In a fictional case, at three in the morning, an emergency department nurse calls the internal medicine resident to reassess a 52-year-old man who has been waiting four hours for a bed. He arrived with abdominal pain, has already received analgesia and now walks the corridor, talks loudly, complains about the delay and slaps his hand on the counter. His daughter, who is with him, films the team with her phone and says she will report everyone. When the resident approaches, he points his finger at her face and says that if nobody sorts it out, he will sort it out his own way. Minutes later, he pushes the IV pole, which falls to the floor.

Nobody was hurt. At the end of the shift, the resident asks the nurse whether it is worth reporting, and the answer is that this happens every week and that it never leads to anything. This text organizes what cohorts, reviews and studies from several countries show about violence from patients and companions in the day-to-day of the hospital: the signs before escalation, de-escalation and what should happen afterwards.

Key points

• In a meta-analysis of studies carried out during the COVID-19 pandemic, the pooled prevalence of verbal violence against health professionals was 45.87 percent and that of physical violence was 10.75 percent (Hadavi et al., 2023).
• The Brøset Violence Checklist, with six observable behaviors, was highly specific but poorly sensitive for predicting violence alerts in Australian emergency departments (Mitra et al., 2025).
• Verbal de-escalation is the recommended first line, but the evidence that it reduces violence is still weak and heterogeneous (Seiyad et al., 2026).
• Underreporting is the rule; fear of consequences, disbelief that anything will change and confusing processes appear in several countries (Spencer et al., 2023).
• Exposure to violence or threats at work was associated with more sick leave certified for mental disorder in a Swedish cohort (Wijkander et al., 2023).

• A routine problem, in peacetime
• Predict: the signs before escalation
• Respond: de-escalation
• After the aggression
• Studies in summary
• How NEXO helps train this reasoning
• Frequently asked questions
• Take-home messages

A routine problem, in peacetime

Attacks on hospitals in wars are another phenomenon, discussed in the text on Attacks on health services in armed conflicts. Here the subject is verbal or physical aggression from a patient or companion, in the emergency department and the ward, on an ordinary day.

How often it happens

The numbers vary with the definition, the country and the period. A systematic review with meta-analysis gathered 13 studies carried out in 2020 and 2021 and found a pooled prevalence of workplace violence of 45.80 percent among health professionals, with 45.87 percent for verbal violence and 10.75 percent for physical violence (Hadavi et al., 2023). The data cover only 2020 and 2021, and the heterogeneity between studies was very high, with I² above 97 in all estimates (Hadavi et al., 2023). These numbers serve as an order of magnitude, not as a precise estimate.

Outside the pandemic, routine data tell the same story. In a cross-sectional study with 4862 professionals from 45 services at three levels of care in China, 4.6 percent reported physical aggression and 17.4 percent reported threats in the previous year (Yang et al., 2019). In two Italian public hospitals, after the implementation of a systematic surveillance system, the annual rate was 2.08 reported events per 100 professionals, with emergency departments and psychiatric wards as the highest-risk areas (Veronesi et al., 2023).

A Lancet report described a survey by the International Council of Nurses, the International Committee of the Red Cross, the International Hospital Federation and the World Medical Association, with just over 120 responses from organizations (Thornton, 2022). All reported verbal aggression, and most perceived an increase in violence since the start of the pandemic. The survey measures perceptions and is small, but it shows that the problem appears in countries with very different safety contexts.

Where and when

A ten-year retrospective cohort in a German university emergency department analyzed 859 incidents recorded by staff using the SOAS-R scale (Benning et al., 2024). The events were concentrated in trauma rooms and detox rooms, were more frequent and more severe outside business hours, and about one third were classified as severe. In more than half of the records, the cause of the aggression could not be identified by the staff. The upward trend in rates over the years was not significant, meaning the study cannot say whether violence increased.

In the Italian study, nurses, younger professionals and men had the highest rates, and in about one third of episodes against female professionals the aggressor was a visitor (Veronesi et al., 2023). Language and cultural barriers were cited as a contributing factor by a good part of the professionals.

Predict: the signs before escalation

Aggression rarely comes out of nowhere. The Brøset Violence Checklist sums up clinical observation in six behaviors, marked as present or absent: confusion, irritability, noisy and agitated behavior, verbal threat, physical threat and attacking objects (Mitra et al., 2025). Each item is worth one point; zero indicates low risk, one or two moderate risk and three or more high risk.

What the checklist gets right and what it misses

In two emergency departments in Melbourne, the scale was recommended for all visits for 12 months and ended up filled in for 39.4 percent of the 108,274 entries analyzed (Mitra et al., 2025). With a cutoff of 3, specificity was 99.2 percent and sensitivity was 15.6 percent for unplanned alerts of potential or actual violence (Mitra et al., 2025). In other words, a high score almost always indicates real risk, but most episodes occurred in patients who did not score high.

Filling in the scale was associated with more alerts, with an adjusted odds ratio of 1.37 (95% CI 1.12 to 1.66) (Mitra et al., 2025). The authors interpret that the staff already identified at-risk patients intuitively and filled in the scale precisely for them. The study is retrospective and shows association, without a causal effect.

In practice, the list serves more to give a name to what is seen and to communicate risk to the team than to rule out danger. The useful question before each item is what lies behind it.

Medical causes first

Confusion and irritability can be delirium, withdrawal, intoxication, hypoglycemia, hypoxia or poorly controlled pain. In the German study, detox rooms were among the locations with the most events (Benning et al., 2024). The diagnostic reasoning and clinical management of agitation are in the texts on Hypoactive delirium in the hospitalized older adult: 4AT, CAM and prevention, CIWA-Ar on shift: when the withdrawal scale misleads and Stimulants in the emergency department: differential and what to avoid.

Many aggressions, however, come from lucid people, frustrated with waiting, failed communication or bad news, and sometimes from companions. Context factors count as much as patient factors: long waits, a crowded environment, shift changes and the early hours of the morning.

Respond: de-escalation

A systematic review of 36 studies on verbal de-escalation in hospitals, emergency departments, psychiatric units, long-term care and outpatient clinics found wide variation in how the technique is defined, taught and evaluated (Seiyad et al., 2026). The intervention studies generally showed gains in staff knowledge, confidence and preparedness. The observational and descriptive studies highlighted strategies that recur: empathic listening, a calm tone of voice, validating what the person feels and seeking a solution together.

At the bedside, this translates into simple gestures:

• Introduce yourself, call the person by name and speak slowly, in a low tone.
• Acknowledge the emotion before discussing the content, for example by saying that the wait is being hard.
• Explain what is happening and the next concrete step, with a realistic timeframe.
• Offer small, achievable choices, such as waiting while seated or speaking with the nurse in charge.
• Keep your distance, leave the exit clear for yourself and the patient, and call for support before you are alone.
• Know when to stop talking and call security.

What the evidence allows us to say

The strength of the evidence on prevention is modest. A systematic review with meta-analysis of randomized and quasi-randomized trials found 11 studies, four of them with a positive and significant effect, but with a high or uncertain risk of bias (Okubo et al., 2022). The feasible meta-analysis pooled only two studies of individual skills, with a confidence interval from minus 0.41 to 0.25 for the violence outcome, which includes the null and is therefore inconclusive; the measure is a difference in means between groups, described in the methods as standardized and in the results table as MD (Okubo et al., 2022).

A scoping review of interventions against violence directed at emergency department nurses identified 40 studies in 11 countries (Gammone et al., 2026). Training programs were the most common, followed by risk assessment tools, defense strategies, multidisciplinary meetings and technological solutions. The vast majority of interventions acted before the incident, and few addressed what comes afterward. Changes to the environment, such as layout, lighting and access control, received little attention.

In the German study, most events were resolved with verbal intervention, and more forceful interventions were more frequent in the most severe events (Benning et al., 2024). The presence of professionals trained in de-escalation was associated with higher recorded severity scores, which the authors interpret as greater attention to the problem and to reporting, and not as an effect of training on violence.

When talking is not enough

De-escalation has limits. If the person keeps approaching, holds an object that could become a weapon or has already attacked someone, the goal becomes protecting the staff, the other patients and the aggressor themselves. That means leaving the area, calling hospital security and, where local law provides for it, the police. The steps in this paragraph are a summary by the editorial team, not a protocol tested in the cited studies. The clinical management of agitation is outside the scope of this text.

In the German emergency department, events were concentrated outside business hours, when there are fewer people on staff (Benning et al., 2024). Knowing in advance whom to call, by what means and how long security takes to arrive is part of preparing for the shift, and not something to discover in the middle of a crisis. The scoping review on the emergency department describes multidisciplinary meetings and risk assessment tools among the organizational strategies studied (Gammone et al., 2026).

After the assault

It is at this stage that the response usually fails. In the Chinese study, among those who suffered physical assault, only 29.1 percent reported it to the police, 25.7 percent of episodes were investigated and 72.4 percent of aggressors suffered no consequence (Yang et al., 2019). Among those assaulted or threatened, 59.4 percent said they wanted to leave the post and 76.0 percent reported fear of caring for severe or urgent patients (Yang et al., 2019).

Reporting

Underreporting appears in practically every study. A systematic review with 19 studies from 16 countries grouped nurses' reasons for not reporting into three levels: fear of the consequences, the perception that aggression is part of the job and unfamiliarity with the process; lack of visible change, an unwelcoming culture and absence of punishment for the aggressor; and lack of policies, training and simple reporting systems (Spencer et al., 2023).

In an analysis of 4109 nurses in British Columbia, Canada, the belief that nothing would change after reporting was among the three most cited reasons for not reporting formally, across all sources and forms of violence (Lee et al., 2025). Not knowing the formal process and lack of support from leadership were also frequent. In a qualitative study with 36 nurses from three Saudi tertiary hospitals, the barriers were emotional, organizational and cultural, with emphasis on rigid hierarchy and the normalization of violence (Elsharkawy et al., 2025).

Reporting changes what the institution sees. In the Italian study, before systematic surveillance the ratio between observed and expected events reached 0.27, which the authors interpret as underreporting of up to 73 percent relative to the prevalence expected from the literature (Veronesi et al., 2023). The new system combined a standardized immediate reporting form, cause analysis with a psychologist, an offer of psychological counseling, an online platform and periodic training, and the ratio rose to 1.14 (Veronesi et al., 2023).

In the Chinese study, the absence of an institutional reporting policy was associated with a higher chance of physical aggression, with an adjusted odds ratio of 3.64 (95% CI 2.57 to 5.18), in a cross-sectional analysis that does not establish cause (Yang et al., 2019).

The record should describe what was said and done, with the exact words of the threat when possible, the place, the time, who witnessed it and whether there was injury. Objective descriptions help the institution identify patterns, such as times, departments and higher-risk situations, and decide where to act. In the Italian study, surveillance data made it possible to identify the most exposed departments and professional groups and the role of visitors (Veronesi et al., 2023).

Support

The reviews converge on one point: reporting only makes sense if there is feedback (Spencer et al., 2023). The review on underreporting recommends investigating each report quickly, giving feedback in conversation after the event and offering support to the team (Spencer et al., 2023), and the Italian model included an offer of psychological counseling (Veronesi et al., 2023). The script below is a synthesis by the editorial team from these sources, not a protocol evaluated in a study:

• Medical care for the person who was assaulted, with the injuries recorded.
• Reporting in the institution's system during the same shift, with an objective description.
• A conversation with the immediate supervisor on the same day, to reorganize the schedule if necessary.
• An offer of psychological support, without obliging the person to talk.
• Feedback to the team on what was done with the report.
• Guidance on filing a police report and on labor rights, according to local law.

The Lancet article cites experiences such as a 24-hour video support line in Portugal and a nursing reporting system in Taiwan whose reviewed cases feed training (Thornton, 2022). These are accounts from organizations, without controlled evaluation.

Leave

The impact can last. In a cross-sectional study with 471 nurses from South Korea and Hong Kong, about three in ten reported violence at work, and those who experienced it had worse mental health indicators and more post-traumatic responses (Hong et al., 2023). In a Swedish prospective cohort of 16 339 workers from all sectors of the economy, not only health care, exposure to violence or threats at work was associated with more sick leave certified for mental disorder, with an odds ratio of 1.46 (95% CI 1.17 to 1.82) (Wijkander et al., 2023). Because the sample is not of health professionals, the result holds as a general indication, and the design shows association, not cause.

Stepping away after an assault is therefore a possible clinical response, not a sign of weakness. The decision should be made with the occupational physician or the attending physician, and returning is usually easier when it is gradual and agreed with the supervisor. The effects of the event on reasoning at the scene, guilt, moral suffering and burnout are in the texts on Acute stress and clinical reasoning: what changes at the scene, Second victim: the doctor after the error, Moral injury: the suffering that is not burnout, Burnout in medical school: signs and prevention and Suicide among doctors: risks and barriers to care.

Studies in brief

Study	Design	Population	Main finding
Hadavi, 2023	Systematic review and meta-analysis	13 studies with health professionals, 2020 and 2021	Verbal violence in 45.87 percent (Hadavi et al., 2023)
Mitra, 2025	Retrospective cohort	108,274 visits to two Australian emergency departments	Specificity of 99.2 and sensitivity of 15.6 percent (Mitra et al., 2025)
Okubo, 2022	Systematic review with meta-analysis	11 randomized and quasi-randomized trials	Meta-analysis of two studies, CI from minus 0.41 to 0.25, inconclusive (Okubo et al., 2022)
Veronesi, 2023	Prospective before-and-after study	Professionals at two Italian hospitals	Observed-to-expected ratio of 0.27 to 1.14 (Veronesi et al., 2023)
Yang, 2019	Cross-sectional study	4862 professionals in China	Physical assault in 4.6 percent in the past year (Yang et al., 2019)
Wijkander, 2023	Prospective cohort	16,339 Swedish workers from various sectors	Sick leave due to mental disorder, OR 1.46 (Wijkander et al., 2023)

Sources: prevalences in percent of the professionals studied; specificity and sensitivity in percent, for a cutoff of 3; Okubo's CI is for a difference in means between groups; the observed-to-expected ratio is dimensionless; OR is the adjusted odds ratio. Data from Hadavi et al., 2023, Mitra et al., 2025, Okubo et al., 2022, Veronesi et al., 2023, Yang et al., 2019 and Wijkander et al., 2023.

Back to the case

The man from the opening had irritability, loud behavior, verbal threat, physical threat and an attack on an object, five of the six items on the Brøset Violence Checklist, on a night in a crowded emergency department. First of all, it was worth reviewing whether the pain was controlled and whether there was a medical cause for the change in behavior. Then, de-escalation: introduce yourself, acknowledge the wait, explain the next step, offer a choice, keep your distance and call security early.

The resident's question at the end of the shift has an answer. The report should be filed in the same shift, with feedback from the institution, an offer of support to whoever was threatened and, if needed, reorganization of the schedule. That it happens every week is precisely the reason to report it.

How NEXO helps train this reasoning

In NEXO, the case of the day is fictional, with a real clinical presentation and up to six progressive clues, and you propose the diagnosis by looking for the code in ICD-10. Noticing that an agitated patient may be confused because of a medical cause, and not just irritated by the wait, is the same exercise of putting clues together and reviewing the hypothesis that appears in the text on Signs, symptoms and history: how to interpret what the patient brings.

The limits on using the game are described on the NEXO economy page. If you research the teaching of clinical reasoning, the page for researchers explains how to get in touch. Professionals who write or review cases find information under curation and under creators, and institutions can learn about the proposal on the for partnerships page.

Frequently asked questions

Should verbal aggression also be reported?

Yes. Verbal violence is the most common form (Hadavi et al., 2023), and in the Canadian study informal reporting was less likely precisely in the face of threats, emotional abuse and verbal sexual harassment (Lee et al., 2025).

Does a risk scale replace the team's eye?

In the Australian cohort, the Brøset Violence Checklist was very specific and not very sensitive, and the authors suggest that the team was already intuitively identifying at-risk patients (Mitra et al., 2025). It complements observation.

Does de-escalation training reduce violence?

The studies show more knowledge and confidence among the team (Seiyad et al., 2026), but the meta-analysis of trials was inconclusive for the violence outcome (Okubo et al., 2022).

When should sick leave be considered?

When the symptoms after the event interfere with sleep, work or life. Exposure to violence at work was associated with more sick leave due to mental disorder in the Swedish cohort (Wijkander et al., 2023).

Take-home messages

• Aggression from patients and companions is frequent in peacetime, especially in the emergency department and in psychiatry.
• Confusion, irritability, noisy behavior, threats and attacks on objects are warning signs; medical causes come first.
• Verbal de-escalation is the first line, with distance, a clear exit and early called support, but its effectiveness is still poorly proven.
• Reporting during the same shift, with feedback from the institution, reduces underreporting and guides prevention.
• Psychological support and leave, when necessary, are part of the response.

Play the case of the day on NEXO and train your reasoning with progressive clues.

References

1. Hadavi M, Ghomian Z, Mohammadi F, et al. Workplace violence against health care workers during the COVID-19 Pandemic: A systematic review and meta-analysis. J Safety Res. 2023;85:1-7. [PubMed](https://pubmed.ncbi.nlm.nih.gov/37330859/).
2. Thornton J. Violence against health workers rises during COVID-19. Lancet. 2022;400(10349):348. [PubMed](https://pubmed.ncbi.nlm.nih.gov/35908564/).
3. Yang SZ, Wu D, Wang N, et al. Workplace violence and its aftermath in China's health sector: implications from a cross-sectional survey across three tiers of the health system. BMJ Open. 2019;9(9):e031513. [PubMed](https://pubmed.ncbi.nlm.nih.gov/31542763/).
4. Veronesi G, Ferrario MM, Giusti EM, et al. Systematic Violence Monitoring to Reduce Underreporting and to Better Inform Workplace Violence Prevention Among Health Care Workers: Before-and-After Prospective Study. JMIR Public Health Surveill. 2023;9:e47377. [PubMed](https://pubmed.ncbi.nlm.nih.gov/37955961/).
5. Benning L, Teepe GW, Kleinekort J, et al. Workplace violence against healthcare workers in the emergency department: a 10-year retrospective single-center cohort study. Scand J Trauma Resusc Emerg Med. 2024;32(1):88. [PubMed](https://pubmed.ncbi.nlm.nih.gov/39285387/).
6. Mitra B, Settle K, Koolstra C, et al. Introduction of the Broset Violence Checklist in the emergency department: A retrospective cohort study. Emerg Med Australas. 2025;37(1):e14546. [PubMed](https://pubmed.ncbi.nlm.nih.gov/39668676/).
7. Seiyad H, Nimako J, Singh B, et al. Verbal de-escalation techniques for the prevention and management of workplace violence in healthcare: a systematic review. Front Psychiatry. 2026;17:1819502. [PubMed](https://pubmed.ncbi.nlm.nih.gov/42952582/).
8. Okubo CVC, Martins JT, Malaquias TDSM, et al. Effectiveness of the interventions against workplace violence suffered by health and support professionals: A meta-analysis. Rev Lat Am Enfermagem. 2022;30:e3638. [PubMed](https://pubmed.ncbi.nlm.nih.gov/35976359/).
9. Gammone M, Cattani D, Barbieri M, et al. Interventions to Prevent Patient- and Visitor-Perpetrated Violence Against Nurses in the Emergency Department: A Scoping Review. J Adv Nurs. 2026;82(8):7644-7667. [PubMed](https://pubmed.ncbi.nlm.nih.gov/41178600/).
10. Spencer C, Sitarz J, Fouse J, et al. Nurses' rationale for underreporting of patient and visitor perpetrated workplace violence: a systematic review. BMC Nurs. 2023;22(1):134. [PubMed](https://pubmed.ncbi.nlm.nih.gov/37088834/).
11. Lee J, Havaei F, Hirani S, et al. Nurses' Workplace Violence Reporting Behaviours and Reasons for Not Formally Reporting: A Cross-Sectional Secondary Analysis. J Clin Nurs. 2025;34(10):4235-4245. [PubMed](https://pubmed.ncbi.nlm.nih.gov/39739552/).
12. Elsharkawy NB, Alruwaili AN, Elsayed Ramadan OM, et al. Barriers to reporting workplace violence: a qualitative study of nurses' perceptions in tertiary care settings. BMC Nurs. 2025;24(1):395. [PubMed](https://pubmed.ncbi.nlm.nih.gov/40200356/).
13. Hong S, Nam S, Wong JYH, et al. Post-traumatic responses to workplace violence among nursing professionals: a collaborative and comparative study in South Korea and Hong Kong. BMC Nurs. 2023;22(1):354. [PubMed](https://pubmed.ncbi.nlm.nih.gov/37794376/).
14. Wijkander M, Farrants K, Magnusson Hanson LL. Exposure to work-related violence and/or threats of violence as a predictor of certified sickness absence due to mental disorders: a prospective cohort study of 16,339 Swedish men and women in paid work. Int Arch Occup Environ Health. 2023;96(2):225-236. [PubMed](https://pubmed.ncbi.nlm.nih.gov/36070001/).