Towards the end of 1998 we decided to introduce a facility to lock the doors of the acute psychiatric wards at Bootham Park Hospital, York. At first sight this would appear to be a retrograde step for modern psychiatric care in the UK. But this was a reasoned and planned move which has resulted in improved care for patients overall.

This paper explains how we reached our decisions and outlines the results of a subsequent audit and survey of users' views.

Why?

First of all, a bit of history. Bootham Park Hospital (BPH) was commissioned in 1774 and many of the original buildings are still in use today, albeit with modifications. It was the shocking state of care in 1792 at BPH (then known as The York Asylum) which lead William Tuke to found the Quaker hospital, The Retreat, also still in use today. The Retreat was a pioneer of an early model of care which came to be known as ‘moral treatment’ of psychiatric disorder. This form of treatment emphasised self-determination of the patient and used minimal restraint.

An editorial in the Lancet (Anonymous, 1976) summarised the position in more recent years. The ‘wholesale unlocking’ of mental hospital wards in the 1960s is described as being the most important single change in the milieu of the psychiatric hospital at that time. Unfortunately, this resulted in an increasing reluctance of hospitals to accept responsibility for patients who were difficult to manage. There was a call to reintroduce locked wards.

It is obviously important to prevent the doors revolving full circle in psychiatric care and any planned treatments involving potentially increasing restrictions on patients needs to be explained carefully.

To understand our decision we need to look at what it had been like on the acute psychiatric wards at BPH in the previous few years. Early attempts to operate risk management policies following the national trend had led to high numbers of patients on close observation. Close observation, or level one observation, is defined as having a nurse within sight of the patient at all times. The average number of patients on close observation at this time was eight in a hospital with 58 acute general psychiatry beds. There had been a recent suicide by an in-patient and two serious assaults by patients who were not being closely observed (Reference Adams and KennedyAdams & Kennedy, 1998). The Mental Health Commission had commented about the high levels of close observation and used the term ‘sentinel nurses’ to describe the increasing numbers of nurses sitting and watching patients or watching the door of the ward to prevent detained patients from leaving. It you had walked onto a ward at this time you would have seen two or three nurses watching individual patients. Naturally, this resulted in very few nurses left to provide therapeutic care to the rest of the patients on the ward.

A rise in illicit drug use and cases of drugs being sold on or near the wards led to a need to consider security. A serious incident occurred when a drug dealer came onto a ward and assaulted a patient and several members of staff. Nowadays most people keep their house door locked during the day as well as at night. There is a parallel with the requirements of safety for in-patients.

What was happening at BPH is no different from the national picture when compared with recent surveys by the Mental Health Commission and the Sainsbury Centre for Mental Health (Reference Ford, Durcan and WarnerFord et al, 1998; Sainsbury Centre for Mental Health, 1998; Reference MuijenMuijen, 1999).

After much consideration we decided to introduce procedures enabling the doors of the wards to be locked at certain times (at night) and under certain conditions.

The policy

The first stage involved writing a policy acceptable to all professionals and patient representatives and to obtain approval from the Trust Board.

As a security measure, we decided routinely to lock the ward doors from 10.00 p.m. to 6.30 a.m. Fire exits were all fitted with automatic unlocking systems. Ward doors were all fitted with bells and keypads. A policy was written following certain key principles.

1. (a) Informal patients could request to leave the ward at any time and a notice to this effect would be displayed by the ward door.

2. (b) The decision to lock the door at any time except for during the night was only to be taken as a last resort and in response to an identified and documented risk management plan.

3. (c) It was thought to be good practice to ensure that every effort should be made to inform patients, visitors and other staff of the reason for locking the door (as far as confidentiality permitted).

4. (d) It was decided that the maximum period the door would remain locked would be two hours and then there would be a review by the ward manager in conjunction with medical staff.

5. (e) Full records would be kept explaining the reason for the locked doors. These records would be audited and a report would be provided after an allotted time period to the General Psychiatry Directorate and the Trust Board.

Results

The audit report (October 1999, details available from the author upon request) described above provided the content of these results.

1. (a) All four wards locked doors on occasion, but one ward used the facility less often than others.

2. (b) The trend was that the facility was used less often over time. This needs further monitoring and coincides with the opening of a special care unit in June 1999.

3. (c) Two-hour periods were often exceeded: in one case doors were locked for two weeks. In these cases the assessed risk had not reduced, as the patients concerned had remained severely ill and at high risk of absconding.

4. (d) The main reason given for locking the door was to manage patients at high risk of absconding and of harming themselves, others or more rarely, substance misuse. On two occasions the door was locked to prevent access by others. In some cases the door was locked as a holding procedure pending an assessment by the special care or forensic team.

5. (e) The most significant finding was that close observations fell dramatically from an average of eight during 1997/98 (two per ward at all times) to an average of three in 1999. This result was statistically significant (P=0.01, χ²=326.1, d.f.=1). In practice there was a dramatic reduction in ‘sentinel’ nursing.

6. (f) It was surprising that there were no spontaneous informal or formal complaints by patients or user groups. A different picture emerged, however, when this was formally tested. Questionnaires were sent out to all discharged patients over a two-month period (n=52). Twenty-three questionnaires were returned. Out of these 39% were worried about the doors being locked, 48% did not mind and 13% indicated that it felt safer. A subsequent focus group of six patients who had been discharged did not welcome the practice of locking doors, commenting on the “prison-like” environment. The discussion on close observations, however, concluded that all patients found this procedure equally difficult. Patients commented that staff would often sit and watch the patient without attempting to engage them in conversation. It was suggested that this practice could heighten a susceptible patient's paranoia.

7. (g) The audit findings implied that the doors were not locked to avoid bringing in extra staff, but on a number of occasions they had been locked because an adequate number of staff were not available.

Discussion

What was most striking about the findings was that the amount of close observation reduced so dramatically. We expected a slight reduction but the magnitude was greater than expected. The reduction coincided with the new policy and occurred well before the opening of a special care ward for in-patients presenting with challenging behaviour. This implies that a greater proportion of the patients on the acute wards were being observed to prevent them from absconding rather than as a consequence of challenging behaviour. The development of a better therapeutic culture on the wards may have contributed to the trend to use the facility less frequently. An interesting observation from this was that when nurses found that they had more time for therapeutic activity, they became concerned that they needed more training. Training needs for staff are important if a ward is to become less custodial and more therapeutic.

User views were more critical of the practice of locking doors but were perhaps equally critical of the practice of close observation. What is needed is a balance between the two approaches.

To conclude, there is evidence that developing a policy that enables the doors of acute psychiatric wards to be locked when necessary has not resulted in a return to the old days of restraint and custodial care. In reality the trend has been the opposite. Occasional use of the facility to lock doors has benefited psychiatric care overall, provided, and I think this is very important, that doors are locked rarely.

If any hospital is planning to bring in a locked-door policy, then it is essential that it is thoughtfully prepared and preserves the rights of patients. Operation of the policy must be closely monitored and patients' views actively sought, to ensure that practice follows the policy and that the staffing complements of the wards are not reduced. Locked doors must not become the norm, but are a useful adjunct to improving therapeutic care for patients on acute psychiatric wards on the rare occasions when they are necessary.