There is no Canadian risk matrix.
Not in the Canada Labour Code, not in Ontario’s OHSA, not in Alberta’s OHS Code, not in the BC OHS Regulation, not in Quebec’s LSST. Every one of them requires you to identify hazards and control them. None of them tells you to multiply two numbers together.
That has a practical consequence worth internalising before you use this page: the score is not what gets examined. What gets examined is the control you chose, and whether it sits as high in the hierarchy as was reasonably practicable. A hazard rated 12 and answered with a sign is a finding. A hazard rated 4 and answered by redesigning the task is good practice. The arithmetic does not save you either way.
So use the grid for what it is genuinely good at — ordering a long list of hazards so the worst ones get attention first — and do not mistake it for the assessment.
The grid
Two axes, four levels each, a product from 1 to 16.
Severity — the most likely outcome if harm occurs:
| Level | Value | What it means |
|---|---|---|
| Minor | 1 | First aid only, worker returns to the task |
| Moderate | 2 | Medical treatment or short time loss, full recovery expected |
| Serious | 3 | Long time loss, permanent impairment possible |
| Critical | 4 | Fatality or life-altering injury |
Likelihood — with the controls you already have:
| Level | Value | What it means |
|---|---|---|
| Rare | 1 | Not seen here or in the industry |
| Unlikely | 2 | Has happened elsewhere, not here |
| Likely | 3 | Has happened here, or near misses are being reported |
| Almost certain | 4 | Happens regularly, or exposure is continuous |
And what the product calls for:
| Level | Score | Action |
|---|---|---|
| Low | 1 – 2 | Monitor; revisit at the scheduled review |
| Moderate | 3 – 4 | Plan a control, with an owner and a date |
| High | 6 – 9 | Control now; interim measures immediately |
| Critical | 12 – 16 | Stop the task until it is controlled |
The hierarchy of controls is the real answer
This is the part that matters in Canada, and it is worth more page space than the matrix.
The accepted order, in descending effectiveness:
- Elimination — remove the hazard entirely. Do the work at ground level instead of at height.
- Substitution — replace it with something less hazardous. A water-based product instead of a solvent.
- Engineering controls — isolate people from the hazard. Guarding, ventilation, interlocks, barriers.
- Administrative controls — change how people work. Procedures, permits, rotation, training, signage.
- Personal protective equipment — the last line, and the one that fails quietest.
The reason this ordering carries legal weight rather than merely good sense: the duty in every Canadian jurisdiction is qualified by what is reasonably practicable, and an inspector or a court reads that as a question about the hierarchy. “Why did you go straight to PPE?” is not a rhetorical question. It has a right answer — because everything above was genuinely not practicable, and here is the reasoning — and if you cannot give it, the score in your spreadsheet does not help.
The corollary for this tool: rate the hazard, then write down which level of the hierarchy your control sits at, and why you did not go higher. That single note is worth more than the number beside it.
Due diligence, and why the paperwork is the defence
Canadian OHS enforcement runs on due diligence: taking every precaution reasonable in the circumstances. It is the defence available to a charged employer, and it is a defence made of records.
Which means a risk assessment does three jobs at once:
- It orders your work.
- It proves you looked, on a date, at a hazard.
- It shows what you decided and why.
The third is the one home-made matrices drop. A spreadsheet with columns for severity, likelihood and score, and nothing else, records that you rated something. It does not record that you thought about it. Add a column for the control selected, one for the hierarchy level, one for who owns it and by when — and the same file becomes evidence rather than arithmetic.
Two mistakes that empty a rating of meaning
Rating the worst imaginable outcome
Almost anything can kill given enough bad luck. If every hazard is rated severity 4, the matrix stops ranking anything and the action plan becomes an unordered list — which is a list nobody works through.
Rate the normally expected outcome. Catastrophic-but-remote scenarios do not disappear; they belong in a dedicated analysis, not in a matrix cell.
Rating inherent rather than residual risk
Rating as though your existing controls did not exist produces a map of fears, not a plan: everything is red, including what has been safely managed for a decade.
Rate the residual risk — what is left with the controls in place. It also makes the rating useful over time, because installing a guard makes it fall, and an indicator that moves when you act is the only kind worth tracking. Recording the inherent rating in a second column is legitimate and useful — it shows what the control is buying you, and what happens when it fails — but it is not the main column.
Where the matrix does not belong
Noise, chemical exposure, vibration and ergonomic load all have their own assessment methods, with measurements and exposure limits set out in regulation. Forcing them into a severity × likelihood grid produces a number with no meaning and, worse, one that looks comparable to the others.
The same goes for psychological health and safety, where the CSA has published a dedicated standard and a four-by-four grid is not the instrument.
What this tool does, and does not
It applies the grid, colours the cell, and exports the result with the scales. That last part is deliberate: a rating without its scales cannot be checked by anyone who was not in the room, and a bare “9” in a column means nothing six months later.
It stores nothing, sends nothing, and rates one hazard at a time. A full assessment across a site is a different exercise and needs something that holds several hundred rows.
What to keep
- The assessment, with the method and scales stated.
- The control selected for each hazard, with its hierarchy level and the reasoning for not going higher.
- Owners and dates, and evidence the actions were completed.
- The review trigger: new equipment, changed process, an incident, a worker concern raised through the joint health and safety committee.
- The committee’s involvement. In most jurisdictions the joint committee or health and safety representative has a role in hazard identification, and their input is usually the part a desk-based assessment is missing — what actually happens when the line jams at three in the morning.
If your method is better
This grid is practice, not a standard, and plenty of people have built better: a third axis for existing control effectiveness, a severity scale calibrated to their own claim categories, a decision rule that forces a walk down the hierarchy before any control is accepted.
If you have one that holds up — especially one that survives an inspection in more than one province — that is exactly what this site exists to collect and give back to the trade.