Management of psychiatric in-patient violence in the Anglia region Implications for record-keeping, staff training and victim support

Aimsand method A total of 384 incidentsof violence against the person(six'serious'and 378'mild'), by adult in-patients in general psychiatric units (GPUs) and learning disability units (LDUs)in 10 National Health Service trustsin the Anglia region, were evaluated by interviewswith staff and examination of records. Results The findings,when compared with standards derived from previous recommendations, showed deficiencies in the documentation of incidents (there was no satisfactory written record of physical restraint for 97%of incidents in GPUsand 85%in LDUs). in the training of staff in 'control and restraint' procedures Of two or more staff were involved in physical restraint, for 3% of incidents in GPUs and 100%in LDUs,the staff had received no training within the previous 12 months) and in policies for victim support (there was no written policy that included procedures for victim support in relation to 84% of incidents in GPUsand 44%in LDUs). Clinical implications Trustsshouldconsiderreviewing their policies on the prevention and management of violence, particularly in relation to staff training.


Management of psychiatric in-patient violence in the Anglia region
Implications for record-keeping, staff training and victim support J. H. Dowson, J. Butler and O. Williams Aimsand method A total of 384 incidentsof violence against the person (six'serious' and 378 'mild'), by adult in-patients in general psychiatric units (GPUs) and learning disability units (LDUs) in 10 National Health Service trusts in the Anglia region, were evaluated by interviews with staff and examination of records. Results The findings,when compared with standards derived from previous recommendations, showed deficiencies in the documentation of incidents (there was no satisfactory written record of physical restraint for 97%of incidents in GPUsand 85%in LDUs).
in the training of staff in 'control and restraint' procedures Of two or more staff were involved in physical restraint, for 3% of incidents in GPUs and 100%in LDUs,the staff had received no training within the previous 12 months) and in policies for victim support (there was no written policy that included procedures for victim support in relation to 84% of incidents in GPUsand 44% in LDUs). Clinical implications Trusts shouldconsiderreviewing their policies on the prevention and management of violence, particularly in relation to staff training.
The management of 'routine' violence in psychia tric services, particularly in in-patient units, continues to place severe demands on staff, especially the nursing staff.
The present study evaluated aspects of the management of reported incidents of violence that involved physical aggression to others, in adult psychiatric in-patient units (CPUs) and learning disability in-patient units (LDUs),in the Anglia region. Information about each incident was obtained from records and from interviews with staff involved. The management of each incident then was evaluated by comparing the findings with seven standards derived from previous recommendations.
The aim of this study was to improve the management of incidents involving physical violence to others by assisting the review of unit policies.

The study
The study evaluated incidents involving either 'mild' or 'serious' injury as a result of patients' aggression, using the following definition of physical aggression: action (other than verbal isation) intended to cause pain, bodily harm or death upon another (Kaye et al 1988). 'Mild' injury to others was defined as painful blows, bruises, sprains, welts, attempted strangulation, etc. and 'serious' injury as fracture, loss of teeth, deep cuts, loss of consciousness, rape, etc. (Kaye et al, 1988). This classification involves some judgement, so that, for example, a serious incident of attempted strangulation would be classed as a 'serious injury'. In the present study, all such incidents involving 'mild' or 'serious' injury that were notified were evaluated, includ ing those that occurred during a 'control and restraint' procedure. Each 'incident' involved violence by a specified patient to one or more victims.
Ten National Health Service trusts in the Anglia Region, responsible for units on 15 sites, agreed to participate. Units approached were either general adult CPUs for patients aged 16-65 years or adult LDUs. Units that usually lacked 24-hour staffing by qualified nurses were excluded. For each site a 'link' person was identified who regularly notified the researcher of relevant incidents: over a five-month period for CPUs and a two-month period for LDUs. Staff involved in each incident were then contacted by tele phone or in person and a semi-structured inter view was given by the researcher to evaluate seven standards that had been identified from previous recommendations (Table 1) (Powell et al, 1994;Walker & Siefert. 1994;Atakan, 1995;Mortimer, 1995). At a subsequent visit, medical and nursing notes, prescription sheets, ward reports and other relevant documentation were examined.

Findings
In relation to 106 beds in locked ('secure') units and 725 other beds in CPUs, 301 incidents of physical violence by 133 patients (78 men and 55 women) were reported that resulted in either Satisfactory written record of physical restraint Staff should have received training within the previous twelve months if two or more are involved in physical restraint Physical restraint is achieved without injury to the patient Physical restraint is achieved without injury to staff Availability of a written policy for the management of violent incidents All staff involved in physical restraint procedure should be aware of any relevant hospital policies Availability of a written policy relating to management of violent incidents which includes procedures for victim support n=number of incidents for which relevant data were obtained, i.e. it varied in relation to the standards evaluated, because missing data varied between the standards.  (Haller & Deluty, 1988;Thomas et al, 1995). This appeared to reflect a failure of communication and pressure of work. Nurses were the staff group most commonly in receipt of violence, forming 68.8% (n=247) of the 359 victims, whereas fellow patients were the victims in 26.5% (n=95). The remainder, 4.7% (n=17), were members of the public, medical staff and other staff. Of the 291 (of a total of 301) incidents when the time of day had been recorded, 35.2% of incidents occurred during the early (day) shift, 46.4% during the late shift and 18.4% during the night shift.
In relation to 10 'secure' beds and 61 beds in LDUs, there were 83 violent incidents involving 27 patients and 106 known victims, all with 'mild' injury. As above, most of the victims did not receive any medical treatment: 47.2% (n=50) of victims had no visible injury; 32.1% (n=34) sustained visible injury. Minor first aid was applied to a further 16% (n=17) and 4.7% (n=5) sought further medical advice. Fifty-two per cent (n=53) of the 102 victims whose status was known were nurses, and 48% (n=49) were fellow patients.
When a physical restraint procedure was involved (i.e. any procedure when hands have been laid on a patient with the aim of preventing violence to others), for 68% of incidents in CPUs and 51% in LDUs for which relevant data were available, the persons directly involved in the restraint were not named in written records, and for 65% and 74%, respectively, there was a failure to record roles undertaken by staff during restraint; also, for 11% and 14%, respectively, there was a failure to note all injuries that had resulted. For 93% and 54%, respectively, there was no indication of the duration of the restraint. Table 1 shows that this resulted in only a minority of incidents involving physical restraint, in both CPUs and LDUs, for which the standard for a written record was considered to have been achieved.
Restraint procedures were involved in 171 of the 301 incidents in the CPUs and in 39 of the 83 incidents in the LDUs. Table 1 also shows that, when staff could be interviewed, only a minority of those restraints in CPUs involving two or more staff (and none in LDUs) had been undertaken by staff who had been trained or updated in restraint techniques within the previous 12 months. In both CPUs and LDUs, 22 staff and two patients were injured during restraints when no member of staff involved had been trained or updated within the previous 12 months.
Although Table 1 shows that the majority of incidents involving physical restraint did not result in injury to patients or staff, 39% of all restraints undertaken resulted in reported injury to staff and 61% of these restraints did not meet the standard for staff training; also, of the 11 restraints that involved reported injury to pa tients (all 'mild'), six did not meet the standard for staff training.
All trusts had policies for the use of 'control and restraint' procedures, apart from one where re-drafting was in process. However, of the 66 restraint procedures (out of a total of 210) where it was possible to obtain information about all the staff involved, 41% involved some staff who were not acquainted with trust policy.
Only four out of the 15 sites had a written policy related to the management of violence that included procedures for victim support. Of the staff interviewed who had been physically assaulted, 16% of the 145 in CPUs and 24% of the 33 interviewed in LDUs considered that post-incident support for victims had been inadequate.

Comment
The recording of basic information about in cidents of violence was often omitted or inade quate. Despite the burden of completing forms, a simple and reliable procedure, used routinely, would be a precaution against subsequent complaints or legal action and provide managers with the information needed to review proce dures and resources (Haller & Deluty, 1988;Atakan, 1995;Royal College of Psychiatrists, 1995). A relatively short 'incidents and re sources' form could be used to cover physical aggression and aggression to property, as well as situations that are judged to have an increased risk of violence, such as staffing difficulties or bed occupancy of more than 100%. Many of the questions could be designed as a 'tick-box1 format to enable rapid completion (Royal College of Psychiatrists, 1995).
Most of the sites offered some formal 'control and restraint' training, often within a wider context. This was usually provided 'in house' and members of the hospital's own staff often had been trained as instructors. There were several different course models being used and there were differences of opinion as to which was the most suitable. The lengths of the courses varied and the categories of staff offered 'control and restraint' training varied from unit to unit. Usually the initial training took 5-10 days, whereas 'training update' courses were for 1-3 days. Despite the provision for 'training update' courses, there was usually no formal procedure for ensuring regular updating. Some units, including some 'secure' units, did not meet the standard on training because 'training' updates had not occurred, although most 'secure' units were associated with a high level of initial training.
Many nurses who managed violent incidents did not have any formal training in 'control and restraint' techniques. Although staff, for exam ple those in some 'long-stay' units where the incidence of violence was low, did not always consider that they required training in the full range of techniques, some staff exposed to relatively frequent violence had not been offered any relevant training. However, it has been noted that nurses trained to respond to violence suffer relatively few injuries (Royal College of Psychiatrists, 1998). An appropriate training strategy is likely to vary considerably between different units, depending on the patients involved and the nature of the hospital site (Mortimer, 1995).
The management of violence involves a variety of professional and interpersonal skills practised in an adequately provided setting, whereas the prevention of violence is a constant priority of a psychiatric service. Furthermore, the manage ment of violence must not be equated just with the application of physical techniques to restrain and control the patient (Powell et al, 1994;Atakan, 1995;Lodge, 1997). However, these skills are often an essential part of an adequate response by staff.
The present findings indicate the need for managers to review and implement policies related to routine record-keeping of incidents, staff training and victim support.