Violent behavior is a serious complication in acute psychiatric inpatients. Several risk factors are known that facilitate such behaviors, preventive measures are however difficult to evaluate, since prevented violent events usually are not recorded. The aim of the study was to assess the possibility to predict violent events and efficaccy of preventive measures.

The study is prospective observational study at large 32-bed PICU in University Psychiatric Hospital, that covers 900.000 population with the average of 8 admissions per day. Recording of violent or other unexpected events is done routinelly. The risk for violent events was measured by BVC and by subjective assessment on a 7-point scale, both done three times daily. Using a special form, a number of prevented violent events is recorded with the same frequency. The results were compared with number of violent events before new assessments were introduced in everyday practice. Events were correlated with clinical assessments using CGI, GAS and BPRS.

The number of actual violent incidents dropped significantly with new assessment methods. Subjective assessments of the risk for violent behavior showed superiority in the prediction of events then BVC. The majority of violent incidents were predictable and preventable, that reflected in low frequency of actual behaviors observed in the study.

Simple routine assessments done by educated stuff at PICU predict well unwanted and dangerous behaviors in acutelly ill psychiatric inpatients. Prevention strategies include known risk factors and give some new insights in the extent of the problem.