Mental Illness and Violent Death: Title and subTitle BreakMajor Issues for Public Health

Suicide and homicide are the fourth and fifth leading causes of death for persons aged 10 to 60 years in the United States.¹ Suicide completion and homicide perpetration are associated with previous attempts at self-harm² or violence toward others,³ and both of these predominant causes of violent death are associated with mental illness.

In 2002 there were 31 655 suicide deaths in the United States for a rate of 10.99 per 100 000 population.¹ Persons who attempt suicide are 38 to 40 times as likely to commit suicide as are persons without previous attempts.⁴ The population-based National Comorbidity Survey Replication, conducted in 2001-2003,⁵ reported that 3.3% of US residents aged 18 to 54 years had seriously thought about killing themselves in the past 12 months. Of this group, 28.6% made a plan to kill themselves and 32.8% of those who made a plan carried out a serious attempt to commit suicide. More than 80% of persons reporting these suicide-related behaviors met Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, criteria for mental illness, including mood disorders, anxiety disorders, impulse-control disorders, and substance use disorders.

There were 17 638 homicide deaths in the United States in 2002 for a rate of 6.12 per 100 000 population.¹ We were unable to identify population-based estimates of mental illness and homicide offending for the United States, but a population-based study of homicide offenders in Sweden who had received multidisciplinary psychiatric evaluations found that 54% had a principal or secondary diagnosis of personality disorder, 47.5% had a principal or secondary diagnosis of substance use disorder, and 25.2% had schizophrenia, bipolar affective disorder, or other psychoses.⁶ Persons with substance use disorders are 12 to 16 times more likely than persons without substance use disorders to engage in violent behavior.⁷ This association is strengthened among persons with co-occurring personality disorders (such as antisocial personality disorder) or major mental illness (such as schizophrenia). A review of clinical risk factors for violence⁷ identified 4 personality dimensions associated with violent behavior: poor impulse control, problems with affect regulation, threatened egotism or narcissism defined as an inflated sense of self-worth and entitlement, and paranoid cognitive personality style.

It is clear that mental disorders, including substance use disorders, are common among persons who demonstrate suicidal or violent behaviors. These data confirm that some persons with mental illnesses may be at risk of harming themselves or others. However, mental illness is also common in the general population. A recent report from the National Comorbidity Survey Replication⁸ estimated that 26.2% of US residents had anxiety, mood, impulse control, or substance use disorders in the previous 12 months. Of these, only 41.1% had received some sort of treatment in the past 12 months, including 12.3% treated by a psychiatrist, 16.0% treated by a nonpsychiatrist mental health specialist, and 22.8% treated by a generalist clinician.⁹ The obvious public health question is: If untreated mental illness is so common, how can mentally ill persons at greatest danger to themselves and others be identified and steered into treatment?

Two articles in this issue of JAMA suggest that many of these high-risk persons can be identified by their clinical and criminal records. Brown et al² recruited patients who had attempted suicide from medical and psychiatric emergency departments for a randomized controlled trial. Participants offered cognitive therapy were half as likely to reattempt suicide as participants who received usual care, an effect attributed by the authors to reduced severity of depression and hopelessness. Suicide attempts are much more common than completed suicides,⁵ presenting opportunities for medical intervention.

Also in this issue of JAMA, Cook at al³ estimate that most homicide offenders have previous arrests, many for violent crimes. These encounters with the criminal justice system are opportunities for offenders to receive substance abuse and other mental health treatment services in conjunction with criminal justice sanctions, such as imprisonment, probation, or diversion from punishment to treatment. Treatment programs “leveraged” by criminal justice sanctions are widely available, but evaluations of their effectiveness in preventing relapse or recidivism are methodologically flawed.¹⁰ Moreover, almost one third of homicide perpetrators have never been arrested and almost two thirds have never been arrested for a violent crime. However, other opportunities exist to identify them. Perpetrators and survivors of violence are often seen in emergency departments¹¹ and trauma centers, and perpetrators also may be identified in civil protection orders.¹²

Identifying persons at risk of violence to themselves or others and offering or compelling them to receive mental health treatment services is warranted. Barriers to delivery of these services may be financial, such as lack of access to health care; structural, such as the lack of mental health programs and practitioners; or personal, such as concerns about confidentiality or discrimination.¹³ Another barrier is scientific—lack of randomized controlled trials of therapeutic interventions for suicide and interpersonal violence prevention to guide clinical and systems management.^{5 ,10}

Other strategies for violence prevention address the lethality of weapons¹⁴ or social, pharmacological,¹⁵ and other situational factors that may lower the threshold for violent ideation to progress to violent action. Although these factors are late in the causal pathway from mental illness to violence, they are no less important for the prevention of violent death. A severely anxious, depressed, impulsive, or hopeless person whose violent actions are interrupted before he or she harms himself or herself or someone else would be considered a good outcome from the perspective of violence prevention. But without effective psychiatric treatment, such a person will still be in distress and is likely to remain at risk for violence. From this public health perspective, society should devote adequate resources to developing and evaluating psychiatric treatments and lowering barriers to their delivery.