Secure psychiatric hospitals are estimated to account for around 19% of the overall mental health budget in England, with a spending allocation of over £1 billion (Wilson et al., 2011). This is the single largest component of the mental health budget. The cost per patient in such hospitals ranges from £152,000 per year in a low secure hospital to £273,000 per year in a high secure hospital (Durcan et al., 2011).

Despite this, the evidence for patient benefits within secure psychiatric hospitals are somewhat limited (Davies et al., 2007; Maden et al., 1999). Whilst single studies have examined the outcomes of discharged patients in several countries (Steels et al., 1998; Yoshikawa et al., 2007), there is a need to synthesise these reports, present information on all adverse outcomes and also to provide some comparative information to contextualize these findings. 

The authors of the present study (Fazel et al, 2016) therefore chose to conduct a systematic review of studies that have tracked patients after being discharged from a secure psychiatric hospital for criminal behaviour, readmission to psychiatric hospitals and mortality.

Methods

Study selection criteria

Studies were selected if the following criteria were met:

• The study was a primary study
• The study followed up patients discharged from any secure hospital (including low, medium or high security)
• The study reported on outcomes for any of the following:
  • Death
  • Suicide
  • Repeat offending (including violent behaviour, contact with police, rearrests or convictions)
  • Readmission to hospital

Studies were excluded if they were:

• A validation study for a risk assessment tool
• An evaluation of an intervention
• Unable to provide data that would allow for calculation of rates

Eleven electronic databases (PubMed, Google Scholar, PsycINFO, JSTOR, Global Health, Medline, Web of Knowledge, DART-Europe, E-thesis portal, Networked Digital Library of Thesis and Dissertations and ProQuest Dissertation and Theses) were searched from the start of the database until 13 March 2013.

Study outcomes

The authors calculated crude rates (CRs) for;

• All-cause mortality
• Suicides
• Readmissions
• Reoffending

The authors calculated this by using number of events and person-years at risk, following the methods outlined in a recent meta-analysis of released prisoners (Zlodre et al., 2012). Person-years at risk was calculated by multiplying the number of patients with the median period of patient follow up. Crude rates were then calculated by dividing the number of events by the person-years at risk. The authors then performed random-effects meta-analyses on crude rates in order to calculate pooled estimates for all-cause mortality, suicides, readmissions and re-offending.

Results

The results of 35 relevant studies were analysed including data from 12,056 patients. Of this sample, 75% were male and the average age was 34.5 years. The average length of admission was 3 years. Average follow-up ranged from 1.5 to 13.6 years for mortality, 1.8 to 9.4 years for readmissions and 1.5 to 13.6 years for reoffending.

All-cause mortality and suicide

The pooled estimate crude death rate for all-cause mortality was 1,538 per 100,000 person-years. This rate was slightly lower for studies based in England and Wales (CDR = 1,240). The pooled estimate crude death rate for suicide was 325 per 100,000 person-years.

Comparison data was available for released prisoners, mentally disordered offenders and community psychiatric patients. CDRs for all-cause mortality ranged from 155 to 561 per 100,000 person-years. CDRs for suicide ranged from 850 to 3,344 per 100,000 person-years.

The authors found that mortality rates amongst discharged forensic patients were high in both absolute and relative terms. Specifically, the CDRs for patients released from a secure psychiatric hospital were higher than the CDRs for recently released prisoners (Zlodre et al., 2012), but similar to the CDRs for patients with mental health problems such as schizophrenia-spectrum disorder (Dutta et al., 2012). The authors therefore suggest that:

it is the mental illness component of being in secure care, rather than anything specific to the forensic setting, that contributes to the increased mortality risk.

Readmissions

The pooled estimate crude readmission rate was 7,208 per 100-000 person-years. Crude readmission rates ranged from 2,926 to 16,461 readmissions per 100-000 person-years. This rate was slightly higher for those patients who were classified under the Mental Health Act and also those who had a longer admission (though the latter did not reach statistical significance). Conversely, patients who were diagnosed with psychopathic disorder were less likely to be readmitted. Comparison data was available for community patients, mentally disordered veterans and offenders with mental disorders treated in an outpatient programme. Crude readmission rates in these studies ranged from 3,838 to 55,555 per 100,000 person-years.

The authors found that rates of readmission to hospital varied markedly across the studies included in the meta-analysis. They added that they were unable to identify many comparative studies that reported on readmission rates for psychiatric patients, and therefore comparisons about readmission rates are difficult to make. However, in comparison to the range of crude readmission rates of community patients, mentally disordered veterans and offender outpatients, it appears that the range of readmission is much smaller for those who had been released from a secure psychiatric hospital.

Reoffending

The pooled estimate crude reoffending rate was 4,484 per 100,000 person-years. Comparison data was available for released prisoners, offenders with personality disorders, mentally disordered offenders and offenders with mental illness. Crude reoffending rates in these studies ranged from 4,535 to 36,964 per 100,000 person-years.

The pooled estimate crude violent reoffending rate was 3,901 per 100,000 person-years. The authors conducted a comparison of this data with reoffending rates in prisoners in the same country in a similar year and where possible of a similar age and gender. Prevalence ratios were one or above, which indicates that rates of prisoner reoffending were higher than in forensic psychiatric patients.

The authors therefore found that re-offending rates were much lower in patients who had been released from a secure psychiatric ward in comparison to released prisoners with similar demographics, and offenders with mental health disorders including personality disorders.

Conclusions

• It was suggested that mental illness, as opposed to residing in a forensic setting, contributed to higher mortality and suicide rates in those released from a secure psychiatric hospital in comparison to people in a range of other settings
• There was marginal difference in the rates of readmission to hospital from those in a secure psychiatric hospital and those in other mental health settings
• Secure psychiatric hospital patients had lower rates of reoffending than prisoners of a similar demographic make up.

Strengths and limitations

The authors of this study should be commended for including unpublished studies in their meta-analysis and avoiding the risk of publication bias.

Despite this, the authors acknowledge that there are several limitations within their study. Firstly, patients came from a range of institutions, each of which will differ in admission criteria and will offer a variety of treatments. Similarly, since the systematic review included studies published between 1982 and 2013, it is likely that treatments on offer have changed over time in this period. This therefore makes it difficult to make comparisons between varying cohorts.

The authors also note that there are issues with the quality of the studies included. The majority were retrospective and included information on risk factors from case notes and outcomes from various official databases. Therefore, the quality and breadth of patient information depends solely on the quality and accuracy of the clinical records which are kept.

Summary

To summarise, the authors are keen to highlight two main findings from this study:

1. It provides some evidence that patients discharged from forensic psychiatric services have lower rates of repeat offending than many comparative groups
2. Forensic psychiatric services could consider interventions that would reduce premature mortality in their discharged patients. For instance, follow-up care and better organised and co-ordinated services that comprehensively address the complex causes of mortality, instead of focusing on a single cause.

The authors note that due to the ever expanding nature of forensic psychiatric services, the comprehensive systematic review and meta-analysis that they have conducted should be used to assist in the development of such services in the UK and in other countries.

I would like to question the outcomes included in this study. Are lower reoffending rates alone enough to encourage the detainment of offenders in secure psychiatric hospitals? Conversely, are higher risks of mortality and suicide reason enough to deter from secure psychiatric hospitals? What about outcomes such as improved mental health or an improved relationship with drugs or alcohol?

With the increasing media focus on what happens when people are released from secure psychiatric hospitals, it feels as though there is still a wealth of research to be conducted into exactly how we can improve outcomes for patients in secure forensic settings.

Links

Primary paper

Fazel S, Fimińska Z, Cocks C, Coid J. (2016) Patient outcomes following discharge from secure psychiatric hospitals: systematic review and meta-analysis. The British Journal of Psychiatry Jan 2016, 208 (1) 17-25; DOI: 10.1192/bjp.bp.114.149997

Other references

Davies et al (2007). Long-term outcomes after discharge from medium secure care: a cause for concern. The British Journal of Psychiatry; 191 (1): 70-74.

Durcan et al (2011). Unlocking pathways to secure mental health care. Centre for Mental Health.

Dutta et al (2012). Mortality in first-contact psychosis patients in the UK: a cohort study. Psychological Medicine; 42: 1649-1661.

Maden et al (1999). Outcome of admission to a medium secure psychiatric unit. I. Short- and long-term outcome. The British Journal of Psychiatry; 175 (4): 313-316.

Steels et al (1998). Discharged from special hospital under restrictions: a comparison of the fates of psychopaths and the mentally ill. Criminal Behaviour Mental Health; 8: 39-55.

Wilson et al (2011). The medium-secure project and criminal justice mental health. Lancet; 378: 110-111.

Yoshikawa et al (2007). Violent recidivism among mentally disordered offenders in Japan. Criminal Behaviour Mental Health; 17: 137-151.

Zlodre et al (2012). All-case and external mortality in released prisoners: systematic reviews and meta-analyses: The PRISMA statement. American Journal of Public Health; 102: 67-75.

Photo credits

• Lucie Provencher CC BY 2.0
• Tax Credits CC BY 2.0
• Daniel X O’Neil CC BY 2.0
• Efrén CC BY 2.0
• Colin Zhu CC BY 2.0