Ten years ago the MacArthur Violence Risk Assessment Study sought to understand the relationship between mental illness and perpetration of violence.

The principal finding—that there was no significant difference between the prevalence of violence by patients with mental illness who were without symptoms of substance abuse and the prevalence of violence by others living in the same neighborhoods who were also without symptoms of substance abuse—has been cited by mental health advocates as "proof" that the popular conception that mentally ill individuals are prone to violence is a myth.

Since then, additional articles, book chapters, and a book about the MacArthur study have appeared, adding new data about a topic that engages the attention of professionals, policymakers and the general public.

While many mental health advocates have championed the report for demonstrating that patients are not more dangerous than the general population, the study has also drawn critics. Among these are psychiatrist Fuller Torrey, M.D., and attorney Jonathan Stanley, J.D., who engage in a printed "debate" in this month's Psychiatric Services with MacArthur study authors John Monahan, Ph.D., and Henry Steadman, Ph.D., and their co-authors (see Major Problem Often Overlooked When Linking Violence, Illness).

Torrey and Stanley raised "six points of interest"—about methodology, interpretation, and generalizability of results—to which Monahan and Steadman responded in turn.

In the six points, and during an interview with Psychiatric News, Torrey emphasized that data about the value of treatment were not included in the original study—though they appeared later in book form—and that mentally ill individuals who receive and are compliant with treatment are not more dangerous than others in the population.

But he said that untreated mental illness is a matter of legitimate concern for public safety.

"The authors of the MacArthur study can't be faulted for the way some people have interpreted the results, but it has been popularly cited as evidence that the mentally ill are not more dangerous than the general public," Torrey told Psychiatric News. "In this way, some advocates hope to reduce the stigma surrounding mental illness.

"But in our view, it's the violence that causes the stigma," he said.

Torrey is executive director of the Stanley Medical Research Institute and president of the Treatment Advocacy Center, an organization that advocates for laws permitting involuntary treatment of patients with severe mental illness.

In the Psychiatric Services article, Torrey and Stanley pointed out that in the MacArthur sample of 951 psychiatric patients, three patients killed six people—a rate much higher than the 5.6 homicides per 100,000 in the U.S. general population.

Monahan and Steadman responded that selecting a few individual cases is not a systematic assessment of violence rates.

In an interview with Psychiatric News, Monahan reasserted his confidence in the study's findings, though he believes that some aspects of the study have been "overstated" by those advocating one or another policy position.

"The study has been cited by some advocates as proof that the mentally ill are no more dangerous than the general population, and as proof by other advocates that treatment reduced violence," Monahan said." We think both positions overstate what the study shows."

"What we found is that violence among the mentally ill is mediated by substance abuse," Monahan said. "We have confidence in the validity of our original findings."

Past APA President Paul Appelbaum, M.D., who was also a member of the MacArthur study group, said he believes interpretation of the MacArthur findings are typically a function of one's beliefs about how best to deal with the problem of stigma.

"There are some facts about violence and mental illness that everyone can agree on," Appelbaum told Psychiatric News. "People with mental illness account for only a small percentage of violence in this country—2 percent, maybe 3 to 5 percent. But no one has ever suggested that it might be more than 5 percent. That means at least 95 percent of the violence is not committed by mentally ill people. Some patients become violent but most do not.

"The rate of violence among the population of people who are mentally ill is greater than among the general population, and the risk of violence is greater with substance abuse," Appelbaum said. "Some studies have found that if you control for substance abuse, there is no difference, but other studies have done that and found that while they have some decline, it doesn't completely explain the variance.

"About these facts everyone conversant with the data can agree," he continued. "The disagreement comes from differences of opinion about how the findings should be used for policy purposes. Some people believe we can get more funding for treating mental illness by emphasizing the potential for violence. These people tend to believe that voluntary treatment for severe mental illness is not enough and that we need to have an increase in coercive methods."

Appelbaum said he was "dubious" about the long-term effect of emphasizing violence. "It increases stigma of mental illness and compels people to want to spend money on containment but not on treatment and reintegration."

Regarding the value of treatment, Appelbaum said he believed it was generally true that treatment diminishes the risk of violence.

"But you can't draw that conclusion from the MacArthur study, because the study wasn't a randomized, controlled trial," he said." Although people who got more treatment had lower rates, one can't rule out that what we are looking at is not causal. Those who are 'good actors' are more likely to be compliant, and those who are the 'bad actors'—who may be more likely to be violent—may also be less likely to be compliant."

Data from the MacArthur Violence Risk Assessment Study was first published in May 1998 in "Violence by People Discharged From Acute Psychiatric Inpatient Facilities and by Others in the Same Neighborhoods," posted at <http://archpsyc.ama-assn.org/cgi/content/abstract/55/5/393>.