisoners suffer from negative mood states (Gibbs 1987; Ostfeld et al. 1987; Zamble and Porporino 1988) such as hostility, depression, hopelessness and anxiety. Suicidal gestures, therefore, occur at high rates in prisons and remand centres. Unfortunately, due to their long-term vulnerability, prisoners continue to have high suicide figures even after their release from prisons (156 per 100,000 person-years, Pratt et al. 2006), indicating that we are dealing with persons who have been, are, and will be suicidal: before, during and after their stay in prisons.

Studies on differences between prison suicide victims and general prisoners have yielded equivocal findings with regard to over-representations of males, whites, unemployed, old, and single or unmarried inmates among suicide victims (Blaauw et al. 2005). There is an abundance of studies that found prior incarcerations and current charges of violent crimes to be more common among suicide victims than among general inmates. Furthermore, histories of suicide attempts, psychiatric illnesses, drug abuse and alcohol abuse may be useful for the assessment of vulnerability for suicide in penal institutions. These findings suggest that demographic characteristics may be less useful for screening on suicide risk than criminal characteristics, and that psychiatric characteristics are the most useful characteristics to assess suicide risk.

An Austrian study (Fruehwald et al. 2004) showed that the most important predictors of suicide among pre-trial prisoners were psychopharmacological treatment while incarcerated, known previous suicide attempt, single-cell accommodation and a high level of violence in the last offence. In a sample of sentenced prisoners, the most important predictors of suicide were suicide threat, psychiatric diagnosis, single-cell accommodation, a high level of violence in the last offence and psychopharmacological treatment while incarcerated. Thus, both psychiatric and criminal characteristics were found to be indicative of suicide risk.

A Dutch study (Blaauw et al. 2005) examined combinations of demographic, psychiatric and criminal characteristics for their capability of distinguishing suicide victims from general inmates in the Netherlands. With regard to demographic characteristics, this study found that the suicide victims and a comparison group had different distributions of age, marital status and living situation. Suicide victims were more often over 40 years, separated, divorced or widowed and of no fixed abode, and less often living in a home with other people. The two samples did not differ with regard to gender, country of birth, race and employment. With regard to criminal characteristics, suicide victims more often had a history of only one prior incarceration in a jail or prison, and they were more often charged with (or convicted for) violent offences and less often with alcohol or drug offences or other offences. With regard to psychiatric characteristics, suicide victims more often had histories of hard drug abuse, multiple substance abuse, psychiatric care and suicide attempts. The two samples did not have different histories of alcohol abuse or soft drug abuse. Of 70 suicide victims (25 missing cases), 73 per cent had received a psychiatric diagnosis (excluding alcohol or drug dependence) during their imprisonment, most often a psychotic disorder (44 per cent), personality disorder (43 per cent) or affective disorder (21 per cent). Among the respondents of the comparison group, 12 per cent had received a psychiatric diagnosis.

Regression analysis showed that a good prediction of suicide vulnerability was formed by a model that successively included history of psychiatric care, aged over 40 years, violent offence, homelessness, one prior incarceration and history of hard drug abuse. With each indicator assigned the beta weight and with a specificity of 82 per cent, the model correctly classified 82 per cent of the Dutch suicides. Thus, a combination of two demographic characteristics, two criminal characteristics and two indicators of psychiatric problems proved capable of identifying 82 per cent of the suicide victims in the Netherlands at a specificity of 0.82 in the general inmate population.

The Dutch screening instrument (see Figure 37.1) was constructed by making use of the results of the regression techniques. Each characteristic in the screening instrument was assigned the logistic regression beta weight, multiplied by 100. Hereby, the indicator ‘violent offence’ was replaced by the indicator ‘a history of suicide attempts or self-destructive behaviours’ because this latter indicator proved to be a more powerful indicator in subsamples of the Dutch dataset.

The resulting screening instrument consists of eight questions, each with its corresponding weight, and with six questions directly resulting from the regression analyses. Weighing 27 points, a history of (intramural or outpatient) mental health care has the highest predictive value for suicide risk. It concerns the question whether the prisoner ever received treatment for psychiatric problems (addictions excluded) at the psychiatric department of a general hospital, in a psychiatric hospital, any institute for outpatient mental health, or from an independent psychologist or psychiatrist. The lack of a fixed residence in the months prior to confinement (of which the week immediately preceding confinement is the most decisive), weighing 23 points, ranks as the next strongest indicator of increased suicide risk. Thirdly, belonging to the age group of 40 years and over is a strong indicator of increased suicide risk: this characteristic weighs 17 points. As the fourth characteristic, weighing 14 points, a past with one previous period of confinement distinguishes clearly between the suicides and the non-suicidal prisoners. A multiple addiction to hard drugs and a history of suicide attempts or self-destructive behaviours appear to be equally important in making a clear distinction; both characteristics weigh 13 points. Addiction, as a characteristic, applies when the prisoner used hard drugs (at least once a week) in combination with soft drugs, large amounts of alcohol or non-therapeutic quantities of medication. Examples of previous suicide attempts or self-destructive behaviours are: taking an overdose of drugs or medication, cutting one's wrists, trying to hang oneself, and trying to come to grief in other ways.

In regards to two important characteristics, no statistically determined weight could be assigned. It may be expected, certainly on the basis of the Dutch and Austrian studies, that psychiatric disorders, particularly psychotic disorders, are of a highly predictive value with regards to risk of suicide. Therefore, it was decided to give attention in the instrument to prisoners, who in the previous five years had contact with a psychologist or a psychiatrist who made such a diagnosis—this can be an Axis-1 diagnosis according to the Diagnostic and Statistic Manual of Mental Disorders (DSM) as well as a diagnosis according to the International Classification of Diseases (ICD). In addition, as mentioned before, recent suicidal expressions or self-destructive behaviours have predictive value. In this regard, attention was given to the prisoner who expressed himself in this manner or showed such behaviour during his stay at the police station, at the courthouse, or during his transportation to the prison.

A German study (Dahle et al. 2005) underlined the value of the instrument for the identification of suicide victims in German penal systems. The study consisted of all adult male pre-trial detainees who committed suicide in Berlin between 1991 and 2000 (suicide group N = 30) and inmates who were booked with the following booking number into the same jail (comparison group N = 30). Based on the 24-point cut-off value, 25 of 30 suicidal inmates (83 per cent) would have been correctly classified as high-risk persons, at the expense of 7 false alarms in the comparison group (23 per cent). A cut-off score of 40 would have led to 7 per cent false positive classifications and 68 per cent proper identifications in the suicide group.

At a demarcation value of 24 points, around 18 per cent of all prisoners are placed in high risk of the suicide group. Although the majority of these prisoners will not commit suicide, most of them suffer from serious mental and emotional problems. When the separate predictors of suicide risk are considered, it becomes obvious that each of these characteristics points at special circumstances. During their confinement, prisoners with a history of mental health care will often suffer from the problems that required earlier treatment. Prisoners with a psychotic or other psychiatric disorder require extra care, whereas those with no fixed address, in many cases, experience social or psychological problems. Relatively older prisoners in the institution may feel isolated compared to other, usually much younger, fellow prisoners. Prisoners who were imprisoned once before often do not feel comfortable during their confinement, and prisoners with a history of multiple hard drug abuse still tend to struggle with their addiction problems and withdrawal symptoms. Prisoners with a history of suicide attempts, and those who recently expressed themselves as such or showed self-destructive behaviours, are also a cause of concern. In short, the majority of those prisoners who meet one or more characteristics from the screening instrument probably belong to the group that should usually be referred for a further interview.

It is advisable to have the screening for suicide risk performed by a prison nurse, because the questions of the screening instrument are likely to have a major overlap with the questions that are usually asked during admission interviews in order to assess mental or addiction problems. In addition, it is recommended that the screening instrument be applied immediately on arrival, because many suicides occur during the first hours, nights and weeks of confinement (Crighton and Towl 1997; Frottier et al. 2002; Shaw et al. 2004).

A prisoner scoring 24 points or more on this instrument, and as such belonging to the high-risk group, should be referred immediately to a psychologist, psychiatrist or in psychiatry trained nurse for a further diagnostic interview. Almost all prisoners in the high-risk group have mental problems, many have attempted suicide previously, but it is unknown which prisoners from this group will attempt suicide in the future. In order to prevent suicides, this entire group should immediately receive extra attention, to begin with a diagnostic interview to be carried out by trained staff. In the diagnostic interview, insight must be acquired into the prisoner's suicide risk. It is important to explore the thoughts, intentions and preparations in connection with possible previous suicide attempts and possible intended future suicide attempts. The degree of suicide ideation may differ in the course of time and can be influenced by certain events. For instance, some prisoners have a higher than normal chance of becoming the victim of bullying, which is related to heightened suicide risk in prisoners (Blaauw et al. 2001b; Ireland 2002; Blaauw et al. 2002). Therefore, regular follow-up interviews are necessary to assess the suicide risk at later moments in time. It is important to record all findings on a registration form. This form should be included in the prisoner's medical file. It is of the utmost importance that the findings of the suicide risk assessment are shared with the prison officers who are responsible for the daily care of these prisoners as these officers usually do not have access to the medical files of prisoners. Inclusion in the medical file promotes the transfer of information between the various members of the medical and mental health staff, and will contribute to the continuity of care in case of relocation to other penal institutions. However, after relocation it is necessary to screen prisoners again (Blaauw et al. 1997, 2001a).

Early identification of suicidal prisoners is an important first step to reduce the number of suicides in penal institutions. Therefore, it is promising that the Dutch research project resulted in a screening instrument that has proved to be easy and requires little time and effort. It is also an instrument that is unobtrusive as it makes use of a few historical characteristics that can be easily obtained, possibly even from files. As such, working with the screening instrument will most likely not conflict with other strategies to prevent suicide. It is even more promising that this instrument was found to be accurate in the prediction (actually post-diction) of suicides in prison systems in The Netherlands, England and Wales, the United States and Germany. Suicides in different countries share more similarities than differences (Blaauw et al. 2005).

As said before, the prevention of suicide should start as soon as possible, immediately following the reception of inmates in the facility. Attention should be paid immediately to the high vulnerability levels of many prisoners. Since many prisoners have psychiatric problems, the immediate routine provision of psychiatric care, including medication and, if necessary, psychotherapy should be self-evident. However, many European prison systems have only very limited mental health care provisions (Themeli et al. 1999). Anxiety, depression, psychosis, suicidal feelings—all these common mental disorders should receive standard treatment in the same way as ordinary people would receive outside the prison walls. When a prisoner breaks his leg, he will be immediately transferred to a hospital. Likewise, when a prisoner develops psychotic symptoms, he should be transferred to a psychiatric ward within or outside the prison. There is no excuse for withholding adequate mental health care. Since many addicted inmates enter the system, detoxification units and sustained medication (e.g. opiate-antagonists) should be available. After detoxification, psychological treatment should be available to help prisoners remain abstinent. The principles of Community Reinforcement Approach would be appropriate here, adapted to the prison system (Roozen 2005). Since many prisoners have problems in self-regulation or lack capacities for stress management, simple stress reduction techniques should become available. Prisoners have high levels of pathological worrying (Kerkhof and Van 't Veer 2004). Worrying may be one of the prodromes for suicidal behaviours. Simple exercises fighting excessive worrying are available and could be handed over upon reception in the form of leaflets. In such leaflets, worry regulation techniques can be proposed such as to postpone worry to specific worry episodes (30 minutes in the morning, 30 minutes in the evening), to spend time thinking about positive memories, about skills and characteristics one is proud of, to think about possible positive future prospects, and to challenge over-generalized negative self-statements.

Because suicidal vulnerability waxes and wanes over time, prison staff should assess suicidal ideation regularly. Prison staff, therefore, should be trained accordingly.

For the many prisoners with borderline personalities, who regularly engage in suicidal threats, gestures and acts, dialectical behaviour therapy (Linehan et al. 2006) is available and is already applied in some prison settings. Adequate treatment of psychiatric disorders during incarceration may as well prevent post-release suicides.

All of these recommendations are self-evident, and they are shared by all international experts in the field (Konrad et al. 2007; Daigle et al. 2007). However, there is not yet much empirical evidence that these recommendations are effective, because prevention programmes are extremely difficult to evaluate due to methodological difficulties. Only very few methodologically sound studies have been performed. The status, therefore, is a widely shared consensus that applying these recommendations would lower the suicide rates in correctional facilities. Directors of prison systems tend to point to this lack of empirical evidence and use this to justify doing nothing to prevent suicides.

Correctional officers and prison staff have to be educated and trained in suicide assessment and management. They have to know the motivation behind suicidal behaviour. In Australia, a specific suicide awareness training for all custodial staff has been implemented, including work supervision by professionals (Eyland et al. 1997). The State of New York developed a comprehensive training programme within its upstate local remand facilities, including a training manual, a videotape, an officer handbook, and pre- and post evaluations leading to certification (Cox et al. 1988, 1989; Cox and Morschauser 1997). Despite a nearly 100 per cent increase in the jail population, there has been more than a 150 per cent decrease in jail suicides since programme implementation. In the Netherlands, suicide prevention courses for correctional officers were set up by the training institute of the Dutch prison system, almost all psychologists working in Dutch remand centres and prisons receive postgraduate courses in suicide prevention. All experiences with suicide awareness programmes and risk-assessment course point to the excellent evaluations by the correctional officers and prison staff, who were happy to receive help in dealing with behaviour they often did not understand.

Architectural design is important in suicide prevention. Safe cells provide safe conditions for the observation of inmates who require special supervision (no potential ligature points, no electrical outlets, no toxic materials, rounded corners for the walls, no exposed pipes, hooks, hinges, door knobs, etc.). However, suicide-proof cells may increase the feelings of loneliness and despair of suicidal inmates. It is also advisable to minimize the amount of time suicidal prisoners are in segregation units.

Close and constant observation at different levels of supervision should be geared to the risk of individual prisoners. Closed-circuit television is not enough, since there are numerous examples of inmates committing suicide in full view of television equipment. Sometimes buddies are trained to observe suicide risk in peers and provide peer counselling in crisis. Monitoring, however, can only be successful in combination with face-to-face contact.

The most powerful deterrent for suicide is personal contact. Communication between inmates and prison officers and mental health care workers may relieve feelings of isolation, hopelessness and crisis. Prison officers should be attentive to signs of excessive worrying, depression and despair, and should communicate with inmates about these signs. They should be able to provide emotional support. A common finding in prison suicides is a recent breakdown of communication between inmate and correctional officers. Prison officers can be attentive to the sense of hopelessness about the future, a characteristic that will only grow stronger in the absence of personal communication. Listeners scheme's including the availability of telephonic contact with Samaritans, or buddy schemes, trained by organizations such as the Samaritans, seem to help (selected) prisoners to survive their crises (Schlosar 1997). Confidential reporting of suicidal communications by fellow inmates should be encouraged.

Prison officers and psychologists and psychiatrists should have a routine communication of inmates' suicidal ideation. All prison officers in consecutive shifts should be aware of the pending danger. The oscillation of suicidal intent should be registered on a daily to weekly base.

Essential feature of suicide prevention is the involvement of family members and friends, if applicable. Suicidal inmates can be helped enormously by granting additional contact with next of kin. Several suicide prevention programmes do offer this extra contact for suicidal inmates. Relatives are the connection to the future. They embody future perspectives for the period upon release. As such, they are the first to help suicidal inmates re-orient themselves regarding the hopelessness of their future perspectives.

Suicide prevention in remand centres and prisons and other correctional facilities should start as soon as possible when inmates enter the system. Thorough and systematic screening is the cornerstone, followed by immediate and adequate mental health care. Suicide prevention is only a small part of the more general mental health care remand centres and prisons offer. In this respect, the quality of suicide prevention is a performance indicator for the more general management of the emotional needs of prisoners. There is reason to believe that, in many countries, suicide prevention in detention could be improved considerably (Daigle et al. 2007; Konrad et al. 2007), and the WHO has issued a resource booklet called Preventing suicide—a resource for prison officers (WHO 2000).