Despite any public perception fuelled by pop culture and the media that violent offenders tend to carry a history of mental illness that leads to anti-social behaviour, a new study suggests that the ties between psychosis and violence are actually very weak.

According to researchers at the University of California, Berkeley in the US, only a distinct minority of violent incidents committed by psychiatric patients are preceded by symptoms of psychosis. They say that if we want to prevent criminals from reoffending, we should look at more generalised anger management strategies rather than focusing on specialised treatments for mental illness.

The researchers looked at data collected by the MacArthur Violence Risk Assessment Study - a comprehensive three-year assessment of more than 1,000 patients discharged from acute psychiatric facilities during the 1990s. After being discharged, the patients were interviewed every 10 weeks for one year to monitor whether they were involved in any violent incidents, including assaults that resulted in physical injury, sexual assault, use of a weapon, or threats with a weapon.

The MacArthur study also tracked whether psychosis preceded the violent incidents reported and collected information on other aspects of the patients' psychological profiles including personality traits, disorders, cognitive abilities, and problem behaviours.

The researchers found that the MacArthur data corroborated what experts already know about violence: a small percentage of repeat offenders are responsible for most violent acts. In the data, approximately 10 percent of individuals were responsible for half of the violent incidents committed by the participants in their first year of discharge from psychiatric facilities.

Of these repeat offenders, half reported experiencing episodes of psychosis (such as delusions or hallucinations) during that year, but only 12 percent of violent incidents were preceded by symptoms of psychosis, and only one in 20 violent reoffenders experienced psychotic episodes prior to committing violent acts. Clearly, the data suggests that any belief we may have that most violent acts are fuelled by occurrences of mental instability is out of whack with what's really going on.

The findings, which are published in Clinical Psychological Science, conclude we need to look outside purely psychosis-related patient management to help rehabilitate violent offenders.

"Effective treatment of psychosis will have negligible direct effects on violence for most patients and important but partial effects for the remainder," the researchers write. "These findings suggest that psychosis sometimes foreshadows violence for a fraction of high-risk individuals, but violence prevention efforts should also target factors like anger and social deviance."