Workplace Violence: APNA 2008 Position Statement

Executive Summary Introduction Outpatient Psychiatric Settings Interventions
Recommendations for Inpatient and Outpatient Settings Appendix Acknowledgments References
Download the full PDF

Inpatient Violence in Psychiatric Settings

On September 3, 2006, Wayne Fenton, psychiatrist, researcher, National Institute of Mental Health scholar, editor of the Schizophrenia Bulletin, and decades-long advocate for people with schizophrenia was shot and killed by one of his patients (Washington Post, 9/4/06). Rare and high-profile incidents like this serve as chilling reminders of the service provider’s vulnerability in psychiatric care. Stalking, sexual assault, and threats are also concerns (Love & Elliott, 2002; Sandberg, McNeil, & Binder, 2002).

Inpatient psychiatric settings include freestanding mental health facilities, private and public inpatient units affiliated with medical hospitals, county units, state hospitals, and forensic units in prisons, jails, or hospitals. Care of the most severe developmentally disabled individuals may occur in an inpatient state facility. The populations of these various inpatient settings range from patients with anxiety, depression, psychosis, eating disorders, dementia with psychosis, or behavioral difficulties to crisis-prone, more agitated, and usually repetitively admitted persons. Dangerousness as a patient trait is often seen in acute care environments and state hospitals. Patients who are court admitted for crimes are most often seen in state hospital forensic units. In acute inpatient settings, services include treatment of acute psychiatric symptoms, detoxification, and crisis stabilization.

“Dangerousness” is the standard for both civil and penal involuntary commitment in the United States and most Western countries. Violent behavior toward self or others is typically the precipitant for involuntary admission to an inpatient unit (Doyle & Dolan, 2002). Thus, people remanded to acute care settings have been found to be dangerous just before their admission. Highly dangerous and mentally ill people are brought to inpatient settings because it is not safe for them to live in the community or in less secure settings.

Historical Perspective

Psychiatric violence is a long-standing, complex and multifaceted phenomenon -- the final common pathway of a variety of factors having to do with the immediate environment and the individual’s internal world. For centuries, and in relative professional isolation, inpatient psychiatric nursing staff have cared for people prone to violent and self injurious behavioral emergencies--often resorting to coercive, rudimentary, and primitive measures to maintain safety.

The early inpatient violence research (1970s – 1980s) tends to characterize the violent inpatient as a unitary phenomenon--where patients were either “violent” or “not violent.” We now know that the violent inpatient is a heterogeneous group of people with a variety of precursors and motivations. As far back as 1989, Poster and Ryan studied the prevalence of assault in psychiatric nursing, and in 1992 Lanza initiated a scholarly program of research delineating psychiatric violence and its consequences.


Nurses working in some psychiatric inpatient settings are exposed to violence on a daily basis due to the nature of the populations served, the public protection functions of inpatient settings, the culture and demands of the institutional environment, the reduced number of RNs to lesser-trained mental health workers, and the limitations of the treatment services provided. This last is especially true of state hospitals and, in some localities, in public and private hospital psychiatric units.

Staff surveys show that 75% to 100% of nursing staff on acute psychiatric units have been assaulted during their careers (Caldwell, 1992; Hatch-Maillette et al., 2007). Poster (1996) reported that 75% of psychiatric nursing staff experienced an assault at least once during their career. Caldwell (1992) found that 62% of psychiatric clinical staff and 28% of nonclinical staff reported that patients assaulted them at least once. And from the same study, 28% of clinical staff and 12% of nonclinical staff reported an assault within the last 6 months. Poster and Ryan (1989) found that 94% of Canadian psychiatric nurses reported having been assaulted at least once in their careers and 54% reported having been assaulted more than 10 times.

For psychiatrists, the risk of violence when treating mentally ill patients is more than four times greater than the risks facing other physicians. An article from The Times of London (Ahuja, 2006) reported that the rate of nonfatal, job-related violent crime among general medical physicians is 16.2 per 1,000.

For psychiatrists and mental health professionals, the rate is 68.2 per 1,000. Bjorkly (1999) found that 15% to 30% of hospitalized psychiatric patients were involved in physical assaults during their hospitalization. In a study by Hatch-Maillette and colleagues in a state psychiatric facility, it appeared that a high degree of patient contact placed the staff at the greatest personal risk of an incident of workplace violence (Hatch-Mailette et al., 2007). Gale, Pellett, Coverdale, and Paton Simpson (2002) examined risk factors for violent events reported to managers in psychiatric units in New Zealand. Results of the survey indicated the following: (a) Inpatient care poses a higher risk to staff for all violent events than does community care; (b) geriatric psychiatry has a higher risk of assault, attack, and sexual harassment than other forms of care; and (c) child and adolescent psychiatric units, alcohol and substance abuse units, and other specialty units were no more violent than other units.

The literature consistently reports that a small number of individuals are responsible for most of the assaults in institutional settings (Convit, Isay, Otis, & Volavka, 1990). Bjorkly (1999) found that four patients were associated with 77% of the violence in a Norwegian secure unit. Using data from a large inpatient forensic hospital, Love and Hunter (1996) found that the habitually violent patients were younger, had a length of stay in excess of 3 years, and averaged 20 more assaults than the average inpatient. Focusing resources to prevent violence in this group has the potential to substantially reduce overall levels of institutional aggression.

Forensic Settings and State Hospitals

People are in forensic settings because they have come to the attention of both the mental health and criminal justice systems. Forensic inpatients generally receive treatment in prison psychiatric facilities or state hospitals. The forensic populations in the United States are burgeoning. Prisons and jails are overcrowded with people who have serious psychiatric problems, trauma histories, and addictions.

In the forensic mental health nursing literature, the tensions between criminal elements and psychiatric phenomena have given rise to the widely debated “custody and care” dilemma (Scales, Phillips, & Crysler, 1989). The tension created between the police power of the state (i.e., custody and security) can sometimes seem at odds with the clinical mandate to provide care and treatment and to meet the requirements of external medically oriented regulatory bodies (Love & Morrison, 2003).

State mental health hospitals and psychiatric units serve people who experience difficult-to-manage psychotic disorders and severely limiting personality disorders. People in state hospitals require long-term treatment, either for their own protection or the protection of others. Most state hospital inpatients are involuntarily committed. It is not unusual for civil commitment (state hospital) populations to contain a large proportion of people who have criminal offenses in their background.

As public sector institutions, state hospitals are “the end of the line” for many severely impaired and habitually violent people. State hospitals house an accumulation of dangerous people who have failed many treatment trials in various settings. Public sector patients tend to either not respond or only partially respond to psychotropic medications. This accumulation of severely impaired people can lead to cultures of hopelessness, coercion, and violence.

Many states have enacted “sexual violent predator” legislation, which has led to a large number of the most dangerous sex offenders being civilly committed to psychiatric hospitals after they have served their prison terms (Sreenivasan, Weinberger, & Garrick, 2003). The sexually violent population tends to be difficult for mental health staff to manage, especially since legal issues may be involved in their care. As the number of highly dangerous, hospitalized sex offenders increases, the frequency of sexual stalking and verbal abuse increases.

Targets of Inpatient Aggression

Targets of inpatient violence can include staff, visitors, and fellow patients. In a large forensic state hospital in the United States, Love (1995) found that patient-to-patient aggression was more common than patient-to-staff aggression. In a review of British psychiatric hospitals, nursing staff were the most frequent targets of assault (Trenoweth, 2003), as was the case in a study of London hospitals (Gournay, Ward, Thornicroft, & Wright, 1998).

Due to long exposure and numerous security functions, psychiatric nurses are assaulted more frequently than other members of the interdisciplinary team (Hatch-Maillette et al., 2007). Among U.S. nursing staff, injury rates resulting from inpatient violence are twice as high as injuries from all causes in the traditionally high-risk industries of mining, lumber, and heavy construction (Lipscomb & Love, 1992). Love and Hunter (1996) found that, in a 1-year period, in five different public sector hospitals, 14% of registered nurses and 25% of nursing aides experienced an OSHA-reportable injury due to patient violence. In these five settings, staff members were injured most often when physically containing an individual during a crisis. Injuries from direct assaults were less frequent. These alarmingly high numbers are twice as high as the occupational health industry standard (Lipscomb & Love 1992).

Some reports suggest that assaults occur less frequently to female staff members than to male staff. Hunter and Carmel (1993) noted that male staff were overrepresented in their 1-year sample of staff injuries in a forensic hospital. In a recent report men and women were equally at risk of violence in inpatient settings, and typically the perpetrator was the same gender as the victim (Flannery, Marks, Laudani, & Walker, 2007). Among 328 mental health professionals from a state psychiatric facility with forensic, inpatient, and community-based patients, female staff reported more verbal and physical assaults and more frequent sexual threats compared with male staff (Hatch-Maillette et al., 2007).

Risk Management and Prediction of Violence

Nursing staff and other mental health professionals are expected to be able to identify imminent dangerousness and predict the potential for future violence. In the early 1980s, researchers undertook the challenge of measuring the accuracy of clinician’s dangerousness predictions. These early studies, dubbed the “first generation” of violence prediction technology, relied on clinical judgments and repeatedly found that clinical impressions were inaccurate to the extent that they tended to be wrong more often than they were right (Monahan, 1981, 1984). The limitations of this first-generation technology led to the identification of important distinctions among various violent inpatients across a variety of settings.

The “second generation” of prediction research employed actuarial methods to measure the relative contributions of specific perpetrator-based variables categorized as either “static” (i.e., fixed and historical factors) or “dynamic” (changeable). Static factors include items such as gender, history of violence, childhood experiences, and behaviors. Dynamic factors include items such as symptoms, age, setting, and degree and type of substance impairment. This actuarial approach to risk assessment vastly improved the accuracy of violence prediction. Static historical factors alone are consistently found to be more accurate than dynamic factors. A history of violence remains the single most important predictor of violence in inpatient settings, particularly a history of institutional violence (Love & Elliot, 2002).

Psychiatric nurses play an important role in the identification of risk factors for violence and in the implementation of interventions that promote and maintain safety. Much of the prediction research has been directed at community placement predictions. Only recently have prediction efforts focused on prediction of violence in institutions (Johnson, 2004). Interactions typical on inpatient units, such as limit setting, denying a request, gaining compliance, involuntarily medicating someone, and de-escalation are associated with violent incidents and emphasize the importance of mental status assessment skills, therapeutic communication competency, unit environments, and nurse-patient alliance.

Trenoweth (2003) found that nurses tend to rely extensively on their personal knowledge of their patients when assessing for dangerousness. The study indicated that nurses perceive the development of nurse-patient relationships and working in a supportive team as protective factors against risk. Protective factors specific to the patient relationship included knowing the patient, understanding the patient’s frame of reference, recognizing the impact the mental health problem has on the patient, being aware of the patient’s history of violence, observing the situation, and identifying behavior chains leading to aggression. When faced with a potentially violent situation, the nurses were able to draw on specific knowledge of the patient to effectively intervene. The study demonstrated that nurses believe the level of risk in a potentially violent scenario does not stem solely from factors within the patient but also reflects external factors, such as the skills of staff, the ability to work effectively in a team, the presence of others who could escalate aggressive behavior, and the availability of weapons.

Several instruments have been tested relative to efficacy for inpatient violence prediction. Bowers, Nijman, and Palmstierna (2007) compared the Attempted and Actual Assaults Scale (Attacks) with Modified Overt Aggression Scale (MOAS)—instruments designed to record the nature and severity of inpatient assaults. The Attacks scale is completed after a violent incident has taken place, and captures details of inpatient violent events. The MOAS measures verbal aggression, property damage, and violence toward self and others. To compare the instruments, the researchers had staff subjects view videotapes of inpatient assaults and then rate the severity of the assault using both instruments. In their study, the Attacks scale was considered superior to the MOAS.

Another tool, the Dynamic Appraisal of Situational Aggression (DASA), was developed and tested by Ogloff and Daffern (2006). The DASA is a 16-item scale drawn from research and other scales including the Broset Violence Checklist and the Historical, Clinical, Risk Management-20 (HCR-20). DASA items include; irritability, impulsivity, unwillingness to follow directions, sensitivity to perceived provocation, easily angered when requests denied, negative attitude, and verbal threats, with irritability as the strongest predicting factor. The DASA scale predicted inpatient violence more accurately than nurses’ clinical judgments alone. Tools such as the items mentioned above can provide greater accuracy in identifying the risk of violence in psychiatric settings and can potentially augment clinical prevention efforts.

Flannery, Marks, et al. (2007) examined the impact of assaults over a 15-year period and produced several cost-effective risk management strategies. From an organizational perspective, risk management strategies should include the development of a quality management database to assess patient and staff victim characteristics, provide clear definitions of assaults to increase reporting, develop and uphold workplace violence policies, create restraint-free environments that incorporate trauma, informed care, post-crisis counseling for victims, and improve law enforcement liaison links within the community. Staff development strategies should include training in behavioral warning signs of impending loss of control; conflict resolution skills; coping skills; nonviolent self-defense, restraint and seclusion procedures; alternatives to restraint and seclusion; psychological trauma; and de-escalation skills.

In another study, staff from 111 different psychiatric units were queried about the preventive and safety measures in their settings. In 45% of the units, personal alarms were provided, 41% had panic buttons installed, 27% had security alarms, 2.7% had metal detectors, 30% had deescalation teams in place, and 44% had security staff on site (Gale et al., 2002).

Weapons screening is a preventive measure. In a 5-year retrospective study of weapon use in a large forensic hospital, Love and Hunter (1996) found that the most common weapons used by patients in assaults were objects readily available in the environment (e.g., chairs, food trays, trash cans, tables, pens and pencils, and silverware).

“Manufactured weapons” (e.g., shanks, blackjacks and garrotes) were rarely used in attacks, but were confiscated from patients who were carrying the handmade items for self protection or self harm. In this study, staff and patients were targeted equally in weapon incidents. The authors endorsed weapon-screening practices and careful adaptations of the hospital environment to reduce the number of items that patients could use as weapons.

Patient Factors

The violence research that focuses on the patient tends to attribute aggression to biologically based emotional regulation, especially the serotonergic system (Blair & Charney, 2003). Link, Stueve, and Phelan (1998) studied the effects of various delusional systems (“threat-control override”). Bjorkly (1999) focused on the role of emotional distress as a mediator between the individual’s thoughts and violent behavior. Command hallucinations calling for the patient to hit another, although relatively rare, have been associated with an increased risk of violence, particularly when the individual’s emotional mediators are unable to override the command voices.

The literature suggests that internal events—the evidence of symptoms such as delusions and hallucinations—are risk factors for assaults in newly admitted patients. External events— primarily a conflict with staff or another patient—were more commonly perceived as precipitants by patients (Crowner, Peric, Stepcic, & Lee, 2005). Interpersonal factors, such as abrasive words, property struggles, disagreements, and invasion of personal space were more commonly offered as reasons for assaults than internal factors (Crowner et al., 2005). When queried about the provocations of violent incidents, patients often report coercive behavior on the part of staff as the key provocation (Morrison, 1992, 1993).

Other researchers conducted a study to establish whether there are temporal and causal relationships between psychotic symptoms and assaults (Nolan et al., 2003). The authors found that 20% of the assaults were directly related to the presence of positive psychotic symptoms. Patients with psychosis-motivated aggression reported delusions and hallucinations with threatening content more frequently than command hallucinations.

A study conducted by Bowers, Allan, Simpson, Nijman, and Warren (2007) found a relationship between adverse incidents and psychiatric admissions, particularly of male patients. The results indicate the following:

In a geriatric population, Wystanski (2000) conducted a study to evaluate the relationship between psychosocial stimulation and changes in medications, with the emergence of assaultive behavior and the course of behavior, in a 24-hour period. Assaultive behavior occurred during 222 of 1,396 observations. Patients with organic brain disorders displayed more assaultive behaviors than those with nonorganic conditions. The study also found that with no modifying factors, such as changes in medications or psychosocial stimulation, the proportion of patients who became assaultive in the first 24 hours of admission was higher in the organic group than the nonorganic group. In the presence of modifying factors, a 20% decrease in aggression occurred in patients in the organic group, and the group with nonorganic conditions experienced a 9% decrease in aggression (Wystanski, 2000). This study confirmed that patients with organic conditions display more aggression and that the use of psychotropic medications is effective in the management of aggression in the inpatient geriatric population. Impulsivity, which is a common symptom of organic brain disorders, may be a factor in this group.

Crowner and associates (2005) conducted a study that focused on episodes of assaults between chronically ill patients in a long-term psychiatric care facility. The study grouped behavioral cues into three categories: threatening behaviors (yelling, arguing, physical contact), intrusive behaviors (following, touching, kissing, or placing a body part within 6 inches of someone), and mixed cues (approach or initiation, competition, immediate threat, and the use of a karate or boxing stance). Results indicate 60% of assaults were preceded by at least one threatening or intrusive behavior; however, the earliest cues of assault were noted just 2 minutes before the assault (Crowner et al., 2005).

Personality and Violence

While acute symptoms such as delusions and hallucinations have been implicated in various violence prediction schematics, the importance of enduring personality traits is also an important clinical consideration. Features such as lifelong difficulties in regulating emotions, impulse control problems, narcissistic and entitlement cognitive style, and a tendency toward paranoia and a hostile attributional bias are all factors relevant to violence management in hospital settings. The cluster B personality disorders, most notably antisocial personality disorder and borderline personality disorder, include violent behavior in their diagnostic criteria and have been associated with inpatient violence and habitual violence.

Psychopathy is a damaging personality syndrome that includes emotional, interpersonal, and behavioral characteristics such as a callous disregard for others, shallow emotion, egocentricity, lack of empathy, pathologic lying, proneness to boredom, thrill seeking, and a propensity to highly impulsive and irresponsible behavior and criminal versatility (Hare, 1991, 2003). It has been noted that while only about 1% of the general population has the psychopathic syndrome, these people are associated with a markedly disproportionate frequency of violent crime, white collar crime, and other forms of social distress (Forth, Kosson, & Hare, 2003). Compared to the general population, people with psychopathy are vastly overrepresented in the forensic population (Hare, 2003). The gold standard in measurement of psychopathy is the Psychopathy Checklist Revised (PCL-R). A high score on the PCL-R is a highly robust predictor of future violence. It has been estimated that between 15% and 20% of forensic patients score high on the PCL-R. Among sex offenders, especially rapists, the percentage of high scorers has been reported to be between 35% and 50%. The presence of psychopathy has important implications for assessing risk of dangerousness and for risk management in helping relationships (Love & Hunter, 1996).

Staff Factors

Patient characteristics alone cannot completely explain the violence that occurs in psychiatric units; certain staff member characteristics, attitudes, or communication styles may result in staff being the target of violence (Ray & Subich, 1998). Mackay, Paterson, and Cassells (2005) conducted a qualitative study of mental health nurses to examine the process of observation of violent patients, nursing skills used during the observation, and benefits and drawbacks for the observed patients. The process of close observation of acutely disturbed psychiatric patients was more than “just watching”; it was described as both caring and interactive. Six effective nursing roles included intervening, maintaining safety of the patients and others, preventing de-escalation and managing aggression and violence, assessment, communication, and therapy. Nurses skilled in these roles, along with years of experience, were thought to have an impact on the success of interventions in practice.

The immediate environment can either raise or lower an individual’s level of dangerousness, and nursing staff behaviors function as antecedents and consequences to aggression (Love & Hunter, 1996). Whittington and Wykes (1994) found that 86% of assaults on staff occurred immediately after the nurse intervened with a patient frustration or requested a patient do some activity.

The recent mental health consumer movement has made it clear that service providers tend to overuse coercive measures. The terms psychiatric survivor and sanctuary harm appear in the lay and scientific literature (Robins, Sauvageot, Cusack, Suffoletta-Maierle, & Frueh, 2005). The degree to which a patient feels coerced (and therefore powerless or cornered) depends largely on the manner in which staff communicates with the individual, according to Monahan and colleagues (1995). In his classic study of asylums, Goffman (1961) noted that a basic split exists between staff and patient realities—two world views that essentially run parallel, rarely intersecting. His observations apply to contemporary institutions. Opinions vary widely between consumers and providers when asked about the same violent incident. Typically, patients perceive staff behavior to be more coercive than the staff members think it is. In a review of the literature on consumer perceptions, Abderhalden, Hahn, Bonner, and Galeazzi (2005, p. 74) concluded that “among all research samples, users consider staff behaviors as a central cause of violence describing it as ‘provocative and disrespectful.’” Power struggles between staff and patients commonly take place in secure settings. Power struggles, if not handled delicately and diplomatically, can activate a violent incident.

Researchers now understand that violence in psychiatric hospitals is highly contextually based. The experiences of both parties—staff and patients—can be useful to deepen the staff’s understanding of violence across many inpatient settings. Units with high predictability, order, and organization have fewer violent incidents than units that are haphazard, disorganized, and unpredictable (Katz & Kirkland, 1990).

The 2007 APNA Position Paper on Seclusion and Restraint discusses prevention of aggression by (a) maintaining a presence on the unit and noticing early changes in the patient and the milieu (Delaney & Johnson, 2006; Johnson & Delaney, 2007); (b) assessing the patient and intervening early with less restrictive measures such as verbal and nonverbal communication, reduced stimulation, active listening, diversionary techniques, limit setting, and as-needed medication (Canatsey & Roper, 1997; Johnson & Delaney, 2007; Johnson & Hauser, 2001; Lehane & Rees, 1996; Maier, 1996; Martin, 1995; Morales & Duphorne, 1995; Richard, Trujillo, Schmeizer, Phillips, & Davis, 1996); and (c) changing aspects of the unit to promote a culture of structure, calmness, negotiation, and collaboration rather than control (Cahill, Stuart, Laraia, & Arana, 1991; Delaney 1994; Harris & Morrison, 1995; Johnson & Morrison, 1993; Whittington & Patterson, 1996).


Relative to violence prevention, the role of staffing is consistently cited as a critical factor. To date, the relationship between staffing and skill mix is poorly understood. Bowers and colleagues (2007) found a significant relationship between lack of staff availability (and/or lack of adequate staffing levels) and the incidence of violence.

In many psychiatric settings, the least-trained staff members (e.g., ancillary staff, aides, mental health technicians, behavior specialists, etc.) spend the most time with the patients. The training requirements for ancillary mental health workers vary widely from state to state (Quinn & Love, 1993). Many of these care providers are unlicensed and have limited on-the-job training. Many studies from medical nursing settings demonstrate that the proportion of registered nurses to other nursing personnel is directly related to quality indicators. That is, the higher the nurse-toancillary staff ratios, the better the care. In her study of five inpatient settings, DeLacy (2005) found a negative correlation between the proportion of nurses to other nursing personnel and the number of hours and events of seclusion or restraint.

Staff Education

To date, there are no national standards for staff training in the prevention and management of inpatient violence. Given the scope and significance of inpatient violence, it is both curious and troubling that staff training programs vary widely in content and process and tend to lack a consistently scholarly and empirical base. Training in the management of assaultive behavior varies from setting to setting and can be as short as one hour and as long as ten days.

There are no standards or evidence compelling enough to draw conclusions about which patient containment tactics are safest. Historically, psychiatric workplace safety programs have been “home grown” programs designed by local staff involving physical tactics for containment, restraint and seclusion, and basic physical self-defense techniques. Farrell and Cubit (2005) evaluated the content of 28 violence prevention and management training programs. They found that the use of restraints, pharmacologic management of aggression, and seclusion were common features. Most programs did not address the psychological and organizational costs associated with aggression in the workplace. Although the authors concluded that the programs they reviewed did tend to cover basic content necessary to educate staff, Farrell and Cubit recommend further research to determine the effectiveness of these training programs.

Many violence prevention training programs focus on de-escalation skills for staff, and they teach that observable autonomic arousal (e.g., loud voice, muscle tension, pacing, pupillary dilation) is a precursor to inpatients’ violent incidents. Violence accompanied by a cycle of escalation is linked to emotional arousal. The patient is visibly upset and getting worse by the moment. Staff members learn to stay calm, lower their voices, and avoid crowding or unreasonable demands. They make every effort to assist the individual in crisis by offering reassurance, negotiating, and arranging choices to allow the individual to “save face.”

Signs of emotional arousal do not precede all violent events, however. Certain kinds of events, by certain patients, may appear “unprovoked” or are carried out in an unemotional, cold, instrumental manner. The belief that some violent events are unprovoked is a fallacy. Behaviorally speaking, all human behavior has provocations and antecedents. Rather than being unprovoked, it is more accurate to say that provocations were hidden or unobserved. Thoughts and emotions can be provocations or antecedents to violent behavior. Predatory or instrumental violence (violence toward some end) tends to be associated with severe personality disorders, particularly antisocial and psychopathic traits. Training programs tend to miss this less frequent but potentially more dangerous subtype of violence. Being able to identify an individual’s psychopathic traits is a competency for all psychiatric nurses as they assess individuals who are at high risk of violence.

While most training programs emphasize de-escalation skills, it should be noted that deescalation is a relatively late intervention. De-escalation implies that the crisis is already beginning. Efforts to establish rapport and to “know the patient” are early interventions. Inpatient staff should be trained to establish rapport with patients immediately so that, when a crisis begins, staff will be more influential and effective in preventing an emergency. Although people in psychiatric settings vary widely in their need and tolerance for alliance building, nursing staff should, whenever possible, engage with the people they serve and convey trust, empathy, consistency, and fairness.

Love and Morrison (2003) noted in their white paper on workplace violence that staff training programs in hospitals should include how to remain safe in a hostage situation. Although hostage-taking situations in heath care are rare, they can be lethal. Hospitals should have an emergency hostage-response plan in place and maintain a close alliance with local police.

Prosecution for Assault

Prosecution for criminal acts committed in a psychiatric hospital pose inherently complicated ethical, practical, clinical, and occupational health consequences (Coyne, 2001). Victims of an assault (patient or staff) have the right to report the incident to police. Staff have the obligation to report instances of dependent adult abuse, even if a fellow staff person or patient committed the abuse. The factors influencing the decision of whether to take action against an individual for an assault include the severity of the assault, the mental state of the patient, the context in which the assault occurred, and opinions about the therapeutic value (or lack thereof).

For many reasons, individuals in psychiatric settings often are not charged for their violent behavior. For example, district attorneys may be reluctant to prosecute assault cases for people who are already receiving treatment because of a criminal offense or people who are being restored to competency for a crime. Prosecution may affect the staff victim’s recovery from an assault. If, on the one hand, the hospital is unwilling to submit the case for prosecution, this can lead to righteous indignation and anger. If a case is submitted and accepted, the staff victim or victims most likely will be involved in protracted legal proceedings and may have to face the assailant in court. This process can reactivate trauma responses.

When faced with a patient’s criminal behavior, mental health agencies may resort to other forms of action such as discharge from the facility and transfer to a more secure, restricted care environment or other changes in the treatment regimen (Coyne, 2001).

Mental health agencies may not pursue prosecution for other reasons, such as negative publicity, concerns that the provision of service will be scrutinized, assurance of patient confidentiality, or the belief that staff provoked the attack. On the other hand, ignoring criminal behavior in the hospital undermines hospital security and creates the impression that antisocial behavior is acceptable. Many antisocial individuals go through life believing rules are for other people. If the hospital does not prosecute certain individuals for criminal behavior, the hospital witlessly reinforces this cognitive distortion. Inpatient psychiatric settings should not be “felony-free-zones” that tolerate any and all forms of criminal behavior. Hospitals should have clear policies regarding who may prosecute whom for what and under what circumstances, as well as the process for making decisions on a case-by-case basis. Hospital bioethics committees may be involved. When violent instances go unreported to the police the instances never become part of the public record. Prosecution ensures that the aggressive act becomes an accessible part of the public record.

Post Violence Staff Needs

It is generally accepted that education and social support should be available for staff after they experience inpatient violence. The trauma and victimology literature provides a theoretical basis when studying the post-assault needs of nurses. For the past two decades, a large body of theoretical, clinical, and empirical literature has focused on the psychobiological effects of trauma. Various trauma survivor studies have empirically validated the following key points:

  1. Most victims of violence benefit from talking about a violent incident.
  2. The social environment of assaulted staff members can affect their recovery for better or worse.
  3. The victim’s proximity to the assailant after the incident has implications for recovery.
  4. The victimized staff member has to go about establishing a sense of personal safety and predictability in the workplace.

A Swedish study recommended group discussion for victims of patient violence (Arnetz & Arnetz, 2000). In Taiwan, Lu, Wang, and Liu (2007) sought to determine how occupational experience, in-service education, and social support can lessen the psychiatric nurse’s response to assault by a patient. The most common physical reaction was soreness in the area where the nurse was hit. Fear of the assaultive patient was the initial emotion, which was followed by feelings of anger. Other common reactions included the need to talk about the event, discomfort in caring for assaultive patients, suppression of the unpleasant feeling, desire to keep the incident a secret, and desire to retaliate. The greater the number of years of experience as a psychiatric nurse, the fewer the number of physical reactions reported. The more responses a nurse experienced in regard to the assault, the longer the time they required to return to work. Nurses who received more social support after the assault had less severe emotional and social reactions (Lu, Wang, & Liu, 2007).

Lanza, Demaio, and Benedict (2005) implemented an extensive educational program for nurses who have experienced assault. The program consisted of twelve 1-hour sessions, each focusing on a specific content area. Sessions covered introduction and sharing of the assault experience, statistics and victim response, analysis of the assault experience (two sessions), relationship with patients, coworker relationships, relationship with family, placement of blame, role conflict, interventions, coping strategies, and personal plan development. The group was limited to 10 members. Results of the study indicate that participants had positive experiences with the program, rating the success of educational objectives as 4.9 or higher on a 5.0 scale. The program increased participant knowledge about the various aspects of assault (e.g., how the family of the victim as well as the victim reacts) and gave participants a supportive network, which helped them recognize that there were other individuals who have experienced similar emotions after such events (Lanza, Demaio & Benedict, 2005).

Unguided education may not be enough after a workplace assault. A study conducted in the United Kingdom (Nhiwatiwa, 2003) found a high risk of repeated assaults at medium-secure psychiatric facilities. Nhiwatiwa suggested this made staff more vulnerable to denial, which could slow recovery time. The author evaluated an educational booklet aimed at reducing the effects of trauma and improving coping measures after violence. The nurses who were assaulted in four medium-secure hospitals in England and Wales and who received the booklet showed greater distress scores 3 months later than those who did not get the booklet. Nhiwatiwa found that the booklet alone, without directions, did not help the nurses to cope with the stress of the assault. The author suggested additional study and different use of the booklet.

Critical Incident Stress Debriefing (CISD) is an intervention designed to support healthy coping and to reduce the immediate and long-term reactions of trauma victims, witnesses, and responders. It focuses on prevention and early intervention and may reduce the risk of chronic, disabling emotional and physical consequences (Antai-Otong, 2001). The components of CISD include immediate emotional support, education about normal stress reactions, symptom reduction, and referrals for further intervention (Antai-Otong, 2001). In recent years many articles have claimed that CISD does not prevent PTSD. Some published reports have suggested that debriefing can actually make participants worse, by exposing them to graphic details from other participants. The studies from which these conclusions were drawn had no control over the quality and methods of the debriefing processes and were drawn from victims of car accidents, natural disasters, and violence in the community. Based on the available information, it is advised that (a) in extremely graphic incidents, first responders and eyewitnesses be debriefed separately from staff members who were not present at the traumatic incident; and (b) providers should refrain from claiming that CISD prevents PTSD, until the research findings become conclusive.

Practical recommendations derived from a study by Privitera and colleagues (2005) include the following: (a) multidisciplinary personal safety training to enhance team-building, improve communication, and help prevent violent events and (b) establishment of post-event protocols to assist staff victims and administrators in navigating complex issues occurring after a violent event (Privitera et al., 2005).

Administrative Policies

Hospital administration, staff, and even nurses have tended to view violence as part of the job. Expecting that patients will be violent renders everyone—staff, patients, and administrators— inured, or less concerned, when violent incidents occur and result in an institutional tolerance for violence (Blair, 1991). Whenever a violent event occurs, the administration should view it as a system failure and should evaluate it as such. If a patient assaults a staff member, the patient owns the violent behavior. Staff should not be blamed for the behavior of an individual patient or group of patients.

It should be made clear, however, that staff members are accountable for their competence in the workplace; that is, they are competent to perform therapeutically and safely in the work environment. In some situations, the manner in which the staff interacts with the individual may be provocative and tends to escalate the problem rather than defuse it. Administration must not tolerate staff behaviors that are nontherapeutic; these behaviors demonstrate incompetence, and processes are necessary to correct these individual’s competence.


Some states have added nurses to their list of protected job classifications. So-called “hit a nurse, go to prison” legislation makes it a felony to assault a nurse. It is now a felony to assault a nurse in Alabama, Arizona, Illinois, Massachusetts, Nevada, and New Mexico. In other states it is a felony to assault a police officer, an emergency medical technician (EMT), an animal regulations officer, and a little league umpire, but not a physician or nurse. At the time of this writing, New York nurses are lobbying for the same type of legislation (Press and Sun-Bulletin, April 8, 2008). Consumer groups, however, have lobbied to prevent this type of legislation from passing, for fear of “criminalizing the mentally ill.”

Executive Summary Introduction Outpatient Psychiatric Settings Interventions
Recommendations for Inpatient and Outpatient Settings Appendix Acknowledgments References
Download the full PDF