Printer Friendly

Advances reported in predicting violence.

Advances reported in predicting violence

When the courts recently decided thatpresidental assailant John Hinckley Jr. should not be allowed to lease a psychiatric hospital to visit his family, the key piece of evidence was that Hinckley had written letters to convicted serial murderer Theodore Bundy. The question remains: "If Hinckley had not written those letters, would he have been considered less dangerous to the community? Perhaps, but the fact that psychiatrists can only guess at such things points up how little is known, even among "experts'

"Psychiatrists should get out of thelegal prediction system,' says psychiatrist Antonio Convit of New York University (NYU) Medical Center in New York City. Perhaps because of the great fear of turning loose someone who will go on to assault or kill, behavioral scientists historically have erred on the side of "false positives'--predicting someone will become violent who, in actuality, does not. This has been a primary failing of hospital- and community-based predictive studies in recent years, according to Convit and others.

But Convit and his NYU colleagues nowreportviolence-prediction results that he says are among the "strongest' to date. Convit is quick to point out that the studies were done in a closed psychiatric ward, where there is little contact with outside society. Nevertheless, if this type of predictive model could be adapted to include the "risks' of the outside world, such as drugs, social stresses and homelessness, it might conceivably be applied to neighborhoods or communities, Convit suggested last week in Chicago at the annual meeting of the American Psychiatric Association.

In a three-pronged study, Convit andhis colleagues first evaluated 69 patients on the "violent ward' at NYU's Manhattan Psychiatric Center and compared them with 40 nonviolent controls elsewhere in the hospital. This yielded four key "risk factors': conviction for violent crime; history of violent suicide attempt; a measurable (although not necessarily dramatic) neurologic abnormality; a "deviant' family environment, including a broken home, child abuse or a parent's substance abuse.

Then, in a preliminary predictive test,the researchers devised a scale of such factors and applied it retrospectively to 51 violent and nonviolent patients. They identified 35 of the patients correctly, Convit reports, with only eight false positives.

Finally, the researchers applied the testprospectively to 79 male schizophrenic patients admitted to the hospital. In a three-month follow-up of these patients (80 percent of violent hospital incidents take place during the first three months, according to the researchers), Convit and his colleagues were correct in predicting violent behavior in 52 of the cases; the results included 19 false positives.

These results are striking, Convit says,particularly since psychiatrists have historically been right in violence prediction only 25 percent of the time; and the majority of those errors have been false positives. And of the few studies that have yielded statistics comparable to NYU's, he says, none has done so with just four predictive variables--a great advantage in applying it to many people at a small cost. In addition, built into the NYU studies was a complex "odds ratio' factor, which takes into account the relative risk of a population--something lacking in many other studies, Convit says. There is a long way to go, says Convit, before psychiatrists can achieve the same degree of accuracy in predicting the actions of a released patient. But these studies, he says, show that "prediction of violence on a psychiatric ward can be done.'
COPYRIGHT 1987 Science Service, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1987, Gale Group. All rights reserved. Gale Group is a Thomson Corporation Company.

Article Details
Printer friendly Cite/link Email Feedback
Author:Greenberg, JOel
Publication:Science News
Date:May 23, 1987
Previous Article:Coming: the big chill?
Next Article:Schizophrenia: new hope from an old drug.

Related Articles
10 activists or terrorists? Judge weighs arguments.
Back off veto threat.
ARABS-ISRAEL - May 22 - If Rockets Don't Stop We Will Not Stop, Says Israel.
AFGHANISTAN - May 20 - Suicide Blast Kills 10 In Afghanistan.
LEBANON - May 20 - Lebanon Clashes With Islamic Militants.
LEBANON - May 21 - Lebanese Army Pounds Refugee Camp.
New program focuses on preparing and trying nursing home cases.
Ten years of summer reading success: the eastern and central reading encouragement and development network (E.C.READ'N).
Risk factor: throat cancer linked to virus spread by sex.

Terms of use | Copyright © 2017 Farlex, Inc. | Feedback | For webmasters