Mental Illness and Violence: Proof or Stereotype?

Next | Previous | Table of Contents

2 Summary of Key Findings

Researchers have approached the study of mental illness and violence from a number of different perspectives depending on their access to community, mental health, or criminal justice populations. This has resulted in a large body of research. The first goal will be to summarize the results of these studies in order to understand the statistical associations between mental illness and violence which have been reported in these various populations. The second goal, addressed in the next chapter, will be to critically evaluate to what extent these findings can be used to support an etiologic (i.e. causal) relationship between mental illness and violence.

Community-Based Studies:

General Population Samples:

Two studies have conducted population surveys of representative samples of adults living in the community—one Canadian, studying 1,200 Edmontonians (Bland and Orn, 1986), and one American, studying 10,059 subjects from three of the five Epidemiological Catchment Area sites (Swanson, Holzer, Ganju, and Jono, 1990). Both studies use what has come to be known as the ‘ECA Methodology'. This methodology characterizes the Epidemiological Catchment Area Surveys conducted in five sites in the United States.  Currently considered to be state-of-the-art in psychiatric epidemiologic surveys, the ECA approach employs large and representative samples, a structured diagnostic interview administered by lay interviewers (The Diagnostic Interview Schedule or DIS), and computer scoring of the results to arrive at standardized DSM-III-R diagnostic categories. Both studies used questions from the DIS to measure physical violence, such as hitting or throwing things at a spouse or partner, spanking or hitting a child, fist fighting since age 18 with someone other than a spouse, using a weapon since age 18, and getting into a physical fight while drinking.

Both studies report statistical associations between violence and mental disorders, although neither study is able to disentangle the temporal ordering of factors. In Canada, three diagnostic categories were studied: antisocial personality disorder, major depression, and alcohol abuse/drug dependence. Altogether, 54.5% of those with a diagnosis were involved in violent behaviour compared to 15.5% of those with no diagnosis. Persons with one or more of these diagnoses were almost seven times more likely to be involved in violence than those without one of these disorders. In particular, the risk of violence was greatly elevated among those diagnosed with a comorbid alcohol abuse disorder. When alcohol was combined with antisocial personality and/or depression, 80-93% were involved in violence. In the United States of America, more than half of the individuals reporting violent behaviour in the preceding year met the criteria for a psychiatric disorder compared to 19.6% of non-violent respondents. The highest percentages of violence were reported among substance abusers, ranging from 19.2% to 34.7% depending on the type of substance abuse. Individuals in the community meeting the criteria for any psychiatric disorder were more likely to engage in assaultive and violent acts compared to those who did not meet the criteria for a psychiatric disorder. However, certain diagnostic categories, such as non-comorbid anxiety disorder, affective disorder, or schizophrenia showed no or only slightly elevated risk of violence. Conversely, those with substance abuse disorders were at greatly elevated risk and also appeared to commit more severe acts of violence. These findings indicate that the public's fear of persons with schizophrenia living in the community is largely unwarranted, although not entirely groundless. Citizens are more likely to be assaulted by someone suffering from a substance abuse disorder than a major mental illness such as schizophrenia.

Both studies suggest that individuals are at greater risk of being assaulted by someone who abuses substances rather than someone who is suffering from major mental illness such as affective disorder, anxiety disorder, or schizophrenia. Using the same ECA data, Swanson (1993) further tested the hypothesis that the relationship between mental disorder and violence could be largely explained by the association between alcohol abuse and violence. Mental illnesses that were uncomplicated by alcohol abuse were associated with some increased risk of violence. However, the apparent large increase in violence among younger, lower socio-economic males was found to be largely due to the increased prevalence of alcohol abuse and comorbidity in this group. A history of arrest and psychiatric hospitalization was found to be associated with an increased probability that a person would be violent.

Police-Citizen Encounters:

Police have considerable discretionary powers in responding to persons with mental illness who may be acting in a disordered or disorderly fashion while in the community. The police may convey an individual to a psychiatric facility for assessment and treatment, or they may proceed with an arrest. Monahan, Caldeira, and Friedlander (1979) have shown that in 30% of police-initiated commitments, the police could have proceeded with an arrest. They did not because they believed that the individual lacked criminal intent or would benefit from treatment. Deinstitutionalization and legislative changes have increased the central role of the police in responding to persons with mental illness who come into contact with the criminal justice system. Bonovitz and Bonovitz (1981) show that the number of mental illness-related incidents handled by police increased over 200% between 1975 and 1979 after legislative changes permitted officers to expedite the removal of individuals with mental illness from the community. Teplin (1985) observed a random selection of 283 police officers in their day-to-day interactions with the public. A symptom checklist was used to assess the presence and severity of psychiatric impairment among those coming into contact with the police. Police encounters with individiuals with a mental disorder occurred infrequently (in 4% of 2,122 persons encountered). Persons with a mental disorder were only slightly more likely than persons without a mental disorder to be considered suspects in crimes, and for those who were considered to be suspects, the type of crime was not found to be related to the presence or absence of mental disorder. Those with a mental disorder did not commit serious crimes at a rate that was disproportionate to their numbers. The pattern of crime among mentally ill suspects was substantially similar to non-mentally ill suspects.

Similar findings are reported by Arboleda-Flórez and Holley (1988) who studied police-citizen encounters in Calgary, Canada. This study involved 350 persons who came into contact with police during a two-week period. Police rated the observable behaviour of these individuals on a continuum from normal to severely abnormal. The circumstances under which the encounters occurred were taken into account. The police officers were then asked to provide some judgement as to the cause of the abnormal behaviour: alcohol, drugs, mental illness, or other. Those persons identified by police as having a mental illness did not record a greater number of crimes against persons,  property, or other crimes compared to those identified as non-mentally ill.

Studies of Psychiatric Patients:

Physical violence in hospital has been reported in approximately 20% of samples studied (e.g. Lagos, Perimutter, and Saexinger, 1977; Binder, McNeil, and Binder, 1988). Typically, a small number of patients (e.g. 5%) are found to be responsible for just over half of all violent incidents and more than half of the serious injuries (Convit, Isay, Otis, and Volavka, 1990; Fottrell, 1980). Patients with psychotic symptoms, particularly paranoia, have been found to be at higher risk of physical aggression toward others (Noble and Rodger, 1989; Kennedy, 1993; McNeil and Binder, 1994). Among psychogeriatric patients, dementia has been shown to be related to aggressive and violent behaviour (Patel and Hope, 1992). Studying an outpatient psychiatric population, Tardiff and Koenigsberg (1985) report that 5% of subjects had been physically assaultive toward others in the few days prior to the evaluations and family members accounted for over half of those assaulted. Assaultive behaviour was associated with being male, younger than 20 years of age, and having a diagnosis of childhood or adolescent disorders or mental retardation.

Straznickas, McNiel, and Binder (1993), found that 19% of patients (113 of 581) in a university- based, locked, short-term psychiatric inpatient unit had physically attacked someone in the two weeks prior to their admission, and thirty-one of those patients who were assaultive attacked more than one person. Of the 113 patients who attacked someone, 50 assaulted people outside of the family, 10 patients assaulted both family members and individuals outside of the family, and 53 assaulted family members. Unfortunately, no comparison groups were used. Therefore, it is not clear whether the relatives of persons with a mental illness were more likely to be the targets of violence compared to the relatives of non-mentally ill.

Violence and fear-inducing behaviour have been found to be characteristic of the acute exacerbations of chronic conditions such as schizophrenia or mania which lead to a hospitalization. Binder et al. (1988) found that 21% of randomly selected inpatients (N=150) in a university psychiatric unit had attacked persons and 25% had engaged in fear-inducing behaviour in the two weeks just prior to their admission. This was especially true for patients suffering from schizophrenia or mania. In addition 13% of patients attacked others during their admission and 32% engaged in fear-inducing behaviour. Patients with a diagnosis of mania were more likely to attack others while those with a diagnosis of schizophrenia were more likely to engage in fear-inducing behaviour. These findings highlight  the importance of context as a factor influencing the expression of violent behaviours.

There is some evidence to suggest that the rate of violence among inpatients may be increasing. For example, Noble and Rodger (1989) report an increase in violent incidents occurring in hospitals between 1976 and 1984. Similarly, Volavka et al. (1995) report an increasing trend in the prevalence of arrests of psychiatric patients for incidents committed while in hospital.

A number of studies have examined the relationship of specific diagnoses to violence within populations of psychiatric inpatients. The major issue addressed by this avenue of research is what kinds of mental illnesses predict violence and criminality among mentally ill populations, not whether mental illness, per se, predicts criminality and violence.

Perhaps the most consistent and striking finding is the association of substance abuse disorders (alcohol and/or drug) with violence and criminality, and the lack of or small association between other disorders (e.g. schizophrenia, affective disorders, or anxiety disorders) with violence. As early as 1974, Guze, Woodruff, and Clayton demonstrated that felony convictions were reported by patients with diagnoses of sociopathic personality disorder and substance abuse leading the authors to conclude that sociopathy, alcoholism, and drug-dependence were the principle psychiatric disorders associated with serious crime. Similar findings are reported in more recent studies (Holcomb and Ahr, 1988; Modestin and Ammann, 1995).

The importance of substance abuse as a predictor of violence was found to hold even when studies were restricted to a single diagnostic group, such as patients with schizophrenia. For example, Cuffel, Shuway, Choulijian, and MacDonald (1994) studied only patients who had been diagnosed with schizophrenia to determine whether a comorbid substance abuse diagnosis increased risk of subsequent violence. Data were gathered from a retrospective record review of 103 outpatients who had been involved for six months in randomized clinical trials at a schizophrenia research clinic in San Francicso, United States. Violent behaviour included both property damage and acts against persons such as verbal threats to harm others, nonverbal threats to harm others, physical assaults, altercations, brandishment of weapons, using a weapon, starting a fire, or destroying property. Patients who were polysubstance abusers (alcohol and drugs) were significantly more likely to commit a violent act, although the risk diminished considerably three months into the study. These findings are consistent with the findings showing a statistical relationship between substance abuse disorders and violence in representative samples of adults.

A second interesting finding from this body of research has been the importance of prior violence and criminality in predicting subsequent violence and criminality (e.g. Asnis,  Kaplan, van Praag, and Sanderson, 1994; Klassen and O'Connor, 1988a, 1988b; Lundy, Pfohl, and Kuperman, 1993). This relationship is important in light of the significant percentage of mentally ill patients who report prior criminal and violent acts (e.g. Grossman, Haywood, Cavanaugh, Davis, and Lewis, 1995; Holcomb and Ahr, 1988).

Cirincione, Steadman, Clark-Robbins, and Monahan (1992) assessed the extent to which a diagnosis of schizophrenia was predictive of criminal violence, after controlling the effects of arrest history, among two cohorts of patients admitted to a New York State facility, one in 1968 (N=255) and the other in 1978 (N=327). The New York State Division of Criminal Justice Services provided data on violent crimes committed for the 11 years following the index psychiatric admission. Violent crimes included murder, manslaughter, rape, attempted rape, assault, kidnapping, and sodomy. Prior arrest history significantly correlated with violent crime in both cohorts. In 1968, diagnosis was a significant predictor of violent crime. However, this was not the case for the 1978 cohort. For those without a prior arrest, diagnosis did not predict subsequent violent crime. Similarly, Wessely, Castle, Douglas, and Taylor (1994) demonstrated a small increased risk of criminality among those with schizophrenia, but this was overshadowed by the much larger effects of prior criminality and substance abuse.

Buckley et al., (1990) provide information on the likely target of violence. They studied 698 patients who were diagnosed with schizophrenia and admitted to a psychiatric inpatient department in Dublin, Ireland between 1983 and 1988. Sixteen percent of patients had engaged in a physically violent or destructive act (i.e. to property) since the onset of their illness. Patients with no history of violence were found to be similar to those with a history of violence with respect to positive and negative symptomatology, and a family history of psychiatric illness. Violence was more common among males. Most of the incidents of community violence occurred in the home and involved episodes of damage to furniture or minor assaults to relatives.

Studies of Incarcerated Offenders:

Prevalence Studies:

A number of studies have assessed the prevalence of mental illness among samples of incarcerated offenders. Interpreted within the broad context of criminalization of the mentally ill, these studies have been cited to support the hypothesis that large numbers of the mentally ill have been diverted from mental health to criminal justice systems. However, it is not clear from these studies what proportion of the mental illnesses reported predated the criminal behaviour (e.g., Allodi and Montgomery, 1975) or what proportion is a result of the “psychiatrization” of criminality (e.g., Davis, 1992).

Two large Canadian studies (Arboleda-Flórez, 1994; Bland, Newman, Dyck, and Orn, 1990)  provide compelling evidence that a significant proportion of incarcerated persons suffer from substance abuse disorders and serious mental illness. Although different methods of measuring mental disorder were used, both studies reveal consistent results. In the most recent study (Arboleda-Flórez, 1994), forensic psychiatrists conducted structured clinical interviews of a randomly selected sample of 1,200 admissions to the Calgary Remand Centre. Subjects were examined during the first 24 hours of detention. A principal diagnosis on either Axis I or Axis II of the Diagnostic and Statistical Manual of Mental Disorders (DSM-III-R) was made in 728 of the 1200 interviewees (60.7%). The one-month prevalence was 49.5% for females and 56% for males for an Axis I disorder and 3.6% for females and 5.5% for males for an Axis II personality disorder. Substance abuse disorders occurred in 35.7% of women and 47.3% of men. Schizophrenia was not found among women (but this may have been due to their smaller numbers in the sample) and occurred in 1.2% of the males. Similarly, Bland, Newman, Dyck, and Orn (1990) studied a smaller, systematic sample of 180 males in custody at the Edmonton Remand Centre. The sample included remanded offenders and provincially sentenced prisoners (i.e. sentenced to two years less one day). Using the Diagnostic Interview Schedule (DIS) for lay interviewers, these authors found a lifetime prevalence of any psychiatric disorder of 92%. The most frequently occurring diagnosis was substance abuse (87%). Antisocial personality disorder occurred in 57%, affective disorder in 23% and schizophrenia in 2% of cases. The larger proportion of personality disorders identified in this study may be a function of the DIS which uses information about criminality in the diagnostic criteria. Studies showing a high prevalence of substance abusers among incarcerated offenders also have been reported in the United States (Barton, 1982; Lamb and Grant, 1982) and the United Kingdom (Taylor and Gunn, 1984).

Despite the high prevalence of mental disorder among incarcerated offenders, Teplin (1990) has shown that the overall rate of detection by prison personnel may be very low, with only 32.5% of personnel indicating poor provision of treatment for mentally disordered offenders. Persons are most likely to be identified for treatment if they had a prior treatment history (91.7% detected).

Analytic Studies of Mental Disorder and Violence:

A number of studies have examined whether violent criminality can be associated with mental illness in general or a particular diagnosis within offender populations (McKnight, Mohr, Quinsey, and Erochko, 1966; Nichol, Gunn, Gristwood, Foggitt, and Watson, 1973; Siomopoulos, 1978; Ashford, 1989; Brownstone and Swaminath, 1989; Côté and Hodgins, 1992; Beaudoin, Hodgins, Lavoie, 1993; Coid, Lewis, and Reveley, 1993; Raine, 1993). However, these studies suffer from multiple methodological problems including small or unrepresentative samples, selected  offender groups such as violent inmates or those remanded for a psychiatric assessment, or lack of a comparison population. Consequently, results are inconsistent and any conclusion linking diagnosis to violence within incarcerated populations would go well beyond the scope and quality of the data.

In a large investigation, Toch and Adams (1989) used record linkage technology to study the relationship between mental illness and criminality in New York State (United States). Computer records of 8,379 inmates were matched to New York State Mental Health Services records. Inmates were considered to be suffering from a mental illness if they appeared in the Mental Health files. Of those without a history of mental illness (including substance abuse), 13.8% had a history of recent and remote violence compared to 17% with a history of mental illness or substance abuse. However, 5.8% of those with a combination of psychiatric history and substance abuse committed apparently unmotivated violent acts, compared to only 1.2% of those without a psychiatric history or history of substance abuse. Although the differences reported are actually quite small, the authors concluded that inmates with a history of mental illness or substance abuse were more prone to commit acts of recent violence (occurring within 3 years), and remote violence (occurring after 3 years or more), as well as unmotivated violent acts.

Rice and Harris (1995) studied violent recidivism among matched cohorts of 685 persons who had been referred for a brief forensic psychiatric assessment. Psychopathy, schizophrenia, and alcohol-abuse were the main independent variables of interest. Violent recidivism occurred in 31% of the subjects. Those meeting the criteria for psychopathy (using the 20-item Psychopathy Checklist) were more likely to have an alcohol problem and this combination was related to violent recidivism. Alcohol abuse in isolation was also linked to violent recidivism and persons diagnosed with schizophrenia were less likely to recidivate. Unfortunately, the authors did not control for previous violence or other known risk factors such as age or socio-economic status. At best, therefore, results are only suggestive of a relationship.

Conversely, Valdiserri, Carroll, and Hartl (1986) examined the relationship of psychosis to criminality among persons referred to an on-site mental health clinic in an American prison. Psychotic inmates were four times more likely than non-psychotic inmates to have been charged with a minor offense. There was no difference between the study groups with respect to number of violent offenses. Similarly, Hodgins and Côté (1993) studied the relationship of mental disorder to violent criminality in a representative sample of 461 subjects being held at penitentiaries in Quebec (Canada). A total of 107 individuals were defined as mentally disordered based on the Diagnostic Interview Schedule. These individuals were not found to have a history of more convictions, or more violent  convictions, compared to non-mentally ill offenders.

Follow-up Studies of Offenders Released into the Community:

The relationship of mental illness to subsequent community adjustment among released offenders has received scrutiny in two large studies (Abram and Teplin, 1990; Feder, 1991; Teplin, Abram, and McClelland, 1994). Neither demonstrates a strong relationship between mental illness and post-release adjustment or recidivism.

Abram and Teplin (1990) were specifically interested in whether persons with dual diagnoses (mental illness and substance abuse) would commit more violent crimes than persons who abused drugs but were not mentally ill. A random sample of 728 released offenders from Cook County Corrections in America were followed for 3 years. Measures of mental illness and substance abuse were collected using the Diagnostic Interview Schedule (DIS) and subsequent arrest data were obtained from the Chicago Police Department, the Federal Bureau of Investigation, and the Illinois Bureau of Investigation. Previous arrests for violent crimes and time at risk (number of days out of jail) were positively associated with subsequent commission of violent crimes. An opiate disorder diminished the probability of a future arrest for a violent crime. However, the authors point out that data were collected prior to the cocaine epidemic in the United States. Mental disorders (schizophrenia, depression, and alcohol disorder) did not predict subsequent arrest for a violent crime after controlling variables such as age or education.

Subsequently, in 1994, Teplin, Abram, and McClelland again studied whether the post-release arrest rates for violent crime for these offenders were related to mental disorder in a six-year follow-up. Those with a severe mental illness, defined as schizophrenia or major affective disorders, had a probability of re-arrest of .43. Those with substance abuse disorders had a probability of .46 of being re-arrested. These differences were not statistically significant and held when prior criminal history and age were statistically controlled. In every diagnostic group, persons with a prior history of violent crime were twice as likely to be re-arrested during the follow-up compared to those with no prior history. Persons with a history of hallucinations or delusions did not have a higher probability of subsequent arrest. Persons with hallucinations and delusions did have a slightly higher number of arrests for violent crimes however this was not statistically significant. This carefully conducted and well-reported study provides compelling evidence against the hypothesis that re-arrest for a violent crime is related to psychiatric diagnosis.

Similar results are reported by Feder (1991) who compared the post-prison adjustment of mentally ill offenders (N=147) to a comparable group of non-mentally ill offenders (N= 400) over an 18-month period. When statistical techniques were used to control for group  differences in criminal history, the only significant factors distinguishing the groups in subsequent arrests were age and prior arrests. Psychiatric status was not significant. Sixty-four percent of mentally ill offenders and 60% of non-mentally ill offenders were re-arrested at least once during the follow-up; 19% of mentally ill and 15% of non-mentally ill for violent crimes. Mentally ill offenders were less likely to receive a sentence involving time and were more likely to be diverted into the mental health system.

A number of smaller or less controlled investigations have also been conducted showing a high prevalence of recidivism (e.g. Guze, Goodwin, and Crane, 1969; Grunberg, Klinger, and Grumet, 1977; Pasewark, Bieber, Bosten, Kiser, and Steadman, 1982; Hodgins and Hébert, 1984; Lindqvist, 1986; Lamb, Weinberger, and Gross, 1988; Menzies and Webster, 1987; McMain, Webster, and Menzies, 1989; Martell and Dietz, 1992), but which cannot be used to support a causal relationship between mental illness and violence.

Summary of Key Findings:

The strongest predictor of violence and criminality has proved to be past history of violence and criminality. This was true for persons with schizophrenia as it was for those with substance abuse disorders.

As yet, there is no consistent evidence to support the hypothesis that mental illness (e.g. schizophrenia or depression) that is uncomplicated by substance abuse is a significant risk factor for violence or criminality, once past history of violence is controlled.

Whether persons with schizophrenia are at risk of violence depends, in part, on the context and the presence of psychotic symptoms. For example, persons with schizophrenia have been found to be at somewhat increased risk of committing violent acts when in the community, especially when they are experiencing psychotic symptoms. Conversely, violent behaviour has been found to be low among hospitalized patients with schizophrenia who are receiving appropriate neuroleptic medication.

The occurrence of violent incidents among hospitalized mentally ill may be increasing. However, a small number of patients, typically those with acute psychotic symptoms or dementia, have usually been found to be responsible for the majority of violent incidents. Most violent incidents leading to hospitalization occur in the home, involve episodes of damage to furniture or minor assaults to relatives.

Formerly hospitalized mental patients may be at high risk of arrest and violence when released into the community, particularly if they have a history of prior arrests or violence or if they experience psychotic symptoms.

Family members (not the general public) are the most likely targets of violence from formerly hospitalized patients in the community.

Substance abuse appears as a significant risk factor for violence and criminality among community, patient, and offender populations. It is unlikely that a member of the public would be at risk of violence from someone with a non-substance abuse disorder.

Studies of police-citizen encounters in both Canada and the United States show that the pattern of criminality of mentally ill and non-mentally ill persons coming into contact with the police is similar; mentally ill are no more likely to be charged with a violent crime compared to non-mentally ill.

The prevalence of substance abuse disorders and mental disorders is high among remanded and provincially incarcerated offenders. Yet, the overall rate of detection of mental illness by correctional staff appears to be quite low. Explanations for the high prevalence of mental illness among incarcerated offenders have included the criminalization of mentally disordered behaviour, the “psychiatrization” of criminal behaviour, and the pathogenic nature of incarcerated settings.

In general, offenders' post-release adjustment does not appear to be related to major mental illnesses (such as schizophrenia or depression) or substance abuse disorders when prior criminality and age are controlled.

Most generally, individuals who are younger are at higher risk of violence and criminality.

Next | Previous | Table of Contents