No easy way to tell which patients may be violent

Physicians know it's their duty to report patients who threaten to harm others. What, if anything, can they do when they only suspect a patient might harm someone?

Reply: Concern after the tragic shootings in Tucson, Ariz., reflects deep-rooted public belief that such tragedies should be preventable. Since people early in the course of serious mental disorders are more likely to be seen by primary care physicians than by psychiatric specialists, it is reasonable to ask whether primary care can play a role in identifying and intervening in these cases.

Research suggests that schizophrenia and other psychotic disorders increase the risk of violent behavior, particularly early in the illness. Even so, only a minority of people with such illnesses commit acts of violence. Conversely, only about 3% to 5% of violence in our society is due to serious mental illness. Thus, eliminating violence committed by people with mental illnesses still would leave society with almost all the violence it now experiences.

Research on the predictors of violence among people with mental illnesses has identified at least two groups of patients associated with differing patterns of behavior. One subgroup tends to be young, male and substance abusing, with unstable personal histories, early onset of antisocial acts and records of previous violence. Their violence is likely to be independent of symptoms of psychosis and not particularly responsive to treatment. A second subgroup without an antisocial history displays violence more closely linked to positive psychotic symptoms such as delusions and command auditory hallucinations; negative symptoms of psychosis (e.g., social isolation, apathy) may lower violence risk. Treatment may help reduce violence risk in this group.

Identification of variables that heighten risk of violence does not imply that accurate prediction is possible in individual cases. After decades of over-predictions that often led to prolonged involuntary hospitalization, progress has been made in developing structured assessment tools based on variables known to correlate with violence. Achieving reasonable degrees of accuracy, however, appears to be possible only when classifying patients into approximate levels of risk (e.g., low, average and high), with false positive predictions more prevalent as the degree of predicted risk increases.

Early prediction of violent behavior is complicated by the turmoil often associated with late adolescence and early adulthood, periods of peak incidence for some of the most severe conditions. Young people adopt alternative lifestyles, often involving extensive drug use, rejection of social norms and endorsement of extreme views about society. Although these behaviors indicate a need for help in completing the developmental tasks of this life-stage, only a small number of these people will develop serious mental disorders -- and even fewer will commit an act of violence.

It is unrealistic to ask busy primary care physicians with no specialized training in violence assessment to identify those rare patients who will commit horrendous acts of violence. As much as we would like to prevent such tragedies, there is no reason to believe we can identify such people with an acceptable degree of accuracy.

That is not to say that primary care physicians have no role to play. They can spot someone who is having a hard time and refer him or her to a mental health professional before the more extreme manifestations of illness appear. This is likely to be useful whether the behaviors are the early signs of serious mental disorders or just reflect a need for help with the difficult task of adjusting to adult life. And preventive treatment may help deter the rare person who otherwise might commit a serious act of violence.

Paul S. Appelbaum, MD, Dollard Professor of Psychiatry, Medicine and Law; director, Division of Psychiatry, Law and Ethics, Columbia University/New York State Psychiatric Institute

Reply: Since primary care physicians make up the wide base of the health care delivery pyramid, we see more patients with mental illness than psychiatrists do. Unfortunately, identifying mentally ill patients at risk for violence is difficult. Most never make it to a physician's office. Many men come to the doctor only if "forced" to by family or a friend. An ongoing relationship helps one identify mental health problems over a few visits as physical symptoms suggest distress or mental disorder rather than other forms of disease. But antisocial or substance-abusing patients may not follow up.

Family members frequently notify a primary care physician of mental health symptoms. The violent patient may not have a significant other to provide information. A person might not feel empowered to report on the patient's behalf or might fear reprisal.

Though public awareness of mental health disorders has increased, most people still believe mental health disorders reflect character flaws, weakness or lack of faith. This stigma makes it hard for many men to acknowledge a mental health problem. Men prone to violence may be more wary of disclosing any sign of vulnerability.

Younger patients with early onset antisocial behavior (usually male, substance-abusing) often come to light outside the family for the first time in schools and the criminal justice system rather than the doctor's office. Antisocial behavior alienates the people -- concerned adults with the time and organizational skills needed -- who might have helped the patient access health care resources: health insurance, physician services, appointments and transportation.

If a patient with a pattern of problems or an event of particular concern comes to the attention of a primary care physician, finding an affordable psychologist or psychiatrist is a major hurdle. Many insurance programs do not cover mental health care, or they require a high patient co-pay, excluding many patients from lower and middle socioeconomic groups. If the patient is insured, finding a covered psychiatrist in the area with open appointments is the next hurdle. The national shortage of psychiatrists puts our most vulnerable (and potentially dangerous) patients at risk.

Many mental health disorders are thought to involve a combination of genetic vulnerability and environmental triggers. The media unveiling of the pervasive presence of male sexual abuse in society may be part of the answer. While girls who are sexually abused often respond with depression, anxiety and dissociation, boys often respond with aggression and "acting out." Reducing the number of adults damaged in childhood who, in turn, inflict damage through antisocial and criminal behavior should be a national health priority.

Society is sailing into a perfect storm for tragedies like the Tucson killings. Significant barriers to physician identification of mentally ill people at risk for violence include: challenges of economic and logistical entry into the health care system, social isolation, stigma of mental health diagnoses, shortage of psychiatrists, and social factors that induce mental illness and distress.

Rachel Bramson, MD, senior staff physician, Scott & White Clinic; associate professor, Dept. of Family and Community Medicine and Dept. of Humanities in Medicine, Texas A&M University System Health Science Center