The First Incident Was The Control Review

Two Queensland prosecutions released in Jul 26 provide useful lessons about how organisations manage work outside their normal operations and how they respond when an existing control system has already failed. One involved an apprentice losing an eye during an unfamiliar ancillary task, while the other involved a dozer operator spending approximately eight hours trapped upside down beneath a coal stockpile.

The industries, work and legislation were different, but the underlying failure was similar. Work continued without the organisation establishing that the risk was understood and adequately controlled.

When An Add-On Job Falls Outside The System

On 06 Jul 26, a construction company was fined $90,000 after an apprentice suffered the near-complete loss of vision in one eye. The company normally installed portal frames but accepted an additional job involving filling fire doors with grout, which was outside its ordinary business activities.

Two young apprentices were assigned to complete the task. The company had not undertaken a risk assessment, developed a safe work instruction, provided adequate personal protective equipment, instructed the apprentices in how to perform the work safely or provided adequate supervision.

The grout hardened inside the guns being used by the apprentices, and they reported the problem to the company. One apprentice shook a gun to loosen the material while it was pointed towards his face, causing the gun to discharge grout into his eye. He was left with approximately five per cent vision and subsequently elected to have the eye removed. The company pleaded guilty to breaching sections 19(1) and 32 of the Work Health and Safety Act 2011 (Qld).

It would be easy to focus on the apprentice pointing the gun towards his face, but that would miss the organisational failure. He had been placed into an unfamiliar task without a defined method, adequate instruction, suitable protective equipment or competent supervision.

The risk was created when the business accepted work outside its established capability and treated it as a minor addition to the main job. The physical size or duration of a task does not determine its risk, and calling something an add-on job does not make the hazards disappear.

Before unfamiliar work is accepted or allocated, the business should establish whether it has the competence, equipment, information and supervision necessary to complete it safely. Where those things are absent, the options are to obtain competent advice, subcontract the work to somebody capable of doing it, develop and verify an appropriate method or decline the task.

Assigning two apprentices does not solve a competence gap. Without a competent person directing the work, it merely places two inexperienced workers in the same uncontrolled situation.

Task Novelty Is A Risk Indicator

Risk assessments commonly focus on recognised high-risk activities, plant and substances. Less attention is given to whether the task is new to the organisation, new to the workers or outside the processes for which the business has established controls.

Task novelty should trigger additional planning. An organisation should know when work falls outside its existing procedures, training, equipment or supervisory capability, because unfamiliar tasks are precisely where assumptions are likely to replace verified controls.

This does not require a new 40-page procedure every time somebody picks up a different tool. It requires a proportionate decision about whether the business understands the work well enough to perform it safely.

The relevant questions are practical. Has the task been performed before, who understands the equipment and material, what can go wrong, what does the manufacturer require, what protective equipment is necessary, and who is competent to supervise the work?

Where those questions cannot be answered, the work is not ready to commence. The operational schedule does not get a casting vote over competence.

The First Dozer Incident Was The Control Review

On 03 Jul 26, a Queensland coal mine operator was fined $95,000 after a dozer fell backwards into a void, rolled onto its roof and became almost completely engulfed in coal. The operator remained trapped upside down for approximately eight hours before being extracted and escaped without significant physical injury beyond mild dehydration.

The dozer was working on a stockpile approximately 20 metres high. Its onboard GPS system, which operators relied upon to identify coal valves beneath the stockpile, was not functioning properly, and the relevant valve did not have a physical marker showing its location on the surface.

The ground beneath the dozer gave way while the operator was attempting to move crusted coal above a suspected void. The machine fell backwards into the void and was buried, with the operator remaining inside until approximately 0730h the following morning.

What makes the case particularly significant is that a materially similar event had occurred on the same stockpile four months earlier. That event also involved a dozer becoming trapped near the same valve while the GPS system was not functioning correctly and the valve was not physically marked.

Recommendations from the earlier incident included installing physical valve markers and ensuring that the GPS systems were fully operational. Those recommendations had not been implemented before the second event.

The first incident was the control review. The organisation did not need another dozer in the void to establish that the combination of an unreliable GPS system, an unmarked valve and continued stockpile operations presented an unacceptable risk.

The operator had already received the warning in operational form. The second event was not new information, it was confirmation that the first response had failed to change the conditions under which the work continued.

Recommendations Are Not Controls

A recommendation does not reduce risk merely because it appears in an investigation report. Until it has been implemented, verified and incorporated into the way work is managed, it remains a proposed control.

A recommendation without an accountable owner, an implementation date, interim risk controls and evidence of closure is not a control. It is minutes with aspirations.

The mine operator subsequently spent approximately $2.41 million on measures that included remote dozer technology, physical indicators above stockpile valves, GPS exclusion zones, revised audit processes, safety management system changes and breathing apparatus in dozers. Approximately $1.9 million of that expenditure related to the remote dozer system.

Those measures demonstrate that substantial controls were technically and financially possible. The governance question is why work was allowed to continue under materially similar conditions while the known deficiencies remained unresolved.

A permanent engineering solution may take time to design and implement, particularly in a complex mining operation. That does not authorise normal operations to continue while the known risk remains effectively unchanged.

Interim controls should have been established and verified. Depending on the circumstances, that could have involved removing dozers with unreliable GPS from the stockpile, physically identifying every valve, establishing exclusion zones, limiting stockpile or bench heights, restricting activities above suspected voids or suspending the relevant operation until an acceptable system was in place.

The specific control belongs to the mine and its competent people. The requirement to do something effective while the permanent solution is being developed should not have been controversial.

Corrective Action Governance

Corrective action registers often create an appearance of control because recommendations have been entered, assigned a status and discussed at meetings. The register records administrative activity, but it does not establish that risk has been reduced.

Known high-consequence risks require stronger governance than a monthly review of overdue actions. The organisation should determine what must change immediately, what interim conditions apply, who has authority to suspend the work and what evidence will be required before an action is closed.

Closure should mean that the control is operating and has been verified in the workplace. It should not mean that equipment has been ordered, a procedure has been drafted or somebody has changed the status field to green five minutes before the executive meeting.

Senior leaders and officers should also be asking whether serious incidents and high-potential events are being treated according to what could reasonably have occurred.

The absence of injury does not turn a failed critical control into a low-priority administrative matter.

In the mine incident, the first operator’s survival did not demonstrate that the existing controls were adequate. It demonstrated that the organisation had received a warning without paying the full human price.

The Executive Questions

Officers and senior leaders should know which operations are continuing while important controls remain defective or incomplete. They should also know whether previous incidents have identified materially similar conditions and whether interim controls have been implemented while permanent solutions are pending.

A report stating that an action is in progress does not answer those questions. Executives need to understand what exposure remains today, who accepted that exposure, what operating restrictions apply and how control effectiveness has been verified.

They should also examine work being accepted outside the organisation’s normal capability. Small ancillary tasks can bypass established planning, procurement, competency and supervision systems because they are not recognised as a change to normal operations.

The construction company should have recognised that the grout work was outside its established method and competence. The mine operator should have recognised that the first dozer event had already tested and disproved the reliability of its existing control arrangements.

The Safety Jon Take

The first incident was the control review. Everything after that was a decision about whether the organisation would act on what it had learned.

Incident investigations do not control risk, recommendations do not control risk and executive reporting does not control risk. Risk is controlled when the conditions of work change and somebody verifies that the change is effective.

The construction case shows the cost of accepting unfamiliar work without first establishing competence and a safe method. The mining case shows the cost of treating a materially similar incident as the beginning of an action list rather than an immediate operating constraint.

Neither outcome arose because the risk was unknowable. The information required to act was available before the serious event, which is precisely why these cases deserve attention.