When 87% of your ED nurses have been assaulted by a patient — and your wellbeing program still hasn’t changed

This is the conversation taking place in executive meetings across the sector, even if it is no longer being voiced as often because the response has become predictable. Patient aggression across our EDs and acute wards is at record levels and still climbing. Our most experienced nurses — the ones everyone turns to when a shift starts to unravel — are not transferring; they are leaving the profession. And we have run resilience training every year, on schedule, and it has produced no observable shift in any number we report on.

The critical point is not that the training was delivered. It is that nothing changed afterwards. The organisation invested in the initiative, scheduled it, measured participation and reported completion rates. Yet the incidents, absences and resignations continued to trend upwards. That is not a delivery failure. It is a diagnosis failure. The intervention was aimed at the wrong category of problem.

Why doesn't resilience training move the assault numbers?

Because resilience training asks the person being harmed to absorb the harm better. It is a control applied to the victim, not to the prevent the hazard.

The scale of the hazard is no longer arguable. In a survey reported by Edith Cowan University in 2024, 87% of nurses said they had experienced patient-related violence. In-hospital assaults rose 60% in Victoria, 48% in Queensland and 44% in New South Wales between 2015 and 2018 (Edith Cowan University, 2024). These are not difficult patients having a hard day. This is a documented, accelerating exposure to physical violence as a routine feature of the job.

Run that through the hierarchy of controls — the framework HSE teams already apply to every other workplace hazard. Training sits near the bottom: an administrative control, one of the least effective measures, used when the hazard cannot be eliminated, substituted or engineered out of the system. No safety manager would respond to a significant increase in a physical hazard by booking exposed workers into an annual awareness session and declaring the risk managed. Yet when the hazard is patient aggression and assault, many organisations continue to rely on resilience training as a primary control.

So the honest answer to "why hasn't the training worked" is that it was never the right control for this risk. That is the wrong question. The right one is: what are we doing upstream of the moment a nurse is hit?

Why is Code Black a leading indicator many organisations are already overlooking?

Every Code Black is a data point. Every behavioural emergency call is a psychosocial hazard event that has already been logged, time-stamped and located — by ward, by shift and by hour. The organisation already knows where the exposure is occurring. It holds a real-time map of workforce risk. Yet in many health services that information dies in a security log, never translated into the language of psychosocial hazard management.

Viewed as a leading indicator, the picture changes. A cluster of Code Blacks on a particular ward, during particular shifts, is not merely an operational nuisance to be staffed around. It is an early warning signal for the resignation that will be processed months later, the WorkCover claim that follows, and the psychosocial complaint that arrives when the service can least afford it. The assault is the visible event. The turnover, compensation costs and regulatory exposure are the less visible consequences trailing behind it. Making that invisible risk visible is precisely the work organisations are expected to do.

This matters now because the regulatory frame has shifted. Psychosocial hazards are not simply a wellbeing issue; they are a work health and safety obligation. Officers have a personal due diligence duty to understand organisational hazards and ensure that reasonably practicable controls are implemented. Where assault data, turnover data and workforce risk indicators are known, a reliance on annual resilience training alone may become difficult to defend as a reasonably practicable response in a regulatory, investigative or coronial context.

What does workforce attrition and workforce damage actually cost a health service?

The workforce consequences are already reflected in organisational financials, whether they have been explicitly recognised or not. Across Australian healthcare, 84% of workers report symptoms of burnout and 68% of staff absences are linked to burnout or mental health concerns (Brightstar Nursing Australia, 2025). Those impacts appear in unplanned leave costs, agency staffing expenditure and escalating overtime budgets — all of which can often be traced back to workforce risks that remain insufficiently controlled.

And the problem is self-reinforcing. The projected national nursing shortage is expected to reach 85,000 by 2025 and 123,000 by 2030 (Health Workforce Australia). In that labour market, the loss of an experienced nurse due to workplace violence or chronic exposure to aggression is rarely offset by an equivalent replacement. Instead, services often rely on agency staff, junior clinicians and reduced skill mixes. The result can be higher exposure to behavioural incidents, increased operational pressure and further workforce attrition. The assault problem and the workforce problem are, in many respects, the same problem feeding itself.

Why won't more violence-management training close the gap?

Because training changes how staff respond to aggression; it does not reduce the conditions that create it. Violence-management training is necessary, but it is not sufficient. The research reflects this reality. ECU researcher Brendan Johnson reported in 2024 that aggression and violence against emergency department workers are "on the rise" and that "current strategies in place to manage the issue are perceived to be insufficient" (Edith Cowan University, 2024). When assault rates continue to rise despite training, the issue is no longer staff capability. It is whether the organisation has implemented enough controls to reduce exposure to the hazard itself.

Violence-management training equips a nurse to handle the moment of aggression. It does nothing for the slower harm: the senior nurse who is hit on Tuesday, absorbs it, says nothing, becomes the person every junior offloads to in the tearoom, and quietly decides over six months that this is no longer a profession she will stay in. That nurse has become an accidental counsellor — carrying the team's distress with no structure, no support and no visibility to the system. There is a known ceiling on how many people any one person can hold that way; past Dunbar's number, around 150, informal support simply does not scale, and on a large multi-site service it breaks long before that.

What is missing is the layer between the incident and the EAP. The EAP is real and it matters, but it is reactive support — it activates after someone is already in distress and self-refers, which the people most affected rarely do in time. There is a gap between "assaulted on shift" and "ringing a confidential line weeks later", and right now nothing structured lives in that gap. That is the pre-EAP space.

What does the structural layer look like instead?

It is wellbeing infrastructure, built as an operating model rather than an event. At the core of the model is a trained wellbeing champion network: trusted peers who recognise concerns early, respond appropriately and connect colleagues with support before issues escalate. It formalises the accidental counsellor role, providing the training, support, boundaries and escalation pathways that are often missing.

When implemented effectively, the model shifts the response upstream. Rather than waiting for distress to become a crisis, it creates earlier identification, earlier conversations and earlier pathways to support. The result is two complementary outcomes: fewer situations escalating to the EAP, and more people accessing the EAP at the right time. Fewer crises. Better-directed support.

It also provides something annual training alone cannot: a defensible organisational response. A standing champion network, aligned to hazard data and integrated with the risk register, is a visible demonstration of due diligence — a stronger answer to the question of what was reasonably practicable than an attendance sheet. It is people-risk infrastructure that sits upstream of the EAP, where emerging concerns can be identified before they become harm.

For a COO or CNO leading a Local Health District or multi-site private health service, the practical task is straightforward: treat Code Black data as a workforce risk signal, map it against absence and turnover trends, and assess the controls operating between the incident and the exit interview. If the strongest control remains an annual resilience session, the evidence suggests the organisation is unlikely to be addressing the risk at its source.