The relationship between mental illness and violence is complex and
unsettled, and it is likely to remain so, even after publication of a new
study drawing on nationwide survey data.
Tragedies like recent mass shootings on university campuses, especially
when perpetrated by people described as formerly in psychiatric treatment,
connect violent acts with mental illness in the public's mind.
Yet Sally Johnson, M.D., a professor of psychiatry at the University of
North Carolina (UNC) at Chapel Hill, told Psychiatric News,"
the issue of dangerousness is more complicated than meets the
A decade ago the MacArthur Violence Risk Assessment study reported that
patients discharged from psychiatric facilities were no more likely to commit
violent acts than anyone else. Other scholars disagreed, and the published
record is divided. Some research concluded that there is a causal connection
between some mental illnesses and violent behavior while other studies say
that the link appears only when mental illness is accompanied by drug or
alcohol abuse, or other risk factors.
Now, Johnson and her colleague Eric Elbogen, Ph.D., an assistant professor
of psychiatry at UNC, have contributed a new study to the debate. Drawing on
data from the National Epidemiologic Survey on Alcohol and Related Conditions
(NESARC), the two researchers, both affiliated with UNC's Forensic Psychiatry
Program and Clinic, found that severe mental illness alone did not predict
"[P]eople with mental illness did report [engaging in] violence more
often, largely because they showed other factors associated with
violence," wrote Elbogen and Johnson in the February Archives of
They used NESARC data collected by the National Institute on Alcohol Abuse
and Alcoholism in two waves from people living in the United States in
2001-2003 and 2004-2005. There were 43,093 respondents in the first wave, and
34,653 who completed the second wave. The sample was weighted to reflect the
general population. Respondents were asked if they had lifetime or recent (in
the prior 12 months) diagnoses of a major mental disorder or substance
dependence, as well as if they had engaged in violence.
The NESARC data allowed researchers to look at respondents at two time
points and see if their earlier psychiatric status predicted later
About 10.9 percent of respondents said they had been diagnosed with
schizophrenia, bipolar disorder, major depression, or substance abuse disorder
alone, while 9.4 percent had co-occurring mental and substance abuse
disorders. The researchers also included other factors in their analysis.
These included dispositional factors (age, race, gender, education, income),
historical factors (parental criminal history), and contextual factors
(victimized in past year, divorced or separated, unemployed), as well as
The incidence of violent behavior was significantly higher among persons
with severe mental illness only when other factors—particularly
substance abuse or dependence—were present, reported Elbogen and
"[T]he current results show that if a person has severe mental
illness without substance abuse and a history of violence, he or she has the
same chances of being violent during the next three years as any other person
in the general population," they wrote. "Multivariate analyses
confirmed that severe mental illness alone did not significantly predict
committing violent acts; rather, historical, dispositional, and contextual
factors were associated with future violence."
Overall, this study is fairly consistent with previous reports, said
Renée Binder, M.D., a professor of psychiatry at the University of
California at San Francisco, who has also studied dangerousness.
"The study is strengthened because it is based upon data from more
than 34,000 individuals, but it has severe limitations," she said. Those
include the self-report format of the survey, the absence of personality
disorder or traumatic brain injury in the original survey, and lack of
information on the stage of mental illness during which violent behavior
"A person diagnosed with schizophrenia may have a greater risk for
violence when delusional, but much less when stabilized on medication,"
said Binder in an interview.
Elbogen agreed, noting that the NESARC survey was primarily designed to
gather data about alcohol use and consequences, but its size and the scope of
its questions could still be useful.
For one, it allowed Elbogen and Johnson to go beyond previous
cross-sectional research and make use of a survey that was both longitudinal
Binder also said the inclusion of "perceived threats of
violence" by a mentally ill person as a factor was valuable, since it
went beyond simple paranoia to include elements of the respondent's life that
posed genuine threats to safety.
"We started with no preconceptions in the context of the
dangerousness debate," said Elbogen. "We're trying to talk to both
camps, and we hope it helps clarify the discussion."
Johnson noted that there were potential policy implications to their
research, especially in an age when public mental health budgets are
"We have to look more closely at how we assume that treating people
for mental illness is the only or even the best intervention needed to help
them and to lessen the risk of danger," she said. "More broadly
based social interventions like job training and placement, family therapy,
housing, and so on might also lessen the risk of violence."
"The Intricate Link Between Violence and Mental
Disorder" is posted at<http://archpsyc.ama-assn.org/cgi/content/full/66/2/152>.▪