Violence Risk Assessment Among People With Mental Disorder

Background:The occurrence of aggression or violence among people with mental disorder has brought about considerable economic loss and caused harm to the victims, including patients themselves, relatives and psychiatric staff, both physically and psychologically. Therefore, violence or aggression has always been a focus in psychiatric field. Studies have proved that not all people with mental problem carry the potential of aggression but only a minority does. And the aggressive patients manifested a group of characteristics. Then, the identification of the violence risk factors, timely assessment and intervention own great significance. As professional staffs who contact patients most frequently, clinical nurses play an important role in violence risk assessment. The improvement of assessment and predictive accuracy are quite essential for the ongoing advancement of the present clinical nursing practice.Objectives:(1) Exploring the prevalence of patients’violence perceived by psychiatric nurses and methods of violence risk assessment adopted at present;(2) Exploring the problem in present clinical nursing practice in violence risk assessment;(3) Advancing violence risk assessment in clinical practice continuously through close cooperation with clinical nurses in action.Methods:There are three parts in this study. Part I:184clinical nurses in a Beijing psychiatric hospital were surveyed by Perception of the Prevalence of Aggression (POPAS) developed by Oud and a self-developed questionnaire to explore the prevalence of perceived patients’aggression and violence risk assessment method adopted. Part II: Violence Risk Screening-10(V-RISK-10) developed by Bj(?)rkly was validated among367admitted and289discharged patients with mental disorders respectively. Validity, reliability, sensitivity, specificity, positive and negative predictive value of Chinese version of V-RISK-10was measured. Advice about using the instrument was gathered from the nurses as well as their opinions for the following study. Part Ⅲ:Action research method was adopted to improve violence risk assessment continuously through close collaboration with nurses. A short-term violence risk assessment tool Br(?)set Violence Checklist (BVC) developed by Almvik was employed. The assessment was done twice a day for the first seven days of patients’hospitalization. A total of3707assessments for281newly admitted psychiatric patients was collected. Validity, reliability, sensitivity, specificity, positive and negative predictive value of BVC was assessed. Meanwhile, difference in violence occurrence between wards employed and not employed the instrument was compared and analyzed along with the feedback from the nursing staff. The reflection was taken again for the ongoing improvement in violence risk assessment.Results:Part I:Verbal aggression and mild physical violence were perceived most frequently by the nurses and over90%of the nurses surveyed were attacked by such violence during the last year. No unified violence risk assessment instrument was employed in the researching hospital, therefore most assessments were done based on nurses’observation and clinical experience. However, most (80.4%) of the nursing staff surveyed were inclined to use violence risk assessment tool. They also suggested that such tool should be violence-targeted, easy-to-use and time-saving. Part Ⅱ:V-RISK-10was proved applicable for Chinese population with satisfied acceptance. Content validity (CVI=0.94) and inter-rater reliability (ICC=0.89) was favorable. For admitted patients, the receiver operating characteristics (ROC) yielded an area under the curve (AUC) of0.63(95%-CI:0.57-0.69). At the cut-off point of8, its sensitivity/specificity was79.6%/38.4%and the corresponding positive/negative predictive value was34.4%/82.4%. For discharged patients, the receiver operating characteristics (ROC) yielded an area under the curve (AUC) of0.62(95%-CI:0.51-0.73). At the cut-off point of5, its sensitivity/specificity was79.26%/33.3%and the corresponding positive/negative predictive value was9.9%/94.5%. The predictive properties, with moderate sensitivity and low specificity, were somehow lower comparing to the results of the original V-RISK-10. Still, the introduction of the instrument was regarded as necessary and helpful by the participated nurses. Meanwhile, it was pointed out that the tool deserves careful modification before application as some of the items were unsuitable in Chinese culture. Besides, they also suggested the importance of dynamic violence risk assessment among newly admitted patients. Part III:The results demonstrated that BVC, easy-to-apply and time-saving (costing2-3minutes per assessment), was regarded as quite useful by the nurses. Content validity (CVI=0.96) and consistency among different raters (ICC==0.67～0.94) were satisfactory. A total of3,707assessments for281patients were collected. Receiver operating characteristics (ROC) yielded an area under the curve (AUC) of0.85(95%-CI:0.80-0.91). At the cut-off point of2, its sensitivity/specificity was69.9%/93.2%and the corresponding positive/negative predictive value was20.8%/99.2%.The predictive accuracy of BVC was promising, rather similar to the results of Norwegian original and those administered to other countries in their native language. Apart from mean time for physical restriction (26.64h vs44.05h, U=4527.00, P<0.01) and the total score of POPAS (45.09vs36.32, t (44)=8.769, P<0.01), there was no significant difference between wards adopted and not adopted BVC scale, including the incidence rate of violence (19.6%vs22.7%) and utilization rate of physical restriction (38.8%vs40.4%). The instrument seems to be a promising tool as suggested by the nurses participated in the study. Still, the modification in score marking system was adviced in the future study.Conclusion:It was revealed that the nurses were attacked by their clients frequently. However, the lack of effective violence risk assessment tools made them at disadvantage in clinical practice. They also expressed their intention in using violence risk assessment tools and gave suggestion in instrument choice. After problem diagnosis, the researcher and clinical nurses collaborated and cooperated closely in order to complement each other’s advantage. Through two circles of’plan-action-observation-reflection’, two violence risk assessment tools were introduced, validated and evaluated. The clinical nursing practice in violence risk assessment got advanced continuously. Violence risk assessment instrument was proved helpful for clinical practice. And action research, with its close cooperation between researcher and practitioner, was certified to be an effective way to improve and enhance clinical practice and deserve further exploration in nursing field.