Near-miss reporting rises and injuries fall

The situation

A general contractor running two mid-rise sites had a near-miss reporting box in the site trailer and, in the year before this comparison, six reports in it. That same year the firm recorded four injuries meeting the reportable threshold. Reporting was technically available; almost nobody used it.

The problem

A box in a trailer asks a worker to stop, find paper, write, and walk it somewhere — friction that near-miss reporting does not need to have. CCOHS's own guidance describes reporting methods plainly: hazards “can be reported verbally, electronically, or by filling out company-specific forms that should be available at bulletin boards or other conspicuous places.” The firm had exactly one of those channels, in exactly one location, and it went almost unused.

The following year the firm added a phone-in line and a one-line report field on every toolbox-talk sign-in sheet, so reporting no longer required a special trip. Both changes cost almost nothing to implement.

The numbers

Year 1: 6 near-miss reports, 4 recordable injuries. Year 2, after the two low-friction reporting channels went live: 41 near-miss reports, 1 recordable injury. No Canadian regulator publishes a formula connecting those two counts, and this page does not claim one — the widely repeated “300 near-misses to 1 injury”-style ratios circulating in general safety literature trace to a 1930s American study, not a Canadian source, and are not cited here. What is verifiable is this site's own before-and-after count.

The rule that decided it

CCOHS's incident-investigation guidance is explicit that investigating a near-miss is not optional busywork: “most importantly, to find out the cause of incidents and to prevent similar incidents in the future,” and that principle “appl[ies] to an inquiry of a minor incident and to the more formal investigation of a serious event.” Every one of the 41 year-two reports got the same brief look-back a lost-time injury would have gotten, which is what turned raw report volume into fixed hazards rather than a filing cabinet.

That same guidance corrects a common myth worth stating plainly: it does not set a 24-hour investigation deadline, or any fixed number of hours. It says only that “as little time as possible should be lost between the moment of an incident and the beginning of the investigation,” and that jurisdictional timing requirements vary — there is no single Canada-wide clock, contrary to a figure that circulates informally on some job sites.

What it would have cost otherwise

A lost-time injury does not end at the claim. Its cost enters WSIB's six-year rolling claims-cost window, which sets the risk band — and therefore the premium rate — the employer sits in going forward: “we'll use your insurable earnings, claims costs and the number of allowed claims, over a six-year period to set your premium rates,” with roughly “five per cent” separating each risk band. Trading three avoided injuries a year for forty-one five-minute conversations is not just a safety outcome; it is a multi-year premium exposure the firm did not carry into the following years.

The outcome

The firm kept both low-friction channels permanently and added the near-miss count, not just the injury count, to its monthly site safety review — treating a rising report count as a leading indicator worth celebrating rather than a sign the site had gotten more dangerous.

A low-friction reporting channel is the same fix behind how an orientation gap was caught during a routine site visit, and the investigation discipline behind it is the same one that turned a work refusal into a 33-minute fix rather than a dispute in a different case on this site.

The tell

Read together, the 41 year-two reports clustered around three recurring hazard types — housekeeping and trip hazards, unguarded floor openings, and PPE not worn on short tasks — and those were the same three hazard types behind three of the four year-one injuries. The near-misses were not random noise; once anyone had a large enough sample to look at, they were pointing at the exact same failure points the injuries came from, just before someone got hurt instead of after.

Takeaways

• • Reporting friction, not worker attitude, is usually why a near-miss box goes unused — add channels people already have open (a phone, a sign-in sheet).
• • There is no published Canadian near-miss-to-injury ratio; do not borrow the American pyramid figures that circulate informally.
• • CCOHS's own guidance sets no fixed investigation-time rule (no “24 hours”) — the standard is “as little time as possible,” and jurisdictional rules vary.
• • A rising near-miss count is a leading indicator, not a sign a site got more dangerous — track it in the same review as injuries, not instead of them.
• • Claims cost sits inside WSIB's six-year rolling window and moves the employer's risk band roughly 5% per band — the real, multi-year cost of the injuries this programme avoided.