Mental Illness and Violence: Proof or Stereotype?

Previous | Table of Contents | Next

Appendix A: Annotated Bibliography cont'd

Studies of Psychiatric Patients cont'd

Holcomb, W.R. and Ahr, P. R. (1988). Arrest rates among young adult psychiatric patients treated in inpatient and outpatient settings. Hospital and Community Psychiatry, 39(1), 52-57.

Purpose or Objective: The purpose of this study was to determine the prevalence of criminal behaviour among a statewide sample of young adult patients with major psychiatric diagnoses who were treated in both outpatient and inpatient settings, and to identify factors contributing to arrests.

Research Design: Although the authors did not specify, the data collection procedures are consistent with a retrospective cohort design.

Setting: Inpatient, outpatient, and community residential care settings in Missouri, United States of America.

Study Subjects: Subjects were a random sample of 611 young adult patients from a total of 32,000 who: (1) had one of five major diagnoses (drug or alcohol abuse, schizophrenia or other psychosis, major affective disorder, personality disorder, organic brain syndrome); (2) were between the ages of 18 and 35 years; and (3) received services during 1982. At last contact with a clinician, 51% were inpatients, 47% were outpatients and 1.7% were admitted to community residential facilities. The average number of prior psychiatric inpatient admissions was 3.27.

Measures: The Missouri Highway Patrol Department was used to collect information on history of arrests. This database included a record of arrests within the state and from all other states. Arrests for violent arrests were distinguished from non-violent arrests. Level of functioning was assessed by the clinician at last contact. Ratings were obtained for 327 (53%) of the total sample.

Main Findings: Thirty-eight percent of the sample were arrested at least once as adults. The data showed that 19% of the sample committed only non-violent crimes, 4.4% committed violent crimes, and 14.5% committed both. Significant differences were found by diagnosis. Patients diagnosed with alcohol or drug abuse were more likely to be arrested for crimes. There were no statistical differences in the association between diagnosis and violent crimes. The relationships of various demographic and mental health variables to the 13 specific arrests categories were examined with correlation coefficients. Few significant relationships emerged.

Conclusions: The authors concluded that the  restriction to young adults (18-35) and the inclusion of outpatients probably underestimated the lifetime prevalence of arrests. The authors argued that public mental health and criminal justice administrators must put a high priority on programs to address the needs of this young adult group. The authors claimed that the arrest rate for their sample was 17 times greater than the arrest rate for the same-aged general population.

Methodological Critique: There was a problem with missing data. Level of functioning was assessed at last contact from the clinician but only 53% of the clinicians were able to complete the study form. A further 167 forms were sent back incomplete because the clinician could not recall the patient. In addition, 62 forms were returned because the clinician was no longer employed at the facility, and 55 forms were not returned. This level of missing data is unacceptable and makes the findings uninterpretable. Even the forms that were returned relied on the clinician's memory and it is unclear whether the clinician would have classified patients correctly. Finally, an excessively large number of statistical comparisons were conducted with the result that the few significant ones are difficult to interpret because they could have occurred by chance alone.

Causality: A causal inference cannot be made on the basis of these findings.

Kennedy, M.G. (1993). Relationship between psychiatric diagnosis and patient aggression. Issues in Mental Health Nursing, 14, 263-273.

Purpose and Objective: The purpose of this study was to examine the relationship between psychiatric diagnosis and patient aggression within the context of routine intrahospital relocation.

Research Design: This study was a secondary analysis of existing data. While the authors fail to mention the study design, the data collection procedure is consistent with a historical follow-up study.

Setting: The original research was conducted at a large state hospital in the Pacific Northwest, United States of America. At the time of the study, there were over 1,000 patients in residence, with an average of 250 admissions per month.

Study Subjects: The records of 201 patients admitted over a five-month period were selected for review. Subjects ranged in age from 19 to 96 years. There were 108 (53.7%) males and 93 (46.3%) females. Caucasians made up 93.5% of the sample. The number of previous admissions to the state hospital ranged from one to 22 (mean = 3.4).

Measures: Data from patient records were collected using an adaptation of the Overt Aggression Scale (OAS). The OAS divides aggressive behaviour into four categories: verbal aggression; physical aggression against objects; physical aggression against self; and physical aggression against other people. Inter-rater reliability of the adapted OAS was tested by having a second rater use the  same scale to rate 28 of the 201 patient records. A Pearson Correlation of .77 was obtained between the two raters indicating good interrater reliability.

Main Findings: The total number of aggressive incidents recorded for the sample was 2,555. The majority of these were categorized as verbal aggression and physical aggression against other people. No significant differences were noted among DSM-III diagnoses in total patient aggression or the individual scale scores of physical aggression toward self, objects, or verbal aggression. The only significant difference noted was between individuals diagnosed with paranoid schizophrenia in the category of physical aggression toward other people.

Conclusions: There are different kinds of patient aggression. The way in which this behaviour is expressed is only partly related to diagnosis. Given that not all aggressive behaviours were the same, different interventions may be needed, depending on the behaviour expressed.

Methodological Critique: As most patients were admitted involuntarily, the generalizability of findings are quite limited.

Causality: A causal inference cannot be made on the basis of such a selected sample.

Klassen, D. and O'Connor, W. A. (1988). Crime, inpatient admissions, and violence among male mental patients. International Journal of Law and Psychiatry, 11, 305-312.

See also: Klassen, D. and O'Connor, W.A. (1988). Predicting violence in schizophrenic and non-schizophrenic patients: A prospective study. Journal of Community Psychology, 16, 217-227 and Klassen, D. and O'Connor, W.A. (1988). A prospective study of predictors of violence in adult male mental health admissions. Law and Human Behaviour, 12, 143-158.

Purpose and Objective: This study examined the relationship between hospitalizations, arrests, and violence in a sample of adult male patients whose admitting complaints suggested risk for violent behaviour. The objective of this study was to make comparisons between mental patients and their demographic peers, that is, young, male, minority, and low socio-economic status individuals.

Research Design: Although the authors do not indicate the study design, the data collection is consistent with a longitudinal follow-up of a case series. Arrests and admission histories were obtained from existing records, and subsequent arrests and admissions were obtained for a one-year follow-up period post-release.

Setting: An urban community health centre in Kansas City, Missouri, United States of America.

Study Subjects: The sample consisted of 304 adult males admitted as inpatients to an urban community mental health centre. They represented 91% of all patients selected. Only those men who had a history of violent  behaviour or indications of violence in the presenting problem were included in the study.

Measures: Arrest data were obtained from the Kansas City, Missouri Police Department and covered the greater Kansas City area as well as outlying counties in Western Missouri and Eastern Kansas. Inpatient admission data were obtained from the mental health centre's records.

Main Findings: The data showed a strong positive association between previous arrests and previous admissions. The authors reported that prior arrests and prior admissions were associated with subsequent violence. The best predictor of arrests during the follow-up period was the number of prior arrests. Diagnosis of substance abuse and number of prior admissions were also statistically significant in predicting arrests. Arrests for violence were predicted by a diagnosis of substance abuse, prior arrests for violent crimes, and age. Admissions for violence were predicted by number of prior admissions and age at first admission.

Conclusions: The authors concluded that arrests and hospital admissions were highly correlated in this sample, both prior to admission and after discharge. Furthermore, subsequent arrests were predicted by prior admissions controlling for the effects of prior arrests. The authors recognized that the nature of this data does not permit the conclusion of an association between crime and mental disorder.

Methodological Critique: Only men who had a history of violent behaviour or indications of violence in the presenting problem were included in the study. Therefore, it is not surprising that the sample as a whole had high rates of violent recidivism.

Causality: A causal inference cannot be made on the basis of these findings.

Lafave, H.G., Pinkney, A.A. and Gerber, G.J. (1993). Criminal activity by psychiatric clients after hospital discharge. Hospital and Community Psychiatry, 44(2), 180-181.

Purpose and Objective: The authors studied arrest rates for psychiatric rehabilitation patients discharged to the community after lengthy hospitalization.

Research Design: Prospective follow-up study with a one year follow-up period.

Setting: Brockville Psychiatric Hospital, Brockville, Ontario, Canada.

Study Subjects: Study subjects comprised 55 of 67 patients discharged from the Brockville Psychiatric Hospital between 1986 and 1988. More than two-thirds had a discharge diagnosis of schizophrenia.

Measures: Data were collected in face-to-face interviews with patients following their discharge using structured interview schedules. Ontario arrest rates were used to form the population comparisons.

Main Findings: Only 2 of the 55 patients  interviewed were charged with a criminal offense during their first year following discharge, giving an unstandardized arrest rate of 3.84 per 100 persons versus 11.35 per 100 persons for the general population.

Conclusions: The frequency of criminal charges in the study cohort was lower than expected based on general population rates for Ontario indicating that persons with a chronic mental illness, particularly schizophrenia, are not more prone to engage in criminal activities than the general population.

Methodological Critique: Although the sample size is quite small, this is an interesting study because it focuses on long-term rehabilitation patients who are presumed by other authors to be at high risk of arrest upon their re-entry into the community. A second important feature is that this study was conducted in Canada. The authors argue that availability of community mental health programs, housing, and social opportunities for released patients improves their chances of successful re-entry. Findings from this study raise an important issue. Given universal access to health care in Canada, and a different community mental health system, the relationship between mental hospital discharge and subsequent arrest noted in previous U.S. studies may not be generalizable to Canada.

Causality: Because of the selected nature of this study population, findings cannot be used to infer etiological relationships between mental illness and violence.

Lagos, J.M., Perlmutter, K. and Saexinger, H. (1977). Fear of the mentally ill: Empirical support for the common man's response. American Journal of Psychiatry, 134(10), 1134-1137.

Purpose and Objective: The purpose of the study was to determine how frequently people behave violently when they become mentally ill.

Research Design: While the authors failed to specify the study design, the data collection procedures are consistent with a retrospective cohort design. Subjects were identified for inclusion in the study based on admission to a psychiatric facility.

Setting: Four psychiatric facilities in New Jersey, and New York State, United States of America.

Study Subjects: One hundred records of patients admitted during 1974 to each of four psychiatric facilities were randomly selected.

Measures: Admission notes for 400 patients were reviewed for any description of patient behaviour that might be construed to be violent.

Main Findings: Results revealed that 37.7% of the 400 admissions sampled manifested some form of violent behaviour prior to admission. Twenty percent of the sample were specifically described as acting in a physically violent way toward persons or objects, and another 11% were described as acting violent but in vague terms.

Conclusions: The authors concluded that by  broadening the scope to include less serious acts of violence, approximately 36% of 321 psychiatric admissions were preceded by some form of violent fear-inducing behaviour. According to the authors, evidence suggests some factual basis for fearing the mentally ill.

Methodological Critique: Given that dangerousness is one of the criteria for hospitalization, it is not surprising that a large number of hospitalized patients have exhibited violence just prior to their admission. Because these findings focus on a violent-prone group, they do not support the conclusion that mentally ill, in general, are prone to violence.

Causality: A causal inference cannot be made on the basis of these findings.

Lindelius, R. and Salum, I. (1973). Alcoholism and criminality. Acta Psychiatrica Scandinavica, 49, 306-314.

Purpose and Objective: To study the relationship between the frequency of criminality and the severity of alcoholism.

Research Design: A consecutive series of patients were studied. Criminality was compared among three groups of alcoholics classified according to severity.

Setting: A special ward for the treatment of alcoholism at a general hospital in a large metropolitan area in Sweden.

Study Subjects: All male alcoholics (1026) admitted to the ward from 1956 to 1961.

Measures: No special measures and no mention made of type of diagnostic instrument. On the basis of clinical manifestations, patients were classified to one of the three groups. Information on criminality was obtained from the General Criminal Register.

Main Findings: More subjects in the low severity group were registered as criminals (45%) than those in the most severe group (30%). Younger alcoholics, below age 40, had a higher risk of being involved in criminality. Most offenses involved property or driving under the influence.

Conclusions: There is no simple relation between alcohol and criminality. Other factors, such as age or personality deviations, may be as important.

Methodological Critique: The main strength of this study is the correlation of criminal activity to levels of severity of alcoholism. However, the study is a case series of individuals admitted to a special unit because of their alcohol problem. Hence this is a highly selected sample with no comparison groups and the diagnostic approach is not specified. It is not clear whether other factors could be accounting for the study findings.

Causality: A causal inference cannot be made on the basis of these findings.

Lindqvist, P. and Allebeck, P. (1990). Schizophrenia and crime: A longitudinal follow-up of 644 schizophrenics in Stockholm. British Journal of Psychiatry,  157, 345-350.

Purpose and Objective: The main goals of this study were to determine if persons with schizophrenia commit crimes more frequently than the general population and if so, to what extent, and to describe the types of crimes committed by schizophrenic patients, compared to the general Swedish population.

Research Design: Historical cohort design with a fifteen year variable follow-up period.

Setting: Stockholm, Sweden.

Study Subjects: Subjects were 644 persons with schizophrenia diagnosed according to ICD-8 who had been discharged from inpatient psychiatric facilities in Stockholm county during 1971.

Measures: Both measures of schizophrenia and criminal activity were obtained from central registry data. The authors have estimated that 85% of their ICD-8 diagnoses meet DSM-III criteria. Files were linked via a unique personal identification number that is given to all Swedish subjects and follows them from birth to death. Criminal activity was measured by ‘principal offenses' committed during one year. The principal offense corresponds to the most severe penalty, but it is not clear whether this corresponds to an arrest or a conviction. Appropriate statistical analyses were used to account for the variable follow-up (e.g. rates per person years of observation) and rates were age- and sex- standardized to the population for comparison purposes. Given that socio-economic status is a likely intervening variable in the disease process, it was not controlled.

Main Findings: The crime rate among women was twice that expected whereas the crime rate among men was similar to the general population. Thirteen percent of the violent offenders were women (4 of 32) whereas the corresponding general population figure was 5%-10%. Assault was the most common violent offense committed and the most serious crime recorded was aggravated assault. The study cohort committed four times as many violent crimes as the general population. Given small numbers, these were not analyzed by age or sex.

Conclusions: Although discharged schizophrenic patients did not, in general, commit more crimes than the general population, they did commit four times as many violent crimes.

Methodological Critique: It is not clear whether the principal offenses of the study cohort were compared to principal offenses in the general population, or whether they were compared to the general population crime rates. As principal offenses measure only the most serious crimes recorded in any given year, a comparison based on the crime rate (all offenses recorded) would bias findings toward finding no effect. Therefore, findings from this study may reflect a conservative estimate of the difference between the criminal activity of the study cohort compared to the general population. Because of small numbers,  findings for violent crimes were not standardized to the population for age and sex, thus, there could have been confounding by these factors.

Causality: Because the study focused on criminal activity among discharged patients, findings cannot be interpreted to reflect causal inferences regarding mental illness and violence among all persons with schizophrenia.

Link, B.G., Andrews, H. and Cullen, F.T. (1992). The violent and illegal behaviour of mental patients reconsidered. American Sociological Review, 57, 275-292.

This empirical study contains an excellent review of prior literature, therefore it has been summarized in some detail. Link addresses the question of whether stereotypical portrayals of persons with mental illness as dangerous are correct by reviewing empirical literature on the criminality of former mental patients. He notes that early studies tended to find no difference between cohorts of mental patients and the general population with respect to criminality, while more recent studies have indicated that former mental patients have significantly higher criminality. The authors indicate that the strength and consistency of more recent findings suggests a causal mechanism and suggests that active psychosis is a likely factor to be considered. Studies have demonstrated that the crimes committed by persons with mental illness are often in response to hallucinations or delusions. Previous research indicates that 20% to 40% of crimes may be directly motivated by psychosis. These findings are supported by studies of violent behaviour in mental hospitals which find that the active phases of psychoses are the periods of greatest risk.

Link then identifies several alternative explanations for these findings. The process of criminalization may result in mentally ill persons being shunted into the criminal justice system. Or, mentally ill persons receiving treatment may be a sub-group of the mentally ill who may be at higher risk of criminality by virtue of other socio-cultural factors. No arrest rate studies to date have successfully connected the higher arrest rates among mentally ill samples to symptoms of mental illness. Thus, it is not known whether higher rates of crime and violent crime noted among mentally ill samples are a function of the mental illness or other non-causal influences. It is also possible that deviance has been medicalized and that the high rates of criminality among mentally ill is a result of this broader trend. This is consistent with findings from studies showing higher proportions of persons in psychiatric hospitals with prior histories of arrest.

Purpose and Objective: This study was designed to address the question of whether former mental patients have higher rates of violent and illegal behaviour than non-patients and whether any differences noted could be explained by factors other than mental illness.

Research Design: Cross sectional surveys  conducted in 1979 and 1982 as part of another study.

Setting: Washington Heights section of New York City, United States of America.

Study Subjects: Study subjects were 521 randomly selected community residents and 232 patients recruited from outpatient psychiatric clinics and inpatient services located in the Washington Heights neighbourhood. Patients who were not neighbourhood residents were excluded. Four groups were constructed for analysis: first treatment contact patients; repeat treatment contact patients; former patients; and never-treated community residents.

Measures: Patients were defined on the basis of their admission to a local facility and DSM-III diagnoses were retained for descriptive purposes. Sixty-three percent had major mental illnesses characterized by psychotic symptomatology (depressive disorders, schizophrenia, or other psychotic disorders). Official arrest data were obtained as well as self-report data concerning arrests, hitting others, fighting, weapon use, and hurting someone badly. Census data were also used to describe the socio-demographic context of the community. First treatment reflected first treatment within the year prior to the interview (for community residents), or first-ever contact with the clinic at the time of study. Psychotic symptoms were measured using a standardized scale.

Main Findings: Former mental patients scored higher on official measures of arrest as well as self-reports of violence compared to community controls. Within the patient group, violence was associated with psychotic symptoms.

Conclusions: The authors conclude that assertions that mental patients and former mental patients are on average no more violent than non-patients is incorrect even when socio-demographic and community context variables are controlled. But, the excess risk posed by patients is modest and only apparent among patients with current psychotic symptoms.

Methodological Critique: The statistical analysis is technical and highly complex, and is not presented in a user-friendly manner. For example, logistic regression coefficients and standard errors are presented without an attempt to translate these into odds ratios and 95% confidence intervals which may be more easily understood.

The authors provide evidence that persons receiving treatment for mental illness are at higher risk of being violent than non-patients. Given that treatment services are oriented toward a dangerousness standard, this finding is not surprising. Because the study was cross-sectional, temporal ordering of factors could not be established. The authors ask the reader to dismiss the possibility of reverse-causation because their results are consistent with prospective studies of former mental patients in which temporal ordering of factors has been  clearly established. This is not warranted as prior prospective studies have not purified their study cohorts to ensure that persons with prior histories of violence are excluded. Given that prior arrests/violence are known to predict future arrests/violence, temporal ordering of factors remains a thorny issue that cannot be so easily dismissed.

Causality: Because former mental patients are known to comprise a sub-group of mentally ill that are more prone to violence, this study cannot be used to arrive at a causal determination of whether the mentally ill in the community are more dangerous and violent than the non-mentally ill.

Lundy, M.S., Pfohl, B.M. and Kuperman, S. (1993). Adult criminality among formerly hospitalized child psychiatric patients. Journal of the American Academy of Child and Adolescent Psychiatry, 32(3), 568-576.

Purpose and Objective: The goal of this study was to investigate risk factors for adult criminality among children who required psychiatric hospitalization at a relatively young age.

Research Design: A case-control design was used. Cases were defined as 23 individuals known to have an adult prison record and controls were 115 males who had no prison record. All of the cases proved to be male.

Setting: The University of Iowa Psychiatric Hospital, Iowa, United States of America.

Study Subjects: Subjects were 138 children (under the age of 12 years at discharge) who received inpatient psychiatric care between 1970 and 1982. Persons with mental retardation, defined as an IQ score of less than 70, were excluded from the study because of difficulties in obtaining informed consent from this group.

Measures: ICD-9 diagnostic categories were used to assign each child to a single, main diagnostic group: organic disorders (organic mental disorder, infantile autism, and developmental delay), attention deficit disorder, emotional disorders (affective disorder, neurotic disorder, neurotic depression, eating disorder, and personality disorder), and adjustment disorders. A miscellaneous grouping was also used. Presence or absence of an adult prison record was obtained from the Iowa Department of Corrections and was used as the main outcome of interest. Childhood risk factors were assessed via a detailed chart review. These factors included assaultive behaviour (excluding that which could be construed as developmentally normal such as altercations with peers or siblings which was not a focus of concern for the caregiver or parent), criminality in a biological parent, psychiatric illness in a biological parent, psychotropic medication use on discharge (to serve as a proxy for severity of disturbance), adoptive status, and treatment responsiveness.

Main Findings: No diagnostic category was statistically related to adult imprisonment. No statistical relationships were also found between adult imprisonment and  presence of multiple psychiatric diagnoses, adoptive status, age at index hospitalization, IQ, gap in verbal-performance IQ of more than 15 points, length of stay, multiple admissions, or parental psychiatric illness. Assaultive behaviour in childhood was associated with 5 times the risk (95% CI, 1.8 to 13.8), parental criminality (4.6 times the risk, 95% CI, 1.43 to 16.41). Race was also found to be an important predictor in a subsample of 74 for whom this information was available.

Conclusions: The authors concluded that the failure of diagnosis to predict outcome is not inconsistent with previous studies. The authors highlight the importance of childhood assaultive behaviour, in the absence of a diagnosis of conduct disorder, and parental criminality as the most important predictors of a poor outcome.

Methodological Critique: A minor difficulty with this study are the small cell sizes resulting in large confidence intervals for some comparisons and less precise estimates than would have been desired under ideal conditions.

Causality: Because all study subjects received psychiatric intervention, results cannot be used to derive a general causal explanation regarding the relationship between mental illness and violence, per se. No comparison group of non-psychiatrically ill was used. Nevertheless, it is interesting that no diagnostic category, not even conduct disorder, predicted adult criminality.

Lurigio, A.J. and Lewis, D.A. (1987). The criminal mental patient: A descriptive analysis and suggestions for future research. Criminal Justice and Behaviour, 14(2), 268-287.

Purpose and Objective: This study was designed to extend earlier findings by yielding a comprehensive and prospective look at the movement of the mentally ill within and between the criminal justice system and mental health systems with a focus on the criminality and dangerousness of patients.

Research Design: Although the authors do not specify, data collection is consistent with a longitudinal follow-up of a case-series. Current arrest and admission data were obtained along with lifelong criminal and psychiatric histories.

Setting: Chicago-area state psychiatric institutions, United States of America.

Study Subjects: A random sample of approximately 320 psychiatric patients, aged 18-65 years, was selected from a population of inpatients from Chicago-area state psychiatric institutions. Sampling was stratified on the basis of age, sex, race and previous admissions, and executed in a series of weekly waves over the course of three months.

Measures: Four sources of information pertaining to the criminality, dangerousness, and victimization experiences of the patients were used. First, the Chicago Police Department's Bureau of Investigations  Unit's records were searched to ascertain whether study subjects were arrested within a recent six-month calender period. Second, official police incident data were reviewed to assess the degree to which the circumstances surrounding the arrests were characterized by violent or dangerous activities. Third, official criminal histories or “rap sheets” were reviewed but only for those patient arrestees identified from the Chicago Police Department. Fourth, the authors assessed the patients' self-reports of violent behaviour via a personal interview. Information on current admission and psychiatric history were collected through hospital charts. In addition, time at risk was estimated.

Main Findings: Based on police records, 10% of the random sample was arrested within the six-month period prior to hospitalization. Patients were involved in a total of 58 police encounters. A significantly greater percentage of arrests (60%) appeared in the aftermath of hospitalization rather than in the pre-admission stages of hospitalization. Of the 58 arrest report narratives, 50% of the encounters were characterized by an interpersonal conflict or altercation. Arrestees were compared to psychiatric patients in the general sample on a number of demographic variables. The data showed that persons aged 18-34 years who were black were over-represented in the arrested group when compared to their prevalence in the larger sample. The group of arrestees had a significantly greater number of prior admissions than the overall sample. Of those arrested, 85% of the sample possessed criminal histories ranging from 1 to 30 previous arrests. During the study's six-month period of reference, patient arrestees were admitted to the hospital on 119 separate occasions, yielding a mean of approximately 4 hospitalizations per patient. Patient arrestees were hospitalized significantly more often than patients in the general sample.

Conclusions: The authors concluded that a relatively small percentage of released psychiatric patients are involved in criminal conduct extending beyond innocuous or “nuisance-type” activities. The patients who were likely to pass through the criminal justice system in the short six month period under investigation were those with chronic readmissions to state psychiatric facilities.

Methodological Critique: The use of arrests in this study may have underestimated the prevalence of criminality among released patients if they were diverted out of the criminal justice system to mental health treatment centres by police.

Causality: A causal inference cannot be made on the basis of these findings.

McNeil, E. and Binder, R.L. (1994). The relationship between acute psychiatric symptoms, diagnosis, and short-term risk of violence. Hospital and Community Psychiatry, 45(2), 133-137.

Purpose and Objective: This study examines the relationship between acute psychopathology and short-term risk of violence in a sample of newly hospitalized acute patients. The goal was to evaluate whether the pattern of symptoms associated with short-term risk of violence varied depending on diagnosis.

Research Design: Although not indicated, it appears that a case-control design was used.

Setting: The study was conducted on a university-based, locked, short-term psychiatric inpatient unit with a mean length of stay of 18 days.

Study Subjects: The sample consisted of 127 diagnostically heterogeneous patients admitted during 1988-1989 and 203 patients who had been admitted to the unit between 1989 and 1990. The total sample included 330 subjects. Fifty-four percent of the 330 patients were male.

Measures: The Overt Aggression Scale (OAS) was used to evaluate violent behaviour exhibited by the patients in hospital. The OAS is a widely used measure with documented reliability and validity as an index of inpatient aggression. It is a behavioural checklist that nursing staff complete at the end of each eight-hour shift to indicate if patients have engaged in physical aggression against other people, against objects, or against themselves, or have engaged in verbal aggression. The Brief Psychiatric Rating Scale (BPRS), a widely used measure of psychopathology with good interrater reliability, was used to evaluate each patient at admission.

Main Findings: Ratings on the OAS indicated that 23% of the patients engaged in physical aggression against other people during their hospitalization. Assaultive patients were over-represented in the diagnostic categories of schizophrenia, mania, and organic psychotic conditions. Compared to non-assaultive patients, patients who became assaultive had an admission mental status characterized by significantly higher levels of thinking disturbance, hostile suspiciousness, agitation, and excitement.

Conclusions: This study found an association between diagnoses such as schizophrenia, mania, and organic psychotic conditions and imminent risk of assaultive behaviour among a sample of acutely ill, newly hospitalized patients.

Methodological Critique: It is not clear whether nursing staff would be more likely to record incidents of violence in the health record of patients who appear to be most agitated with thought disturbances. Thus, it is not clear to what extent underreporting of violence in some diagnostic groups may have resulted in these findings.

Causality: A causal inference cannot be made on the basis of these findings.

Previous | Table of Contents | Next