HEALTH AT WORK

Workplace Safety

When Work Turns Hostile: Preventing and Managing Workplace Violence

22 July 2026 · Health at Work · Reviewed by Dr. Ackhmez Chuttoo, MD

Photo: Pavel Danilyuk / Pexels

Workplace violence covers a wider spectrum than most people expect. At one end sits physical assault — a punch, a shove, a weapon. At the other end, and far more common, sits behaviour that is just as damaging but far easier to dismiss: a customer who screams obscenities at a cashier, a patient who threatens a nurse with what they will do after discharge, a manager who systematically humilirates a subordinate in team meetings until that person stops sleeping properly. All of it falls under the occupational health definition of workplace violence, and all of it carries documented health consequences. The problem is not that organisations deny the serious end of the spectrum — it is that they routinely fail to treat the everyday end as a hazard at all.

What the Hazard Actually Is

Workplace violence is physical or psychological harm that arises from work or the conditions in which work is performed. It has three distinct routes of exposure, and conflating them leads to poor prevention.

The first is external violence — harm inflicted by people who have no employment relationship with the organisation: customers, patients, clients, members of the public, or in rare cases, strangers. A retail worker absorbing a torrent of verbal abuse from a customer demanding a refund, a bank teller threatened during a robbery, an emergency responder attacked while providing care — these are all external violence events, and they are the most commonly reported category.

The second is client-on-worker violence, which overlaps with external but deserves separate attention in high-care settings. A person with acute psychiatric illness who strikes a support worker, a dementia patient who bites a carer, a child in crisis who throws furniture at a teacher — these incidents occur within a care relationship, which creates specific complications around reporting, response, and the worker's own sense of whether they are 'allowed' to feel harmed.

The third is internal violence — harm from colleagues, supervisors, or the organisation itself. Persistent bullying, harassment based on protected characteristics, deliberate isolation, public humiliation, threats tied to performance reviews. This category is the least reported and, in terms of cumulative psychological damage, often the most destructive.

The physical mechanism of harm is not always a blow. Prolonged exposure to verbal aggression activates the same stress-response pathways as a physical threat. Cortisol and adrenaline release repeatedly over weeks and months produces measurable physiological damage: elevated blood pressure, disrupted sleep architecture, suppressed immune function, and structural changes in the brain regions associated with fear regulation. A worker who handles three hostile customer interactions per shift, five days a week, is not experiencing minor annoyance — they are accumulating a physiological load that will manifest clinically if the exposure continues.

Who Is Exposed, and in Which Specific Situations

Exposure is not random. Certain job characteristics predict risk with enough consistency that they should trigger automatic assessment.

• Direct contact with the public, especially in emotionally charged transactions. Retail workers handling returns and complaints. Hospitality staff managing intoxicated guests. Transport workers enforcing ticketing rules. The common factor is that the worker holds a boundary — a policy, a price, a rule — that the other person does not want to accept.
• Delivery of care to people in distress or with impaired cognition. Hospital emergency departments, psychiatric wards, residential care homes, addiction services. The worker is proximate to people at the worst moments of their lives, sometimes people who cannot regulate their own behaviour due to illness, substances, or pain.
• Lone working. A security guard doing a night patrol alone, a social worker conducting a home visit without a colleague, a delivery driver entering unfamiliar buildings. Lone workers are at higher risk not only because there is no physical backup but because incidents go unwitnessed and unreported more often.
• High-value asset handling. Cash-handling roles, jewellery retail, pharmacy staff who manage controlled drugs. The asset creates a motive for external aggression that other roles do not carry.
• Enforcement and compliance roles. Traffic wardens, building inspectors, environmental health officers, immigration officials. These workers deliver decisions that people experience as punitive, in encounters where the worker has authority but limited physical protection.
• Remote and digital customer contact. Often overlooked. Customer service workers managing complaints by phone or online chat absorb abusive language at high volume, sometimes hundreds of interactions per week, with no physical cues to help them anticipate or de-escalate.

Within each of these categories, specific tasks elevate risk further: working at the end of a shift when public frustration has built, working during peak demand periods when wait times are long, working in understaffed teams where a worker cannot step away from a deteriorating situation.

Early and Late Signs — What They Look Like in Practice

One of the reasons workplace violence stays under-reported is that its health effects look like other things. A worker developing early-stage psychological injury from chronic aggression exposure does not typically walk into a manager's office and say 'I have been harmed.' They say — or show — something else entirely.

Early signs include a reluctance to take on tasks that previously felt normal: avoiding certain shifts, finding reasons not to handle the phone queue, arriving late to avoid a particular colleague. Physically, the early period often produces headaches, gut disturbance, and disrupted sleep — the first thing to go is usually the ability to fall asleep easily, followed by waking in the early hours. Mood changes are also early: a worker who was previously even-tempered becomes short, tearful at low-stimulus moments, or flat in a way that colleagues notice but cannot name. Hypervigilance is a key early marker — the worker startles easily, scans exits when entering a room, or reports being 'on edge' without being able to explain why.

Later signs, if exposure continues or if a single severe incident has occurred, are more pronounced and more disabling. Intrusive memories — not simply 'thinking about' an incident but involuntary, vivid re-experiencing that stops the person mid-task. Avoidance that has expanded beyond specific tasks to whole areas of life: not going to public spaces, stopping social contact, withdrawing from family. Numbing — the person reports feeling nothing, which they sometimes describe as being 'fine' or 'over it' when they are not. Sleep at this stage is severely fragmented, and concentration is impaired enough to affect work performance, which creates a secondary cycle of anxiety about job security. Anger disproportionate to triggers is common and often confusing for the person experiencing it — they may describe 'snapping' at a partner or child over something trivial and not understanding why. At this stage, the clinical threshold for post-traumatic stress disorder or adjustment disorder is frequently met, and self-referral to professional support is overdue.

In a colleague, the signs to watch for specifically are: withdrawal from team communication, a marked change in appetite or weight, alcohol use that has visibly increased, coming in early or staying late to avoid peak-contact periods, and physical complaints with no clear medical cause — stomach problems, persistent headaches, recurring illness — that cluster around the period following an incident or a period of sustained aggression.

What Employers Are Required to Do

Under typical occupational health and safety regulations in most jurisdictions, an employer's duty is not limited to responding to violence — it extends to identifying and controlling the risk before harm occurs. The legal framework usually requires the following, regardless of which specific regulatory system applies.

First, a risk assessment specific to violence. This is not satisfied by a generic workplace risk assessment. It should identify which roles are exposed, which tasks within those roles carry highest risk, what the history of incidents in that workplace looks like, and what control measures are already in place and whether they are adequate. The assessment must be reviewed after any significant incident and at regular intervals even without incidents.

Second, a written policy that defines workplace violence in terms the organisation will actually apply — including verbal abuse and psychological harm, not only physical assault — and that specifies what reporting mechanisms exist, who is responsible for investigation, what protections exist for the reporting worker, and what the consequences of substantiated violence are. A policy that exists only on an intranet page and has never been communicated is not a policy in any meaningful operational sense.

Third, adequate information, instruction, and training. In most jurisdictions, this is a discrete legal duty. Workers cannot be expected to manage violent incidents by instinct. Training should cover: recognising early behavioural cues, de-escalation techniques appropriate to the specific role, what to do when a situation cannot be de-escalated, how to report, and what support is available afterwards.

Fourth, post-incident support. Sending a worker back to their task an hour after a violent incident without follow-up is an exposure to secondary harm, and in most jurisdictions it is a failure of the employer's duty of care. There should be a documented, consistent process for the immediate period after an incident and for the following days and weeks.

Controls That Work — and Controls That Are Overrated

Effective control of violence risk follows the same hierarchy as other occupational hazards: eliminate the hazard first, then substitute, then engineer, then administer, then protect. In practice:

What works well:

• Elimination of the situation that produces violence. If a particular task — such as enforcing a disputed policy on the shop floor — consistently generates aggression, removing the frontline worker from that enforcement role and routing disputes to a dedicated team or a manager is more effective than training every worker to handle it individually. Changing a process is more powerful than training people to survive a bad process.
• Environmental design. Service windows that increase physical distance without creating a sense of hostility, seating arrangements in waiting areas that reduce frustration, clear signage that manages expectations about wait times, controlled access to back-office areas, lighting that removes hidden corners. These are not cosmetic — poorly designed physical environments measurably increase aggression frequency.
• Staffing and rostering. Ensuring no worker is alone in a high-risk situation, scheduling experienced staff during peak risk periods, and building handover processes that do not leave incoming workers uninformed about a deteriorating situation from the previous shift.
• Genuine de-escalation training, practised regularly. The evidence distinguishes between training that covers techniques theoretically and training that involves repeated role-play of realistic scenarios. The latter produces durable behaviour change; the former fades within weeks. Annual classroom sessions on conflict management are largely ineffective without regular refresh and practice.
• A reporting culture that workers actually trust. Most organisations have under-reported violence not because workers do not experience it but because previous reports went nowhere, or the worker was made to feel they should have handled it better. Tracking what happens after a report is submitted — did an investigation occur, was the worker updated, was any change made — is the only honest measure of whether a reporting culture exists.

What is weak or overrated:

• Lone panic buttons without a guaranteed response. A panic button that alerts a colleague who is also alone, or that generates a response time of several minutes, provides limited real protection in a fast-moving incident. It may be legally required and is worth having, but it should not anchor a prevention strategy.
• Zero-tolerance policies stated but not enforced. The most common finding in organisations with chronic violence problems is a policy that says zero tolerance and a practice that tolerates a great deal. Staff observe this gap immediately and conclude, correctly, that the policy is performative.
• Generic resilience training offered to workers in high-violence roles. Teaching a healthcare worker who is assaulted regularly to 'build resilience' is addressing the wrong end of the problem. Resilience training has a place in overall wellbeing support, but it is not a substitute for reducing the hazard.

What a Worker Should Do If They Are Already Affected

If you have experienced a violent incident at work, or if you recognise the early signs described above in yourself, the following applies regardless of how serious the event seemed at the time.

Report the incident through whatever mechanism your employer has — written, electronic, or verbal to a manager — and do this as soon as practically possible. The purpose is not only to trigger a response but to create a documented record that protects you if symptoms develop later and you need to demonstrate that an occupational incident occurred. Do not edit your account of what happened based on how you think it will be received. Describe it factually and completely.

Take the offer of post-incident support if your employer provides it. If your employer does not offer it, request it explicitly. If you are offered a return to the same environment within a very short period after a significant incident, you are entitled to raise with your employer whether a temporary adjustment is possible — most occupational health frameworks recognise temporary redeployment or modified duties as a reasonable response to a violent incident.

Seek professional help without waiting until symptoms are severe. A general practitioner or occupational health professional can assess whether what you are experiencing crosses a clinical threshold, and early intervention for psychological injury following a violent incident significantly improves long-term outcomes. The research on this is consistent: people who receive psychological first aid or structured support within the first few weeks after a significant incident recover faster and more completely than those who wait until the condition has become entrenched. Waiting to see if it resolves on its own is a reasonable instinct but a poor strategy.

If your employer responds to a report with dismissal, minimisation, or implicit blame — suggesting you provoked the incident, that this is 'part of the job,' or that you are overreacting — this response is itself a harm, and in most jurisdictions it may constitute a failure of the employer's duty of care. Document what was said and when, and consider whether an occupational health referral, a trade union representative, or an external advice service is the appropriate next step.

What Is Commonly Misunderstood

'It is part of the job' is the most damaging misunderstanding in this area. It is heard most often in healthcare, retail, and hospitality, and it functions as a mechanism for normalising harm rather than preventing it. The fact that an occupational group is routinely exposed to violence does not make that exposure acceptable, legally permissible, or inevitable. Employers in high-exposure industries have the same duty to control violence risk as employers in any other industry. Workers in those industries have the same right to work without being harmed. High incidence of violence in a sector is evidence that controls are inadequate, not evidence that violence is an inherent feature of the work.

Workplace violence is primarily a security problem. Security measures — guards, cameras, access control — address only the physical dimension of violence and primarily the external type. The majority of workplace violence by frequency is verbal, psychological, or internal, and it requires occupational health responses: risk assessment, policy, training, reporting systems, and mental health support. Treating it as a security issue means most of the problem goes unaddressed.

If no one is reporting it, there is no problem. Under-reporting of workplace violence is so consistent across industries and research populations that low incident-report numbers are, paradoxically, a signal worth investigating rather than a reassurance. Workers under-report because they believe nothing will change, because they fear being seen as unable to cope, because the process is burdensome, or because they have internalised the message that what happened to them was normal. An organisation that has not had a violence report in twelve months should ask why, not assume safety.

Verbal abuse does not cause lasting harm. It does. Repeated exposure to verbal aggression produces documented neurological and physiological changes. A customer service worker who absorbs abusive language across hundreds of interactions per week is not experiencing an inconvenience — they are experiencing a chronic occupational exposure with measurable health consequences. The absence of a bruise does not mean the absence of an injury.