Predicting Violence Risk in Mental Health Inpatients
Author Information
Author(s): Moscovici Mario, Farrokhi Farhat, Vangala Lavanya, Simpson Alexander I. F., Kurdyak Paul, Jones Roland M.
Primary Institution: Centre for Addiction and Mental Health
Hypothesis
Different clinical settings may benefit from different cut-off scores for the Dynamic Appraisal of Situational Aggression (DASA).
Conclusion
The DASA is a moderately good predictor of violent incidents in inpatient psychiatric settings, with proposed new risk categories that may improve its clinical utility.
Supporting Evidence
- 17% of patients had at least one violent incident during the study.
- 88,124 DASA scores were analyzed, showing a significant association with violence.
- 82% of violent incidents occurred within the first two weeks of admission.
- The AUC for DASA scores was 0.73, indicating moderate predictive validity.
Takeaway
This study looked at how well a tool called DASA can predict if patients in mental health hospitals might become violent. It found that the tool works pretty well, especially if we use new ways to categorize risk.
Methodology
The study included all consecutive mental health hospital admissions from 2016 to 2019, analyzing DASA scores and violent incidents using time-to-event analysis and Receiver Operating Characteristics.
Potential Biases
The retrospective nature of the study may introduce biases related to the reliability of DASA scores recorded by staff.
Limitations
The study only included the first violent incident and did not assess clinical interventions applied to high-risk patients, which may underestimate the association between DASA scores and violence.
Participant Demographics
The mean age of participants was 39.5 years, with approximately two-thirds male; over half had a diagnosis of schizophrenia spectrum or psychotic disorder.
Statistical Information
P-Value
p<0.001
Confidence Interval
HR 1.79 (95% CI 1.73-1.85)
Statistical Significance
p<0.001
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website