Prisons in England and Wales are full to capacity and have been unable to find sufficient places to accommodate all the prisoners being sent by the courts. Many prisoners have mental disorders (Reference Singleton, Meltzer and GatwardSingleton et al, 1998) and significant numbers are in need of urgent transfer to psychiatric hospital. Between 5% and 8% (1300 to 2000 per annum) of all people detained annually under section in psychiatric hospitals in England come from court or prison (Information Centre, 2006).

In April 2006 the transfer to the National Health Service (NHS) of responsibility for healthcare in prisons was completed with services being commissioned by local primary care trusts. Acutely unwell prisoners may be moved to healthcare centres in prisons, where they receive care from general practitioners (GPs) and mental health inreach teams, but under the current Mental Health Act 1983 they cannot be treated in prison without consent. Previous studies have found unacceptable delays in transferring mentally ill patients to hospital (Reference Isherwood and ParrottIsherwood & Parrott, 2002). For those refusing treatment while waiting for transfer the only option is to consider treatment without consent under common law (Reference Earthrowl, O'Grady and BirminghamEarthrowl et al, 2003). In spite of this fact, delays in hospital transfer of up to 3 months have been considered within the bounds of the acceptable. The number of prisoners waiting more than 12 weeks for transfer has to be entered on the quarterly Prison Health Star Rating/Performance Assessment Form and the information is included in the 6-monthly Report on the State of Healthcare across the Prison Estate, which is submitted by the Head of Prison Health to the Minister of Justice. Under the provisions of the Mental Health Act, no time limits are specified in relation to sections 47 and 48, for the transfer to hospital of sentenced and unsentenced prisoners respectively. In the case of court orders for hospital transfer, the Act specifies a time period of 7 days (sections 35, 36) or 28 days (sections 37, 38) from the date of the order within which the transfer should be made. This contrasts with the time period of 14 days from the second medical recommendation for civil sections (sections 2, 3) or 24 h from the medical recommendation in the case of an emergency application under section 4.

New initiatives are under way to try to speed up transfers from prison to hospital. In November 2005, a joint Prison Health (Department of Health) and Mental Health Unit (Home Office) working party issued Procedure for the Transfer of Prisoners to and from Hospital under Sections 47 and 48 of the Mental Health Act 1983 with the aim of reducing ‘unacceptable delays’ (Department of Heath, 2005). This included a ‘best practice flow chart’ for carrying out the steps involved in a transfer, and recommendations to report delays to the mental health commissioner in the responsible primary care trust. Subsequently mental health trusts in England were invited by the Department of Health and the National Offender Management Service to participate in a 6-month pilot to identify barriers and obstacles to the national roll-out of a waiting-time limit for the transfer of acutely mentally ill prisoners, and to shape what that limit could realistically be. Participating trusts were asked to admit a prisoner with acute mental illness within 14 days of receipt of a referral from a prison inreach team psychiatrist; 16 trusts participated in the initial phase of the pilot, which has recently been extended for further 6 months. The present study was unaffected by this pilot, which was initially scheduled to run from 8 January 2007 to 9 July 2007.

This paper presents findings of times taken to transfer prisoners to hospital before and after the implementation of new procedures in a busy remand prison in north London, with the aim of assessing whether the new procedures have led to a reduction in transfer times. A secondary aim was to assess how many prisoners awaiting transfer to hospital were refusing medication, as such people may face lengthy periods without treatment of their mental illness.

Method

Transfers to a psychiatric unit from HMP Pentonville during two 6-month periods were compared before and after the implementation of new procedures based on the government guidance Procedure for the Transfer of Prisoners (Department of Health, 2005). Target times for completion of each step of the best practice flowchart were included, for example 72 h from identification of an inmate needing transfer by an inreach team psychiatrist to the referral being faxed to the consultant in the appropriate transfer unit, and 28 days from identification to approaches being made to the commissioning primary care trust if the transfer had not occurred.

Pentonville is a category B local prison with capacity for 1200 male prisoners. There are approximately 7000 new receptions into the prison per year, of which approximately 60% are remand prisoners. The new procedures were implemented from 1 April 2006, and the 6-month periods chosen for the study commenced on 1 July 2005 and 1 July 2006. Patients transferred to hospital were identified from LIDS (the prison service inmate database), an electronic database of hospital transfers kept by the inreach team, and from paper-based hospital transfer records. Patient clinical information was obtained from the Inmate Medical Record.

Results

A total of 75 patients were transferred to hospital, 33 in the first period and 42 in the second period. The two groups were similar in age, ethnicity and offending behaviour: 41% were White, 40% Black, 15% from an Asian background and 4% other. The mean age was 35 years (range 21 to 63 years). Regarding the main offence: 52% were violent in nature, including 3 charges of murder and 1 of attempted murder, 23% were sexual, 12% were acquisitive, 4% involved arson, and 9% were other offences such as criminal damage.

On mental state examination at time of identification for transfer, 81% of the total number of men were found to have a psychosis, 7% had mania and 7% had severe depression. Others had predominant features of cognitive impairment or behavioural disturbance. In all, 71% had a past psychiatric history including previous diagnosis, and 17% had no past history. In 12% of cases it was unclear whether there was a past psychiatric history. Of the 53 patients with a previous diagnosis, the primary disorder was for 60% schizophrenia or schizoaffective disorder, for 13% affective disorder and for 17% other non-affective psychoses. The remaining 10% included two men with intellectual disability, one with Wilson's disease and one with Asperger syndrome.

Many individuals awaiting transfer refused to agree to treatment. Table 1 shows how many of those patients who had been offered medication were refusing when first identified and at time of transfer.

Table 1. Patients refusing medication

	Group 1			Group 2
Time	Prescribed, n	Refusing, n	Prescribed, n	Refusing, n	Overall refusing, %
At identification	29	21	32	13	56
At transfer	31	14	37	15	43

1. Group 1 were studied from July to December 2005

2. Group 2 were studied from July to December 2006

Of those transferred, 32% were transferred on a court order (section 35, 37 or 38) and 68% on a section 47 or 48 hospital order; 59% went to a low secure psychiatric intensive care unit, 33% to a medium secure unit and 7% to a general adult ward. One patient was sent to a high secure unit.

The delay from identification to eventual hospital transfer ranged from 15 days to 301 days for the first group and from 8 days to 148 days for the second group. The longest delays in each group were for medium secure placements. Table 2 shows the delays in transfer by unit type and Table 3 by section of the Mental Health Act 1983. Overall there was a reduction in mean waiting time from 77 days to 53 days (24 days; P=0.07, 95% CI -2 to 50).

Table 2. Mean transfer delay (days) by unit type

	Group 1			Group 2
Unit type	n	Delay, days	n	Delay, days
General ward	4	62	1	120
PICU	19	58	25	44
MSU *	10	118	15	64
HSU	–			1	27
All units	33	77	42	53

PICU, psychiatric intensive care unit; MSU, medium secure unit; HSU, high secure unit; n, number of patients

1. Group 1 were studied July to December 2005

2. Group 2 were studied July to December 2006

* P<0.05

Table 3. Mean transfer delay (days) by section of Mental Health Act 1983

	Group 1 1			Group 2 2
Section	n	Delay, days	n	Delay, days
Court order	8	48	16	45
Hospital order	25	86	26	56
All cases	33	77	42	53

n, number of patients

1. Group 1 were studied July to December 2005

2. Group 2 were studied July to December 2006

Discussion

We found a trend towards shorter delay times for transfer to hospital, from 11 weeks to 7 weeks, following the introduction of new transfer procedures. For both groups there was a shorter delay in transfer following a court order, where the Mental Health Act 1983 specifies a period of 7 or 28 days following the court order, during which the transfer should be made. Delays were found to be longer for transfer to medium secure units than to low secure units.

Some factors contributing to the reduction in delay times for transfer were directly within the control of the inreach team. These included ensuring that the referral to the admitting unit was made as soon as possible after identification of the need for transfer, and that the medical recommendation by the team psychiatrist and the documentation required by the mental health unit at the Home Office were sent within the times specified in the new procedure. The biggest reduction in transfer times was to medium secure units, where delays were reduced by nearly 50% (P=0.04). Enlisting the assistance of the mental health commissioner within the responsible primary care trust after 28 days was one contributing factor to this reduction. Nevertheless, time to transfer to a medium secure unit remained considerably longer than to a psychiatric intensive care unit. Shortage of beds and lack of step-down facilities were the reasons most often given for delays. In some cases the inreach team was able to influence the decision as to whether a recommendation for transfer should be sought under a court order or a hospital order, and in these cases a court order was the preferred choice. This was because experience suggested that, in spite of the need to allow time for the production of court reports and to wait for the next court appearance, the total delay time would be less using this route.

We found that approximately 50% of men identified for hospital transfer refused medication during the period they are waiting to be transferred. These people with a treatable mental illness were kept in prison without treatment for about 2 months on average, since they could not be given treatment against their will in prison under the current provisions of the Mental Health Act 1983.

This study is an early assessment of the impact of new government guidelines designed to cut down delays in the transfer of acutely unwell prisoners to hospital, at a time when final discussions of the new Mental Health Bill were taking place. No other study to our knowledge has assessed the rate of non-adherence to medication among acutely unwell prisoners awaiting transfer to hospital. A consequent weakness of the study is the relatively small sample size, restricted to one prison.

We conclude that unacceptable delays in transfer to hospital of mentally ill prisoners are likely to remain, even after the introduction of new procedures by the Department of Health and the Prison Service. We recommend that time limits should be specified for hospital transfers from prison comparable to norms under civil sections. Ideally this should be included in an amendment to the new Mental Health Bill that recently completed its passage through Parliament.

Declaration of interest

None.