Journal Logo



Violence and Mental Illness

Swanson, Jeffrey W. PhD; Special Issue Editor

Author Information
Harvard Review of Psychiatry: 1/2 2021 - Volume 29 - Issue 1 - p 1-5
doi: 10.1097/HRP.0000000000000281
  • Free

Interpersonal violent behavior by any measure is a top-tier public health problem in the United States, accounting for nearly 180,000 fatalities and an estimated 16 million nonfatal traumatic injuries treated in emergency departments over the past decade.1 Homicide ranks as the leading cause of death among black men under age 452 and is the third leading source of mortality among all U.S. adolescents and young adults.3 Violence is thus a pervasive and quotidian reality in many communities, but in recent years it has captured public attention at recurring moments of terror in our national life—when we hear the news of a senseless shooting rampage in a common space,4,5 and we learn that the shooter was mentally ill.6,7

Serious mental illnesses amount to a daunting public health challenge in their own right, of course. More than 11 million U.S. adults suffer disabling symptoms of schizophrenia or major mood disorders, and one out of three of them received no treatment at all in the previous year;8 an estimated 300,000 or more were arrested and booked into U.S. jails,9 mostly for miscellaneous minor offenses like trespassing and public intoxication.10 But what do these illnesses have to do with violence? What do we know, scientifically speaking, about any connection between mental illness and violent behavior, or about how these two species of trouble are not related? What do psychiatrists and other mental health professionals have to offer at the slim intersection of these quite different public health problems?

This special issue of the Harvard Review of Psychiatry addresses the vexing topic of violence and mental illness with a set of reviews and essays written from diverse intellectual and practice perspectives; the articles reflect both the multifaceted nature of the problem and the importance of interdisciplinary research to inform effective interventions and policies to try to solve it. The topics range from the pharmacological treatment of hostility in schizophrenia, to an innovative legal tool for clinicians to effectuate risk-based temporary firearm removal from patients in crisis; from the treatment of aggression in autism spectrum disorder, to the role of psychiatric research into mass shootings; and from the challenges of violence risk assessment in clinical and legal settings, to the intimate relational context of violence directed at family members and caregivers. What emerges across the broad range of these scholarly inquiries is a complex picture of intentionally inflicted human trauma—a socially conditioned health problem with many facets and contributing risk factors, but no unifying, theoretically derived or empirically validated explanation. Much work remains to be done.

The idea that an objective malady of the mind could “make” a person act violently is at least as old as the Biblical story of an “evil spirit” that overtook the Old Testament King Saul, filling him with unwarranted fear and causing him suddenly to hurl a javelin at his harp-playing protégé and future son-in-law, David.11 We have only to substitute “mental illness” for “evil spirit” in this ancient tale, and it finds resonance with a widely accepted modern narrative of psychosis-driven violence; we read into the story that it was Saul’s delusional perception of threat from David that enraged the king and animated his irrational attempt to kill his young friend. And given this simple formulation of the violence problem, the obvious solution is to remove its one causal agent—the “evil spirit”—by means of some effective intervention. Antipsychotic pharmacotherapy as a remedy for violence is thus the modern analog to exorcism.

But does this narrative of disorder-driven violence fit the evidence from scientific research, and not just the conceptual legacy of mind-body dualism from the ancient world? We have known empirically for three decades that adults residing in the community with diagnosable schizophrenia, bipolar disorder, or major depression are approximately three times more likely to engage in any violence, defined as physically assaultive behavior ranging from simple battery to injurious acts with weapons.12 But we have also known that a mere 3%–5% of violent acts occurring in the community are attributable to mental illness.13 Even if serious mental illnesses were to be cured altogether, more than nine out of ten violent acts would still occur.

Said differently, despite a significant relative risk of violence associated with serious mental illnesses, people with these conditions represent a very small subgroup of a larger population in which violence occurs, for a whole variety of reasons unrelated to psychopathology. Thus, the large majority of the perpetrators of violent crimes do not have a diagnosable mental illness, and conversely, most people with psychiatric disorders are never violent. The National Inmate Survey conducted by the U.S. Department of Justice identified likely current mental illness in only 17%–29% of incarcerated violent offenders.14 And a national study of 1410 schizophrenia patients found that only 3% had engaged in any recent violence causing injury or using a weapon, while an additional 16% had engaged in less serious assaultive behavior such as pushing or slapping someone without causing injury and with no weapon in hand.15

But if mental illness rarely causes violent behavior, then what does? What is the nature of violence, really? A young man shoves his girlfriend against a bedroom wall. An Army veteran shoots his estranged wife and two small children before turning the gun on himself. Two drinking buddies get into a fist fight in a bar. An urban gang member shoots a rival gang member over a drug-market turf dispute. A driver runs another driver off the road after being cut off in traffic. A member of an extremist political group sends explosive devices in the mail to prominent government leaders. An exasperated mother vigorously spanks her toddler for having a tantrum. A new mother drowns her infant child. A former psychiatric hospital patient punches his mother in the arm. Another former psychiatric hospital patient randomly shoots several strangers in a crowded movie theater. A young teenage boy spends hours every day enacting scores of virtual murders in a brutally violent video game.

Do the stark differences between these scenarios tell us more than their similarities? It would seem difficult indeed to explain all of these acts within a single, overarching theory of violence, and virtually impossible to prevent them with a common intervention. Insofar as psychopathology plays an important role in at least some of the examples, it stands to reason that effective psychiatric treatment could plausibly have prevented a number of these harmful acts. But the diverse scenarios also make clear that our understanding of any link between mental illness and violent behavior—and of the corresponding configuration of mental health interventions to mitigate violence and its sequelae—must be qualified and contextualized. To build such an understanding of violence requires a conceptual scaffolding of multiple dimensions: severity, lethality, relationships, social settings, mental states, instrumental and reactive motivations, age, sex, life-course events and development, and cultural norms.

When a person with a mental disorder commits a violent act, the symptoms of the illness itself often fail to explain why he or she did it; an analysis from the landmark MacArthur Violence Risk Assessment Study of discharged psychiatric patients found that only 12% of violent incidents were immediately preceded by psychosis.16 What, then, might have contributed to the violent behavior of the other patients? Maybe it was rooted in childhood trauma—the angry, resentful, and impulsive recapitulation of physically abusive behavior learned from a parent. Perhaps violence is a means to an end, borne of the demands of grinding poverty, having no regular home, or the need to survive in a predatory environment where violence marks the daily rhythm of life. Or violence could result from taking drugs that distort perception, disinhibit aggression, and destroy relationships, and whose procurement requires contact with dangerous associates. Some vectors of violence are rooted in individual development and personality and in learned patterns of action, others in the provocations of an unhealthy social surround, and still others in the neurochemical ingredients of impulsivity or intoxication. At the end of the day, interpersonal violence is a complex human behavior caused by many factors that compound each other and interact to increase or moderate risk.

There is much work remaining to fully comprehend the problem of interpersonal violence in all of its tentacular manifestations and contexts. A great deal of knowledge has already accumulated about violence and mental illness, of course, but there is a need for synthesis and cross-pollination across disciplines—from psychiatry and behavioral sciences, to criminology and the law, to epidemiology and public health. As an enhanced scientific understanding becomes incrementally available, we must also meet the ongoing translational challenges of testing and implementing multilevel interventions in the real world, and of determining who is most likely to benefit. Emerging, evidence-based solutions range from tailored therapies designed to modify aggressive behavior in individuals, to community-level programs to reduce the social and economic determinants of violence, to legal policies that restrict access to lethal means for people at high risk of causing harm to others.

Starting at the individual clinical level, two articles in our special issue provide comprehensive reviews of the state of research and practice regarding pharmacotherapies and behavioral interventions to mitigate hostility and aggressiveness—putative precursors to violence—among patients diagnosed with schizophrenia and autism spectrum disorder (ASD). Leslie Citrome and Jan Volavka17 review evidence for an anti-hostility effect attributable to treatment of schizophrenia patients with a range of atypical antipsychotic medications; they find “robust evidence” of such an effect for clozapine, followed by olanzapine, and weak evidence for other antipsychotics. David Im18 reviews research on pharmacologic and nonpharmacologic treatments targeting aggressive behavior in adults with ASD, summarizing evidence from controlled clinical trials suggesting benefits from a range of medications, including risperidone, propranolol, fluvoxamine, and dextromethorphan/quinidine, but also, to a lesser degree, from behavioral interventions such as multisensory environments. Among more traditional and holistic approaches, yokukansan—a herbal treatment used in Japanese Kampo medicine—and vigorous aerobic exercise appear to lessen aggression in adults with ASD, according to some studies.

As informative as they are, both of these reviews invite the same question: to the extent that various treatments can reduce hostility and aggression in patients with these particular disorders, how do the treatments work? Do the described therapies work indirectly by ameliorating core symptoms of illness that increase the risk of violence? In the case of schizophrenia, this could mean mitigating paranoid delusions or command hallucinations, and in ASD it could involve ameliorating deficits in such areas as social-emotional reciprocity, misinterpreting communication cues as threatening, rigid thinking patterns, and hyperreactivity to sensory input. Or, rather, do the treatments work by directly reducing persistent hostility and aggressive behavior—psychological states and behavioral patterns that could incline anyone to violence, even if they do not have a diagnosable psychiatric disorder. And considering evidence from both reviews together, what do we make of the suggestion that antipsychotic medications could reduce violence risk in patients with diagnoses as different as schizophrenia and ASD? What underlying causal model of violence could account for a common therapeutic effect in such a heterogeneous clinical population?

Im’s review18 addresses the question of underlying causes of violence by summarizing early evidence for a variety of neurobiological processes implicated in aggressive behavior and in different manifestations of aggression such as impulsive and reactive assaultiveness versus more deliberative or instrumental acts of violence. Some evidence implicates the dopamine system. High levels of aggressive behavior in young healthy subjects are correlated with a specific mutation in a gene for an enzyme involved with dopamine elimination, which increases stimulation in neural networks that regulate emotional arousal and impulsivity. Along these same lines, a study found that levels of a particular dopamine metabolite in cerebrospinal fluid were significantly lower in impulsively aggressive violent offenders with antisocial personality than in non-impulsively aggressive offenders with other personality disorders. Other evidence implicates the serotonin system. Im cites a study that found a significant correlation between neurotransmitter abnormalities in the serotonin system and particular types of aggression. In particular, the study compared levels of a serotonin metabolite in the cerebrospinal fluid of homicide offenders, discovering that those who acted impulsively had significantly higher levels than those whose crimes were premeditated.

Could impulsive aggression, at least, be explained at a biological level by variations in connectivity within different neural networks? If so, what might such explanations contribute to a larger, cross-disciplinary understanding of violence as a public health problem with social and legal contours? What might it imply regarding criminal culpability and sanctions, and also regarding the design of interventions to prevent violence and rehabilitate violent offenders? As intriguing as such questions might be, they outrun the current state of science. Im18 points out that many of these studies have been criticized for their small sample sizes and inability to rule out the possible effects of other psychopathology that correlates with these same kinds of neurotransmitter abnormalities. With respect to ASD, in particular, Im notes that “a well-studied and established mechanism for the development of aggression in adults with ASD has yet to be elucidated.” The same appears to be true for violence in schizophrenia—and for violent behavior in general, for that matter—leaving much room for development of neurobiological as well as behavioral, cognitive-emotional, and social-ecological theories of violence, and for well-designed empirical studies to test and integrate them.

Just thinking of schizophrenia and ASD together highlights the elastic and contested boundary of what is meant by “mental illness,” especially in the fraught context of assessing violence risk and attributing causality for violent acts. Schizophrenia is perhaps the paradigmatic mental illness in this regard, whereas autism is considered by many stakeholders to be something else: a neurological, developmental, or otherwise “medical” condition but not a mental illness. This tension is nowhere more apparent than in the autism community’s painful reaction when, in the frenzied and finger-pointing aftermath of a mass shooting, the national media reports that the shooter was “mentally ill” with the elaboration that he had been diagnosed at some point with autism or ASD.19 Whatever their different experiences and illness careers, people living with schizophrenia and with ASD share a common challenge that must be addressed along with other features of these disabling health conditions: the burden of stigma and scorn that flows from the public’s attribution of dangerousness to those affected.

Travis Labrum, Lisa Dixon, and their colleagues20 write with clarity and sensitivity about a complex set of challenges involving persons with serious mental illnesses who may engage in violent behavior directed at family caregivers and other relatives. These scholars review evidence suggesting that at least one out of five family members in close contact with a person with a serious mental illness was a victim of some form of violence from that person in the past year. Perhaps not surprisingly, studies find that family violence involving people with psychiatric disorders is associated with various risk factors that epidemiological research has identified in community surveys of violence and mental illness in general—factors such as co-occurring substance use disorder, nonadherence to prescribed medication, and experience with violence and victimization earlier in the life course.

Labrum, Dixon, and collaborators20 review other studies—some involving rich qualitative research—that illuminate particular relational dynamics that can cause conflict and raise the negative emotional temperature in families where an adult member has a psychiatric disability. These troublesome issues include financial dependence and perceived burden, family members setting limits, and living in a climate of persistent criticism and hostility. The review also offers hopeful evidence for promising strategies to prevent family violence in these situations, including engagement in recovery-oriented treatment, community-based services to lessen reliance on family members for financial and other tangible support, and specific conflict-management techniques.

Debra Pinals21 writes an informative column addressing the challenging task of assessing violence risk in clinical and forensic settings. She presents a series of brief case scenarios both to illustrate the complexities of the problem and to suggest how mental health practitioners can use formal and informal risk assessment to guide their clinical judgment. Pinals’s column shows how inherently difficult it is to get violence risk assessment right—balancing individual and social costs in the trade-off of false positives and false negatives—but also how important such assessments can be in guiding consequential decisions about hospital admission and discharge, civil commitment, and other matters. Doing what is best for patients while also protecting public safety is a weighty task, laden with uncertainty and made heavier by the often unrealistic expectations of legal actors in the psychiatric arena, other stakeholders, and the general public. Pinals’s essay and case-based scenario approach should be most helpful to clinicians facing real-world situations in which they must balance ethical and clinical concerns at the intersection of safety with coercion and fairness.

Jonathan Metzl and his collaborators22 write an eloquent Perspectives essay in which they tackle the “Rubik’s cube” of mass-casualty shootings in relation to mental illness. They outline a multipronged strategy to guide future research into the complex causes of these tragic events, and describe a broadened collaborative role for psychiatry in trying to prevent them. These scholars make a compelling case that mental health–oriented researchers in this area must deliberately reject the stigmatizing assumption that psychopathology is the main driver of a mass-casualty shooting. Rather, the destructive motivations of mass shooters must be situated “within larger social structures and cultural scripts.” Metzl and his colleagues also argue for a new alertness to potential racial bias in the social systems that officially define and measure the parameters of mass shootings, and how this may obscure features of the larger public health crisis of gun violence. They suggest that access to firearms—a key ingredient in the recipe for a mass shooting—must be understood more broadly within a framework that incorporates social, cultural, legal, and political, as well as psychological, aspects of private gun ownership in the United States. In the end, these authors urge psychiatry to align with complementary disciplines to seize new opportunities for funded interdisciplinary research into the many causes and consequences of mass shootings and gun violence—but to do so in a way that incorporates authentic community engagement, especially with respect to marginalized communities that are often most affected.

Finally, and turning to the policy arena of gun violence prevention, Swanson and colleagues23 present a Disruptive Innovation, arguing the case for state legislatures to authorize clinicians to directly petition courts for civil restraining orders to temporarily remove firearms from patients who pose an imminent risk of harming others or themselves. A growing number of states in recent years have enacted extreme risk protection order (ERPO) laws, providing law enforcement with the clear legal authority to remove guns—typically for up to 12 months—from people in the community who manifest behavioral indicators of imminent risk. Most of these laws provide an avenue for family members, too, to effectuate the temporary removal of firearms from an individual of concern for gun violence or suicide. However, only three jurisdictions to date allow clinicians to initiate ERPO petitions themselves in relation to their own patients.

Swanson and his collaborators23 discuss the potential advantages, disadvantages, and challenges in including clinicians as authorized legal petitioners for ERPOs. Would such a policy create a legal liability for psychiatrists and other health practitioners? If so, would statutory-immunity provisions allay such concerns and encourage the widespread use of ERPOs in health care settings? Could ERPOs undermine therapeutic relationships that depend on confidentiality, and if so, are such concerns outweighed by the potential of ERPOs to save lives? Definitive answers to these key questions await solid empirical evidence from the experience of states willing to put such a policy to the test.

Taken together, the articles in this special issue serve to summarize important facets of the complex connection between mental illness and violent behavior, and to illuminate the potential for mental health practitioners and researchers to play a more productive role in preventing violence. At the same time, and perhaps in spite of themselves, these reviews and essays illustrate how much remains to be learned, as they highlight gaps to be filled and questions to be answered by future studies.

Of equal importance to evidence-based practice is practice-informed research. The hope is that these articles will serve as a guide to future scholarship and scientific inquiry in this area, and that such endeavors will pursue answers to the right questions about violence and mental illness, which could then be applied to achieve two complementary goals: to meaningfully reduce the toll of violent injury and mortality in the population, and to safely and respectfully integrate people with mental illnesses into community life, with the acceptance and support that will allow them to thrive.


1. Centers for Disease Control and Prevention. Fatal injury and nonfatal injury reports, 2018.
2. Centers for Disease Control and Prevention. Leading causes of death, males, non-Hispanic black, 2017.
3. Centers for Disease Control and Prevention. Ten leading causes of death by age group, 2018.
4. Burns R, Crawford C. School shootings, the media, and public fear: ingredients for a moral panic. Crime Law Soc Change 1999;32:147–68.
5. American Psychological Association. Stress in America, 2019.
6. McGinty EE, Webster DW, Barry CL. Effects of news media messages about mass shootings on attitudes toward persons with serious mental illness and public support for gun control policies. Am J Psychiatry 2013;170:494–501.
7. McGinty EE, Webster DW, Jarlenski M, Barry CL. News media framing of serious mental illness and gun violence in the United States, 1997–2012. Am J Public Health 2014;104:406–13.
8. Substance Abuse and Mental Health Services Administration. National Survey on Drug Use and Health 2019: detailed tables.
9. Steadman HJ, Osher FC, Robbins PC, Case B, Samuels S. Prevalence of serious mental illness among jail inmates. Psychiatr Serv 2009;60:761–5.
10. Swanson JW, Frisman LK, Robertson AG, et al. Costs of criminal justice involvement among persons with serious mental illness in Connecticut. Psychiatr Serv 2013;64:630–7.
11. Swanson JW. On thoughts, threats, and throwing the spear. Schizophr Bull 2016;42:883–4.
12. Swanson JW, Holzer CE, Ganju VK, Jono RT. Violence and psychiatric disorder in the community: evidence from the Epidemiologic Catchment Area Surveys. Hosp Community Psychiatry 1990;41:761–70.
13. Swanson JW. Mental disorder, substance abuse, and community violence: an epidemiological approach. In: Monahan J, Steadman H, eds. Violence and mental disorder. Chicago: University of Chicago Press, 1994;101–36.
14. Bronson J, Berzofsky M. Indicators of mental health problems reported by prisoners and jail inmates, 2011–2012. 2017. National Inmate Survey, U.S. Department of Justice, Office of Justice Programs, Bureau of Justice Statistics.
15. Swanson JW, Swartz MS, Van Dorn RA, et al. A national study of violent behavior in persons with schizophrenia. Arch Gen Psychiatry 2006;63: 490–9.
16. Skeem J, Kennealy P, Monahan J, Peterson J, Appelbaum P. Psychosis uncommonly and inconsistently precedes violence among high-risk individuals. Clin Psychol Sci 2016;4:40–49.
17. Citrome L, Volavka J. Specific anti-hostility effects of atypical antipsychotics in persons with schizophrenia: from clozapine to cariprazine. Harv Rev Psychiatry 2020;29:20–34.
18. Im DS. Treatment of aggression in adults with autism spectrum disorder. Harv Rev Psychiatry 2020;29:35–80.
19. Park A. Don’t blame Adam Lanza’s violence on Asperger’s. Time 2014;11 Mar.
20. Labrum T, Zingman MA, Nossel I, Dixon L. Violence by persons with serious mental illness toward family caregivers and other relatives: a review. Harv Rev Psychiatry 2020;29:10–9.
21. Pinals DA. Violence risk assessment in clinical settings: enduring challenges and evolving lessons. Harv Rev Psychiatry 2020;29:90–3.
22. Metzl J, Piemonte J, McKay T. Mental illness, mass shootings, and the future of psychiatric research into American gun violence. Harv Rev Psychiatry 2020;29:81–9.
23. Swanson JW, Nestadt PS, Barnhorst AV, MD, Frattaroli S. Risk-based temporary firearm removal orders: a new legal tool for clinicians. Harv Rev Psychiatry 2020;29:6–9.
© 2021 President and Fellows of Harvard College