Violence, dissatisfactionand rapid tranquillisation in psychiatric intensive care

We examined the associations of violence, patient dissatisfaction and occurrence of rapid tranquillisation in psychiatric intensive care, using an on-line nurse- based computerised database over a two-year period. Non-Caucasians were over-represented in violent incidents with physical threat, and previous forensic history was associated with more violent means of attack. Dissatisfaction related to non-understandable provocation and the total number of violent incidents. There was no correlation between rapid tranquillisations or side-effects and dissatisfaction. Remedial action and education in the psychiatric intensive care unit may reduce violence, and better prescribing habits, avoiding anti-psychotic polypharmacy in rapid tranquillisation, should be encouraged. Using a case note study, this paper presents a longitudinal survey of the effect of psychiatric in-patient care on benzodiazepine prescribing. Standards were proposed to assessthe quality of this prescribing. Basedon these standards, the study shows inappropriate use of benzodiazepines. Following admission, there was an increase in the number of patients prescribed benzodiazepines and in the number of benzodiazepines prescribed. Of the benzodiazepines withdrawn, most were contrary to the proposed standard. The quality of drug history showed little emphasis being placed on rationalising benzodiaze pine prescribing. The issue of how benzodiazepines should be handled during psychiatric admission is discussed. in

The aims of the present study were to audit the usage of a psychiatric intensive care unit and the occurrence of violent incidents over a two-year period, and to measure patient dissatisfaction with the service and correlate it with violent incidents, rapid tranquillisation episodes and total side-effects.

The study
The psychiatric intensive care unit in South Manchester, established in 1992, has 12 beds and serves a population of 175 000 (aged 16-65) with 80 other general adult beds on a district general hospital site. Black and ethnic minorities make up 11% of the catchment population.
To identify and measure violent incidents a nurse-run interactive computer system was established, the Psychiatric Studies in Aggress ion Database. Daily ratings were made for the first seven days, then weekly, on symptom, general aggression and side-effect rating scales. Specific violent incidents are recorded using the Staff Observation of Aggression Scale (SOAS)of Palmstierna & Wistedt (1987). A patient/user dissatisfaction questionnaire was included for use on discharge or transfer back to the base ward. It had three groups of scores on dissatis faction with the ward environment, information made available to people on admission and nurse/doctor contacts.
Rapid tranquillisation events were also re corded through the SOAS and patients' sideeffects during their stay were rated on a simple seven-item three-point rating scale. Items com prised tremor, rigidity, dystonic reaction, akathisia, visual difficulties, tardive dyskinesia and 'other' ratings. All side-effect ratings were to talled over the duration of stay. Rapid tranquil lisation was mainly by intramuscular zuclopenthixol acetate or haloperidol, often with the addition of lorazepam intramuscularly. The medication used was down to individual con sultant team choice with no overall unit policy at this time. Doses used generally followed British National Formulary guidelines, although higher doses were occasionally used in very disturbed or resistant cases.

Findings
Two hundred and fifty-one patients were ad mitted between September 1992 and November 1994. The mean age was 34.6 years, range 16-71 years. One hundred and fifty-seven admissions (62%) were male and 94 (38%) female. The median length of stay was nine days with a range of 0-131. Thirty-eight (15%) had a pre vious forensic history. In terms of ethnic group ing, 212 (85%) patients were Caucasian, 15 (6%) were Afro-Caribbean and 12 (5%)were of Indian sub-continent origins. One hundred and se venty-seven patients (70%) were detained under the Mental Health Act 1983.

Violent incidents and their associations
Of the 251 patients admitted, 121 (48%) had at least one violent incident during their stay and there were a total of 391 violent incidents rated through the SOAS scale. Of the 121 patients having violent incidents, 59 (49%) were asso ciated with a non-understandable provocation on the SOAS (Table 1 and 2). Twenty-two (18%) had an incident involving self-harm. Patients with violent incidents had significantly longer stays than those who did not (median (inter quartile range) 14 (7-26) days versus 6 (2-14) days, P<0.001). Those patients with a nonunderstandable provocation of incidents stayed significantly longer compared with those with no such violent incidents (median (interquartile range) 16 (8-30) days versus 7 (3-16) days, P=0.001). A significantly higher proportion of non-Caucasian people (n=21) were involved in verbal and physical threat violent incidents than Caucasian people (n=100). For non-Caucians 21

5(4)
22 (18) 1. Cups, ties, cigarettes, keys, knees or head-butts. (100%) had verbal and physical threat whereas only 76 (76%) of Caucasian people had such incidents (Fisher's exact test; P=0.01) ( Table 3). The total SOAS score would put the overall violence at the level of mild, however nursing injuries included a number of severe punches, a fractured nose, a severe bite wound and the use of objects such as knives and tools in a minority of assaults during this study period.

Patient dissatisfaction scores
Patient dissatisfaction scores were obtained on the last day of stay for 170 of the 281 patients (68%). The scale consisted of 12 items, including dissatisfaction with rooms, information on ad mission, availability of nurses, ward rounds, meals, ward booklets, explanations of treatment, welcome to the ward, recreational facilities, answers to questions, preparation for discharge and ward decor and surroundings. It could be rated from being totally satisfied to not at all satisfied on a six-point scale. The maximum score for dissatisfaction was 60, and the mean score was 25.4 (s.d. 11.5, range 0-60). There was no significant difference in dissatisfaction scores between those who experienced rapid tranquillisation events (n=76) and those who did not (n=84) during their stay, and there was no significant correlation with rated side-effects. There were, however, two significant associ ations; with the occurrence of at least one violent incident of an unprovoked type (n=40, mean dissatisfaction score 29.3) no violent incident of an unprovoked nature (n=130, mean score 24.2) (P=0.013). Also there was a borderline significant rank correlation with patient dissatisfaction score and the total number of violent incidents (P=0.058).
Examination of differences between individual items of dissatisfaction for patients with at least one non-understandable violent incident indi cated that four items showed significant differ ences between those with no such violent incidents and those with incidents (see Table  4). Typical comments made were: favourable -"it was a pleasant stay, where the staff let me do my own thing and helped me to relax in the surroundings", "the staff make you feel more homely, especially the female staff; unfavour able -"all I want to know is when do I get released from this establishment", "the food was the pits, there was never enough sent from the kitchen, the staff seemed embarrassed to give it to the patients"; bizarre -"need more colour, love the sound system, I have met some lovely staff and patient mortals", "I did not read the booklet, decor should be Deep Diana Pink". A similar pattern of significant differences (P<0.05) can be found for information and explanation with at least one violent incident of any type during stay.

Side-effects and rapid tranquillisation
Overall, 194 of 251 patients reported side-effects during their stay (77%). The total score during the stay was a median of 2 (interquartile range 1-5; range 0-48). The percentage zero score was 48 (19%). There was no significant association with patient dissatisfaction.

Comment
Various factors affecting the risk of violence have been the subject of recent work (Sheehan et al 1995;Thomas et al, 1995) and suggestions have been made as to what can be done about it (Hyde & Harrower-Wilson, 1994: Atakan, 1995.
Our study demonstrated the possibility of experienced nursing staff in a psychiatric in tensive care unit running a comprehensive database over a two-year period. Problems with reliability were addressed by regular teaching, meetings and videotape work.
Findings of interest included the association of violent incidents, including those with nonunderstandable provocation, with longer stay; manic or mixed states with early violence; selfharm with youth, severe depression and the combined diagnoses, schizophrenia/personality disorder/drug misuse not uncommon in cur rent clinical practice. Non-Caucasians were significantly over-represented in physical threat but this was not specific to any single ethnic group. A previous forensic history was associated with particularly violent means of attack.
Dissatisfaction scores on discharge were avail able for 68% of admissions and were significantly related to at least one violent incident of nonunderstandable provocation and the total num ber of violent incidents. Of particular interest was the significant difference between those with and without incidents concerned with items of information, explanation, recreation and pre paration for discharge. This suggests areas of remedial action and education in the psychiatric intensive care unit milieu may reduce the occurrence of violence. Consequently we have recently introduced day-time adult education sessions on the psychiatric intensive care unit to apparent good effect.

Violence, dissatisfaction and rapid tranquillisation
There was no correlation between the occur rence or number of rapid tranquillisation events and patients' general dissatisfaction scores. Side-effects from potent antipsychotic drugs used in rapid tranquillisation incidents did not seem to influence dissatisfaction, although rapid tranquillisation was significantly associated with higher total side-effects over the patient stay. Within those groups who received rapid tran quillisation the polypharmacy group who re ceived a number of different antipsychotics had significantly raised side-effects compared with those receiving monotherapies. either with shortacting haloperidol or intermediate-acting zuclopenthixol acetate alone.
Further research needs to examine the role of ethnic grouping in violence, factors influencing dissatisfaction and remedial and educational aspects of patients' stay on psychiatric intensive care unit. Better prescribing habits avoiding antipsychotic polypharmacy in rapid tranquil lisation should also be encouraged. Benzodiazepine prescribing in a psychiatric hospital

J. Summers and K. W. Brown
Using a case note study, this paper presents a longitudinal survey of the effect of psychiatric inpatient care on benzodiazepine prescribing. Standards were proposed to assessthe quality of this prescribing. Based on these standards, the study shows inappropriate use of benzodiazepines. Following admission, there was an increase in the number of patients prescribed benzodiazepines and in the number of benzodiazepines prescribed. Of the benzodiazepines withdrawn, most were contrary to the proposed standard. The quality of drug history showed little emphasis being placed on rationalising benzodiaze pine prescribing. The issue of how benzodiazepines should be handled during psychiatric admission is discussed.
In 1988, the Committee on Safety of Medicines declared that the use of benzodiazepines should be decreased as dependence was becoming a subject of increasing concern (Committee on Safety of Medicines, 1988). Benzodiazepine mis use is a further issue which has raised concern. In 1993, temazepam was the most commonly misused drug in Scotland (Scottish Health Service Common Services Agency, 1993). Furthermore, there is evidence from the Home Office Drugs Inspectorate that "the vast majority of misused benzodiazepines are obtained on prescription" (National Medical Advisory Committee, 1994).
Although anxiolytic benzodiazepine use has decreased to about one-quarter of its use 15 years ago (Royal College of Psychiatrists. 1997), the above concerns remain relevant and empha sise the need for high standards in benzodiaze pine prescribing.
Psychiatrists should be