Making sense of mental health service changes

Specific interventions for people with mental disorders attract detailed evaluation, in a recognisable pattern – first, asking whether the intervention is better than no treatment, then how much intervention, for how long, and finally questions about best fit between individual and intervention. By contrast, major innovations of service or law reform are rarely examined so stringently, but tend to follow from concepts of human rights or preferences, and consensus on the general good. They are commonly introduced across the board rather than through a more conservative, experimental approach. We suggest that every change should be evidence based and trialled adequately, alongside rigorous, contemporaneous process evaluation.

Transition from a mainly inpatient treatment model to a wider community based one is a case in point, still creating concerns. This was mostly driven by perceptions that long periods in large hospitals created or contributed to impairments [1–3], while being ‘in the community’ could prevent this. It was soon clear that it did not necessarily do so [4], but other worries seem to dominate the literature. One which keeps recurring is that reducing psychiatric hospital bed numbers is coterminous with a rising prison population, and some highlight a range of adverse parallels [5]. Juriloo and colleagues [6] are the latest to claim such ‘trans-institutionalisation’, but this bears more examination. Penrose [7], to whom the original observation is attributed, took numbers of patients recorded as residents in mental illness and handicap institutions in several European countries as an index of mental health service provision and numbers of sentenced prisoners as an indicator of crime rates. He then correlated these two crude variables, without allowing for moderators or mediators.

Evidence that reducing hospital bed numbers need not trouble the criminal justice system comes from sound community-based studies. Wallace et al. [8], for example, compared violence rates between people with schizophrenia and the general population before and after major reduction in psychiatric hospital bed numbers in Australia. Small increases in the proportion of people with schizophrenia with a record of violence to others were about the same as small increases in the general population over the same period.

Steadman et al. [9] conducted a sophisticated study of men with a history of mental hospitalisation entering prisons in six USA states in 1968, when there were 64,400 state hospital beds, with the number in 1978 when there were just 24,731. First, they observed that, although the number of beds had fallen, admission numbers were less affected, so hospital beds were being used differently rather than not at all. Secondly, they found variation in the number of prisoners with a history of psychiatric hospitalisation but, in the state with the largest influx of ex-mental hospital patients into the prison, such people accounted for only 16.5% of the increase in the prison population.

Similar cohort studies have followed. Winkler et al. [10], in a systematic literature review investigating outcomes for patients discharged after at least a year in hospital for mental disorder (except dementia or intellectual disability), found 171 studies worldwide (although not the Steadman study). Eighteen of them included imprisonment as an explicit outcome; 11 reported that no patient had been imprisoned. The seven other studies, with cohort sizes between 73 and 737, found only 0.03–4.1% of discharged patients were subsequently imprisoned. Some argue that as austerity has hit many Western countries this situation has changed. If so, it is more likely to be due to community service failures than bed loss, as most of the latter long preceded any fiscal crisis.

Worldwide, we have a problem, with large numbers of people with mental disorders in prison, but repeated epidemiological studies show that their proportions there remain remarkably constant [11,12]. The real concern is that whatever we do seems to have little impact. There is an urgent need for research that shows how to be more effective in reaching and treating these people, and for government agencies to be more responsive to such evidence.

Testable hypotheses should focus on appropriate community services – and not just mental health services. Many prisoners with mental disorders are low level offenders who fall through all service nets, their period in prison too brief to allow adequate treatment arrangements, but long enough to disrupt existing fragile social and health care – whether under pre-trial detention [13] or post-sentencing [14]; many are homeless, so attending mental health appointments is consequently low on their list of priorities. There are specialist forensic mental health services in most European countries; these are expensive and generally kept for the most serious offenders. Staff in other specialties, though – general adult, learning disability, child and adolescent or, increasingly, older age services – tend to feel ill equipped to deal with the criminal justice system and offender risk management as well as mental disorder. In the United Kingdom, use of court-ordered community treatment is falling [15], yet evidence of the effectiveness of such frameworks for offenders with mental disorder is good [16]. If courts are not given this option, even low-level offenders with mental disorder will be sent to prison.

Perhaps a few people are being denied full specialist forensic mental health examination [6], but inferring this from falling numbers of such examinations is not valid. In Finland, for example, this is much more likely to follow from fewer seriously violent offenses than from service reduction. Similarly, offenders – or people on a pathway to offending – may not be able to access a hospital bed when needed, but we must not infer that only from falling bed numbers. Community service provision should be reviewed. In countries with a specialist forensic mental health system, should that be extended to offenders with mental disorder but low risk of seriously harming others, or should general psychiatrists be helped to take a more active role?

Prisons are not good places for people with mental disorders, but before we rush to endorse concepts of translocation, it is worth remembering Goffman’s [17] caution:

If all the mental hospitals in a given region were closed down today, tomorrow, relatives, police and judges would raise a clamor for new ones, and these true clients of the mental hospital would demand an institution to satisfy their needs.

We are not convinced by Juriloo and colleagues’ conclusion about ‘transinstitutionalisation’, but we do applaud their call for more research in this area. We need accurate evidence of real need in the systems and solutions that are evidence based, then scientifically tested.

Disclosure statement

No potential conflict of interest was reported by the authors.