Risque de violence et troubles mentaux graves

In opposition to a widespread belief in psychiatry, mental disorders per se, without alcohol or drug abuse, represent a risk for violence against others, including homicide, that is significantly higher than that of the general population. The present paper extensively reviews the scientific literature during the past 15 years and present their findings in a summary table. The contradictory results of the notorious McArthur study are also presented and critically discussed. Elevated risks of assaults are more closely associated with a subgroup of patients showing one or more of the following characteristics: a history of previous violence; non compliance with antipsychotic medication and follow-up; alcohol and/or illegal drug abuse; violent fantasies; acute psychotic symptoms; and cerebral damage. Although violence associated with mental disorders is significantly higher than that observed among the general population, the absolute number of assaults committed by patients is still low. Between 85 and 97% of violent offenders are not mentally ill. Certain aspects of the deinstitutionalization, such as lack of outpatient resources and a modification of the law reflecting more concerns for the rights of the patients might provoke health care interruptions. However, a clinical commitment relayed by legal interventions in case of danger generate significant reduction of the risks for violence.