Dangerousness and mental illness:
The research and implications for nursing practice

By Richard Lakeman Dip Nsg., B.N., BA Hons., FANZCMHN

Prepared for the Hawke's Bay Nurses Forum.

Presented on the 31st of July 1996.

Key Points About Dangerousness:

The Concise Oxford Dictionary (1985) defines danger as "liability or exposure to harm, risk or peril". It is of great concern to nurses to identify danger in their workplaces and to take action to minimize it. A hospital corridor is not innately dangerous but add a volume of water to make the floor slippery and you have a source of danger. Add foot traffic and you have the ingredients for an accident - the environment may be said to be dangerous. The astute nurse cognizant of the danger will have the water cleaned up or sign posted, thus minimizing the danger.

Labeling an object or situation dangerous involves a prediction that given the presence or absence of certain conditions an object poses a likely threat of harm to another. This brief discussion will examine the reliability and implications of identifying those with mental illness as dangerous to others and some of the conditions which may predispose someone to act in a dangerous manner.

The association between mental illness and dangerousness:

For some the association between mental illness and dangerousness is automatic. Media portrayals of mental illness give rise to beliefs that the sole conditions necessary for violence towards others by a person are mental illness and the individuals liberty. This belief is further reinforced by sensationalist reporting on occasions when someone with a diagnosed mental illness does commit a violent act (Levey and Howells, 1995). On these occasions any history of contact with mental health professionals is taken as evidence that the person must have been "sick" at the time of the act.

For many people violent acts themselves are seen as signs of illness, hence such expressions as "he must have been sick" in response to hearing of a particularly brutal physical or sexual assault on another. Such a statement is a reflection of how entrenched medicine has become in our daily lives. Conrad and Shneider (1980) point out that it is a relatively recent phenomenon that such problems as violence and addiction have been defined as medical problems. Medicine has ascended to a position of authority in labeling and treating all manner of social problems and the authority of medicine in the social domain has increasingly gained legal legitimacy.

Psychiatric predictions of future dangerousness represent the single most significant factor in the commitment of the mentally ill in the United States (Fisher, 1989) and this is increasingly becoming so in most countries (Mulvey and Lidz, 1995). In New Zealand the Mental Health Act 1992 introduced the concept of "dangerousness " in an explicit way, as a central concept to consider in assessing and compulsorily treating those with mental disorder. This was a significant shift from the focus of the old Act where people could be detained because they needed care (Bridgeman, 1994). The grounds for compulsory treatment are a mental disorder of such a degree that it "Poses a serious danger to the health or safety of that person or of others..." or "Seriously diminishes the capacity of that person to take care of himself or herself..." (MH(CAT) Act, 1992). Nurses have specific powers under the act to detain people whom they believe pose a danger due to mental disorder for the purposes of a medical assessment. Contention exists over what constitutes a danger and when danger becomes serious.

When the court receives an application for a compulsory treatment order the judge must consider whether the person poses a serious danger. According to Cooper (1994) serious danger has been defined by precedent as an imminent, demonstrable risk of no less than severe physical violence which is not reasonably preventable or avoidable without compulsory treatment. Using this narrow definition of dangerousness it would not be surprising if there is an association between violence and the compulsorily detained mentally ill. Indeed a history of violence will be a defining feature of this subgroup as a history of violence or a recent violent act may be necessary to satisfy the criteria of dangerousness. In some cases health professionals may be impotent to compel someone to receive assessment (even though they might benefit from it) until the individual has reached a stage where they pose an imminent danger to others. Violence and mental illness may not even be independent terms. Swanson et al (1990, p. 762) point out that "... assaultive behavior lies at the core of what is labeled as psychiatric disorder". Violence and mental disorder are to varying extents seen as synonymous by the general public, the law and the psychiatric profession.

Nurses who work in inpatient psychiatric units are likely to encounter people who are deemed a danger towards others. This is also true for other settings. It has been estimated that at least half of all health professionals will be assaulted during their careers (Blair & New, 1992). Many hundreds of research studies and articles have been published on the issue of violence from a broad range of perspectives. Monahan (1992) concludes on the basis of an extensive review of the research on violence and mental disorder that there is a limited connection between mental disorder and violence.

Probably the most extensive research examining the relationship between violence and mental disorder was undertaken in the 1980s in what is known as the National Institute of Mental Health's Epidemiological Catchment Area Surveys involving close to 10,000 interviews in various parts of the United States. Swanson et al (1990) found that 2.1 % of those who did not meet DSM-III criteria for mental disorder reported committing a violent act in the previous year. This compares with 12.7% for those who met criteria for schizophrenia, 11.7% for major depression, 11% for mania or bipolar disorder, 24.6% for alcohol abuse/dependence and 34.7% for drug abuse/dependence. "Violence was most likely to occur among young, lower class men, among those with a substance abuse diagnosis, and among those with a diagnosis of major mental disorder" (Monahan, 1992). Monahan (1992, p.517) identifies further analysis based on this data which suggests that when "... patients were not experiencing psychotic symptoms, their risk of violence was not appreciably higher than a demographically similar members of their home communities...".

Violence and specific symptoms of mental illness:

Epidemiological research serves to confirm the assumption that alcohol or drug abuse may be one of the conditions that may predispose one towards violence and tentative links may be drawn between the symptoms of mental illness and violence. However further research is necessary to provide precision in the clinical decision making process. For example most people with a diagnosis of schizophrenia are not violent in fact they may be more likely to be victims of violence than perpetrators (Bridgeman, 1994).

A frequent method of inquiry used to illuminate the nature of violence within the mentally ill as a subgroup has been to examine the characteristics of those psychiatric patients who have been violent and compare them with groups who have not. Blomhoff, Seim and Friis (1990) found that the best single predictor of violence was a history of previous violence. Other significant findings were a high degree of violence in the family of origin and a high level of aggression on admission accompanied by an absence of anxiety.

An assumption of a great deal of research is that if a greater number of symptoms and demographic characteristics are explored a more precise model might evolve to inform predictions of dangerousness. Paranoid schizophrenia is one diagnostic subgroup which has long been associated with violence (Blair and New, 1991). Junginger (1995) explored compliance with command hallucinations, a relatively common phenomena which may accompany this disorder. Based on self reporting by psychiatric patients he found that people were at risk of violence if they experienced command hallucinations. People who could give an identity to the hallucinated voice were more likely to comply with the commands and most significantly, the dangerousness of the commands appeared to be a function of the environment, that is people tended to experience less violent commands in hospital than those experienced elsewhere.

Neurobiological explanations of violence:

Yet another approach to understanding aggression and violence is the quest for neurobiological correlates. If structural and functional changes in the brain are found to be associated with violent behavior then the labeling and treatment of dangerousness could truly be claimed as the legitimate domain of medicine.

Damage to certain centres of the brain such as the limbic structures, temporal, and frontal lobes have been found to be associated with aggressiveness and rage (Garza-Trevino, 1994). Certain medical conditions such as temporal lobe epilepsy have also been linked with violent behavior in humans and other species (Garza-Trevino, 1994).

It is doubtful whether abnormalities in genetic structure, endocrine function or neurotransmitter balance can fully account for most violent behavior. Kalat (1992, p. 433) for example points out that the highest incidence of violence, as measured by crime statistics, is in men 15 to 25 years old, who also have the highest level of testosterone in the blood. However not every young man with elevated testosterone is violent. While biological factors may account for violent behavior in a small number of individuals, and a predisposition in others it is more likely that the physical and social environment plays at least a modifying role.

Violence and the environment:

The research by Junginger (1995) is particularly interesting because it suggests that the environment has a modifying effect on symptoms of illness, in this case the "dangerousness" of command hallucinations. Other research has focused on environmental or temporal factors such as overcrowding, staffing levels, noise and time of day with mixed and sometimes contradictory results (Blair and New, 1991). More consistent and reliable results may be found in the research on violence and the social environment.

Violence and human interaction:

Violence towards other does not occur in an interactional vacuum. At least one other person must be involved and a pattern of interaction usually precedes violence. Sheridan et al (1990) examined perceived events leading up to the use of restraints in a psychiatric unit from the patients' perspective. Contrary to expectation most patients perceived events leading up to use of restraints as external rather than internal to themselves. Thus delusions and hallucinations were seen as events that triggered aggression in only a minority of cases, whereas patient-staff conflict, enforcement of rules and conflict with other patients were more commonly perceived as precipitants (Sheridan et al, 1990). Such research tends to confirm the idea that it is not the symptoms of mental illness which tend to trigger violence but rather it is the response of others to the person or a the individuals response to others.

The views that most people who are violent are out of control and that staff need to control patients are outdated and may in themselves precipitate unnecessary conflict (Morrison, 1993). Watson's (1991, p.14) study of care and control in psychiatric settings suggests that "...controlling practices contribute to the stress of mental illness and provoke the very behaviors they are designed to contain". Violence then, may become a self fulfilling prophesy. Staff may create distance between themselves and those whom they perceive as dangerous. Under the umbrella of "safety" controlling practices may ensue creating conflict, culminating in violent behavior. The person may then be said to have a history of violence justifying further controls and possibly reinforcing that violent behavior.

The notion that violent behavior can be learned and reinforced is validated by research by Morrison (1994, p.249) who found that of psychiatric patients who had been violent most utilised what she called a "coercive" interactional style", that is "...using others for self gain". She found that this style of relating overshadowed even a history of violence as a predictor of latter violence (Morrison, 1993). Morrison (1993) proposed that most violence among those with mental illness was similar to violence among wider society. The key to preventing violence is to eliminate the rewards for violence and to reinforce non-coercive styles of relating. The rewards for violence will be idiosyncratic. For some it may be seclusion or time-out, for others merely the sense of power that one has over another, for others it may be the focused attention of health professionals.

The conditional nature of violence:

As nurses and clinicians we are equipped with an armory of statistics on who might be violent towards others e.g. the young, intoxicated, male patient with schizophrenia, We also have a responsibility to interpret and use these findings with caution. While useful in telling us who might be violent, the research is far from useful in telling us who is dangerous with any specificity. The reality of skilled clinical decision making is such that clinicians take into account the conditional nature of violence (Mulvey and Lidz, 1995). Anyone may be dangerous under certain conditions. For example some people tend to be violent in their home environment and otherwise present a respectable face to the world, others are violent only when intoxicated. The symptoms of mental illness for most people are not necessary conditions in themselves for violence.

Implications for practice:

When in hospital nurses are the frequent targets of violence. It is therefore important that nurses have an understanding of why people in general are violent and most importantly the conditions under which the individual patient may be violent. This means being involved with patients and attempting to understand them as individuals. This will be evidenced when we see care plans in action which takes into account the conditions under which the individual has been and is likely to be violent rather than the generic "potential for violence" that is so common. Interventions are needed which set explicit boundaries for socially acceptable behavior and consequences for unacceptable behavior, while also reinforcing non-coercive styles of relating.

It will now be evident that comparing the dangerousness of a person with an object such as a floor is flawed. Unlike a floor people can and do use violence in a purposeful way to get what they want. Labeling a person as dangerous may have far reaching effects for that individual. The label may travel with that person affecting their relationships with other people to such an extent that in some instances future violence becomes a function of the effect of the label. On the other hand the effect of labeling a floor as dangerous is likely to be the minimization of danger. One is unlikely to hear "Be careful.. that floor was once dangerous, it may be dangerous again" but the opposite is true for people. There is a potential that people will relate to the individual who is labeled as dangerous in a manner which increases that person's sense of alienation and which may reinforce and even provoke violent behavior. When aware of this dynamic health professionals are challenged to respond to the individual in a way which validates and empowers.

Nurses must challenge coercive and controlling institutional practices that may reinforce and provoke violent behavior. Opportunities for patients to get what they want and need without recourse to violence must be provided. At the same time the risk of violence must be acknowledged and the victims of violence supported. These are major challenges for nurses and the institutions in which they work. Health professionals face the dilemma of balancing a societal mandate to control those with mental illness and a mandate to care for them. All who work with the mentally ill, in whatever setting must ensure that caring takes precedence.

Some questions for reflection:


Blair, D.T., & New, S.A. (1992). Patient violence in psychiatric settings: Risk identification and treatment as provocation. In S.A. Smoyak, & D.T. Blair (Eds.). Violence and abuse. (pp. 36-53). Thorofare, NY: Slack.

Blair, D.T., & New, S.A. (1991). Assaultive behavior: Know the risks. Journal of Psychosocial Nursing, 29(11), 25-29.

Blomhoff, S., Seim, S., & Friis, S. (1990). Can prediction of violence among psychiatric inpatients be improved? Hospital and Community Psychiatry, 41(7), 771-775.

Bridgeman, G. (1994). Dangerousness and psychiatric illness. Mental Health News, Autumn, 10-12.

Conrad,P., & Schneider, J.W. (1980). Deviance and medicalisation: From badness to sickness. St Louis: C.V. Mosby.

Fisher, A.A. (1989). The process of definition and action: The case of dangerousness. Doctoral dissertation, University of California, San Francisco.

Garza-Trevino, E.S. (1994). Neurobiological factors in aggressive behavior. Hospital and Community Psychiatry, 45(7), 690-699.

Junginger, J. (1995). Command hallucinations and the prediction of dangerousness. Psychiatric Services, 46(9), 911-914.

Kalat, J.W. (1992). Biological psychology (4th ed). Belmont, California: Wadsworth, Inc.

Levey, S., & Howells, K. (1995). Dangerousness, unpredictability and the fear of people with schizophrenia, Journal of Forensic Psychiatry, 6(1), 19-39.

Monahan, J. (1992). Mental disorder and violent behavior: perceptions and evidence. American Psychologist, 47(4), 511-521.

Morrison, E.F. (1994). The evolution of a concept: Aggression and violence in psychiatric settings. Archives of Psychiatric Nursing, 8(4), p. 245-253.

Morrison, E.F. (1993). Towards a better understanding of violence in psychiatric settings: Debunking the myths. Archives of Psychiatric Nursing, 7(6), 328-335.

Mulvey, E.P., & Lidz, C.W. (1995). Conditional prediction: A model for research on dangerous to others in a new era. International Journal of Law and Psychiatry, 18(2), 129-143.

Sheridan, M., Henrion, R., Robinson, L., & Baxter, V. (1990). Precipitants of violence in a psychiatric inpatient setting. Hospital and Community Psychiatry, 41(7), 776-780.

Swanson, J.W., Holzer, C.E., Ganju, V.K., & Tsutomu Jono, R. (1990). Violence and psychiatric disorder in the community: Evidence from the epidemiologic catchment area surveys. Hospital and Community Psychiatry, 41(7), 761-770.

1996 Richard Lakeman