A near miss can be easy to dismiss precisely because nobody was hurt. But the absence of injury does not prove that the controls worked. A credible near miss can reveal the same underlying failure that later causes a serious accident, giving managers an opportunity to act before the outcome depends on luck again.
A warning came before people were injured
In September 2026, the Health and Safety Executive reported a prosecution involving a scrap-metal site where an excavator-related near miss occurred just ten days before two members of the public were injured in a later event.
HSE found that the underlying failures in the earlier near miss were the same. Although the case arose in waste and recycling rather than construction, the management lesson transfers directly to construction sites where excavators, telehandlers, temporary access routes, lifting operations and changing layouts create comparable high-potential events.
Not every near miss is reportable under RIDDOR
Internal near-miss reporting and statutory reporting are not the same thing. Under the Reporting of Injuries, Diseases and Dangerous Occurrences Regulations 2013, specified dangerous occurrences must be reported by the responsible person.
HSE’s dangerous-occurrence guidance sets out the categories. A site event may deserve urgent internal investigation without meeting one of those statutory categories.
That means the first management question should not simply be “Do we have to report this to HSE?” It should be “What allowed this to happen, and could the next occurrence seriously hurt somebody?” RIDDOR screening then sits alongside the investigation rather than replacing it.
A near miss may show that the risk assessment is no longer valid
Regulation 3 of the Management of Health and Safety at Work Regulations 1999 requires risk assessments to be reviewed where there is reason to suspect they are no longer valid or where there has been a significant change.
A near miss can provide exactly that reason. If a lifting plan assumes pedestrians will be excluded but someone enters the lifting zone because a barrier has been moved, the fact nobody was struck does not demonstrate success. It may show that one of the assumptions on which the plan depended is not reliable in practice.
The potential consequence should influence the response. A falling object missing somebody by two metres and the same object striking somebody can originate from exactly the same control failure.
Investigate causes, not just the person closest to the event
HSE’s HSG245, Investigating accidents and incidents, provides a practical framework based on gathering information, analysing it, identifying risk-control measures and developing and implementing an action plan.
Weak investigations often stop at the first human action: “the operator reversed without checking”, “the worker entered the exclusion zone” or “the supervisor did not notice”. Those facts may matter, but they are not always the underlying cause.
A stronger investigation asks why the behaviour was possible. Was physical segregation ineffective? Had the sequence changed? Was the method unrealistic for the actual site? Had a shortcut become normal practice? Was production pressure influencing decisions? Had the issue been reported previously? Was supervision appropriate?
Useful evidence may include photographs, CCTV, equipment data, witness accounts, RAMS, lifting or traffic plans, inspection records, competence information, permits and records of recent changes.
Corrective actions only count when they are implemented and tested
Many businesses investigate reasonably well and then lose control at the action stage. A report identifies actions, they are entered on a spreadsheet, a toolbox talk is delivered and the event is marked closed.
Corrective action should match the failure identified. If pedestrians can physically enter a plant operating area, changing the layout or providing effective segregation may be stronger than another verbal reminder. If a worker misunderstood a new procedure, targeted instruction or demonstration may be appropriate.
Each significant action should have an owner and a realistic completion date. For higher-potential events, an effectiveness check should follow: is the revised barrier still in place, is the new sequence being followed, has the same unsafe condition appeared elsewhere, and have supervisors challenged deviations?
Closing a spreadsheet cell is not the same as closing the risk.
What should you do now?
- Make near-miss reporting simple enough that workers will use it.
- Record the potential consequence, not only the fact nobody was injured.
- Screen RIDDOR requirements separately.
- Preserve useful evidence before site conditions change.
- Identify immediate, underlying and management-system causes where appropriate.
- Review the risk assessment and method if the event challenges existing assumptions.
- Use the hierarchy of control when selecting corrective action.
- Give actions named owners and deadlines.
- Check whether the same weakness exists on other projects.
- Verify afterwards that the corrective action has actually worked.
Our professional view
The phrase “near miss” can make an event sound less significant than it was. What matters is not how close somebody literally came to injury; it is what the event tells management about the controls.
A high-potential near miss is valuable information. The strongest organisations do not celebrate simply because nobody was hurt. They ask what would have happened if the timing, distance or position had been slightly different, and they use that warning to improve the system before the next event.
All Star Safety’s health and safety consultancy service includes accident and incident investigation, root-cause analysis, RIDDOR support and corrective-action planning. Our IOSH Managing Safely courses also cover incident investigation as part of practical risk management.