In this issue, Gavin et al (Reference Gavin, Parson and Grubin2003) present a study of the uses of a screening instrument for mental illness in remand prisoners. This is the latest in a long series of epidemiological studies concerning mental disorders in prisoners. They have produced a screening instrument for mental illness in remand prisoners that deserves to be more widely known. The same group has shown that routine screening methods miss substantial numbers of mentally-ill prisoners and on follow-up, even those identified seldom receive appropriate treatment (Reference Birmingham, Mason and GrubinBirmingham et al, 1998). This was often because of the disruption to interventions caused by the prison regime and the unforeseen actions of the courts.

Why so many prison morbidity surveys?

Fazel & Danesh (Reference Fazel and Danesh2002a) published a systematic review of 62 surveys from 12 countries, including 22 790 prisoners. They report a 6-month prevalence of psychosis in 3.7% of men and 4% of women, and major depression in 10% of men and 12% of women. The differences between sentenced and remanded groups were small, as were differences between countries and differences over time. This contrasts with an increasing prevalence of mental illness over time in some series of surveys (Reference GunnGunn, 2000). The more outlying results appear to be accounted for by differences in methodology, particularly in the time period used (Reference Fazel and DaneshFazel & Danesh, 2002b).

This is somewhat surprising. It is generally assumed that many people with mental illness enter the prison system as a direct or indirect consequence of their mental illness, thereby accounting for the large excess of psychiatric disorder in this population. If this were the case, different jurisdictions might accumulate different proportions of severely mentally-ill individuals in their prisons, because of differing approaches to forensic mental health law across jurisdictions and differences in the overall imprisonment rates. The actual rate of serious violence due to mental illness does not appear to change much over time, even when the population rates of serious violence such as homicide increase substantially (Reference Taylor and GunnTaylor & Gunn, 1999). There is also no evidence that rates of homicide by mentally-ill people vary substantially across countries, with a few exceptions (Reference CodCoid, 1983). It is also assumed that the accumulation of mentally-ill people in prison is due to the substantial change in the form of mental health service available in most countries today (Reference TorreyTorrey, 1995; Reference GunnGunn, 2000). However, there is little evidence for change over time in Fazel & Danesh's review.

There are so many published surveys of mental illness in prisons because of the obvious affront to humanitarian sensibilities and the apparent dangers to mentally-ill prisoners. Yet the most obvious danger, suicide in prison, appears to be accounted for not by the prevalence of severe mental illness, but by the prevalence of substance misuse problems - particularly opiate use and dependence in prisoners (Reference GoreGore, 1999). Studies of psychiatric morbidity in prisoners seldom examine the serious physical illnesses prevalent in prisons and comorbid with mental disorders, such as blood-borne infections (Reference Allwright, Bradley and LongAllwright et al, 2000). A recent trend in this direction can be detected, however (Reference Fazel, Hope and O'DonnellFazel et al, 2001).

Fazel & Danesh (Reference Fazel and Danesh2002a) also report systematic review figures for personality disorder (65% of men, 42% of women), including antisocial personality disorder, but here the consistency in the results across surveys breaks down. Others have recently turned their attention to the prevalence of drugs and alcohol problems in prison populations (e.g. Reference Allwright, Bradley and LongAllwright et al, 2000). Here, the reported prevalence rates vary substantially between surveys, though this might simply reflect the highly-variable methodologies used. Reliable studies of intellectual disability in prison populations are rare. Studies reporting comorbidity of mental illness and substance abuse problems are also very rare.

Is it a universal law that wherever a prison population exists, about 4% of men and women will have a 6-month prevalence of psychosis, while 10% of men and 12% of women will have major depression? Early text-books reported a unique prison psychosis probably caused by the sensory deprivation of the early reformatory/penitentiary regimes (Reference Scott, Hirsch and ShepheardScott, 1974). Although a specific prison psychosis is no longer recognised, it seems that psychosis in prisons is a very constant phenomenon. Psychiatrists who do not work with prisoner populations often assume that the patients referred on to them by forensic psychiatrists are somehow qualitatively different from those they find in the community. Gunn et al (Reference Gunn, Maden and Swinton1990) showed that the majority of these prisoner patients are already well-known to local services. Many others have repeated this observation. But perhaps we are missing something about the natural history of mental illness in prisoners that needs a different type of study.

Future directions

The studies reviewed by Faizel and Danesh (Reference Fazel and Danesh2002a) may represent a resource for further analysis and a prompt for future studies. Perhaps useful information can be gleaned about the relationship between psychiatric morbidity rates and the regimes in prisons, where spatial density, social density and the experience of crowding (Reference Baum, Paulus, Stokols and AltmanBaum & Paulus, 1987; Reference Canter, Bottoms and LightCanter, 1987) may amount to toxic factors (Reference SommerSommer, 1979; Reference Cox, Paulus, McCain, Baum and SingerCox et al, 1982).

Will the availability of enhanced screening instruments such as that used by Gavin et al (Reference Gavin, Parson and Grubin2003) lead to better health outcomes? Only if there is a real change in attitudes to mentally disordered offenders in mental health services generally. Apart from a general willingness to acknowledge the needs of prisoners with psychosis, the delivery of services to prisoners with anxiety and affective disorders, drugs and alcohol problems, brain injury, learning disability, challenging behaviour and repetitive self-harm has changed little or worsened. It could be argued that screening at reception in prison is no longer worth the effort and instead, systematic screening earlier in the pathway through the criminal justice system is a better strategy (Reference Shaw, Creed and PriceShaw et al, 1999; Reference JamesJames, 2000). Obviously, better cooperation between all stages of the process is likely to be the most effective solution, and there is some evidence for this (Reference Pierzchniak, Purchase and KennedyPierzchniak et al, 1997). Perhaps with the appearance of this screening tool to help quantify mental illness, we should now move on to other disorders in prisoners and other ways of asking what the problem is.