At some point, some well-intending person who wanted what we all want (less violence) promoted the hypothesis that people with a mental illness are more likely to be violent. I say hypothesis here because that's what it was, an idea, a suggestion, something that was not yet supported by evidence. The trouble with people sharing their hypotheses is that the media doesn't seem to understand the difference between hypotheses and facts. As a result, people in positions of power are allowed to share their ideas as though they are facts, and people encode them that way.
My soap box is in the shop this week, so I'll just get to the point. Recently, Dr. Eric Elbogen, University of North Carolina at Chapel Hill, and his colleagues started with the question, "What predicts acts of violence?" They hypothesized that having a serious mental illness (SMI) would be associated with acts of violence, but they also hypothesized that there were other factors that would be more strongly associated with acts of violence. These included being physically abused during childhood, witnessing domestic violence, having a parent that went to prison, level of education, sex (male or female), use of substances, tending to perceive people as threatening, tendency to feel anger toward others, being divorced, and significant financial distress.
To answer this question the research team used data from the National Epidemiologic Survey on Alcohol and Related Conditions (NESARC). In this study they conducted face-to-face interviews with more than 40,000 individuals in the United States. They conducted initial interviews in 2001-2003 and then follow-up interviews in 2004-2005 and ended up with data from 34,653 participants. In these interviews, participants answered questions about their mental health, including whether they currently (past 12 months) or had ever had schizophrenia, bipolar disorder, any psychotic disorder, or depression. These disorders comprise a cluster of disorders called a "Serious Mental Illness" or SMI. Finally, participants answered questions about whether they have committed acts of violence or physical aggression. Their main question was: what predicts whether a person will engage in violent behavior from the first assessment to the next.
The sample was 42% male and the average age of participants was 46 years. 20% of the people in this study had schizophrenia, depression, bipolar disorder, or a psychotic disorder either in the past 12 months or ever in their lives. That's 1 in 5 people. Some estimates even suggests that at least 1 in every 3 people will suffer from a psychiatric disorder at some point in their lives. I will come back to this point later. In terms of the other predictors, 17% of participants had not completed high school, 4% were physically abused as children, 11% were exposed to domestic violence, and 7% had parents with a criminal history. Between the 1st and 2nd assessments, 355 participants (1%) reported engaging in severe violence and 837 participants (2.4%) reported engaging in acts of physical aggression.
Now for the results. Indeed, individuals with SMI were more likely to report acts of violence than those without SMI. However, SMI was not the best predictor of violent behavior. The best predictors of severe violent behavior were: being male, having a parent with criminal history, being physically abused as a child, having a tendency to perceive others as threatening, and being a recent victim of violence. The best predictors of acts of physical aggression were: alcohol abuse, being physically abused as a child, being male, being recently divorced/separated, and having a parent with criminal history. Actually, the likelihood of engaging in violent behavior for these risk factors was 2-4 times the risk observed for having SMI. Even more importantly, when you account for these risk other factors, there was no association between SMI and violent behavior.
This study is meaningful, informative, and important in many ways, but like all science it has limitations. Most importantly, everything in this study was self-reported. So, the truth that is uncovered from the study is only as good as people are at reporting their mental health and violent tendencies. It's not unreasonable to think that some people may have lied when asked about whether they have ever engaged in violent behavior. This may be a necessary evil in this type of research because as soon as you use something objective, such as police records, the study loses the ability to recruit so many participants. Indeed, with a low prevalence outcome like violent behavior that occurs in only 1-2% of people, you need thousands of participants to study it.
The point here is that the community has come to believe that mental illness and violence are intimately linked, which just isn't the case. There is a small association between SMI and violence, but actually people with SMI are more likely to be victims of violence or be dangerous to themselves than anything else. The false belief that mental illness is associated with violence will not only keep us from actually preventing violence, but also comes at an incredible cost to society.
Like I mentioned earlier, mental illness will affect at least 1 in 3 people. This isn't someone else's problem, and this isn't about people you don't know. This is about you, your friends, your family, your colleagues, and your neighbors. When the media and the public adopt the false belief that mental illness = danger to the community, people who need help are stigmatized. Despite the high rates of mental illness, most people with mental illness (predominantly depression, anxiety, and substance use disorders) will never seek help. The average person will wait, and suffer, for 10 years before seeking help. Yet, mental health professionals know that we can effectively treat at least 65% of those cases within 3-6 months. Stigmatizing mental illness turns what could be a flu into a plague. So let's stop.
Elbogen, E. B., Dennis, P. A., & Johnson, S. C. (2016). Beyond Mental Illness Targeting Stronger and More Direct Pathways to Violence. Clinical Psychological Science, 2167702615619363.