A female shift worker at Inghams' chicken processing factory was attacked by an ex-employee in the car park late at night. She won $150,000 at trial. The Court of Appeal overturned it — not because the risk of harm wasn't foreseeable or significant, or because there was no breach of duty, but because causation failed. The Court held Inghams' breach of duty was not proven to be a necessary condition of the plaintiff's injury.
This case sits alongside Coca Cola v Pareezer as one of the most important Australian decisions on the limits of causation — and why proving a breach isn't the same as proving it caused your injury.
Kim Yen Tat was a shift worker at Inghams' chicken processing factory at Murarrie in Queensland. Late one night, after finishing her shift, she was attacked in the factory's car park by an ex-employee. She sustained significant PTSD and sued Inghams for damages under the Workers Compensation and Rehabilitation Act 2003 (Qld) — the same legislation that governs claims handled by workers compensation lawyers Gold Coast wide.
Before attacking Tat, the ex-employee had approached three other female workers in the car park, spinning various versions of the same story in what appeared to be an attempt to coax them into his vehicle. None of those three workers felt his behaviour was menacing enough to report to the security guard who was stationed in an office nearby.
Inghams had a security guard present but had not provided security awareness training to its staff, and its security measures were not geared toward addressing the risk of third-party violence to employees leaving after shifts.
Read the full judgment on Queensland Judgements Further reading — Hall & Wilcox analysisThe trial judge found Inghams had breached its duty of care by failing to train staff to report suspicious activity and by failing to gear its security measures toward the risk of third-party violence. That finding of breach was not disturbed on appeal. However, the Court of Appeal held the plaintiff failed to establish causation.
Where the case fell apart was causation. The trial judge's reasoning depended on a sequence of three conclusions: that a trained worker would have reported the attacker's behaviour, that the security guard would have located and directed him to leave, and that he would have complied and not returned. None of these conclusions were supported by evidence. The Court of Appeal found they amounted to mere conjecture.
In particular, there was no evidence at all that the attacker — a man who had already approached three workers and was intent on violence — would have complied with a security guard's direction to leave. Without that link, the chain of causation broke. The but-for test was not satisfied, and the damages were overturned.
The breach was never in dispute on appeal. What split the courts was whether the trial judge's causal reasoning was inference or conjecture.
Inghams failed to train staff to report suspicious activity and failed to gear security toward third-party violence. Had it done so, a trained worker would have reported the attacker, the guard would have located and removed him, and he would not have returned to attack Tat.
The post-incident warning email Inghams sent to staff — advising them to report suspicious behaviour — was treated as evidence of what proactive measures should have been in place before the attack.
Each step in the trial judge's causal reasoning — that a worker would have reported, that the guard would have acted, that the attacker would have complied and not returned — was unsupported by evidence. That is conjecture, not inference.
Critically, there was nothing to suggest a determined attacker who had already approached three workers would have simply walked away if asked to by a security guard. The causation could not be established on the evidence.
The trial judge's causation reasoning depended on three sequential conclusions. The Court of Appeal found none of them were supported by evidence — each was conjecture rather than reasonable inference.
Three workers had already encountered the attacker before the assault. None of them felt his behaviour was menacing enough to report. The trial judge assumed that with security awareness training they would have — but there was no evidence to support that assumption. The three workers who were actually there made a different judgment in real time.
Even if a report had been made, the trial judge assumed the guard would have found the attacker in the car park and directed him off the premises. There was no evidence about what the guard would actually have done in that situation, how quickly he would have responded, or whether he would have found the right person.
This was the most critical gap. The attacker had already approached three workers and was clearly intent on finding a victim. There was no evidence at all that he would have simply complied with a direction to leave — let alone that he would not have returned afterward. The Court of Appeal found this assumption in particular was entirely unsupported.
This case reinforces the difference between inference and conjecture — and why a chain of causation is only as strong as its weakest unproven link.
Proving a breach of duty is not the same as proving negligence — all elements of the tort must be established. In this case, the finding that Inghams breached its duty of care was not disturbed on appeal. However, because causation was not established, the negligence claim failed. This distinction is one of the most important — and most misunderstood — aspects of any personal injury lawyers Queensland practitioners deal with.
The starting point for causation is the but-for test: but for the defendant's breach, would the harm have occurred? If the answer is yes — if the harm would have happened regardless — causation fails. Here, the Court of Appeal was not satisfied the attack would have been prevented even if Inghams had done everything it should have. The attacker's behaviour was the decisive variable, and his response to any intervention was entirely speculative.
Courts can draw inferences from evidence. What they cannot do is assume facts that are not supported by any evidence at all. The distinction between inference and conjecture is critical. A reasonable inference is one the evidence actually supports. A conjecture is a possibility — one of many — without evidential foundation. Multiple conjectures stacked together do not become reasonable simply because they are sequential.
The Court of Appeal confirmed that Inghams' post-incident warning email to staff — advising them to report suspicious behaviour in the car park — was properly used as evidence of what it should have been doing before the attack. Taking precautions after an incident does not make an employer liable, but it can confirm what reasonable precautions were available and practical before the event.
This case and Coca Cola v Pareezer sit together as the two clearest illustrations of how causation can defeat an otherwise strong negligence claim. In both cases the employer was found to have breached its duty of care. However, in both cases the claim ultimately failed because the plaintiff could not establish the necessary causal link between the breach and the injury.
The third link in the chain is the one courts will always scrutinise hardest when the harm is caused by a third party's criminal act. Would the attacker have complied? Would they have come back anyway? These questions go to the behaviour of someone who has already demonstrated an intent to commit a crime. Assuming they would have been deterred by a security guard's direction is a significant leap — and courts require more than a hopeful assumption to satisfy the but-for test.
The practical lesson: in cases involving third-party criminal acts, causation evidence needs to specifically address what the attacker would have done differently. General arguments about what a trained security response "would have" achieved are not enough if they can't be grounded in something more than inference.
One conversation. No pressure. Tell us a bit about your situation and we'll be in touch.