Safe Spaces: Mitigating Potential Aggression in Acute Care Psychiatry

Patients experiencing mental health crises in psychiatric emergency and inpatient settings may present risk to themselves, health care staff, and other patients. In this article, we summarize relevant literature and provide clinical considerations on minimizing harm to vulnerable patients and staff.

Clinical Context

Health care workers have a uniquely increased risk of workplace violence. According to the Bureau of Labor Statistics, of 20,870 workers in the private industry who experienced an intentional nonfatal occupational injury by another person in 2019, 70% were employed in “[health care] and social assistance” (1).

Emergency settings particularly raise concerns for patient aggression, with a significant portion of emergency staff perceiving workplace violence as inevitable (2, 3). Contributing factors may include long wait times, acute intoxication, overcrowding, inadequate staffing, and psychiatric decompensation (4, 5).

Although much has been written about the problem of workplace violence in emergency and crisis settings, underreporting and varying definitions of violence may make it difficult to appreciate the full scope of the problem. The Occupational Safety and Health Administration has defined workplace violence as “any act or threat of physical violence, harassment, intimidation, or other threatening disruptive behavior that occurs at the work site” (6). This definition could be interpreted to include a wide range of behaviors, from demeaning or threatening comments to physical assault resulting in significant physical injury or disability.

Similarly, psychological and physical sequelae vary widely, making workplace violence and its implications challenging to fully grasp. Regardless of the specifics, exposure to patient aggression has the potential to lower morale, impair performance, contribute to compassion fatigue, and increase turnover in an already strained health care system. Sequelae for staff may include feelings of helplessness, burnout, and even posttraumatic stress disorder (PTSD) (3, 7).

The proportion of violence attributable to mental health crises in general (nonpsychiatric) emergency departments (EDs) is not known, and little has been published about violence in psychiatric EDs or crisis centers. A 5-year study by Lawrence et al. (8) compared physical assaults in a psychiatric ED versus those occurring in two inpatient psychiatric units. Lawrence et al. identified 60 patients with incidents in the ED and 124 on the inpatient unit. The authors specifically used the word “incident” to avoid attributing intent to these events. Among patients with incidents versus those without incidents, higher rates of the following characteristics were identified: younger age, involvement in assisted outpatient treatment (i.e., outpatient commitment), history of violent behavior, thoughts of harming others, and use of cannabis (including synthetic forms). Other diagnostic features included schizophrenia, schizoaffective disorder, and mania. However, the authors cautioned against placing too much emphasis on diagnoses, to avoid stigmatizing or overly restricting patients on that basis alone.

Interestingly, Lawrence et al. noted that the ED and inpatient unit results were quite similar, and posited that existing data from inpatient settings may be generalizable to psychiatric EDs. One key difference was the timing of the assaults; whereas assaults were least likely to occur between 12 a.m. and 6 a.m. on the inpatient unit, no temporal relationship was found in the ED. Otherwise, in both settings, patient paranoia and disorganization were common themes in the incidents. Two-thirds of the assaults involved a staff victim, about one-third involved other patients, and very few involved visitors (8).

In addition to implications for staff, patients who are either directly involved in violence as perpetrators or victims, or who witness violence, are at risk of a variety of sequelae. Much has been written about the impact of coercive measures, such as restraint or seclusion, on patients; less has been written about patients who have been directly or indirectly victimized by peer violence while seeking care. Feeling unsafe on psychiatric units because of the risk of victimization can interfere with the therapeutic milieu and can trigger trauma reactions.

For example, in 2005, Robins et al. (9), after interviewing a small sample of patients with severe and persistent mental illness, all of whom had multiple admissions, used the terms “sanctuary trauma” and “sanctuary harm” to describe these phenomena. Robins et al. defined sanctuary trauma as an experience that meets the DSM-IV definition of a trauma for the purposes of diagnosing PTSD. The authors defined sanctuary harm as incidents that “do not meet the formal criteria for trauma but that involve insensitive, inappropriate, neglectful, or abusive actions by staff or associated authority figures and invoke in consumers a response of fear, helplessness, distress, humiliation, or loss of trust in psychiatric staff” (9).

Approaches to Mitigating the Risk of Violence in Acute Care and Crisis Settings

Restraint and Seclusion

Although at times restraint and seclusion remain necessary to protect patients and staff, the risk to both groups must be carefully weighed against the benefit in any given case. Staff and patients alike are at risk of injury during such interventions.

Adverse effects of restraint and seclusion.

Estimates of PTSD in patients who have been restrained vary from 25% to 47%, according to a 2019 systematic review by Chieze et al. (10). The same review noted that most patients’ emotional reactions to restraint and seclusion were negative, particularly “feelings of punishment and distress,” although there were some reports of positive reactions, such as feelings of “safety, help.” Additionally, Chieze et al. found that those with a history of trauma prior to the restraint had higher rates of PTSD than those without such history. Thus, the potential for confounding variables cannot be ignored; for instance, patients may already have PTSD. Therefore, awareness of possible preexisting PTSD and re-traumatization is important when considering restrictive measures.

In addition to the aforementioned potential adverse effects, recent studies have demonstrated that patients belonging to vulnerable groups, including those with public or no insurance, African American men, and homeless patients, are more likely than their counterparts to be restrained (11, 12). Such social determinants of health, and the potential role of systemic racism and unconscious bias, are similarly important reasons to carefully consider approaches to minimize use of restrictive measures and to maximize collaborative approaches to preventing aggression.

Several professional organizations have recognized a need to limit the use of restraint and seclusion, including the American Psychiatric Nurses Association (13) and the American Association for Emergency Psychiatry (14). The Joint Commission standards (15), as of 2009, state that restraint and seclusion should be used “only when it can be clinically justified or when warranted by patient behavior that threatens the physical safety of the patient, staff, or others,” and one of the key elements of performance for the standard requires that such measures be used “only when less restrictive interventions are ineffective.”

Strategies for reducing restraint and seclusion.

Some strategies for reducing restraint and seclusion include unit or department culture and education, administrative review and oversight, changes to the physical structure of treatment spaces, early identification of at-risk patients, staff training in de-escalation techniques, drills or simulations involving standardized patients, and debriefing of staff and patients after incidents.

The Safewards model (16), for example, identifies six domains: patient community, patient characteristics, regulatory framework, staff team, physical environment, and outside hospital. These each can have an impact on conflict (countertherapeutic behaviors, including self-harm, aggression, and elopement) and containment (steps taken to prevent or mitigate conflict). These domains are used to identify individual and group characteristics of staff and patients, as well as regulatory frameworks, physical layout, and factors outside the hospital, that give rise to “flashpoints.” Flashpoints are circumstances that may give rise to conflict (e.g., involuntary commitment, denial of exit from unit, receipt of bad news, conflict with visitors, disagreement among patients). In this model, patient and staff modifiers are proposed that can mitigate the risk of the above factors giving rise to flashpoints and conflict. For example, if the patient is experiencing paranoia and referential thinking, staff modifiers may include providing appropriate medication and/or psychoeducation and helping the patient plan for triggers, such as having too many people around (e.g., the patient may be able to voluntarily use a comfort room or other space when the unit is crowded) (16).

Several programs aimed at reducing the use of restrictive interventions also have been based on the “Six Core Strategies” outlined by Kevin Huckshorn (17): leadership toward organizational change; data-informed practice; workforce development (e.g., education, training, appointment to a steering committee); tools to reduce seclusion and/or restraint (e.g., de-escalation techniques and assessments of risk factors for aggression and for death or injury caused by restraint use); consumer roles in inpatient settings (e.g., giving choices where possible and using peer specialists to empower consumers); and debriefing techniques.

Programs based on the Six Core Strategies have shown progress in reducing use of restraint and seclusion across several studies. In fact, a 4-year study (18) involving Connecticut Valley Hospital, a 650-bed inpatient public sector psychiatric hospital, was found to decrease annual restraint hours by 89% (from 5,300 baseline restraint hours to 570) and annual worker’s compensation medical costs by 24% (from $780,937 baseline compensation to $527,715). A review (19) of restraint and seclusion programs created on the basis of the Six Core Strategies, and conducted in various countries, also showed promising preliminary results.

Simulated patient scenarios are another way to provide an opportunity to increase staff comfort with their roles on an interdisciplinary response team. A multimodal curriculum (20) developed to educate emergency medicine residents, nurses, and hospital police officers yielded encouraging results. Staff were recruited to act as standardized patients with agitation in two clinical scenarios. Learners were assigned to multidisciplinary teams and received didactic education in de-escalation techniques, proper application of restraints, and crisis management principles. They also participated in simulations with standardized patients, with each learner assigned a defined role on the response team. Participants’ survey responses indicated improved attitudes toward many factors affecting patient care. The authors also noted that the curriculum generated ideas for multiple quality improvement projects, which were subsequently carried out. Limitations of the project included the significant resources required to train staff and to conduct simulations (20).

Post seclusion/restraint conferences, also known as post incident reviews or debriefing, are often used in psychiatric settings. These can take many forms, and they may involve patients as well as staff or just staff. Although there have been mixed reviews on the efficacy of such practices, the diverse array of specific post incident review protocols has made finding definitive results difficult. However, a large-scale review of the literature (21) examining post incident reviews/conferences of varying types indicated that debriefing can significantly decrease subsequent seclusion and restraint. The Goulet and Larue literature review (21) included studies from several countries, including the United States, Canada, Australia, and the United Kingdom. The authors described post seclusion/restraint review as “highly recommended and vital to improving the care experience for both patients and staff, developing best practices, and reducing the incidence of seclusion/restraint.” However, they also noted a wide range of practices, and suggested a typology stratified by whether an intervention focused on staff, the patient, or both. They suggested further study to identify the weighted impact of the various components.

Another study (22) focused on interviewing staff from two programs in Norway—one university based and one community based—about their experiences with post incident reviews. In that study (22), Hammervold et al. interviewed 19 multidisciplinary providers. Overall, those interviewed felt that such reviews had the potential to improve quality of care, but that the potential for improvement would be greater if they were able to more effectively elicit patient perspectives during such reviews. The study authors cited difficulty in engaging with some of the patients, in part because of patient “passivity” during the post incident reviews. The authors suggested the potential for including input from family members, peers, or advocates, in addition to providing training for staff in how to engage patients in the process. Similar skills could also be used to engage the patient prior to an incident, as a means of more effective de-escalation and/or prevention (22).

The similarities in approaches to post-incident debriefing or review across several countries underscore the global nature of psychiatry’s goals of reducing restrictive interventions and maintaining the safest possible environment for staff and patients. Furthermore, there is significant overlap between strategies for reducing restraint and seclusion and to prevent violence. As we review additional strategies for managing aggression and violence, these alternatives can be thought of as opportunities to intervene earlier in the treatment course.

Early Identification of At-Risk Patients and Situations

Assessing a patient’s risk of becoming violent affords the treatment team the opportunity to mitigate risk factors and to prevent violence, thereby improving patient and staff well-being. Diagnostic categories associated with aggression include bipolar disorder (e.g., manic phase), psychotic disorders (especially symptoms of paranoia and/or referential thinking), substance use disorders (especially during intoxication or withdrawal), and cognitive impairment (e.g., disinhibition). As previously noted, however, aggression should not automatically be assumed on the basis of diagnoses, and aggression may be present independent of diagnoses.

Symptoms rarely occur in a vacuum, and intersecting clinical features increase risk additively. For example, a study of 216 patients with bipolar disorder (23) experiencing their first psychotic episode indicated significant increases in risk associated with recent suicide attempt, alcohol use disorder, learning disabilities, and initial manic episode. Interestingly, the study did not demonstrate significant association with age, gender, or personality disorder.

Precipitating factors for violence include unsatisfactory patient-staff interactions, such as limit setting, real or perceived power struggles, and a patient being informed of an involuntary commitment. Peer-to-peer disputes are also frequent triggers. Mood or psychotic symptoms, for example, can cause or contribute to negative interactions or cause a patient to misinterpret a neutral exchange. Demographic and historical characteristics associated with higher risk include male gender, younger age, history of trauma, and history of aggression in the treatment setting (24).

Use of validated tools, such as the Dynamic Appraisal of Situational Aggression (DASA) and the Broset Violence Checklist, may reduce incidence of aggression, and by extension, reduce the use of restrictive interventions (25–27). Such practices provide clinicians with the capacity to influence whether a person becomes aggressive. Knowledge of these tools can increase staff members’ sense of control and safety and may lead to lower rates of burnout and turnover.

Nonpharmacological Approaches to Mitigating Risk of Aggression

A large-scale review (28) of available literature aimed to create a glossary of techniques from interventions and quality improvement projects aimed at reducing restrictive practices, such as restraint and seclusion, chemical sedation, and constant observation. The authors classified the components of each intervention and explored the evidence regarding clinical efficacy and cost-effectiveness. Although Baker et al. (28) identified 16 clusters of techniques, they noted that four clusters were present in more than two-thirds of the interventions. Those clusters were “goals and planning” (staff setting goals, such as reducing occurrence of a type of restrictive intervention and implementing a plan to meet the goal); “shaping knowledge” (providing education about the antecedents of restrictive interventions and the performance of specific techniques); identifying “antecedents” (anticipating and attempting to mitigate factors that could result in patient distress or agitation), and providing “feedback and monitoring” (analyzing practices and outcomes).

The Baker et al. (28) review identified some promising findings and underscored the need for more rigorous and structured research in this area. For example, changes to the physical environment, such as reducing noise; adding comfort objects, such as weighted blankets; and creating more private spaces for people to engage in de-escalation, were used in a number of interventions associated with decreases in restrictive interventions. Other commonly used and potentially effective techniques included instructing staff on how to perform interventions, conducting problem solving, and providing feedback on outcomes (28).

Studies have also been done on the effects of music in acute inpatient settings. One such study (29), conducted across 3 months, with the prior 3 months serving as a control, showed promising findings of reduction in seclusion, assault, and as-needed medication administrations. In this relatively affordable intervention (less than $400), patients were given wireless headphones and were able to listen to their choice of a musical genre for 30 minutes (29). Limitations of the study included a low rate of survey response from the patients. The authors also noted an incidental finding of a modestly reduced length of stay, and they recommended further research on this finding. Further study of this intervention seems warranted and achievable. Of note, this intervention was distinct from formalized music therapy, another nonpharmacologic intervention.

A Cochrane review of standardized music therapy for patients with schizophrenia spectrum disorders (30) identified low-to-moderate evidence for positive effects of music therapy given in addition to standard care. Data specifically examining patient behavior indicated that, in the medium-term trials (3–4 months), patients could benefit significantly. A comparison study on an inpatient unit, before and after implementation of music therapy and specifically studying aggressive or disruptive behavior, may prove useful.

The aforementioned studies have primarily been conducted in inpatient psychiatric settings, and there is need for further study in ED settings. One group of authors (31) described “an unambiguous gap in the literature regarding the efficacy of interventions for ABD [acute behavioral disturbance] management in EDs [emergency departments].”

Overall, the literature clearly indicates the potential for nonpharmacologic interventions to reduce severe agitation and aggression without the use of restrictive interventions. Further research is needed to systematically evaluate specific intervention programs and methods. Ideally, these interventions should be tested separately in general ED, psychiatric ED, and inpatient hospital settings.

Pharmacological Approaches to Mitigating Risk of Aggression

Pharmacological considerations vary depending on the patient’s age, etiology of symptoms, and individual factors, such as prior response to medications. For example, in a patient with a known history of severe dystonic reactions, first-generation antipsychotics may not be the clinician’s first choice. Certain underlying conditions, such as arrhythmias, may affect medication choice, particularly if repeated administrations are needed. In short, medication and dosage choices should be tailored to the individual whenever possible.

Furthermore, whereas interventions should target the underlying cause of agitation, and there may be overlap between long- and short-term management, this discussion focuses on the management of acute agitation. Long-term approaches differ and are beyond the scope of this article.

Project Best Practices in the Evaluation and Treatment of Agitation (BETA) was the product of a 2012 collaboration of five interdisciplinary work groups led by the American Association of Emergency Psychiatry. As described by Roppolo et al. (32), the BETA groups defined five principles for the management of acute agitation. When possible, the first approach to managing agitation should be verbal de-escalation. Pharmacotherapy aimed at treating the underlying cause of the agitation should be started. Similarly, each patient should receive an appropriate medical evaluation to recognize potential organic causes, such as electrolyte abnormalities, toxidromes, and head injuries. Identification of such phenomena is essential to the safety of patients as well as staff. A psychiatric evaluation should also be done as soon as feasible in order to try to understand the underlying cause or illness spectrum and to help guide thinking regarding pharmacotherapy or nonpharmacological interventions that may be useful. For example, a patient with acute mania who is willing to take oral medications may benefit from prompt initiation of lithium or valproic acid, or even benzodiazepines. Finally, the authors indicated that restraint and seclusion should be a last resort, and if needed, optimally should be done by a team of five trained staff.

The Project BETA team assembled a list of medications commonly used to manage agitation and the risks associated with each; this list has been updated by Roppolo et al. (32). Some key points follow:

Benzodiazepines should be avoided among patients with suspected or confirmed delirium caused by an underlying medical condition and among patients who have taken a central nervous system depressant. For patients with psychosis or undifferentiated aggression, benzodiazepines can be helpful. These agents are typically required for the treatment of alcohol or benzodiazepine withdrawal.

The term “chemical restraint” is often used when medications are given in response to agitation. However, the intent of the administration must be considered. The Centers for Medicare and Medicaid Services define chemical restraint as “a drug or medication when it is used as a restriction to manage the patient’s behavior or restrict the patient’s freedom of movement and is not a standard treatment or dosage for the patient’s condition” (33). For psychosis or mania that results in aggression, however, administration of an antipsychotic may be the first step in treating the underlying cause as well as the symptom (agitation).

Indeed, medications given for acute agitation often overlap with those routinely used to treat the underlying psychiatric conditions. For example, administration of 5–10 mg of olanzapine to a patient for agitation caused by acute mania or psychosis likely does not meet the definition of chemical restraint. In this context, the medication is not intended to restrict the patient and may in fact help the patient stabilize and reintegrate to a less restrictive setting. Similarly, patients experiencing psychiatric emergencies are entitled to safety and relief of their symptoms. Zun et al. (34) eloquently summarized this point: “The appropriate goal of treatment is to calm the patient, and this treatment should be tailored to the underlying cause of the agitation.”

Conclusions

In this article, we summarized several themes in the literature related to managing the agitation that can occur during psychiatric emergencies. Key points include the toll that violence at work takes on health care staff, costs to health care systems in terms of workman’s compensation, and the trauma—from restraint and/or seclusion or witnessing and/or experiencing violence by a peer—that patients in acute care settings may encounter at their most vulnerable times.

The cost of violence in these settings underscores the need for mitigation. Strategies supported by available evidence include incident debriefing or review; interventions that are based on structured models (such as Safewards or the Six Core Strategies); practice of team-based responses to behavioral emergencies through simulated scenarios; early identification of patients at risk for violence (through tools such as the DASA); and use of complementary approaches (such as providing music therapy or making music available for patients as a means of de-escalation).

Controversies about treatment remain, such as whether restrictive interventions are ever appropriate, and whether complete elimination of such interventions while maintaining safety is possible. Moreover, recent evidence has shown disparities in restrictive practices, raising the possibility of perpetuating structural racism and causing further harm.

Although there is a significant amount of literature on mitigating risk with a variety of interventions, there appears to be a lack in specific, consistently used, structured interventions across studies. Programs that have been based on the structured models previously mentioned have varied widely in their implementation. Larger, multicenter studies or more manualized approaches to designing interventions would afford mental health professionals and health care systems the opportunity to replicate findings and to work from a stronger evidence base. Fortunately, the weight of the evidence supports our ability to significantly reduce violence, aggression, and the use of restrictive interventions, giving us a valuable opportunity to positively affect staff and patient well-being, reduce stigma, and more effectively use health care resources.