Are people with psychiatric disorders violent?

“Guns don’t kill people, the mentally ill do.”

Ann Coulter, social and political commentator, Jan. 16, 20131

“Mental health is often a big problem underlying these tragedies.”

Paul Ryan, Speaker of the US House of Representatives, Feb. 15, 20182

“…sicko came to a school with bad intentions.”

Donald Trump, US President, Feb. 22, 20183

Are people with psychiatric disorders violent? The relevant data are influenced by reporting biases (e.g., changing willingness to report assaults) and the varying likelihood of being caught, convicted, and diagnosed, but there is compelling evidence that those with mental health problems are more likely than others to commit violent crimes. Analyses of the population-wide Swedish National Crime Register suggest that, compared to the general population, violent crime rates are doubled for patients with schizophrenia and bipolar mood disorders and tripled for those with unipolar depression.4–6 These are nontrivial differences triggering attention-grabbing headlines and fears of people with mental illness.7,8

The converse question is worth asking also: Are violent individuals more likely to have psychiatric disorders? Here again, the data are influenced by multiple factors, but most evidence suggests that they have elevated rates of psychopathology, with mood and psychotic disorders occurring in prison populations at up to 4 times that seen in community samples.9,10

It is informative to consider these figures in context. For example, homicide rates in Manitoba are double the Canadian average. Homicide rates in the United States are triple those in Canada.11,12 Despite this, few would take seriously headlines fearfully declaring “American Tourists Swarming Our Town!” Indeed, people with mental illnesses are 10–20 times more likely to commit suicide than homicide, and they are more likely than most to be victims of crime.13 These numbers provide yet more reasons to deliver adequate care to those with mental health problems. They also indicate that fears of the mentally ill are no better founded than fears of Manitobans.

What about personality disorders?

Almost by definition, people with antisocial and related personality disorders commonly exhibit socially objectionable behaviours.14 Despite this, personality disorder diagnoses provide poor predictive value for violent behaviour. The two best predictors are actuarial strategies,15 which focus on diverse demographic and behavioural histories, and use of the HCR-20 (Historical, Clinical, Risk Management), a 20-item checklist designed to estimate the probability of future violence.16 In common practice the HCR-20 is administered by clinicians and given in conjunction with the Psychopathy Checklist (PCL). Perhaps surprising to some, the predictive value of the HCR-20 is not diminished when it is administered without item H7 (personality disorder) and the PCL.16

Can levels of violence be lowered by treating more psychiatric patients?

Recent mass shootings in schools have prompted calls to increase mental health care funding as a strategy to decrease violence. If the money becomes available, much good might be accomplished. However, no more than 3%–5% of interpersonal violence is attributable to serious mental disorders.17–19 Paranoid, depressive and grandiose personality features are seen in some mass murderers, but these events are rare, accounting for less than 1% of homicides; there is no compelling evidence that the perpetrators display elevated levels of more serious pathology.8 In the absence of these associations there are better reasons for improving mental health care, including those that do not stigmatize patients.* There are also better strategies for reducing violence.

How can violence levels be lowered?

As a start, it might help to remind ourselves that homicide rates (the most reliable of violence figures) are at historic lows — lower than two decades ago, lower than two centuries ago, lower than five centuries ago.11,12,20 To decrease interpersonal violence further, greater access to mental health care can be one component, but it should not be the only one. Other factors that at least covary with high homicide rates (and, often, mental illness) include low average age of the community, high levels of unemployment, high income inequality gaps, low proportions of immigrants, poor living conditions, and cultural norms that valorize violence as a response to frustration.8,21–24 Effective strategies are challenging to identify, but include adequate neighbourhood policing accompanied by fast and fair sentencing, early childhood education programs, more equitable access to socioeconomic opportunities, and greater support for both victims and perpetrators.25–28

The role of clinicians is challenging too. The ability to predict violent behaviour is poor, and policies that putatively promote public health can harm individual patients.† This noted, some things help. Accidental shootings and suicide rates can be decreased by gun safety practices (e.g., safe storage),19,29 and clinicians working in communities with high gun ownership levels might find themselves in a position to advise. In the rare case when a patient expresses violent intent toward a specific person, clinicians in the United States have a legal obligation to report it.30 In Canada, reporting is required by many professional organizations, though not by the criminal code.31

Gun control laws have a role to play too, and they receive much public support. Unfortunately, many have been poorly designed and implemented, sufficiently so that their efficacy is less than might be supposed, particularly when compared to what could have been accomplished if the expended money and effort had been directed elsewhere. This noted, there is modest evidence that some laws can help. This includes high-quality background checks that focus on risk (e.g., history of impulsive angry behaviour) rather than mental illness, temporary restrictions of gun access during periods of crisis (e.g., family member–initiated, court-mandated gun-violence restraining orders), and banning firearms in settings where the possibility of violence is elevated (e.g., bars, demonstrations, holidays, weekends).19,23,32–34 “Stand your ground” laws, in comparison, can increase the risk of homicide.34 Each of these contributing factors is aggravated by substance use,35,36 though in at least some populations, this association might be driven by pre-existing behavioural tendencies.37

Conclusion

Violent behaviour reflects the confluence of many, often intricately interacting, factors. Despite this complexity, the steady decrease in homicide rates provides optimism that progress can be made. Many of the contributing factors are within the domain of psychiatry. This includes obtaining a better understanding of biological, psychological, legal and other sociocultural factors that influence problematic behaviours, and using this information when making decisions about patients and policy. Our communities will be best served, it is proposed, if we focus on these features while avoiding the temptation to use fears of the mentally ill to obtain more funding.