The Mental Illness-Violence Link

The recent shooting in Jacksonville has predictably renewed the public debate on violence and mental illness. The same sequence plays out after every horrible incident: First, there is the accusation, often from right-leaning lawmakers, that mental illness is what is driving our dizzying rate of mass shootings and that more needs to be done to protect the public from dangerous people. Soon following are emphatic statements from mental health advocacy groups that no real association between mental illness and violence exists and that those with mental health problems are far more likely to be victims of violent actions then perpetrators.

These arguments vigorously bounce around for a while on the news cycle and in social media, thoroughly confusing any well-meaning individual who is honestly trying to understand the issue, until the debate just fizzles out (until the next shooting).

Is there any way we can move this discussion forward to a place of real understanding? Actually there is, but it requires a little more time than a soundbite, and a shift from more dogmatic political positions to one that requires a bit of nuance and acceptance of complexity.

Working in this more complicated space, however, can be a little uncomfortable for everyone. I know it is for me. As a practicing psychiatrist, I cringe when officials call someone who commits an act of mass violence “psycho” or “sick,” and I’m disturbed by the speed at which mental illness becomes the go-to explanation for violent acts instead of other painfully obvious issues like gun access. Yet I am also aware how frequently questions of risk and aggression come up with the population we serve. Indeed, several official psychiatric diagnoses, such as antisocial personality disorder, intermittent explosive disorder, and conduct disorder, do include aspects of violence in their diagnostic criteria.

So what does the research actually show? Interestingly, there’s been good data on this topic for over 30 years, and its remarkable consistency stands in stark contrast to the debate that tends to persist at the extremes. Back in 1990, a large epidemiological study found that meeting criteria for schizophrenia, bipolar disorder, or major depression was related to a threefold increase in the rate of being involved in a minor or more major violent incident within the past 12 months—with substance use increasing the risk further. However, the vast majority of people (88%) did not act violently the past year, and the study calculated that only 4% of the violence perpetrated could be attributed to mental illness. Furthermore, there is some evidence that people with major mental illness alone, in the absence of factors such as poverty, exposure to neighborhood violence, and substance use, have levels of violence no different from the general population. More recently, in 2020, another important study was published from Sweden that found similar conclusions, namely:

• Most people with mental illness were not perpetrators of violence, but...
• People with mental illness were 4 times more likely to be violent than people without mental illness.
• Substance use made this link considerably stronger.

What this and other studies suggest is that all of the talking points described above in the debate about mental illness and violence are correct, sort of. People with mental illness are somewhat more likely to commit acts of violence—and most people who suffer from mental illness are not violent.

The articles also illustrate some more subtle but incredibly important angles that often get lost in the politically-driven debates and that can lead to further confusion. First, the term “mental illness” is probably too broad for this discussion, as the research indicates that different types of psychiatric disorders have different relations to violent behavior. Some, like psychotic disorders such as schizophrenia, are indeed related to higher levels, while others may be associated with reduced risk, or no risk at all. Second, risk of violence appears to be much higher when other factors such as substance use or antisocial personality are part of the equation. Here one might reasonably ask, however, aren’t these “other factors” not psychiatric disorders themselves? The answer is yes, but with an asterisk. One the one hand, substance use disorders and diagnoses such as antisocial personality disorders are very much parts of official lists of psychiatric conditions found in the Diagnostic Statistical Manual of Mental Disorders (DSM). At the same time, these kinds of diagnoses are often excluded in research studies looking at the association between mental health and violence and usually are treated differently by the courts when a person’s level of criminal responsibility is assessed.

Another important caveat is that most of this research has to do with violence defined quite broadly. Mass shootings may be altogether different, and for this other studies are required. In 2023, the Secret Service completed a report that reviewed mass attacks from 2016 to 2020. Among their many findings were that 34% had a history of drug use (with 12% considered to be actively under the influence when committing the act) and 58% experienced “mental health symptoms” prior to their attack; psychotic symptoms that were present in 28%. A little over half of the attackers had received some kind of mental health treatment in the past but this was “often not sustained.”

In total, there seems to be a compelling case to make that the commonly heard extreme positions stating either that people with mental health conditions are inherently dangerous or that mental illness is a non-factor when it comes to violent behavior are both wrong. The reality is somewhere in the middle. This is admittedly a somewhat awkward conclusion for mental-health and public-safety advocates alike, but ignoring this reality in the face of decades of quite consistent research serves little purpose other than to continue to divide people into polarized camps. From a public-policy standpoint, there seems to be little to justify more radical positions that people with mental illness should be locked away or automatically stripped of their 2nd-Amendment rights, any more than any other group that might be statistically more likely to act violently—like, for example, men. At the same time, well-intentioned groups that try to allay the considerable stigma already attached to mental health conditions by denying any link between mental illness and violence lose their credibility as accurate brokers of scientific information.

From my perspective, it seems abundantly clear that it’s beyond time to abandon both alarmist and dismissive positions on the link between mental illness and violence. The reality is more subtle, and its messaging has been made more confusing than it probably needed to be in part due to the mental health field’s own difficulty in coherently articulating what exactly counts as mental illness. But while appreciating these complexities might be difficult to capture in a tweet or soundbite, it very much can help provide a pathway toward meaningful, politically viable, and more scientifically-informed policy.