Sylvia Seegrist: Guilty But Insane

There were concerns about the fact that someone had been treated and released back into the community, considered no longer a danger to self or others. How many more might there be?

However, Seegrist's case defied this simplistic idea. She had a history of strange and aggressive behavior and had been through several treatment programs. She could not be treated against her will, although her parents had insisted to the court that she was not competent to make decisions about her own treatment.

In 1989, spurred by the Seegrist case, Congress approved a plan to revise the mental health system to allow involuntary commitment on the basis of clearly expressed threats to people or property. The bill also required training for caseworkers, more funding, and improved communication between state and local mental health services. Dissenting voices called it a "blueprint for a nightmare," unnecessarily overburdening an already overwhelmed system.

The real issue was whether mental health professionals could accurately predict who might be dangerous. Studies were launched to determine this. A key consideration in terms of how much an agency could actually step in was the tension between the rights of the state vs. the rights of the individual.

Cathy Young points to the case of Michael Laudor, a man with schizophrenia who had overcome many of the debilitating effects of his illness to graduate from law school. He became a hero to mental illness advocates, but when his drugs stopped being so effective and further stressors sent him spiraling downward, he ended up stabbing his pregnant fiancée to death. Could anyone have foreseen this?

The American Psychiatric Association has protested the unreliability of testimony that purports to be able to predict risk. Fortunately, risk assessment is improving with further research, and the predictions now utilize both clinical judgment and statistical data. The best predictions, however, are for short-term rather than long-term risk.

Mental health experts once used their best clinical judgment to determine whether someone was going to repeat his violent behavior if let out into the community. Those people would be committed involuntarily for their own good. However, research indicated that psychiatrists were right in only one out of three cases. That means that there were many "false positives"—people were committed who would not be violent—and "false negatives"—people were allowed to go free who then committed violence. That error rate was unacceptable.

In the 1980s, a number of studies were undertaken to develop instruments that would improve the percentage of correct assessments of dangerousness, and instead of focusing on dangerousness itself, they emphasized what they called "risk factors."

Interviews and inventories were developed to determine whether a defendant was a psychopath (which had a high correlation for recidivism), whether he was sexually deviant (another good predictor), how impulsive he was, whether he had a character disorder or mental illness, whether he had paranoid delusions, what his school record was, whether he had committed crimes as a juvenile, and what his past history of violence was. Out of these studies came guidelines for making predictions based on facts and logic rather than on intuition or psychoanalytic assumptions.

Risk management, i.e., devising programs that might help a person avoid repeating his crimes, focuses on factors that yield to intervention, such as substance abuse or paranoid delusions. What becomes important in risk assessment is the individual's social support, living arrangements, and access to treatment.

The idea of "dangerousness" has been a central issue in the legal/mental health arena for many years, yet establishing an empirical body of data from which to make accurate predictions has been difficult. The problems include the actual legal definition, confusing research literature, personal biases that creep into the decisions, and a professional's fear of responsibility and liability.

One case in 1981, Estelle v. Smith, indicates a real need for standards. This was a death penalty case in Texas. On the basis of a brief mental status examination, the state's psychiatrist testified that the defendant, Smith, was a "severe sociopath." Based more or less on an intuitive sense of the man's apparent lack of remorse for being an accomplice in a murder (but was not the killer), the doctor stated that Smith would certainly commit other crimes. The psychiatric assessment was poorly rendered and it brought about numerous protests from the mental health community that it was unethical and not representative of responsible assessment. 

According to those researchers who have devoted considerable time to the subject, risk assessment research on which judgments are to be made should meet seven criteria:

• "Dangerousness" must be segregated into component parts: risk factors, harm, and likelihood of occurrence
• A rich array of risk factors must be assessed from multiple domains in the offender's life
• Harm must be scaled in terms of seriousness and assessed with multiple measures
• The probability estimate of risk must be acknowledged to change over time and context
• Priority must be given to statistical research
• The research must be done in large and broadly representative samples
• The goal must be management as well as assessment

Despite protests by mental health groups over the years at the stereotype of the schizophrenic offender, in fact in recent studies there is a moderate association between current diagnosis of active symptoms of a major psychosis — especially one with paranoid delusions or poor thought control -- and violence in the community.  This risk increases with substance abuse and with the refusal to take medication. 

All in all, the mental health community is attempting to refine the methods for knowing when someone like Sylvia Seegrist might become dangerous.