Predicting Violence in Psychiatric Patients
Author Information
Author(s): Abderhalden Christoph, Needham Ian, Dassen Theo, Halfens Ruud, Haug Hans-Joachim, Fischer Joachim
Primary Institution: University Bern Psychiatric Services
Hypothesis
Can combining the Brøset-Violence-Checklist with subjective clinical risk assessment improve the prediction of violence in psychiatric inpatients?
Conclusion
The BVC-VAS is an easy to use and accurate instrument for systematic short-term prediction of violent attacks in acute psychiatric wards.
Supporting Evidence
- The AUCROC of the new BVC-VAS was 0.95 in the derivation study.
- The BVC-VAS was well accepted by the nursing staff and easily integrated into daily routine.
- The original BVC checklist showed stable performance across independent datasets.
Takeaway
This study created a tool to help predict if psychiatric patients might become violent, making it easier for staff to keep everyone safe.
Methodology
Two independent prospective cohort studies were conducted for instrument development and clinical application, involving newly admitted psychiatric patients.
Potential Biases
Potential biases may arise from the subjective nature of the VAS ratings and the influence of provided interpretations on staff assessments.
Limitations
The study design may have affected the performance of the VAS, and the small number of events limited statistical modeling.
Participant Demographics
The derivation sample consisted of 219 patients and the validation sample of 300 patients, with similar gender distribution and lengths of stay.
Statistical Information
P-Value
p = 0.002
Confidence Interval
95%-CI 0.90 to 0.98
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website