Email | Print

The possibility that a person with a mental disorder may physically harm others is an issue in both civil and criminal mental health law in the United States and in many other countries, as well as a cause of public concern. A recent large-scale study of people discharged from psychiatric hospitals, conducted by myself and a number of colleagues as part of the MacArthur Foundation Violence Risk Assessment Study, adds a new perspective to the public debate over this risk. We found:

Patients discharged from psychiatric facilities who did not abuse alcohol or illegal drugs had a rate of violence no different than their neighbors in the community;
The prevalence of violence is higher among people who have symptoms of substance abuse;
People who have been discharged from a psychiatric hospital are more likely than other people living in their communities to have symptoms of substance abuse; and
Substance abuse tripled the rate of violence among non-patients and increased the prevalence of violence among people who have been discharged from a psychiatric hospital five times.
In civil law, involuntary mental hospitalization and, increasingly, involuntary treatment and intensive supervision in the community are often predicated on a clinical judgment of "dangerousness to others." In criminal law, involuntary treatment in a forensic hospital and, also increasingly, involuntary community treatment and monitoring turn on an assessment of undue risk of violence. Liability for clinicians who negligently assess risk, or who fail to take professionally appropriate actions to prevent assessed risk from being realized, has been the law in some U.S. jurisdictions for 20 years.
Despite legal mandates that violence risk assessments be routinely performed, a great deal of research conducted over the past 20 years suggests that the validity of such assessments is -- at best -- only modestly greater than chance. Older studies from the 1950s and 1960s tended to find that discharged patients were no more violent than others in the community. More recent studies from the 1970s and 1980s reported that discharged patients were more violent than their neighbors. Our study addresses these contradictions by recognizing the role of alcohol and drug abuse.

We have identified and taken pains to avoid four problems that severely limited existing studies: a limited range of risk factors being studied, weak measurement of violence in communities, restricted patient samples, and uncoordinated research efforts. The study was designed with three purposes in mind: to improve the validity of clinical risk assessment, to enhance the effectiveness of clinical risk management, and to provide information useful in reforming mental health law and policy.

In our study, nearly 1,000 male and female psychiatric patients between 18 and 40 years old were monitored for violence to others every 10 weeks during their first year after discharge. Patient self-reports of violence were merged with reports from "collaterals" (usually family members or friends) and police and hospital records. Members of a comparison group made up of more than 500 people living in the same Pittsburgh neighborhood as a group of discharged patients were interviewed once about violence in the past 10 weeks. The researchers also interviewed the group members' collaterals, and checked police records. The study was sponsored by the John D. and Catherine T. MacArthur Foundation and the National Institute of Mental Health.

We are now in the process of developing a new statistical risk assessment instrument that can help clinicians more accurately distinguish between "higher violence risk" and "lower violence risk" patients. We hope to help clarify ongoing policy debates and provide a tool for clinicians who care for individuals with mental illnesses and make decisions about hospital admissions and discharge.