Author – Chris Sherwin, Operations Manager at COMET

A dropped object lands two metres from a worker, a train passes a signal at danger with nothing on the line ahead, or a gas release disperses before it finds an ignition source. Each one gets logged as a near miss, and too often each one gets a lighter investigation than it deserves, because no consequence was realised for people, the environment, assets or reputation.

The outcome tells you how lucky you were, but it tells you nothing about how safe you were. In each of those events a barrier failed outright and the gap between the hazard and the person was filled by chance: a metre of distance, an empty section of track, a change in wind direction. Luck is not a control, and it does not appear in your hierarchy of controls for good reason.

Senior HSEQ leaders across energy, rail, construction, maritime and utilities face the same question. Does your organisation investigate near misses for what they reveal about barriers, or does it file them and move on?

Potential severity, not actual severity

The shift that changes near-miss investigation is the move from asking what happened to asking what could have happened. A potential serious injury or fatality (PSIF) is any event where, with a small change in circumstances, someone could have been killed or permanently harmed. The industries that have adopted PSIF thinking have found that a significant share of their near misses sit in this category, and that driving down minor injuries does not reliably drive down serious ones. The events that hurt people most tend to be preceded not by minor injuries of the same type, but by near misses involving the same failed barrier.

This means your near-miss log already contains a map of where your critical controls are failing. Many organisations have not read it, because the triage happens on actual consequence rather than potential consequence, and because investigation effort follows the triage.

Failed barrier analysis, applied to the event that nearly happened

A barrier analysis of a near miss asks the same questions as a barrier analysis of a fatality. Which barriers were supposed to prevent this? Which were in place and effective? Which failed, and which were missing altogether? The answers are identical whether or not the hazard reached a person. The only thing that differs is the willingness to look.

The chain of causation then does the rest. Working back from the failed barrier, the investigation traces both the visible contributing factors (the defective lanyard, the missed inspection) and the hidden ones (the maintenance backlog that deprioritised the inspection, the procurement decision that chose that lanyard, the production pressure that normalised working beneath a live lift). The hidden factors are the ones that will cause the next event, and they are the ones that a light-touch near-miss review never reaches.

Measure your barriers before they fail

A near miss is one form of barrier test, and an uncontrolled one. A mature organisation also tests barriers deliberately. Audits and inspections, when they are structured well, measure whether a control is present, functioning and understood by the people relying on it. The findings from those tests should feed the same analysis as the findings from events. If the inspection shows the control is degrading and the near miss shows the control has failed, you have two data points on the same trend rather than two unrelated records in two systems.

How COMET helps organisations become safer

COMET was built by investigators, and its methodology treats a near miss as a lower-consequence event, not a lesser one. The depth of the investigation should match the potential severity, and COMET gives you two routes to do that.

For near misses with serious potential, such as the dropped object, the signal passed at danger or the gas release, the full COMET investigation applies the depth of analysis you would use after a serious injury, tracing the chain of causation back to the hidden factors before they cause the next event.

For events with low potential severity, such as observations, minor incidents and near misses that could not realistically have caused serious harm, COMET Lite provides a streamlined process that takes a fraction of the time. It still includes barrier analysis and Human Factors, and it uses the same coded root cause taxonomy as a full investigation, so the learning from a near miss sits alongside the learning from a serious incident and counts towards the same trends.

COMET Assured brings audit and inspection findings into that same taxonomy, so your deliberate barrier tests and your accidental ones build one body of evidence. A finding that a control is degrading sits alongside the near miss where it failed, and both point to the same root cause.

If you want to find out whether your organisation is failing safe or failing lucky, book a discovery call.

‍