Violence prevention and management in acute psychiatric care : aspects of nursing practice

Abstract

Aim: The general aim of this thesis was to explore and evaluate different aspects of nursing practice in relation to prediction, prevention and management of patient violence in acute psychiatric care.

The specific aims were: to evaluate the short-term predictive capacity of the violence risk assessment instrument Brøset Violence Checklist (BVC) when used by nurses in a psychiatric intensive care unit (study I), to compare the occurrences of coercive interventions and violence-related staff injuries before and after a two-year nursing development and violence prevention intervention (study II), to describe aspects of the caring approaches used by nurses in acute psychiatric intensive care units (study III) and finally (in study IV) to test the hypothesis that staff training according to the ‘Bergen model’ has a significant positive influence on the violence prevention and management climate in psychiatric inpatient wards, as perceived by patients and staff.

Methods: Both quantitative and qualitative methods were used. In study I, data from the BVC and the Staff Observation Aggression Scale were retrospectively collected from a psychiatric intensive care unit (PICU) and analysed in an extended Cox proportional hazards model. In study II, register-based rates of coercive interventions and violence-related staff injuries were retrospectively collected from the same unit as in study I, and subsequently analysed through Chi-square tests. In study III, qualitative data were collected from 19 individual interviews with nurses working on four PICUs in different parts of Sweden. The data analysis was guided by the interpretive description approach. In study IV, a 13 item questionnaire was developed (called the E13). Each item was related to the violence prevention and management climate on inpatient units. The E13 was distributed to patients and staff on 41 psychiatric wards before the staff had been trained according to the Bergen model and subsequently to patients and staff on 19 wards where the staff had been trained. Data analysis included factor analysis, Fisher’s exact test, Cronbach’s alpha and Mann-Whitney U-test.

Findings: A positive scoring by the nurses on any of the six BVC items resulted in a six-fold increase in the risk for short-term severe violence on the PICU. A negative scoring on all items correctly predicted no risk for severe violence in 99% of all assessments (study I). In study II, an increase in the total rate of coercive interventions was found on the PICU one year after the intervention, while the rate of violence related staff injury remained unchanged. However, during the study period, an unplanned re-organisation of the PICU, including a substantial reduction of beds, meant that the PICU from then on could only admit the most acutely ill patients. In study III, interviews with nurses working on four different PICUs revealed two caring approaches which were metaphorically named the bulldozer and the ballet dancer. The bulldozer approach functioned as a shield of power that protected the ward from chaos, but at the same time involved the risk for engaging in uncaring actions. The ballet dancer approach functioned as a means of initiating relationships with patients and appeared strongly related to caring actions. In study IV, four items of the E13 questionnaire were rated significantly more positive by staff on trained wards. These four items concerned good rules on the ward, the ability of staff to stay calm when approaching aggressive patients, the staff’s interest in understanding why a patient is acting aggressively and the ability of staff to approach aggressive patients at an early stage. One item was rated significantly more positive by patients on trained wards which was the item relating to the interest of staff in understanding why a patient is aggressive. No item was rated more negatively on trained wards.

Conclusions: Violence prevention and management in nursing practice involves a caring approach in all levels of prevention; in the everyday care as well as in coercive situations. It involves protection of the dignity of the patient and the nurse-patient relationship. The BVC has a good predictive capacity but should primarily be used to initiate early preventive interventions. In evaluation studies of violence prevention and management interventions, a mixed methods design should be considered, including the perspective of patients.