A retrospective study describing the characteristics of one Mental Health Trust’s admissions under sections 47 and 48 of the Mental Health Act 1983

Abstract

Sections 47 and 48 of the Mental Health Act 1983 allow prisoners to be transferred from prison to an appropriate health-care setting in order to be treated. There is an awareness that delays exist when transferring prisoners to hospital. However, literature regarding the delay in returning these patients from hospital is limited. The admissions from prison to a Psychiatric Intensive Care Unit (PICU) in South West London were compared to non-offenders on the PICU in order to compare the average length of stay for both groups and the time taken for the discharge from PICU once felt clinically appropriate. The study also compared demographic profiles, reason for admissions, psychiatric diagnosis and index offences. Over six years, there were 18 admissions from prison to a PICU. The control group comprised 37 non-offenders admitted to the same PICU. On average the prison group took longer to be deemed clinically ready for discharge and, even once clinically ready, then took longer to be discharged. The average length of stay in PICU was 77.83 days for prisoners, and 16.46 days for non-offenders. All 55 admissions were between 1 January 2008 and 31 December 2014. The offender pathway and the difference in the length of stay between prisoners and non-offenders in a PICU warrants further exploration. Possible recommendations to reduce the length of stay of prisoners include improved information sharing between prisons and hospital, and clearer guidelines regarding the level of security required.

Keywords

Forensic psychiatry forensic sections prison

Introduction

Mental illness is common amongst the prison population. The Office for National Statistics (ONS) surveyed prison populations for mental ill health in England and Wales in 1997; >90% of prisoners were found to have at least one mental health disorder, including psychosis, neurosis, alcohol misuse, drug dependence and personality disorder.¹

The responsibility for managing health care in prisons was fully transferred from Her Majesty's Prison Service (HMPS) to the National Health Service (NHS) in April 2006. The aim in involving the NHS was to allow prisoners access to the same quality and range of health services as the general public receives in the community.²

Prisoners who suffer from mental disorders and require inpatient treatment within a secure mental health service are only allowed to be transferred to hospital under the Mental Health Act (MHA) 1983 with consent from the Secretary State for Justice.

Sentenced prisoners are transferred under section 47 (s47) of the MHA, and prisoners who are on remand or not yet sentenced are transferred under section 48 (s48). Section 49 (s49) empowers the Home Secretary to attach a restriction direction, imposing restrictions upon the discharge of the offender (prisoner). In the UK, under mental health legislation, patients are not able to receive treatment in prison without consent.³

Studies carried out by the Home Office demonstrate that between 1993 and 1999, the number of prisoners transferred from prison to hospital under s47 or s48 fluctuated between 723 and 785 per year. In 2000, the number decreased to 662 before rising in 2003 to 721. These findings show that from 2000 to 2003, there was a 9% increase in the number of transfers under s47 or s48. Of the 721 transfers under the MHA in 2003, 425 (59%) required urgent treatment, as they were untried or unsentenced, 296 (41%) were transferred after their sentence and 96 (13%) were remitted back to prison to resume their sentence after being treated.⁴

Once the team in charge of the patient’s clinical care deems that the criteria for detention under the MHA is no longer met or that the patient can no longer receive treatment, remission to prison should be achieved with the minimum delay. The responsibilities of coordinating, overseeing and managing the prison transfer to hospital and remission back to prison are shared between the mental health service provider and the respective prison.³ It is important to understand the pathway of transferring prisoners to hospital and the obstacles associated with this process in order to explore potential delays within the remission process.

Despite the Bradley Report highlighting issues surrounding delays transferring prisoner to hospitals, the delays in remitting the prisoners back to prison has not been given consideration. It is imperative that delays in transferring patients back to their respective prisons are minimised. This study focuses on the admissions within a certain catchment area from various prisons to a particular non-forensic Psychiatric Intensive Care Unit (PICU) over the course of six years.

Aims and methods

The study took place in a male-only 13-bed PICU in South West London, serving a population of approximately 1.3 million people. Patients are accepted from any prison in the UK if the prisoner has a general practitioner in the Trust catchment area.

The aims of the study were to establish the length of stay of prisoners in a PICU, the length of time taken for the remission of patients back to prison and the reasons for any delay. The prisoners were compared to a group of non-offenders in the same PICU.

Data were collected retrospectively from the electronic notes of patients who presented from 1 January 2008 to 31 December 2014. Two groups of patients admitted to Ward 1 at PICU Springfield University Hospital were studied. One group included prisoners who were transferred under either s47 or s48 of the MHA 1983, and the other group included non-offenders. The group of non-offenders was chosen as the control variable for the study.

The control-group admissions were selected by locating patient records on the electronic database, RiO, for all those admitted to the PICU from 2008 to 2014, who were not being transferred from a prison (non-offenders); 37 patients were selected based on these criteria. These criteria ensured that there were at least two non-offenders for each offender: one non-offender with a similar date of admission to the date of admission of the patient from prison, and another non-offender with a similar date of discharge in proximity to the date of discharge of the same offender. This was to ensure that the two groups could be compared more effectively by reducing the bias of confounding factors affecting the patient’s stay at the PICU. For example, there was a possibility that personnel difficulties at the ward or prison at that particular point in time may have caused delays.

The sample included all patients who were transferred from prison under either s47 or s48 of the MHA 1983 to Ward 1 of the PICU between 2008 and 2014. The time it took between prisoners being deemed ready to leave the PICU and actually leaving the PICU was compared to the respective times of other patients who were not being sent back to prison.

The data were analysed using Microsoft Excel. In order to determine whether the results were statistically significant, the chi-square test for parametric data and t-test for continuous data were used.

The St George’s Joint Research and Enterprise Office confirmed that there was no need for this study to require an ethical review. The study was a service evaluation and was conducted following ethical research principals. No patients were contacted during the study, and all data were anonymous.

Results

Demographics

During the six-year period, 55 admissions were included in the study. All patients were male. At the time of admission, the age of patients ranged from 18 to 58 years, and the mean age of the total sample was 33.24 years (SD = 11.35). Nearly a third of all patients (n = 16; 29.09%) were between 31 and 40 years of age, 15 were between 21 and 30 years of age (n = 15; 27.27%), and six were between the 51 and 60 years of age (n = 6; 10.91%).

Admissions

There were 18 transfers from prison. However, one prisoner was admitted twice during a different year within the six-year period, and therefore the relevant data were included in the study as separate admissions (17 patients, 18 admissions transferred under s47 or s48). Therefore, 18 admissions from prison (n = 18; 32.73%) were compared to 37 non-offenders (n = 37; 67.27%).

Of these 18 admissions from prison, there were six transfers under s47 with a s49 restriction (n = 6; 33.33%) and the same number of transfers under s48 with a s49 restriction (n = 6; 33.33%). With regard to transfers without such restriction, there were three under s47 as well as three under s48 (n = 3; 16.67%). In total, the implementation of the restriction order s49 by the Secretary State for Justice came to six (n = 6; 33.33%) in the sample obtained of 18 s47 and s48 transfers to the PICU.

Index offence

The offences were categorised depending on the nature of the crime committed: to property, violent crime and sexual offence. Violent crime accounted for more than half of the offences (n = 9; 52.94%), and included assault and robbery. The second most common offence was property crime (n = 5; 29.41%), for which theft was the only charge. There were two patients for whom the nature of their charges could not be ascertained using the electronic database of patient notes and thus were unknown (n = 2; 11.76%; see Table 1).

Table 1.

Prison transfers and non-offenders: index offence and primary psychiatric diagnosis.

	Sections 47 and 48 (n = 17^a)	Non-offenders (n = 37)
Index offence
Property	5	–
Violent	9	–
Sexual	1	–
Unknown	2	–
Psychiatric diagnosis
Paranoid schizophrenia	9 (52.94%)	18 (48.65%)
Catatonic schizophrenia	1 (5.88%)	0
Bipolar affective disorder	2 (11.76%)	5 (13.51%)
Borderline personality disorder	1 (5.88%)	3 (8.12%)
Schizoaffective disorder	0 (0.0%)	2 (5.41%)
Major depressive disorder	1 (5.88%)	1 (2.7%)
Substance use psychosis	2 (11.76%)	7 (18.92%)
Unknown	1 (5.88%)	1 (2.7%)

Seventeen patients were used for measuring the prevalence of psychiatric diagnosis, as one patient was previously admitted.

Psychiatric diagnosis

Using the patient notes on an electronic database, the primary psychiatric diagnosis given to the 17 patients from prison and the 37 patients from the control group of non-offenders were identified. All the diagnoses were according to ICD-10 criteria.

For both the prison transfers (n = 9; 52.94%) and non-offenders (n = 18; 48.65%), paranoid schizophrenia was the most common primary diagnosis.

In the prison group, the joint-second most common diagnoses were bipolar affective disorder (n = 2; 11.76%) and schizoaffective disorder (n = 2; 11.76%). Substance-use psychosis was the second most common diagnosis in the non-offender group (n = 7; 18.92%), with bipolar affective disorder (n = 5; 13.51%) being the third most common in this group (see Table 2).

Table 2.

The mean, median and total length of stay in days for prison transfers and non-offenders.

	Sections 47 and 48 (n = 18)	Non-offenders (n = 37)
Total	1401	609
M (SD)	77.83 (73.08)	16.46 (11.87)
Median (min–max)	55 (3–311)	16 (2–61)

Reason for admission to the PICU

The reasons the prisoners and non-offenders were admitted to the PICU were split into two categories: risk to self and violence. Violence was the most common reason for prisoners (n = 11; 64.71%) and non-offenders (n = 6; 35.30%) being admitted. Examples included assaults on other prisoners and prison staff. Risk to self was the second most common reason for prisoners (n = 6; 23.53%) and for non-offenders (n = 8; 21.62%) being admitted, mainly because prisoners refused treatment, food or drink (see Figure 1).

Figure 1.

The primary reason for the admissions of prisoners and non-offenders to Ward 1 of the PICU at Springfield University Hospital within the six-year period.

Length of stay

The length of stay for the 18 admissions ranged from 3 to 311 days. The total number of days spent under the care of the PICU team by all prisoners was 1401 for 18 offenders, and the mean length of stay for the offending group was 77.83 days (SD = 72.08). In comparison, the length of stay for the non-offending group ranged from 2 to 61 days, with the second highest length of stay in this group being 36 days. The total number of days spent within the PICU was 609 for 37 non-offenders, and the mean length of stay for the non-offending group was 16.46 days. The difference in the mean between the control group and the prison population was 61.37 days (see Table 1).

Independent samples t-tests were conducted to identify any significant differences in mean length of stay between the prison group and the non-offender group. The t-value was 5.02, and the p-value was <.00001.

If a patient was admitted from prison to the PICU under s47 or s48 of the MHA 1983, they were 2.42 times more likely to spend a longer time period in the PICU as an inpatient in comparison with patients not transferred from prison (non-offender group). This relationship was statistically significant (p < .005).

Ready for discharge

Of the 18 prison admissions to the PICU, five were remitted back to prison (n = 5; 27.78%). The time it took for these prisoners from their point of admission to the PICU to the point in which they were ready for discharge from the PICU was compared to the other prisoners and the control group.

For the whole prison group, the mean time taken for the admissions to be decided by the PICU clinical team as being ready to leave the unit was 43.8 days (SD = 27.56). For those admissions who were sent back to prison, the mean time was 56 days (SD = 94.94). Those who were not sent back to prison had a mean time of 41.89 days (SD = 26.57). Finally, for non-offenders (control), the mean length of time was 15.13 days (SD = 13.11; see Table 3).

Table 3.

The mean (SD) number of days taken from the point of admission to being ready to leave, and the mean (SD) number of days taken once ready to leave and then actually to leave.

	All prisoners (n = 18)	Prisoners sent back to prison (n = 5)	Prisoners not discharged to prison (n = 9)*	Non-offenders
From PICU admission to being ready for discharge	43.8 (27.56)	56 (94.94)	41.89 (26.57)	15.13 (13.11)
From being ready for discharge to actual discharge	32.86 (67.46)	14.6 (25.03)	43 (83.87)	1.86 (3.34)

Actual discharge from PICU

The time it took for prisoners and non-offenders to be discharged from the PICU once deemed ready for discharge by the responsible clinical team was measured.

For the whole prison group, the mean time taken was 32.86 days (SD = 67.46). For those admissions who were sent back to prison, the mean time was 14.6 days (SD = 25.03). Those who were not sent back to prison had a mean time of 43 days (SD = 83.87). Finally, for non-offenders (control), the mean length of time was 1.86 days (SD = 3.34; see Table 3).

Discussion and summary

Principal findings

The number of admissions to Ward 1 of the PICU at Springfield University Hospital under s47 or s48 was 18 in the six-year period. Of these 18 admissions, there were an equal number of s47 (n = 9; 50%) and s48 transfers (n = 9; 50%). Statistics from the Home Office report stated that in the year 2003, of 721 prison admissions to a mental health-care setting in the UK, 59% were transferred under s48 and 41% were transferred under s47. The current study found that the main reason patients from prison were admitted to the PICU was because of some form of violence (64.71%). Aggression and violence are very common within psychiatric settings, especially within PICUs. However, if patients with psychotic symptoms are imprisoned, the use of seclusion and the lack of organised activities within the prison can exacerbate their aggression, resulting in unfortunate incidents such as a member of staff or a patient being injuring. Risk to self is also an extremely important issue, as the rate of suicide in prisoners is 15 times higher than in the general population, although the risk was often self-neglect rather than suicidal behaviour in this group.⁵

Non-compliance with treatment is a major problem in prisons and is often the reason for admission to the PICU. There have been suggestions that the MHA 1983 should be revised in order for it to extend to treatment within prisons, but this did not happen in the revision of the MHA in 2007. The assumption is that if patients refuse medication in prison, they requite inpatient care. However, the shortage of available beds has not been addressed in the legislation. In addition, non-compliance with treatment was seen to occur throughout many patient journeys included in this study. This perpetuates a vicious cycle of patients recovering, only to deteriorate again. Thus, psychiatrists may be more reluctant to remit patients back to prison due to the likelihood of these patients being readmitted shortly after discharge. In order to counteract this, psychiatrists logically tend to invest an increased amount of time managing their patients, the purpose being so that patients are discharged in their most optimum mental state and their risk of deterioration in prison is minimal.⁶

There were two surveys carried out in the UK citing the average length of stay as 22 days for acute units and 45 days for forensic units.⁷ Therefore, in the present study, the non-offender group was found to have a relatively lower mean length of stay compared with the national average. However, the study found that the average length of stay for the prison population was extremely high.

There is some suggestion in the literature that extended unnecessary length of stay can lead to management problems,⁸ with Bowers⁷ referring to ‘behavioural contagion’, in particular amongst ‘groups of disturbed patients on the PICU’. Although there is little evidence, it seems likely that there is a risk of these patients’ mental state deteriorating.

The time it took admissions from the prison group to reach the point where they were deemed by the responsible clinical team to be ‘ready for discharge’ or ready to leave the PICU was compared to the control group of non-offenders. The study found that it took on average 43.8 days for it to be decided that prisoners were ready for discharge from the PICU, whereas for non-offenders, it took 15.13 days. Most admissions to a PICU have some sort of forensic history (44%⁹ and 59%¹⁰ are quoted), with the majority of offences being violent in nature. The presence of forensic history therefore seems unlikely to explain this difference. Nevertheless, in this study, a significant difference was found between the patient groups, possibly due to the prison group carrying a greater risk to self or due to the reluctance of psychiatrists in remitting patients back to prison if it was felt they were unlikely to accept medication there.⁶

The increased length of stay for prison admissions to PICU or any other health-care setting is an issue which needs to be addressed and minimised in order to ensure optimum NHS expenditure. The cost per day for a bed in a South London PICU is approximately £560. For this cohort admitted to Ward 1 of the PICU at Springfield University Hospital, the potential savings which the NHS could have acquired was calculated using the mean length of stay for each patient once deemed ready for discharge. The total savings came to be £331,229 for 18 prison transfers, and £38,539 for 37 non-offenders. Thus, the total savings for all 55 admissions of this study to the PICU in the six-year period came to £369,768. The total amount of money spent by the NHS was approximately £790,000 for 18 prison transfers and £340,000 for 37 non-offenders. The study of costs highlight another key reason in tackling delays associated with transferring patients from the PICU, either back to prison or to another health-care setting. This extremely high cost for the NHS can potentially be reduced by implementing a new information-sharing protocol for both the prisons and hospitals in the hope of enabling more efficient communication between the two ends of the offender pathway.

Strengths and limitations

The main limitation of the study was that the data sample only included patients admitted to one PICU at Springfield University Hospital, and therefore the data may not be generalizable. However, the Trust serves a large and diverse population.

The data-collection process was entirely dependent on the quality and the detail of the RiO electronic notes. By and large, the notes were sufficient for the study. However, identifying the time it took for the prison group to be ready to leave the PICU proved difficult due to incomplete documentation.

Conclusions

Sections 47 and 48 of the MHA 1983 allow prisoners deemed to be suffering from a mental illness to be transferred to a hospital. There has been an increase in awareness regarding the delay in transferring prisoners to hospital. However, literature regarding the delay in discharging these patients back to prison is limited.

This retrospective study analysed the 18 prison transfers under s47 and s48 of the MHA 1983, and drew comparisons against the control group of 37 non-offenders. The analysis of the quantitative data provides a respectable representation of the transfers to a PICU from prison.

In the study, the average length of stay in the PICU was found to be much greater for prisoners compared with non-offenders. On average, the prison group also took longer to be deemed clinically ready for discharge, and even once clinically ready, then took longer to actually be discharged.

The study of prison transfers to a secure mental health-care service provider is still an important issue. The potential exists to improve the transfer process, and thus be able to provide an equivalent level of care to all those suffering from a mental disorder. The significant cost for hospitals to keep patients no longer needing treatment from PICU has a tremendous impact on the NHS in terms of finance and bed availability for other patients.

The offender pathway, and the difference in the length of stay between prisoners and non-offenders in a PICU, warrants further exploration to establish where the ‘blocks’ are in delay of remittance of this group to prison.

References

Singleton

Meltzer

Gatward

. Psychiatric morbidity among prisoners in England and Wales, London: Office for National Statistics, 1998.

Department of Health and HM Prison Service. Changing the outlook, London: Department of Health, 2001.

Department of Health. Good practice procedure guide: the transfer and remission of adult prisoners under s47 and 48 of the Mental Health Act, www.rcpsych.ac.uk/pdf/GoodPracticeGuide.pdf. (2011, accessed 14 February 2015).

Prison Reform Trust. Troubled inside: responding to the mental health needs of men in prison, www.prisonreformtrust.org.uk/Portals/0/Documents/Troubled%20Inside_Men%20report%20corrected.pdf (2005, accessed 21 February 2015).

Department of Health. Choosing health: making healthy choices easier, London: Department of Health, 2004.

Royal College of Psychiatrists. Prison transfers: a survey from the Royal College of Psychiatrists, www.rcpsych.ac.uk/pdf/PRISON%20TRANSFERS%20OP81.pdf (2011, accessed 10 February 2015).

O’Brien

Cole

. Mental health nursing practice in acute psychiatric close-observation areas. Int J Mental Health Nurs 2004; 13: 89–99.

Bowers L. Psychiatric Intensive Care Units: a literature review. Report from the Conflict and Containment Reduction Research Programme, 2006.

Pereira

Sarsam

Bhui

. The London Survey of Psychiatric Intensive Care Units: psychiatric intensive care; patient characteristics and pathways for admission and discharge. J Psychiatric Intensive Care 2006; 1: 17–24.

10.

Eaton

Ghannon

. Prediction of violence on a PICU. Med Sci Law 2000; 40: 143–146.