Why near-miss reporting matters as much as reportable incidents

A near miss — an event that didn't cause injury but easily could have — generally isn't RIDDOR-reportable in the way an actual injury is. That doesn't make it unimportant; in many respects it's more useful, because it's a warning about a hazard before anyone actually gets hurt by it.

What counts as a near miss, in practice

It's worth being generous with the definition rather than narrow. A load that slipped but didn't fall, a vehicle reversing into a space someone had just walked out of, a piece of edge protection that turned out to be loose when someone leaned on it — none of these caused injury, and none of them would necessarily meet a strict definition of "dangerous occurrence" under RIDDOR. But each one is telling you something true about a hazard on your site right now, and that's the actual value of logging it.

Why near misses get under-reported

The honest reason near misses often go unrecorded isn't that people don't notice them — it's that reporting one can feel like admitting a mistake, or drawing attention to something that "nothing actually happened" from. A worker who reports that a load nearly fell on them might reasonably worry it reflects badly on how they were working, rather than being seen as flagging a genuine site hazard. If that's the prevailing attitude on a site, near misses simply stop getting mentioned out loud, let alone written down — and the site loses its earliest warning system as a result.

Sites with a genuinely strong safety culture tend to log near misses as readily as actual incidents, on the basis that the underlying hazard — a poorly secured load, a gap in edge protection, a piece of plant behaving unexpectedly — is usually identical whether or not this particular instance happened to cause harm.

Building a culture where they actually get logged

This tends to come down to how a report is received the first few times it happens, not how it's described in a policy document. If the first response to a reported near miss is "what were you doing near there" rather than "tell me exactly what happened," people learn quickly which kind of response to expect and adjust their willingness to speak up accordingly. Making the reporting process quick and low-friction matters too — a near miss that requires the same lengthy form as a serious incident report is less likely to get logged consistently, particularly for the kind of minor, easy-to-shrug-off event that's actually the most common and often the most instructive.

A worked example

A scaffold board shifts underfoot but doesn't give way. No injury, no incident, nothing RIDDOR requires reporting. If it's logged as a near miss and someone actually looks at why it happened — a missing board clip, incorrect loading, a fixing that had worked loose — that's a real finding that can be corrected before the same board gives way properly under the next person who steps on it. If it's not logged, the same underlying fault stays on the scaffold, waiting for the next person to find it the hard way.

This is also directly relevant to the proposed 2026 RIDDOR changes, which include extending near-miss reporting to specific new categories. Sites that already treat near misses seriously, rather than as informal conversations that don't get written down, will find that kind of regulatory shift far less disruptive than sites starting the habit from scratch.

Reviewing the pattern, not just the individual report

A single near miss tells you about one moment. A pattern of near misses tells you about a site. If three separate near misses over a month all involve the same access route, the same piece of plant, or the same time of day, that's a much stronger signal than any one of them individually — and it's the kind of signal that only shows up if someone is actually reviewing near-miss reports as a set, rather than filing each one away individually once it's logged. This is where near-miss data earns its keep: not as a compliance record, but as an early map of where a site's genuine hazards actually cluster.

Near misses involving subcontractors specifically

It's worth paying particular attention to near misses involving subcontractors who are new to a site, or who turn over quickly, since they're less likely to have picked up the informal, unwritten knowledge of a site's quirks that longer-standing workers have absorbed — where the blind spot on a particular access route is, which piece of plant has a known habit of swinging wider than expected. A near miss involving a new arrival is often less about that individual's competence and more about a site-specific hazard that hasn't yet been made explicit to people who haven't been there long enough to learn it informally.

Feeding near misses back into induction

A near miss that's logged, reviewed, and then quietly filed hasn't fully delivered its value if the lesson from it never reaches the people who'd benefit from knowing about it. Feeding recurring near-miss findings back into site induction — "this access route has caught two people out this month, here's what to watch for" — turns an individual incident into shared site knowledge, which is exactly the informal awareness that new arrivals are otherwise missing until they learn it the hard way themselves.

Common mistakes

• –Responding to a reported near miss by questioning the worker's judgement rather than examining the hazard
• –Making near-miss reporting require the same lengthy process as a serious incident, discouraging the small, common ones from being logged
• –Logging near misses but never actually reviewing the pattern across them to spot a recurring hazard
• –Treating a near miss as a closed matter once it's written down, without following through on a fix
• –Assuming near misses aren't worth recording because they're not currently RIDDOR-reportable
• –Not distinguishing near misses that reflect a genuine site hazard from ones that reflect a gap in a new arrival's site-specific knowledge