Factors associated with dangerous behavior in forensic inpatients: results from a pilot study

This study was designed to identify risk factors associated with violence within a forensic inpatient hospital setting. The primary purpose was to develop a screening tool to aid in the rapid identification of patients requiring high versus low security ward placement. Subjects included 232 consecutive admissions during a five-month period to a 300-bed forensic division within a public-sector psychiatric hospital. Demographic, historical, and current clinical variables were collected, and dangerous behaviors were documented by nursing staff on a daily basis. The associations between dangerous outcomes and various risk factors were examined using chi-square or t tests, as appropriate. Stepwise logistic regression analysis was performed to assess the contribution of each risk factor significantly associated with violent behavior. A strong association was found between 10 risk factors and the incidence of violence in our forensic population. These factors included current clinical factors (grossly inappropriate behavior observed on the ward, an assessment of current escape risk); historical factors (psychiatric hospitalization history, criminal history variables, and a self-report of prior violence); and one demographic variable (length of stay). Of the 10 risk factors then included in the stepwise regression analysis, four contributed significantly to the final predictive model. Grossly inappropriate behavior, patient self-report of prior violence, history of psychiatric hospitalization, and a history of 10 or more prior psychiatric hospitalizations were most predictive of a violent outcome. Although situational and environmental variables also must be incorporated into final assessments and decisions, the identified significant variables can provide an additional tool in the rapid assessment of violence potential.