An empirical analysis of specific syndromes of violent behavior

One hundred thirty-eight neuropsychiatric outpatients were rated according to relevant behavioral parameters of violence. A cluster analysis based on these ratings yielded five homogenous subgroups. Two groups, which differed only in severity of violence, were defined by a behavioral pattern closely resembling the episodic dyscontrol syndrome. One group was defined by infrequent but very severe violence, one by relatively infrequent violence of lesser severity, and one by no history of violence. These groups were clearly differentiated by a number of variables including neurological and psychiatric diagnoses, verbal aggression parameters, and neurological and psychiatric history variables. A specific etiology for violence, even within relatively homogenous subgroups, was not supported, with results suggestive of multiple determination by biological and psychosocial factors. On the other hand, different patterns of potential etiological factors were identified for different groups, which holds implications for theoretical understanding of violence and for differential diagnosis, treatment, and prevention. VioLit summary: OBJECTIVE: The aim of this study by Mungas was to describe the complexity of understanding violence and nonviolence through analysis of five homogeneous subtypes of neuropsychiatric outpatients. METHODOLOGY: A quasi-experimental, retrospective analysis was used to analyze 138 neropsychiatric outpatients who had been evaluated at the UCLA Neurobehavioral Clinic during an 18 month period in 1979-80. This clinic specialized in evaluation and treatment of patients manifesting concomittant behavioral, psychiatric, psychological, and neurological abnormalities or those with behavioral abnormalities where neurological dysfunction was suspected. Twenty percent of the patients had a primary complaint of violence; 35-40 percent had attacked property or persons, and 40 percent had no history of violent behavior and were referred to the clinic for neuropsychiatric reasons. This latter group served as a control group of psychiatric but not violent patients. Patients were dropped from the consideration if there was incomplete information. The author stated that this was not a concern because this was a retrospective analysis. The variable of violent behavior rating was defined as "overt physical behavior with clear-cut aggressive intent" and was measured with Likert scales for severity in four behavioral parameters--frequency, severity, appropriateness to the environmental content, and directedness (p. 355). These ratings were done solely by the author and based on thorough, retrospective chart reviews which included written accounts of patient histories from clinical evaluation, results from self-report questionnaires (administered by external observers familiar with the patients) asking about past history of violent behavior. Inter-rater reliability estimates were determined using product-moment correlation coefficient--frequency (.92), appropriateness (.75), severity (.78), and organization/directedness (.76). Ratings for verbal aggression were frequency (.85), severity (.91), and appropriateness (.79). Seven historical variables were rated based upon a thorough chart review; all were composite variables with the presence of several different historical variables contributing in an additive fashion to the overall score. These historical variables were history of behavior disorder as a child, history of development abnormalities, lifelong level of independent functioning, home environment as a child, drug abuse history, results of neurodiagnostic evaluation including EEG and CAT scan, and history of head trauma. Neurological diagnoses and DSM-III psychiatric diagnoses were recorded from the patient's chart. All diagnoses were recorded in the case of multiple diagnoses. The homogeneous subgroups of patients were empirically defined by submitting the ratings on the four parameters of physically violent behavior to a cluster analysis algorithm, the K-means clustering algorithm of the BMDP statistical package. MANOVA was used to analyze these groups once they had been identified. FINDINGS/DISSCUSION: The five groups were differentiated in the following ways. The first group (n=20) was characterized by high frequency, high severity, high inappropriateness, and relatively high organization. The second group (n=15) was characterized by low frequency, high severity violent behavior. The third group (n=15) looked like the first group except severity was more moderate. Group four (n=31) was characterized by relatively infrequent acts of mild to moderate severity that tended to be more provoked and better organized in comparison to the first three groups. Group five (n=57) consisted of nonviolent patients. The differences on the four behavioral parameters were significant (p= AUTHOR'S RECOMMENDATIONS: The author argued that more reseach is needed to clarify the relationship between research aimed at clarifying the relationship between particular behavioral symptoms and specific etiologies of violence. This will, it is claimed, add to the ability to more specifically diagnose and treat violent patients. EVALUATION: Different violent offenders are often treated as a single, homogenous group. However, this study shows that there are several distinct differences between types of violent persons that can be empirically verified, at least in this particular sample. The sample size is large but not very generalizable. This is a very specific group of individuals who are neuropsychiatric patients in Los Angeles, CA. There is clearly a need for a broader sample base in further research to see if distinguishing features still hold up. This study approached its question from a data-first-theory-later standpoint which allows for the questions to be asked, but further deductive study is needed to sort out the relationships and control for confounding factors. Association is helpful, but treatment will need to not only categorize patients but also know what led up to differences in violent persons. (CSPV Abstract - Copyright © 1992-2007 by the Center for the Study and Prevention of Violence, Institute of Behavioral Science, Regents of the University of Colorado) KW - California KW - 1970s KW - 1980s KW - Violence Causes KW - Biological Factors KW - Neurological Factors KW - Psychological Factors KW - Psychosocial Factors KW - Mental Illness KW - Mentally Ill Adult KW - Mentally Ill Offender KW - Mentally Ill Patient KW - Patient Violence KW - Patient Studies KW - Adult Offender KW - Adult Mental Health KW - Adult Patient KW - Adult Violence KW - Mental Health Patient KW - Offender Characteristics Language: en